ViewVC Help
View File | Revision Log | Show Annotations | Download File
Revision: 1.464
Committed: Sun Jan 26 06:04:20 2020 UTC (4 years, 5 months ago) by root
Branch: MAIN
CVS Tags: rel-4_33
Changes since 1.463: +4 -4 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 aio and C<epoll>
111 interfaces, the BSD-specific C<kqueue> and the Solaris-specific event port
112 mechanisms for file descriptor events (C<ev_io>), the Linux C<inotify>
113 interface (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 Via the C<EV_FREQUENT> macro you can compile in and/or enable extensive
165 consistency checking code inside libev that can be used to check for
166 internal inconsistencies, suually caused by application bugs.
168 Libev also has a few internal error-checking C<assert>ions. These do not
169 trigger under normal circumstances, as they indicate either a bug in libev
170 or worse.
175 These functions can be called anytime, even before initialising the
176 library in any way.
178 =over 4
180 =item ev_tstamp ev_time ()
182 Returns the current time as libev would use it. Please note that the
183 C<ev_now> function is usually faster and also often returns the timestamp
184 you actually want to know. Also interesting is the combination of
185 C<ev_now_update> and C<ev_now>.
187 =item ev_sleep (ev_tstamp interval)
189 Sleep for the given interval: The current thread will be blocked
190 until either it is interrupted or the given time interval has
191 passed (approximately - it might return a bit earlier even if not
192 interrupted). Returns immediately if C<< interval <= 0 >>.
194 Basically this is a sub-second-resolution C<sleep ()>.
196 The range of the C<interval> is limited - libev only guarantees to work
197 with sleep times of up to one day (C<< interval <= 86400 >>).
199 =item int ev_version_major ()
201 =item int ev_version_minor ()
203 You can find out the major and minor ABI version numbers of the library
204 you linked against by calling the functions C<ev_version_major> and
205 C<ev_version_minor>. If you want, you can compare against the global
206 symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the
207 version of the library your program was compiled against.
209 These version numbers refer to the ABI version of the library, not the
210 release version.
212 Usually, it's a good idea to terminate if the major versions mismatch,
213 as this indicates an incompatible change. Minor versions are usually
214 compatible to older versions, so a larger minor version alone is usually
215 not a problem.
217 Example: Make sure we haven't accidentally been linked against the wrong
218 version (note, however, that this will not detect other ABI mismatches,
219 such as LFS or reentrancy).
221 assert (("libev version mismatch",
222 ev_version_major () == EV_VERSION_MAJOR
223 && ev_version_minor () >= EV_VERSION_MINOR));
225 =item unsigned int ev_supported_backends ()
227 Return the set of all backends (i.e. their corresponding C<EV_BACKEND_*>
228 value) compiled into this binary of libev (independent of their
229 availability on the system you are running on). See C<ev_default_loop> for
230 a description of the set values.
232 Example: make sure we have the epoll method, because yeah this is cool and
233 a must have and can we have a torrent of it please!!!11
235 assert (("sorry, no epoll, no sex",
236 ev_supported_backends () & EVBACKEND_EPOLL));
238 =item unsigned int ev_recommended_backends ()
240 Return the set of all backends compiled into this binary of libev and
241 also recommended for this platform, meaning it will work for most file
242 descriptor types. This set is often smaller than the one returned by
243 C<ev_supported_backends>, as for example kqueue is broken on most BSDs
244 and will not be auto-detected unless you explicitly request it (assuming
245 you know what you are doing). This is the set of backends that libev will
246 probe for if you specify no backends explicitly.
248 =item unsigned int ev_embeddable_backends ()
250 Returns the set of backends that are embeddable in other event loops. This
251 value is platform-specific but can include backends not available on the
252 current system. To find which embeddable backends might be supported on
253 the current system, you would need to look at C<ev_embeddable_backends ()
254 & ev_supported_backends ()>, likewise for recommended ones.
256 See the description of C<ev_embed> watchers for more info.
258 =item ev_set_allocator (void *(*cb)(void *ptr, long size) throw ())
260 Sets the allocation function to use (the prototype is similar - the
261 semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
262 used to allocate and free memory (no surprises here). If it returns zero
263 when memory needs to be allocated (C<size != 0>), the library might abort
264 or take some potentially destructive action.
266 Since some systems (at least OpenBSD and Darwin) fail to implement
267 correct C<realloc> semantics, libev will use a wrapper around the system
268 C<realloc> and C<free> functions by default.
270 You could override this function in high-availability programs to, say,
271 free some memory if it cannot allocate memory, to use a special allocator,
272 or even to sleep a while and retry until some memory is available.
274 Example: The following is the C<realloc> function that libev itself uses
275 which should work with C<realloc> and C<free> functions of all kinds and
276 is probably a good basis for your own implementation.
278 static void *
279 ev_realloc_emul (void *ptr, long size) EV_NOEXCEPT
280 {
281 if (size)
282 return realloc (ptr, size);
284 free (ptr);
285 return 0;
286 }
288 Example: Replace the libev allocator with one that waits a bit and then
289 retries.
291 static void *
292 persistent_realloc (void *ptr, size_t size)
293 {
294 if (!size)
295 {
296 free (ptr);
297 return 0;
298 }
300 for (;;)
301 {
302 void *newptr = realloc (ptr, size);
304 if (newptr)
305 return newptr;
307 sleep (60);
308 }
309 }
311 ...
312 ev_set_allocator (persistent_realloc);
314 =item ev_set_syserr_cb (void (*cb)(const char *msg) throw ())
316 Set the callback function to call on a retryable system call error (such
317 as failed select, poll, epoll_wait). The message is a printable string
318 indicating the system call or subsystem causing the problem. If this
319 callback is set, then libev will expect it to remedy the situation, no
320 matter what, when it returns. That is, libev will generally retry the
321 requested operation, or, if the condition doesn't go away, do bad stuff
322 (such as abort).
324 Example: This is basically the same thing that libev does internally, too.
326 static void
327 fatal_error (const char *msg)
328 {
329 perror (msg);
330 abort ();
331 }
333 ...
334 ev_set_syserr_cb (fatal_error);
336 =item ev_feed_signal (int signum)
338 This function can be used to "simulate" a signal receive. It is completely
339 safe to call this function at any time, from any context, including signal
340 handlers or random threads.
342 Its main use is to customise signal handling in your process, especially
343 in the presence of threads. For example, you could block signals
344 by default in all threads (and specifying C<EVFLAG_NOSIGMASK> when
345 creating any loops), and in one thread, use C<sigwait> or any other
346 mechanism to wait for signals, then "deliver" them to libev by calling
347 C<ev_feed_signal>.
349 =back
353 An event loop is described by a C<struct ev_loop *> (the C<struct> is
354 I<not> optional in this case unless libev 3 compatibility is disabled, as
355 libev 3 had an C<ev_loop> function colliding with the struct name).
357 The library knows two types of such loops, the I<default> loop, which
358 supports child process events, and dynamically created event loops which
359 do not.
361 =over 4
363 =item struct ev_loop *ev_default_loop (unsigned int flags)
365 This returns the "default" event loop object, which is what you should
366 normally use when you just need "the event loop". Event loop objects and
367 the C<flags> parameter are described in more detail in the entry for
368 C<ev_loop_new>.
370 If the default loop is already initialised then this function simply
371 returns it (and ignores the flags. If that is troubling you, check
372 C<ev_backend ()> afterwards). Otherwise it will create it with the given
373 flags, which should almost always be C<0>, unless the caller is also the
374 one calling C<ev_run> or otherwise qualifies as "the main program".
376 If you don't know what event loop to use, use the one returned from this
377 function (or via the C<EV_DEFAULT> macro).
379 Note that this function is I<not> thread-safe, so if you want to use it
380 from multiple threads, you have to employ some kind of mutex (note also
381 that this case is unlikely, as loops cannot be shared easily between
382 threads anyway).
384 The default loop is the only loop that can handle C<ev_child> watchers,
385 and to do this, it always registers a handler for C<SIGCHLD>. If this is
386 a problem for your application you can either create a dynamic loop with
387 C<ev_loop_new> which doesn't do that, or you can simply overwrite the
388 C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
390 Example: This is the most typical usage.
392 if (!ev_default_loop (0))
393 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
395 Example: Restrict libev to the select and poll backends, and do not allow
396 environment settings to be taken into account:
400 =item struct ev_loop *ev_loop_new (unsigned int flags)
402 This will create and initialise a new event loop object. If the loop
403 could not be initialised, returns false.
405 This function is thread-safe, and one common way to use libev with
406 threads is indeed to create one loop per thread, and using the default
407 loop in the "main" or "initial" thread.
409 The flags argument can be used to specify special behaviour or specific
410 backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
412 The following flags are supported:
414 =over 4
416 =item C<EVFLAG_AUTO>
418 The default flags value. Use this if you have no clue (it's the right
419 thing, believe me).
421 =item C<EVFLAG_NOENV>
423 If this flag bit is or'ed into the flag value (or the program runs setuid
424 or setgid) then libev will I<not> look at the environment variable
425 C<LIBEV_FLAGS>. Otherwise (the default), this environment variable will
426 override the flags completely if it is found in the environment. This is
427 useful to try out specific backends to test their performance, to work
428 around bugs, or to make libev threadsafe (accessing environment variables
429 cannot be done in a threadsafe way, but usually it works if no other
430 thread modifies them).
434 Instead of calling C<ev_loop_fork> manually after a fork, you can also
435 make libev check for a fork in each iteration by enabling this flag.
437 This works by calling C<getpid ()> on every iteration of the loop,
438 and thus this might slow down your event loop if you do a lot of loop
439 iterations and little real work, but is usually not noticeable (on my
440 GNU/Linux system for example, C<getpid> is actually a simple 5-insn
441 sequence without a system call and thus I<very> fast, but my GNU/Linux
442 system also has C<pthread_atfork> which is even faster). (Update: glibc
443 versions 2.25 apparently removed the C<getpid> optimisation again).
445 The big advantage of this flag is that you can forget about fork (and
446 forget about forgetting to tell libev about forking, although you still
447 have to ignore C<SIGPIPE>) when you use this flag.
449 This flag setting cannot be overridden or specified in the C<LIBEV_FLAGS>
450 environment variable.
454 When this flag is specified, then libev will not attempt to use the
455 I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
456 testing, this flag can be useful to conserve inotify file descriptors, as
457 otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
461 When this flag is specified, then libev will attempt to use the
462 I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
463 delivers signals synchronously, which makes it both faster and might make
464 it possible to get the queued signal data. It can also simplify signal
465 handling with threads, as long as you properly block signals in your
466 threads that are not interested in handling them.
468 Signalfd will not be used by default as this changes your signal mask, and
469 there are a lot of shoddy libraries and programs (glib's threadpool for
470 example) that can't properly initialise their signal masks.
474 When this flag is specified, then libev will avoid to modify the signal
475 mask. Specifically, this means you have to make sure signals are unblocked
476 when you want to receive them.
478 This behaviour is useful when you want to do your own signal handling, or
479 want to handle signals only in specific threads and want to avoid libev
480 unblocking the signals.
482 It's also required by POSIX in a threaded program, as libev calls
483 C<sigprocmask>, whose behaviour is officially unspecified.
487 When this flag is specified, the libev will avoid using a C<timerfd> to
488 detect time jumps. It will still be able to detect time jumps, but takes
489 longer and has a lower accuracy in doing so, but saves a file descriptor
490 per loop.
492 The current implementation only tries to use a C<timerfd> when the first
493 C<ev_periodic> watcher is started and falls back on other methods if it
494 cannot be created, but this behaviour might change in the future.
496 =item C<EVBACKEND_SELECT> (value 1, portable select backend)
498 This is your standard select(2) backend. Not I<completely> standard, as
499 libev tries to roll its own fd_set with no limits on the number of fds,
500 but if that fails, expect a fairly low limit on the number of fds when
501 using this backend. It doesn't scale too well (O(highest_fd)), but its
502 usually the fastest backend for a low number of (low-numbered :) fds.
504 To get good performance out of this backend you need a high amount of
505 parallelism (most of the file descriptors should be busy). If you are
506 writing a server, you should C<accept ()> in a loop to accept as many
507 connections as possible during one iteration. You might also want to have
508 a look at C<ev_set_io_collect_interval ()> to increase the amount of
509 readiness notifications you get per iteration.
511 This backend maps C<EV_READ> to the C<readfds> set and C<EV_WRITE> to the
512 C<writefds> set (and to work around Microsoft Windows bugs, also onto the
513 C<exceptfds> set on that platform).
515 =item C<EVBACKEND_POLL> (value 2, poll backend, available everywhere except on windows)
517 And this is your standard poll(2) backend. It's more complicated
518 than select, but handles sparse fds better and has no artificial
519 limit on the number of fds you can use (except it will slow down
520 considerably with a lot of inactive fds). It scales similarly to select,
521 i.e. O(total_fds). See the entry for C<EVBACKEND_SELECT>, above, for
522 performance tips.
524 This backend maps C<EV_READ> to C<POLLIN | POLLERR | POLLHUP>, and
527 =item C<EVBACKEND_EPOLL> (value 4, Linux)
529 Use the Linux-specific epoll(7) interface (for both pre- and post-2.6.9
530 kernels).
532 For few fds, this backend is a bit little slower than poll and select, but
533 it scales phenomenally better. While poll and select usually scale like
534 O(total_fds) where total_fds is the total number of fds (or the highest
535 fd), epoll scales either O(1) or O(active_fds).
537 The epoll mechanism deserves honorable mention as the most misdesigned
538 of the more advanced event mechanisms: mere annoyances include silently
539 dropping file descriptors, requiring a system call per change per file
540 descriptor (and unnecessary guessing of parameters), problems with dup,
541 returning before the timeout value, resulting in additional iterations
542 (and only giving 5ms accuracy while select on the same platform gives
543 0.1ms) and so on. The biggest issue is fork races, however - if a program
544 forks then I<both> parent and child process have to recreate the epoll
545 set, which can take considerable time (one syscall per file descriptor)
546 and is of course hard to detect.
548 Epoll is also notoriously buggy - embedding epoll fds I<should> work,
549 but of course I<doesn't>, and epoll just loves to report events for
550 totally I<different> file descriptors (even already closed ones, so
551 one cannot even remove them from the set) than registered in the set
552 (especially on SMP systems). Libev tries to counter these spurious
553 notifications by employing an additional generation counter and comparing
554 that against the events to filter out spurious ones, recreating the set
555 when required. Epoll also erroneously rounds down timeouts, but gives you
556 no way to know when and by how much, so sometimes you have to busy-wait
557 because epoll returns immediately despite a nonzero timeout. And last
558 not least, it also refuses to work with some file descriptors which work
559 perfectly fine with C<select> (files, many character devices...).
561 Epoll is truly the train wreck among event poll mechanisms, a frankenpoll,
562 cobbled together in a hurry, no thought to design or interaction with
563 others. Oh, the pain, will it ever stop...
565 While stopping, setting and starting an I/O watcher in the same iteration
566 will result in some caching, there is still a system call per such
567 incident (because the same I<file descriptor> could point to a different
568 I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
569 file descriptors might not work very well if you register events for both
570 file descriptors.
572 Best performance from this backend is achieved by not unregistering all
573 watchers for a file descriptor until it has been closed, if possible,
574 i.e. keep at least one watcher active per fd at all times. Stopping and
575 starting a watcher (without re-setting it) also usually doesn't cause
576 extra overhead. A fork can both result in spurious notifications as well
577 as in libev having to destroy and recreate the epoll object, which can
578 take considerable time and thus should be avoided.
580 All this means that, in practice, C<EVBACKEND_SELECT> can be as fast or
581 faster than epoll for maybe up to a hundred file descriptors, depending on
582 the usage. So sad.
584 While nominally embeddable in other event loops, this feature is broken in
585 a lot of kernel revisions, but probably(!) works in current versions.
587 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
590 =item C<EVBACKEND_LINUXAIO> (value 64, Linux)
592 Use the Linux-specific Linux AIO (I<not> C<< aio(7) >> but C<<
593 io_submit(2) >>) event interface available in post-4.18 kernels (but libev
594 only tries to use it in 4.19+).
596 This is another Linux train wreck of an event interface.
598 If this backend works for you (as of this writing, it was very
599 experimental), it is the best event interface available on Linux and might
600 be well worth enabling it - if it isn't available in your kernel this will
601 be detected and this backend will be skipped.
603 This backend can batch oneshot requests and supports a user-space ring
604 buffer to receive events. It also doesn't suffer from most of the design
605 problems of epoll (such as not being able to remove event sources from
606 the epoll set), and generally sounds too good to be true. Because, this
607 being the Linux kernel, of course it suffers from a whole new set of
608 limitations, forcing you to fall back to epoll, inheriting all its design
609 issues.
611 For one, it is not easily embeddable (but probably could be done using
612 an event fd at some extra overhead). It also is subject to a system wide
613 limit that can be configured in F</proc/sys/fs/aio-max-nr>. If no AIO
614 requests are left, this backend will be skipped during initialisation, and
615 will switch to epoll when the loop is active.
617 Most problematic in practice, however, is that not all file descriptors
618 work with it. For example, in Linux 5.1, TCP sockets, pipes, event fds,
619 files, F</dev/null> and many others are supported, but ttys do not work
620 properly (a known bug that the kernel developers don't care about, see
621 L<>), so this is not
622 (yet?) a generic event polling interface.
624 Overall, it seems the Linux developers just don't want it to have a
625 generic event handling mechanism other than C<select> or C<poll>.
627 To work around all these problem, the current version of libev uses its
628 epoll backend as a fallback for file descriptor types that do not work. Or
629 falls back completely to epoll if the kernel acts up.
631 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
634 =item C<EVBACKEND_KQUEUE> (value 8, most BSD clones)
636 Kqueue deserves special mention, as at the time this backend was
637 implemented, it was broken on all BSDs except NetBSD (usually it doesn't
638 work reliably with anything but sockets and pipes, except on Darwin,
639 where of course it's completely useless). Unlike epoll, however, whose
640 brokenness is by design, these kqueue bugs can be (and mostly have been)
641 fixed without API changes to existing programs. For this reason it's not
642 being "auto-detected" on all platforms unless you explicitly specify it
643 in the flags (i.e. using C<EVBACKEND_KQUEUE>) or libev was compiled on a
644 known-to-be-good (-enough) system like NetBSD.
646 You still can embed kqueue into a normal poll or select backend and use it
647 only for sockets (after having made sure that sockets work with kqueue on
648 the target platform). See C<ev_embed> watchers for more info.
650 It scales in the same way as the epoll backend, but the interface to the
651 kernel is more efficient (which says nothing about its actual speed, of
652 course). While stopping, setting and starting an I/O watcher does never
653 cause an extra system call as with C<EVBACKEND_EPOLL>, it still adds up to
654 two event changes per incident. Support for C<fork ()> is very bad (you
655 might have to leak fds on fork, but it's more sane than epoll) and it
656 drops fds silently in similarly hard-to-detect cases.
658 This backend usually performs well under most conditions.
660 While nominally embeddable in other event loops, this doesn't work
661 everywhere, so you might need to test for this. And since it is broken
662 almost everywhere, you should only use it when you have a lot of sockets
663 (for which it usually works), by embedding it into another event loop
664 (e.g. C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> (but C<poll> is of course
665 also broken on OS X)) and, did I mention it, using it only for sockets.
667 This backend maps C<EV_READ> into an C<EVFILT_READ> kevent with
668 C<NOTE_EOF>, and C<EV_WRITE> into an C<EVFILT_WRITE> kevent with
669 C<NOTE_EOF>.
671 =item C<EVBACKEND_DEVPOLL> (value 16, Solaris 8)
673 This is not implemented yet (and might never be, unless you send me an
674 implementation). According to reports, C</dev/poll> only supports sockets
675 and is not embeddable, which would limit the usefulness of this backend
676 immensely.
678 =item C<EVBACKEND_PORT> (value 32, Solaris 10)
680 This uses the Solaris 10 event port mechanism. As with everything on Solaris,
681 it's really slow, but it still scales very well (O(active_fds)).
683 While this backend scales well, it requires one system call per active
684 file descriptor per loop iteration. For small and medium numbers of file
685 descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
686 might perform better.
688 On the positive side, this backend actually performed fully to
689 specification in all tests and is fully embeddable, which is a rare feat
690 among the OS-specific backends (I vastly prefer correctness over speed
691 hacks).
693 On the negative side, the interface is I<bizarre> - so bizarre that
694 even sun itself gets it wrong in their code examples: The event polling
695 function sometimes returns events to the caller even though an error
696 occurred, but with no indication whether it has done so or not (yes, it's
697 even documented that way) - deadly for edge-triggered interfaces where you
698 absolutely have to know whether an event occurred or not because you have
699 to re-arm the watcher.
701 Fortunately libev seems to be able to work around these idiocies.
703 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
706 =item C<EVBACKEND_ALL>
708 Try all backends (even potentially broken ones that wouldn't be tried
709 with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
712 It is definitely not recommended to use this flag, use whatever
713 C<ev_recommended_backends ()> returns, or simply do not specify a backend
714 at all.
718 Not a backend at all, but a mask to select all backend bits from a
719 C<flags> value, in case you want to mask out any backends from a flags
720 value (e.g. when modifying the C<LIBEV_FLAGS> environment variable).
722 =back
724 If one or more of the backend flags are or'ed into the flags value,
725 then only these backends will be tried (in the reverse order as listed
726 here). If none are specified, all backends in C<ev_recommended_backends
727 ()> will be tried.
729 Example: Try to create a event loop that uses epoll and nothing else.
731 struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
732 if (!epoller)
733 fatal ("no epoll found here, maybe it hides under your chair");
735 Example: Use whatever libev has to offer, but make sure that kqueue is
736 used if available.
738 struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
740 Example: Similarly, on linux, you mgiht want to take advantage of the
741 linux aio backend if possible, but fall back to something else if that
742 isn't available.
744 struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_LINUXAIO);
746 =item ev_loop_destroy (loop)
748 Destroys an event loop object (frees all memory and kernel state
749 etc.). None of the active event watchers will be stopped in the normal
750 sense, so e.g. C<ev_is_active> might still return true. It is your
751 responsibility to either stop all watchers cleanly yourself I<before>
752 calling this function, or cope with the fact afterwards (which is usually
753 the easiest thing, you can just ignore the watchers and/or C<free ()> them
754 for example).
756 Note that certain global state, such as signal state (and installed signal
757 handlers), will not be freed by this function, and related watchers (such
758 as signal and child watchers) would need to be stopped manually.
760 This function is normally used on loop objects allocated by
761 C<ev_loop_new>, but it can also be used on the default loop returned by
762 C<ev_default_loop>, in which case it is not thread-safe.
764 Note that it is not advisable to call this function on the default loop
765 except in the rare occasion where you really need to free its resources.
766 If you need dynamically allocated loops it is better to use C<ev_loop_new>
767 and C<ev_loop_destroy>.
769 =item ev_loop_fork (loop)
771 This function sets a flag that causes subsequent C<ev_run> iterations
772 to reinitialise the kernel state for backends that have one. Despite
773 the name, you can call it anytime you are allowed to start or stop
774 watchers (except inside an C<ev_prepare> callback), but it makes most
775 sense after forking, in the child process. You I<must> call it (or use
776 C<EVFLAG_FORKCHECK>) in the child before resuming or calling C<ev_run>.
778 In addition, if you want to reuse a loop (via this function or
779 C<EVFLAG_FORKCHECK>), you I<also> have to ignore C<SIGPIPE>.
781 Again, you I<have> to call it on I<any> loop that you want to re-use after
782 a fork, I<even if you do not plan to use the loop in the parent>. This is
783 because some kernel interfaces *cough* I<kqueue> *cough* do funny things
784 during fork.
786 On the other hand, you only need to call this function in the child
787 process if and only if you want to use the event loop in the child. If
788 you just fork+exec or create a new loop in the child, you don't have to
789 call it at all (in fact, C<epoll> is so badly broken that it makes a
790 difference, but libev will usually detect this case on its own and do a
791 costly reset of the backend).
793 The function itself is quite fast and it's usually not a problem to call
794 it just in case after a fork.
796 Example: Automate calling C<ev_loop_fork> on the default loop when
797 using pthreads.
799 static void
800 post_fork_child (void)
801 {
802 ev_loop_fork (EV_DEFAULT);
803 }
805 ...
806 pthread_atfork (0, 0, post_fork_child);
808 =item int ev_is_default_loop (loop)
810 Returns true when the given loop is, in fact, the default loop, and false
811 otherwise.
813 =item unsigned int ev_iteration (loop)
815 Returns the current iteration count for the event loop, which is identical
816 to the number of times libev did poll for new events. It starts at C<0>
817 and happily wraps around with enough iterations.
819 This value can sometimes be useful as a generation counter of sorts (it
820 "ticks" the number of loop iterations), as it roughly corresponds with
821 C<ev_prepare> and C<ev_check> calls - and is incremented between the
822 prepare and check phases.
824 =item unsigned int ev_depth (loop)
826 Returns the number of times C<ev_run> was entered minus the number of
827 times C<ev_run> was exited normally, in other words, the recursion depth.
829 Outside C<ev_run>, this number is zero. In a callback, this number is
830 C<1>, unless C<ev_run> was invoked recursively (or from another thread),
831 in which case it is higher.
833 Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
834 throwing an exception etc.), doesn't count as "exit" - consider this
835 as a hint to avoid such ungentleman-like behaviour unless it's really
836 convenient, in which case it is fully supported.
838 =item unsigned int ev_backend (loop)
840 Returns one of the C<EVBACKEND_*> flags indicating the event backend in
841 use.
843 =item ev_tstamp ev_now (loop)
845 Returns the current "event loop time", which is the time the event loop
846 received events and started processing them. This timestamp does not
847 change as long as callbacks are being processed, and this is also the base
848 time used for relative timers. You can treat it as the timestamp of the
849 event occurring (or more correctly, libev finding out about it).
851 =item ev_now_update (loop)
853 Establishes the current time by querying the kernel, updating the time
854 returned by C<ev_now ()> in the progress. This is a costly operation and
855 is usually done automatically within C<ev_run ()>.
857 This function is rarely useful, but when some event callback runs for a
858 very long time without entering the event loop, updating libev's idea of
859 the current time is a good idea.
861 See also L</The special problem of time updates> in the C<ev_timer> section.
863 =item ev_suspend (loop)
865 =item ev_resume (loop)
867 These two functions suspend and resume an event loop, for use when the
868 loop is not used for a while and timeouts should not be processed.
870 A typical use case would be an interactive program such as a game: When
871 the user presses C<^Z> to suspend the game and resumes it an hour later it
872 would be best to handle timeouts as if no time had actually passed while
873 the program was suspended. This can be achieved by calling C<ev_suspend>
874 in your C<SIGTSTP> handler, sending yourself a C<SIGSTOP> and calling
875 C<ev_resume> directly afterwards to resume timer processing.
877 Effectively, all C<ev_timer> watchers will be delayed by the time spend
878 between C<ev_suspend> and C<ev_resume>, and all C<ev_periodic> watchers
879 will be rescheduled (that is, they will lose any events that would have
880 occurred while suspended).
882 After calling C<ev_suspend> you B<must not> call I<any> function on the
883 given loop other than C<ev_resume>, and you B<must not> call C<ev_resume>
884 without a previous call to C<ev_suspend>.
886 Calling C<ev_suspend>/C<ev_resume> has the side effect of updating the
887 event loop time (see C<ev_now_update>).
889 =item bool ev_run (loop, int flags)
891 Finally, this is it, the event handler. This function usually is called
892 after you have initialised all your watchers and you want to start
893 handling events. It will ask the operating system for any new events, call
894 the watcher callbacks, and then repeat the whole process indefinitely: This
895 is why event loops are called I<loops>.
897 If the flags argument is specified as C<0>, it will keep handling events
898 until either no event watchers are active anymore or C<ev_break> was
899 called.
901 The return value is false if there are no more active watchers (which
902 usually means "all jobs done" or "deadlock"), and true in all other cases
903 (which usually means " you should call C<ev_run> again").
905 Please note that an explicit C<ev_break> is usually better than
906 relying on all watchers to be stopped when deciding when a program has
907 finished (especially in interactive programs), but having a program
908 that automatically loops as long as it has to and no longer by virtue
909 of relying on its watchers stopping correctly, that is truly a thing of
910 beauty.
912 This function is I<mostly> exception-safe - you can break out of a
913 C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
914 exception and so on. This does not decrement the C<ev_depth> value, nor
915 will it clear any outstanding C<EVBREAK_ONE> breaks.
917 A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
918 those events and any already outstanding ones, but will not wait and
919 block your process in case there are no events and will return after one
920 iteration of the loop. This is sometimes useful to poll and handle new
921 events while doing lengthy calculations, to keep the program responsive.
923 A flags value of C<EVRUN_ONCE> will look for new events (waiting if
924 necessary) and will handle those and any already outstanding ones. It
925 will block your process until at least one new event arrives (which could
926 be an event internal to libev itself, so there is no guarantee that a
927 user-registered callback will be called), and will return after one
928 iteration of the loop.
930 This is useful if you are waiting for some external event in conjunction
931 with something not expressible using other libev watchers (i.e. "roll your
932 own C<ev_run>"). However, a pair of C<ev_prepare>/C<ev_check> watchers is
933 usually a better approach for this kind of thing.
935 Here are the gory details of what C<ev_run> does (this is for your
936 understanding, not a guarantee that things will work exactly like this in
937 future versions):
939 - Increment loop depth.
940 - Reset the ev_break status.
941 - Before the first iteration, call any pending watchers.
942 LOOP:
943 - If EVFLAG_FORKCHECK was used, check for a fork.
944 - If a fork was detected (by any means), queue and call all fork watchers.
945 - Queue and call all prepare watchers.
946 - If ev_break was called, goto FINISH.
947 - If we have been forked, detach and recreate the kernel state
948 as to not disturb the other process.
949 - Update the kernel state with all outstanding changes.
950 - Update the "event loop time" (ev_now ()).
951 - Calculate for how long to sleep or block, if at all
952 (active idle watchers, EVRUN_NOWAIT or not having
953 any active watchers at all will result in not sleeping).
954 - Sleep if the I/O and timer collect interval say so.
955 - Increment loop iteration counter.
956 - Block the process, waiting for any events.
957 - Queue all outstanding I/O (fd) events.
958 - Update the "event loop time" (ev_now ()), and do time jump adjustments.
959 - Queue all expired timers.
960 - Queue all expired periodics.
961 - Queue all idle watchers with priority higher than that of pending events.
962 - Queue all check watchers.
963 - Call all queued watchers in reverse order (i.e. check watchers first).
964 Signals and child watchers are implemented as I/O watchers, and will
965 be handled here by queueing them when their watcher gets executed.
966 - If ev_break has been called, or EVRUN_ONCE or EVRUN_NOWAIT
967 were used, or there are no active watchers, goto FINISH, otherwise
968 continue with step LOOP.
970 - Reset the ev_break status iff it was EVBREAK_ONE.
971 - Decrement the loop depth.
972 - Return.
974 Example: Queue some jobs and then loop until no events are outstanding
975 anymore.
977 ... queue jobs here, make sure they register event watchers as long
978 ... as they still have work to do (even an idle watcher will do..)
979 ev_run (my_loop, 0);
980 ... jobs done or somebody called break. yeah!
982 =item ev_break (loop, how)
984 Can be used to make a call to C<ev_run> return early (but only after it
985 has processed all outstanding events). The C<how> argument must be either
986 C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
987 C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
989 This "break state" will be cleared on the next call to C<ev_run>.
991 It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
992 which case it will have no effect.
994 =item ev_ref (loop)
996 =item ev_unref (loop)
998 Ref/unref can be used to add or remove a reference count on the event
999 loop: Every watcher keeps one reference, and as long as the reference
1000 count is nonzero, C<ev_run> will not return on its own.
1002 This is useful when you have a watcher that you never intend to
1003 unregister, but that nevertheless should not keep C<ev_run> from
1004 returning. In such a case, call C<ev_unref> after starting, and C<ev_ref>
1005 before stopping it.
1007 As an example, libev itself uses this for its internal signal pipe: It
1008 is not visible to the libev user and should not keep C<ev_run> from
1009 exiting if no event watchers registered by it are active. It is also an
1010 excellent way to do this for generic recurring timers or from within
1011 third-party libraries. Just remember to I<unref after start> and I<ref
1012 before stop> (but only if the watcher wasn't active before, or was active
1013 before, respectively. Note also that libev might stop watchers itself
1014 (e.g. non-repeating timers) in which case you have to C<ev_ref>
1015 in the callback).
1017 Example: Create a signal watcher, but keep it from keeping C<ev_run>
1018 running when nothing else is active.
1020 ev_signal exitsig;
1021 ev_signal_init (&exitsig, sig_cb, SIGINT);
1022 ev_signal_start (loop, &exitsig);
1023 ev_unref (loop);
1025 Example: For some weird reason, unregister the above signal handler again.
1027 ev_ref (loop);
1028 ev_signal_stop (loop, &exitsig);
1030 =item ev_set_io_collect_interval (loop, ev_tstamp interval)
1032 =item ev_set_timeout_collect_interval (loop, ev_tstamp interval)
1034 These advanced functions influence the time that libev will spend waiting
1035 for events. Both time intervals are by default C<0>, meaning that libev
1036 will try to invoke timer/periodic callbacks and I/O callbacks with minimum
1037 latency.
1039 Setting these to a higher value (the C<interval> I<must> be >= C<0>)
1040 allows libev to delay invocation of I/O and timer/periodic callbacks
1041 to increase efficiency of loop iterations (or to increase power-saving
1042 opportunities).
1044 The idea is that sometimes your program runs just fast enough to handle
1045 one (or very few) event(s) per loop iteration. While this makes the
1046 program responsive, it also wastes a lot of CPU time to poll for new
1047 events, especially with backends like C<select ()> which have a high
1048 overhead for the actual polling but can deliver many events at once.
1050 By setting a higher I<io collect interval> you allow libev to spend more
1051 time collecting I/O events, so you can handle more events per iteration,
1052 at the cost of increasing latency. Timeouts (both C<ev_periodic> and
1053 C<ev_timer>) will not be affected. Setting this to a non-null value will
1054 introduce an additional C<ev_sleep ()> call into most loop iterations. The
1055 sleep time ensures that libev will not poll for I/O events more often then
1056 once per this interval, on average (as long as the host time resolution is
1057 good enough).
1059 Likewise, by setting a higher I<timeout collect interval> you allow libev
1060 to spend more time collecting timeouts, at the expense of increased
1061 latency/jitter/inexactness (the watcher callback will be called
1062 later). C<ev_io> watchers will not be affected. Setting this to a non-null
1063 value will not introduce any overhead in libev.
1065 Many (busy) programs can usually benefit by setting the I/O collect
1066 interval to a value near C<0.1> or so, which is often enough for
1067 interactive servers (of course not for games), likewise for timeouts. It
1068 usually doesn't make much sense to set it to a lower value than C<0.01>,
1069 as this approaches the timing granularity of most systems. Note that if
1070 you do transactions with the outside world and you can't increase the
1071 parallelity, then this setting will limit your transaction rate (if you
1072 need to poll once per transaction and the I/O collect interval is 0.01,
1073 then you can't do more than 100 transactions per second).
1075 Setting the I<timeout collect interval> can improve the opportunity for
1076 saving power, as the program will "bundle" timer callback invocations that
1077 are "near" in time together, by delaying some, thus reducing the number of
1078 times the process sleeps and wakes up again. Another useful technique to
1079 reduce iterations/wake-ups is to use C<ev_periodic> watchers and make sure
1080 they fire on, say, one-second boundaries only.
1082 Example: we only need 0.1s timeout granularity, and we wish not to poll
1083 more often than 100 times per second:
1085 ev_set_timeout_collect_interval (EV_DEFAULT_UC_ 0.1);
1086 ev_set_io_collect_interval (EV_DEFAULT_UC_ 0.01);
1088 =item ev_invoke_pending (loop)
1090 This call will simply invoke all pending watchers while resetting their
1091 pending state. Normally, C<ev_run> does this automatically when required,
1092 but when overriding the invoke callback this call comes handy. This
1093 function can be invoked from a watcher - this can be useful for example
1094 when you want to do some lengthy calculation and want to pass further
1095 event handling to another thread (you still have to make sure only one
1096 thread executes within C<ev_invoke_pending> or C<ev_run> of course).
1098 =item int ev_pending_count (loop)
1100 Returns the number of pending watchers - zero indicates that no watchers
1101 are pending.
1103 =item ev_set_invoke_pending_cb (loop, void (*invoke_pending_cb)(EV_P))
1105 This overrides the invoke pending functionality of the loop: Instead of
1106 invoking all pending watchers when there are any, C<ev_run> will call
1107 this callback instead. This is useful, for example, when you want to
1108 invoke the actual watchers inside another context (another thread etc.).
1110 If you want to reset the callback, use C<ev_invoke_pending> as new
1111 callback.
1113 =item ev_set_loop_release_cb (loop, void (*release)(EV_P) throw (), void (*acquire)(EV_P) throw ())
1115 Sometimes you want to share the same loop between multiple threads. This
1116 can be done relatively simply by putting mutex_lock/unlock calls around
1117 each call to a libev function.
1119 However, C<ev_run> can run an indefinite time, so it is not feasible
1120 to wait for it to return. One way around this is to wake up the event
1121 loop via C<ev_break> and C<ev_async_send>, another way is to set these
1122 I<release> and I<acquire> callbacks on the loop.
1124 When set, then C<release> will be called just before the thread is
1125 suspended waiting for new events, and C<acquire> is called just
1126 afterwards.
1128 Ideally, C<release> will just call your mutex_unlock function, and
1129 C<acquire> will just call the mutex_lock function again.
1131 While event loop modifications are allowed between invocations of
1132 C<release> and C<acquire> (that's their only purpose after all), no
1133 modifications done will affect the event loop, i.e. adding watchers will
1134 have no effect on the set of file descriptors being watched, or the time
1135 waited. Use an C<ev_async> watcher to wake up C<ev_run> when you want it
1136 to take note of any changes you made.
1138 In theory, threads executing C<ev_run> will be async-cancel safe between
1139 invocations of C<release> and C<acquire>.
1141 See also the locking example in the C<THREADS> section later in this
1142 document.
1144 =item ev_set_userdata (loop, void *data)
1146 =item void *ev_userdata (loop)
1148 Set and retrieve a single C<void *> associated with a loop. When
1149 C<ev_set_userdata> has never been called, then C<ev_userdata> returns
1150 C<0>.
1152 These two functions can be used to associate arbitrary data with a loop,
1153 and are intended solely for the C<invoke_pending_cb>, C<release> and
1154 C<acquire> callbacks described above, but of course can be (ab-)used for
1155 any other purpose as well.
1157 =item ev_verify (loop)
1159 This function only does something when C<EV_VERIFY> support has been
1160 compiled in, which is the default for non-minimal builds. It tries to go
1161 through all internal structures and checks them for validity. If anything
1162 is found to be inconsistent, it will print an error message to standard
1163 error and call C<abort ()>.
1165 This can be used to catch bugs inside libev itself: under normal
1166 circumstances, this function will never abort as of course libev keeps its
1167 data structures consistent.
1169 =back
1174 In the following description, uppercase C<TYPE> in names stands for the
1175 watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
1176 watchers and C<ev_io_start> for I/O watchers.
1178 A watcher is an opaque structure that you allocate and register to record
1179 your interest in some event. To make a concrete example, imagine you want
1180 to wait for STDIN to become readable, you would create an C<ev_io> watcher
1181 for that:
1183 static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
1184 {
1185 ev_io_stop (w);
1186 ev_break (loop, EVBREAK_ALL);
1187 }
1189 struct ev_loop *loop = ev_default_loop (0);
1191 ev_io stdin_watcher;
1193 ev_init (&stdin_watcher, my_cb);
1194 ev_io_set (&stdin_watcher, STDIN_FILENO, EV_READ);
1195 ev_io_start (loop, &stdin_watcher);
1197 ev_run (loop, 0);
1199 As you can see, you are responsible for allocating the memory for your
1200 watcher structures (and it is I<usually> a bad idea to do this on the
1201 stack).
1203 Each watcher has an associated watcher structure (called C<struct ev_TYPE>
1204 or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
1206 Each watcher structure must be initialised by a call to C<ev_init (watcher
1207 *, callback)>, which expects a callback to be provided. This callback is
1208 invoked each time the event occurs (or, in the case of I/O watchers, each
1209 time the event loop detects that the file descriptor given is readable
1210 and/or writable).
1212 Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
1213 macro to configure it, with arguments specific to the watcher type. There
1214 is also a macro to combine initialisation and setting in one call: C<<
1215 ev_TYPE_init (watcher *, callback, ...) >>.
1217 To make the watcher actually watch out for events, you have to start it
1218 with a watcher-specific start function (C<< ev_TYPE_start (loop, watcher
1219 *) >>), and you can stop watching for events at any time by calling the
1220 corresponding stop function (C<< ev_TYPE_stop (loop, watcher *) >>.
1222 As long as your watcher is active (has been started but not stopped) you
1223 must not touch the values stored in it except when explicitly documented
1224 otherwise. Most specifically you must never reinitialise it or call its
1225 C<ev_TYPE_set> macro.
1227 Each and every callback receives the event loop pointer as first, the
1228 registered watcher structure as second, and a bitset of received events as
1229 third argument.
1231 The received events usually include a single bit per event type received
1232 (you can receive multiple events at the same time). The possible bit masks
1233 are:
1235 =over 4
1237 =item C<EV_READ>
1239 =item C<EV_WRITE>
1241 The file descriptor in the C<ev_io> watcher has become readable and/or
1242 writable.
1244 =item C<EV_TIMER>
1246 The C<ev_timer> watcher has timed out.
1248 =item C<EV_PERIODIC>
1250 The C<ev_periodic> watcher has timed out.
1252 =item C<EV_SIGNAL>
1254 The signal specified in the C<ev_signal> watcher has been received by a thread.
1256 =item C<EV_CHILD>
1258 The pid specified in the C<ev_child> watcher has received a status change.
1260 =item C<EV_STAT>
1262 The path specified in the C<ev_stat> watcher changed its attributes somehow.
1264 =item C<EV_IDLE>
1266 The C<ev_idle> watcher has determined that you have nothing better to do.
1268 =item C<EV_PREPARE>
1270 =item C<EV_CHECK>
1272 All C<ev_prepare> watchers are invoked just I<before> C<ev_run> starts to
1273 gather new events, and all C<ev_check> watchers are queued (not invoked)
1274 just after C<ev_run> has gathered them, but before it queues any callbacks
1275 for any received events. That means C<ev_prepare> watchers are the last
1276 watchers invoked before the event loop sleeps or polls for new events, and
1277 C<ev_check> watchers will be invoked before any other watchers of the same
1278 or lower priority within an event loop iteration.
1280 Callbacks of both watcher types can start and stop as many watchers as
1281 they want, and all of them will be taken into account (for example, a
1282 C<ev_prepare> watcher might start an idle watcher to keep C<ev_run> from
1283 blocking).
1285 =item C<EV_EMBED>
1287 The embedded event loop specified in the C<ev_embed> watcher needs attention.
1289 =item C<EV_FORK>
1291 The event loop has been resumed in the child process after fork (see
1292 C<ev_fork>).
1294 =item C<EV_CLEANUP>
1296 The event loop is about to be destroyed (see C<ev_cleanup>).
1298 =item C<EV_ASYNC>
1300 The given async watcher has been asynchronously notified (see C<ev_async>).
1302 =item C<EV_CUSTOM>
1304 Not ever sent (or otherwise used) by libev itself, but can be freely used
1305 by libev users to signal watchers (e.g. via C<ev_feed_event>).
1307 =item C<EV_ERROR>
1309 An unspecified error has occurred, the watcher has been stopped. This might
1310 happen because the watcher could not be properly started because libev
1311 ran out of memory, a file descriptor was found to be closed or any other
1312 problem. Libev considers these application bugs.
1314 You best act on it by reporting the problem and somehow coping with the
1315 watcher being stopped. Note that well-written programs should not receive
1316 an error ever, so when your watcher receives it, this usually indicates a
1317 bug in your program.
1319 Libev will usually signal a few "dummy" events together with an error, for
1320 example it might indicate that a fd is readable or writable, and if your
1321 callbacks is well-written it can just attempt the operation and cope with
1322 the error from read() or write(). This will not work in multi-threaded
1323 programs, though, as the fd could already be closed and reused for another
1324 thing, so beware.
1326 =back
1330 =over 4
1332 =item C<ev_init> (ev_TYPE *watcher, callback)
1334 This macro initialises the generic portion of a watcher. The contents
1335 of the watcher object can be arbitrary (so C<malloc> will do). Only
1336 the generic parts of the watcher are initialised, you I<need> to call
1337 the type-specific C<ev_TYPE_set> macro afterwards to initialise the
1338 type-specific parts. For each type there is also a C<ev_TYPE_init> macro
1339 which rolls both calls into one.
1341 You can reinitialise a watcher at any time as long as it has been stopped
1342 (or never started) and there are no pending events outstanding.
1344 The callback is always of type C<void (*)(struct ev_loop *loop, ev_TYPE *watcher,
1345 int revents)>.
1347 Example: Initialise an C<ev_io> watcher in two steps.
1349 ev_io w;
1350 ev_init (&w, my_cb);
1351 ev_io_set (&w, STDIN_FILENO, EV_READ);
1353 =item C<ev_TYPE_set> (ev_TYPE *watcher, [args])
1355 This macro initialises the type-specific parts of a watcher. You need to
1356 call C<ev_init> at least once before you call this macro, but you can
1357 call C<ev_TYPE_set> any number of times. You must not, however, call this
1358 macro on a watcher that is active (it can be pending, however, which is a
1359 difference to the C<ev_init> macro).
1361 Although some watcher types do not have type-specific arguments
1362 (e.g. C<ev_prepare>) you still need to call its C<set> macro.
1364 See C<ev_init>, above, for an example.
1366 =item C<ev_TYPE_init> (ev_TYPE *watcher, callback, [args])
1368 This convenience macro rolls both C<ev_init> and C<ev_TYPE_set> macro
1369 calls into a single call. This is the most convenient method to initialise
1370 a watcher. The same limitations apply, of course.
1372 Example: Initialise and set an C<ev_io> watcher in one step.
1374 ev_io_init (&w, my_cb, STDIN_FILENO, EV_READ);
1376 =item C<ev_TYPE_start> (loop, ev_TYPE *watcher)
1378 Starts (activates) the given watcher. Only active watchers will receive
1379 events. If the watcher is already active nothing will happen.
1381 Example: Start the C<ev_io> watcher that is being abused as example in this
1382 whole section.
1384 ev_io_start (EV_DEFAULT_UC, &w);
1386 =item C<ev_TYPE_stop> (loop, ev_TYPE *watcher)
1388 Stops the given watcher if active, and clears the pending status (whether
1389 the watcher was active or not).
1391 It is possible that stopped watchers are pending - for example,
1392 non-repeating timers are being stopped when they become pending - but
1393 calling C<ev_TYPE_stop> ensures that the watcher is neither active nor
1394 pending. If you want to free or reuse the memory used by the watcher it is
1395 therefore a good idea to always call its C<ev_TYPE_stop> function.
1397 =item bool ev_is_active (ev_TYPE *watcher)
1399 Returns a true value iff the watcher is active (i.e. it has been started
1400 and not yet been stopped). As long as a watcher is active you must not modify
1401 it.
1403 =item bool ev_is_pending (ev_TYPE *watcher)
1405 Returns a true value iff the watcher is pending, (i.e. it has outstanding
1406 events but its callback has not yet been invoked). As long as a watcher
1407 is pending (but not active) you must not call an init function on it (but
1408 C<ev_TYPE_set> is safe), you must not change its priority, and you must
1409 make sure the watcher is available to libev (e.g. you cannot C<free ()>
1410 it).
1412 =item callback ev_cb (ev_TYPE *watcher)
1414 Returns the callback currently set on the watcher.
1416 =item ev_set_cb (ev_TYPE *watcher, callback)
1418 Change the callback. You can change the callback at virtually any time
1419 (modulo threads).
1421 =item ev_set_priority (ev_TYPE *watcher, int priority)
1423 =item int ev_priority (ev_TYPE *watcher)
1425 Set and query the priority of the watcher. The priority is a small
1426 integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI>
1427 (default: C<-2>). Pending watchers with higher priority will be invoked
1428 before watchers with lower priority, but priority will not keep watchers
1429 from being executed (except for C<ev_idle> watchers).
1431 If you need to suppress invocation when higher priority events are pending
1432 you need to look at C<ev_idle> watchers, which provide this functionality.
1434 You I<must not> change the priority of a watcher as long as it is active or
1435 pending.
1437 Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is
1438 fine, as long as you do not mind that the priority value you query might
1439 or might not have been clamped to the valid range.
1441 The default priority used by watchers when no priority has been set is
1442 always C<0>, which is supposed to not be too high and not be too low :).
1444 See L</WATCHER PRIORITY MODELS>, below, for a more thorough treatment of
1445 priorities.
1447 =item ev_invoke (loop, ev_TYPE *watcher, int revents)
1449 Invoke the C<watcher> with the given C<loop> and C<revents>. Neither
1450 C<loop> nor C<revents> need to be valid as long as the watcher callback
1451 can deal with that fact, as both are simply passed through to the
1452 callback.
1454 =item int ev_clear_pending (loop, ev_TYPE *watcher)
1456 If the watcher is pending, this function clears its pending status and
1457 returns its C<revents> bitset (as if its callback was invoked). If the
1458 watcher isn't pending it does nothing and returns C<0>.
1460 Sometimes it can be useful to "poll" a watcher instead of waiting for its
1461 callback to be invoked, which can be accomplished with this function.
1463 =item ev_feed_event (loop, ev_TYPE *watcher, int revents)
1465 Feeds the given event set into the event loop, as if the specified event
1466 had happened for the specified watcher (which must be a pointer to an
1467 initialised but not necessarily started event watcher). Obviously you must
1468 not free the watcher as long as it has pending events.
1470 Stopping the watcher, letting libev invoke it, or calling
1471 C<ev_clear_pending> will clear the pending event, even if the watcher was
1472 not started in the first place.
1474 See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
1475 functions that do not need a watcher.
1477 =back
1482 =head2 WATCHER STATES
1484 There are various watcher states mentioned throughout this manual -
1485 active, pending and so on. In this section these states and the rules to
1486 transition between them will be described in more detail - and while these
1487 rules might look complicated, they usually do "the right thing".
1489 =over 4
1491 =item initialised
1493 Before a watcher can be registered with the event loop it has to be
1494 initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
1495 C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
1497 In this state it is simply some block of memory that is suitable for
1498 use in an event loop. It can be moved around, freed, reused etc. at
1499 will - as long as you either keep the memory contents intact, or call
1500 C<ev_TYPE_init> again.
1502 =item started/running/active
1504 Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
1505 property of the event loop, and is actively waiting for events. While in
1506 this state it cannot be accessed (except in a few documented ways), moved,
1507 freed or anything else - the only legal thing is to keep a pointer to it,
1508 and call libev functions on it that are documented to work on active watchers.
1510 =item pending
1512 If a watcher is active and libev determines that an event it is interested
1513 in has occurred (such as a timer expiring), it will become pending. It will
1514 stay in this pending state until either it is stopped or its callback is
1515 about to be invoked, so it is not normally pending inside the watcher
1516 callback.
1518 The watcher might or might not be active while it is pending (for example,
1519 an expired non-repeating timer can be pending but no longer active). If it
1520 is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
1521 but it is still property of the event loop at this time, so cannot be
1522 moved, freed or reused. And if it is active the rules described in the
1523 previous item still apply.
1525 It is also possible to feed an event on a watcher that is not active (e.g.
1526 via C<ev_feed_event>), in which case it becomes pending without being
1527 active.
1529 =item stopped
1531 A watcher can be stopped implicitly by libev (in which case it might still
1532 be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
1533 latter will clear any pending state the watcher might be in, regardless
1534 of whether it was active or not, so stopping a watcher explicitly before
1535 freeing it is often a good idea.
1537 While stopped (and not pending) the watcher is essentially in the
1538 initialised state, that is, it can be reused, moved, modified in any way
1539 you wish (but when you trash the memory block, you need to C<ev_TYPE_init>
1540 it again).
1542 =back
1546 Many event loops support I<watcher priorities>, which are usually small
1547 integers that influence the ordering of event callback invocation
1548 between watchers in some way, all else being equal.
1550 In libev, watcher priorities can be set using C<ev_set_priority>. See its
1551 description for the more technical details such as the actual priority
1552 range.
1554 There are two common ways how these these priorities are being interpreted
1555 by event loops:
1557 In the more common lock-out model, higher priorities "lock out" invocation
1558 of lower priority watchers, which means as long as higher priority
1559 watchers receive events, lower priority watchers are not being invoked.
1561 The less common only-for-ordering model uses priorities solely to order
1562 callback invocation within a single event loop iteration: Higher priority
1563 watchers are invoked before lower priority ones, but they all get invoked
1564 before polling for new events.
1566 Libev uses the second (only-for-ordering) model for all its watchers
1567 except for idle watchers (which use the lock-out model).
1569 The rationale behind this is that implementing the lock-out model for
1570 watchers is not well supported by most kernel interfaces, and most event
1571 libraries will just poll for the same events again and again as long as
1572 their callbacks have not been executed, which is very inefficient in the
1573 common case of one high-priority watcher locking out a mass of lower
1574 priority ones.
1576 Static (ordering) priorities are most useful when you have two or more
1577 watchers handling the same resource: a typical usage example is having an
1578 C<ev_io> watcher to receive data, and an associated C<ev_timer> to handle
1579 timeouts. Under load, data might be received while the program handles
1580 other jobs, but since timers normally get invoked first, the timeout
1581 handler will be executed before checking for data. In that case, giving
1582 the timer a lower priority than the I/O watcher ensures that I/O will be
1583 handled first even under adverse conditions (which is usually, but not
1584 always, what you want).
1586 Since idle watchers use the "lock-out" model, meaning that idle watchers
1587 will only be executed when no same or higher priority watchers have
1588 received events, they can be used to implement the "lock-out" model when
1589 required.
1591 For example, to emulate how many other event libraries handle priorities,
1592 you can associate an C<ev_idle> watcher to each such watcher, and in
1593 the normal watcher callback, you just start the idle watcher. The real
1594 processing is done in the idle watcher callback. This causes libev to
1595 continuously poll and process kernel event data for the watcher, but when
1596 the lock-out case is known to be rare (which in turn is rare :), this is
1597 workable.
1599 Usually, however, the lock-out model implemented that way will perform
1600 miserably under the type of load it was designed to handle. In that case,
1601 it might be preferable to stop the real watcher before starting the
1602 idle watcher, so the kernel will not have to process the event in case
1603 the actual processing will be delayed for considerable time.
1605 Here is an example of an I/O watcher that should run at a strictly lower
1606 priority than the default, and which should only process data when no
1607 other events are pending:
1609 ev_idle idle; // actual processing watcher
1610 ev_io io; // actual event watcher
1612 static void
1613 io_cb (EV_P_ ev_io *w, int revents)
1614 {
1615 // stop the I/O watcher, we received the event, but
1616 // are not yet ready to handle it.
1617 ev_io_stop (EV_A_ w);
1619 // start the idle watcher to handle the actual event.
1620 // it will not be executed as long as other watchers
1621 // with the default priority are receiving events.
1622 ev_idle_start (EV_A_ &idle);
1623 }
1625 static void
1626 idle_cb (EV_P_ ev_idle *w, int revents)
1627 {
1628 // actual processing
1629 read (STDIN_FILENO, ...);
1631 // have to start the I/O watcher again, as
1632 // we have handled the event
1633 ev_io_start (EV_P_ &io);
1634 }
1636 // initialisation
1637 ev_idle_init (&idle, idle_cb);
1638 ev_io_init (&io, io_cb, STDIN_FILENO, EV_READ);
1639 ev_io_start (EV_DEFAULT_ &io);
1641 In the "real" world, it might also be beneficial to start a timer, so that
1642 low-priority connections can not be locked out forever under load. This
1643 enables your program to keep a lower latency for important connections
1644 during short periods of high load, while not completely locking out less
1645 important ones.
1648 =head1 WATCHER TYPES
1650 This section describes each watcher in detail, but will not repeat
1651 information given in the last section. Any initialisation/set macros,
1652 functions and members specific to the watcher type are explained.
1654 Most members are additionally marked with either I<[read-only]>, meaning
1655 that, while the watcher is active, you can look at the member and expect
1656 some sensible content, but you must not modify it (you can modify it while
1657 the watcher is stopped to your hearts content), or I<[read-write]>, which
1658 means you can expect it to have some sensible content while the watcher is
1659 active, but you can also modify it (within the same thread as the event
1660 loop, i.e. without creating data races). Modifying it may not do something
1661 sensible or take immediate effect (or do anything at all), but libev will
1662 not crash or malfunction in any way.
1664 In any case, the documentation for each member will explain what the
1665 effects are, and if there are any additional access restrictions.
1667 =head2 C<ev_io> - is this file descriptor readable or writable?
1669 I/O watchers check whether a file descriptor is readable or writable
1670 in each iteration of the event loop, or, more precisely, when reading
1671 would not block the process and writing would at least be able to write
1672 some data. This behaviour is called level-triggering because you keep
1673 receiving events as long as the condition persists. Remember you can stop
1674 the watcher if you don't want to act on the event and neither want to
1675 receive future events.
1677 In general you can register as many read and/or write event watchers per
1678 fd as you want (as long as you don't confuse yourself). Setting all file
1679 descriptors to non-blocking mode is also usually a good idea (but not
1680 required if you know what you are doing).
1682 Another thing you have to watch out for is that it is quite easy to
1683 receive "spurious" readiness notifications, that is, your callback might
1684 be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1685 because there is no data. It is very easy to get into this situation even
1686 with a relatively standard program structure. Thus it is best to always
1687 use non-blocking I/O: An extra C<read>(2) returning C<EAGAIN> is far
1688 preferable to a program hanging until some data arrives.
1690 If you cannot run the fd in non-blocking mode (for example you should
1691 not play around with an Xlib connection), then you have to separately
1692 re-test whether a file descriptor is really ready with a known-to-be good
1693 interface such as poll (fortunately in the case of Xlib, it already does
1694 this on its own, so its quite safe to use). Some people additionally
1695 use C<SIGALRM> and an interval timer, just to be sure you won't block
1696 indefinitely.
1698 But really, best use non-blocking mode.
1700 =head3 The special problem of disappearing file descriptors
1702 Some backends (e.g. kqueue, epoll, linuxaio) need to be told about closing
1703 a file descriptor (either due to calling C<close> explicitly or any other
1704 means, such as C<dup2>). The reason is that you register interest in some
1705 file descriptor, but when it goes away, the operating system will silently
1706 drop this interest. If another file descriptor with the same number then
1707 is registered with libev, there is no efficient way to see that this is,
1708 in fact, a different file descriptor.
1710 To avoid having to explicitly tell libev about such cases, libev follows
1711 the following policy: Each time C<ev_io_set> is being called, libev
1712 will assume that this is potentially a new file descriptor, otherwise
1713 it is assumed that the file descriptor stays the same. That means that
1714 you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the
1715 descriptor even if the file descriptor number itself did not change.
1717 This is how one would do it normally anyway, the important point is that
1718 the libev application should not optimise around libev but should leave
1719 optimisations to libev.
1721 =head3 The special problem of dup'ed file descriptors
1723 Some backends (e.g. epoll), cannot register events for file descriptors,
1724 but only events for the underlying file descriptions. That means when you
1725 have C<dup ()>'ed file descriptors or weirder constellations, and register
1726 events for them, only one file descriptor might actually receive events.
1728 There is no workaround possible except not registering events
1729 for potentially C<dup ()>'ed file descriptors, or to resort to
1732 =head3 The special problem of files
1734 Many people try to use C<select> (or libev) on file descriptors
1735 representing files, and expect it to become ready when their program
1736 doesn't block on disk accesses (which can take a long time on their own).
1738 However, this cannot ever work in the "expected" way - you get a readiness
1739 notification as soon as the kernel knows whether and how much data is
1740 there, and in the case of open files, that's always the case, so you
1741 always get a readiness notification instantly, and your read (or possibly
1742 write) will still block on the disk I/O.
1744 Another way to view it is that in the case of sockets, pipes, character
1745 devices and so on, there is another party (the sender) that delivers data
1746 on its own, but in the case of files, there is no such thing: the disk
1747 will not send data on its own, simply because it doesn't know what you
1748 wish to read - you would first have to request some data.
1750 Since files are typically not-so-well supported by advanced notification
1751 mechanism, libev tries hard to emulate POSIX behaviour with respect
1752 to files, even though you should not use it. The reason for this is
1753 convenience: sometimes you want to watch STDIN or STDOUT, which is
1754 usually a tty, often a pipe, but also sometimes files or special devices
1755 (for example, C<epoll> on Linux works with F</dev/random> but not with
1756 F</dev/urandom>), and even though the file might better be served with
1757 asynchronous I/O instead of with non-blocking I/O, it is still useful when
1758 it "just works" instead of freezing.
1760 So avoid file descriptors pointing to files when you know it (e.g. use
1761 libeio), but use them when it is convenient, e.g. for STDIN/STDOUT, or
1762 when you rarely read from a file instead of from a socket, and want to
1763 reuse the same code path.
1765 =head3 The special problem of fork
1767 Some backends (epoll, kqueue, linuxaio, iouring) do not support C<fork ()>
1768 at all or exhibit useless behaviour. Libev fully supports fork, but needs
1769 to be told about it in the child if you want to continue to use it in the
1770 child.
1772 To support fork in your child processes, you have to call C<ev_loop_fork
1773 ()> after a fork in the child, enable C<EVFLAG_FORKCHECK>, or resort to
1776 =head3 The special problem of SIGPIPE
1778 While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1779 when writing to a pipe whose other end has been closed, your program gets
1780 sent a SIGPIPE, which, by default, aborts your program. For most programs
1781 this is sensible behaviour, for daemons, this is usually undesirable.
1783 So when you encounter spurious, unexplained daemon exits, make sure you
1784 ignore SIGPIPE (and maybe make sure you log the exit status of your daemon
1785 somewhere, as that would have given you a big clue).
1787 =head3 The special problem of accept()ing when you can't
1789 Many implementations of the POSIX C<accept> function (for example,
1790 found in post-2004 Linux) have the peculiar behaviour of not removing a
1791 connection from the pending queue in all error cases.
1793 For example, larger servers often run out of file descriptors (because
1794 of resource limits), causing C<accept> to fail with C<ENFILE> but not
1795 rejecting the connection, leading to libev signalling readiness on
1796 the next iteration again (the connection still exists after all), and
1797 typically causing the program to loop at 100% CPU usage.
1799 Unfortunately, the set of errors that cause this issue differs between
1800 operating systems, there is usually little the app can do to remedy the
1801 situation, and no known thread-safe method of removing the connection to
1802 cope with overload is known (to me).
1804 One of the easiest ways to handle this situation is to just ignore it
1805 - when the program encounters an overload, it will just loop until the
1806 situation is over. While this is a form of busy waiting, no OS offers an
1807 event-based way to handle this situation, so it's the best one can do.
1809 A better way to handle the situation is to log any errors other than
1810 C<EAGAIN> and C<EWOULDBLOCK>, making sure not to flood the log with such
1811 messages, and continue as usual, which at least gives the user an idea of
1812 what could be wrong ("raise the ulimit!"). For extra points one could stop
1813 the C<ev_io> watcher on the listening fd "for a while", which reduces CPU
1814 usage.
1816 If your program is single-threaded, then you could also keep a dummy file
1817 descriptor for overload situations (e.g. by opening F</dev/null>), and
1818 when you run into C<ENFILE> or C<EMFILE>, close it, run C<accept>,
1819 close that fd, and create a new dummy fd. This will gracefully refuse
1820 clients under typical overload conditions.
1822 The last way to handle it is to simply log the error and C<exit>, as
1823 is often done with C<malloc> failures, but this results in an easy
1824 opportunity for a DoS attack.
1826 =head3 Watcher-Specific Functions
1828 =over 4
1830 =item ev_io_init (ev_io *, callback, int fd, int events)
1832 =item ev_io_set (ev_io *, int fd, int events)
1834 Configures an C<ev_io> watcher. The C<fd> is the file descriptor to
1835 receive events for and C<events> is either C<EV_READ>, C<EV_WRITE>, both
1836 C<EV_READ | EV_WRITE> or C<0>, to express the desire to receive the given
1837 events.
1839 Note that setting the C<events> to C<0> and starting the watcher is
1840 supported, but not specially optimized - if your program sometimes happens
1841 to generate this combination this is fine, but if it is easy to avoid
1842 starting an io watcher watching for no events you should do so.
1844 =item ev_io_modify (ev_io *, int events)
1846 Similar to C<ev_io_set>, but only changes the requested events. Using this
1847 might be faster with some backends, as libev can assume that the C<fd>
1848 still refers to the same underlying file description, something it cannot
1849 do when using C<ev_io_set>.
1851 =item int fd [no-modify]
1853 The file descriptor being watched. While it can be read at any time, you
1854 must not modify this member even when the watcher is stopped - always use
1855 C<ev_io_set> for that.
1857 =item int events [no-modify]
1859 The set of events the fd is being watched for, among other flags. Remember
1860 that this is a bit set - to test for C<EV_READ>, use C<< w->events &
1861 EV_READ >>, and similarly for C<EV_WRITE>.
1863 As with C<fd>, you must not modify this member even when the watcher is
1864 stopped, always use C<ev_io_set> or C<ev_io_modify> for that.
1866 =back
1868 =head3 Examples
1870 Example: Call C<stdin_readable_cb> when STDIN_FILENO has become, well
1871 readable, but only once. Since it is likely line-buffered, you could
1872 attempt to read a whole line in the callback.
1874 static void
1875 stdin_readable_cb (struct ev_loop *loop, ev_io *w, int revents)
1876 {
1877 ev_io_stop (loop, w);
1878 .. read from stdin here (or from w->fd) and handle any I/O errors
1879 }
1881 ...
1882 struct ev_loop *loop = ev_default_init (0);
1883 ev_io stdin_readable;
1884 ev_io_init (&stdin_readable, stdin_readable_cb, STDIN_FILENO, EV_READ);
1885 ev_io_start (loop, &stdin_readable);
1886 ev_run (loop, 0);
1889 =head2 C<ev_timer> - relative and optionally repeating timeouts
1891 Timer watchers are simple relative timers that generate an event after a
1892 given time, and optionally repeating in regular intervals after that.
1894 The timers are based on real time, that is, if you register an event that
1895 times out after an hour and you reset your system clock to January last
1896 year, it will still time out after (roughly) one hour. "Roughly" because
1897 detecting time jumps is hard, and some inaccuracies are unavoidable (the
1898 monotonic clock option helps a lot here).
1900 The callback is guaranteed to be invoked only I<after> its timeout has
1901 passed (not I<at>, so on systems with very low-resolution clocks this
1902 might introduce a small delay, see "the special problem of being too
1903 early", below). If multiple timers become ready during the same loop
1904 iteration then the ones with earlier time-out values are invoked before
1905 ones of the same priority with later time-out values (but this is no
1906 longer true when a callback calls C<ev_run> recursively).
1908 =head3 Be smart about timeouts
1910 Many real-world problems involve some kind of timeout, usually for error
1911 recovery. A typical example is an HTTP request - if the other side hangs,
1912 you want to raise some error after a while.
1914 What follows are some ways to handle this problem, from obvious and
1915 inefficient to smart and efficient.
1917 In the following, a 60 second activity timeout is assumed - a timeout that
1918 gets reset to 60 seconds each time there is activity (e.g. each time some
1919 data or other life sign was received).
1921 =over 4
1923 =item 1. Use a timer and stop, reinitialise and start it on activity.
1925 This is the most obvious, but not the most simple way: In the beginning,
1926 start the watcher:
1928 ev_timer_init (timer, callback, 60., 0.);
1929 ev_timer_start (loop, timer);
1931 Then, each time there is some activity, C<ev_timer_stop> it, initialise it
1932 and start it again:
1934 ev_timer_stop (loop, timer);
1935 ev_timer_set (timer, 60., 0.);
1936 ev_timer_start (loop, timer);
1938 This is relatively simple to implement, but means that each time there is
1939 some activity, libev will first have to remove the timer from its internal
1940 data structure and then add it again. Libev tries to be fast, but it's
1941 still not a constant-time operation.
1943 =item 2. Use a timer and re-start it with C<ev_timer_again> inactivity.
1945 This is the easiest way, and involves using C<ev_timer_again> instead of
1946 C<ev_timer_start>.
1948 To implement this, configure an C<ev_timer> with a C<repeat> value
1949 of C<60> and then call C<ev_timer_again> at start and each time you
1950 successfully read or write some data. If you go into an idle state where
1951 you do not expect data to travel on the socket, you can C<ev_timer_stop>
1952 the timer, and C<ev_timer_again> will automatically restart it if need be.
1954 That means you can ignore both the C<ev_timer_start> function and the
1955 C<after> argument to C<ev_timer_set>, and only ever use the C<repeat>
1956 member and C<ev_timer_again>.
1958 At start:
1960 ev_init (timer, callback);
1961 timer->repeat = 60.;
1962 ev_timer_again (loop, timer);
1964 Each time there is some activity:
1966 ev_timer_again (loop, timer);
1968 It is even possible to change the time-out on the fly, regardless of
1969 whether the watcher is active or not:
1971 timer->repeat = 30.;
1972 ev_timer_again (loop, timer);
1974 This is slightly more efficient then stopping/starting the timer each time
1975 you want to modify its timeout value, as libev does not have to completely
1976 remove and re-insert the timer from/into its internal data structure.
1978 It is, however, even simpler than the "obvious" way to do it.
1980 =item 3. Let the timer time out, but then re-arm it as required.
1982 This method is more tricky, but usually most efficient: Most timeouts are
1983 relatively long compared to the intervals between other activity - in
1984 our example, within 60 seconds, there are usually many I/O events with
1985 associated activity resets.
1987 In this case, it would be more efficient to leave the C<ev_timer> alone,
1988 but remember the time of last activity, and check for a real timeout only
1989 within the callback:
1991 ev_tstamp timeout = 60.;
1992 ev_tstamp last_activity; // time of last activity
1993 ev_timer timer;
1995 static void
1996 callback (EV_P_ ev_timer *w, int revents)
1997 {
1998 // calculate when the timeout would happen
1999 ev_tstamp after = last_activity - ev_now (EV_A) + timeout;
2001 // if negative, it means we the timeout already occurred
2002 if (after < 0.)
2003 {
2004 // timeout occurred, take action
2005 }
2006 else
2007 {
2008 // callback was invoked, but there was some recent
2009 // activity. simply restart the timer to time out
2010 // after "after" seconds, which is the earliest time
2011 // the timeout can occur.
2012 ev_timer_set (w, after, 0.);
2013 ev_timer_start (EV_A_ w);
2014 }
2015 }
2017 To summarise the callback: first calculate in how many seconds the
2018 timeout will occur (by calculating the absolute time when it would occur,
2019 C<last_activity + timeout>, and subtracting the current time, C<ev_now
2020 (EV_A)> from that).
2022 If this value is negative, then we are already past the timeout, i.e. we
2023 timed out, and need to do whatever is needed in this case.
2025 Otherwise, we now the earliest time at which the timeout would trigger,
2026 and simply start the timer with this timeout value.
2028 In other words, each time the callback is invoked it will check whether
2029 the timeout occurred. If not, it will simply reschedule itself to check
2030 again at the earliest time it could time out. Rinse. Repeat.
2032 This scheme causes more callback invocations (about one every 60 seconds
2033 minus half the average time between activity), but virtually no calls to
2034 libev to change the timeout.
2036 To start the machinery, simply initialise the watcher and set
2037 C<last_activity> to the current time (meaning there was some activity just
2038 now), then call the callback, which will "do the right thing" and start
2039 the timer:
2041 last_activity = ev_now (EV_A);
2042 ev_init (&timer, callback);
2043 callback (EV_A_ &timer, 0);
2045 When there is some activity, simply store the current time in
2046 C<last_activity>, no libev calls at all:
2048 if (activity detected)
2049 last_activity = ev_now (EV_A);
2051 When your timeout value changes, then the timeout can be changed by simply
2052 providing a new value, stopping the timer and calling the callback, which
2053 will again do the right thing (for example, time out immediately :).
2055 timeout = new_value;
2056 ev_timer_stop (EV_A_ &timer);
2057 callback (EV_A_ &timer, 0);
2059 This technique is slightly more complex, but in most cases where the
2060 time-out is unlikely to be triggered, much more efficient.
2062 =item 4. Wee, just use a double-linked list for your timeouts.
2064 If there is not one request, but many thousands (millions...), all
2065 employing some kind of timeout with the same timeout value, then one can
2066 do even better:
2068 When starting the timeout, calculate the timeout value and put the timeout
2069 at the I<end> of the list.
2071 Then use an C<ev_timer> to fire when the timeout at the I<beginning> of
2072 the list is expected to fire (for example, using the technique #3).
2074 When there is some activity, remove the timer from the list, recalculate
2075 the timeout, append it to the end of the list again, and make sure to
2076 update the C<ev_timer> if it was taken from the beginning of the list.
2078 This way, one can manage an unlimited number of timeouts in O(1) time for
2079 starting, stopping and updating the timers, at the expense of a major
2080 complication, and having to use a constant timeout. The constant timeout
2081 ensures that the list stays sorted.
2083 =back
2085 So which method the best?
2087 Method #2 is a simple no-brain-required solution that is adequate in most
2088 situations. Method #3 requires a bit more thinking, but handles many cases
2089 better, and isn't very complicated either. In most case, choosing either
2090 one is fine, with #3 being better in typical situations.
2092 Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
2093 rather complicated, but extremely efficient, something that really pays
2094 off after the first million or so of active timers, i.e. it's usually
2095 overkill :)
2097 =head3 The special problem of being too early
2099 If you ask a timer to call your callback after three seconds, then
2100 you expect it to be invoked after three seconds - but of course, this
2101 cannot be guaranteed to infinite precision. Less obviously, it cannot be
2102 guaranteed to any precision by libev - imagine somebody suspending the
2103 process with a STOP signal for a few hours for example.
2105 So, libev tries to invoke your callback as soon as possible I<after> the
2106 delay has occurred, but cannot guarantee this.
2108 A less obvious failure mode is calling your callback too early: many event
2109 loops compare timestamps with a "elapsed delay >= requested delay", but
2110 this can cause your callback to be invoked much earlier than you would
2111 expect.
2113 To see why, imagine a system with a clock that only offers full second
2114 resolution (think windows if you can't come up with a broken enough OS
2115 yourself). If you schedule a one-second timer at the time 500.9, then the
2116 event loop will schedule your timeout to elapse at a system time of 500
2117 (500.9 truncated to the resolution) + 1, or 501.
2119 If an event library looks at the timeout 0.1s later, it will see "501 >=
2120 501" and invoke the callback 0.1s after it was started, even though a
2121 one-second delay was requested - this is being "too early", despite best
2122 intentions.
2124 This is the reason why libev will never invoke the callback if the elapsed
2125 delay equals the requested delay, but only when the elapsed delay is
2126 larger than the requested delay. In the example above, libev would only invoke
2127 the callback at system time 502, or 1.1s after the timer was started.
2129 So, while libev cannot guarantee that your callback will be invoked
2130 exactly when requested, it I<can> and I<does> guarantee that the requested
2131 delay has actually elapsed, or in other words, it always errs on the "too
2132 late" side of things.
2134 =head3 The special problem of time updates
2136 Establishing the current time is a costly operation (it usually takes
2137 at least one system call): EV therefore updates its idea of the current
2138 time only before and after C<ev_run> collects new events, which causes a
2139 growing difference between C<ev_now ()> and C<ev_time ()> when handling
2140 lots of events in one iteration.
2142 The relative timeouts are calculated relative to the C<ev_now ()>
2143 time. This is usually the right thing as this timestamp refers to the time
2144 of the event triggering whatever timeout you are modifying/starting. If
2145 you suspect event processing to be delayed and you I<need> to base the
2146 timeout on the current time, use something like the following to adjust
2147 for it:
2149 ev_timer_set (&timer, after + (ev_time () - ev_now ()), 0.);
2151 If the event loop is suspended for a long time, you can also force an
2152 update of the time returned by C<ev_now ()> by calling C<ev_now_update
2153 ()>, although that will push the event time of all outstanding events
2154 further into the future.
2156 =head3 The special problem of unsynchronised clocks
2158 Modern systems have a variety of clocks - libev itself uses the normal
2159 "wall clock" clock and, if available, the monotonic clock (to avoid time
2160 jumps).
2162 Neither of these clocks is synchronised with each other or any other clock
2163 on the system, so C<ev_time ()> might return a considerably different time
2164 than C<gettimeofday ()> or C<time ()>. On a GNU/Linux system, for example,
2165 a call to C<gettimeofday> might return a second count that is one higher
2166 than a directly following call to C<time>.
2168 The moral of this is to only compare libev-related timestamps with
2169 C<ev_time ()> and C<ev_now ()>, at least if you want better precision than
2170 a second or so.
2172 One more problem arises due to this lack of synchronisation: if libev uses
2173 the system monotonic clock and you compare timestamps from C<ev_time>
2174 or C<ev_now> from when you started your timer and when your callback is
2175 invoked, you will find that sometimes the callback is a bit "early".
2177 This is because C<ev_timer>s work in real time, not wall clock time, so
2178 libev makes sure your callback is not invoked before the delay happened,
2179 I<measured according to the real time>, not the system clock.
2181 If your timeouts are based on a physical timescale (e.g. "time out this
2182 connection after 100 seconds") then this shouldn't bother you as it is
2183 exactly the right behaviour.
2185 If you want to compare wall clock/system timestamps to your timers, then
2186 you need to use C<ev_periodic>s, as these are based on the wall clock
2187 time, where your comparisons will always generate correct results.
2189 =head3 The special problems of suspended animation
2191 When you leave the server world it is quite customary to hit machines that
2192 can suspend/hibernate - what happens to the clocks during such a suspend?
2194 Some quick tests made with a Linux 2.6.28 indicate that a suspend freezes
2195 all processes, while the clocks (C<times>, C<CLOCK_MONOTONIC>) continue
2196 to run until the system is suspended, but they will not advance while the
2197 system is suspended. That means, on resume, it will be as if the program
2198 was frozen for a few seconds, but the suspend time will not be counted
2199 towards C<ev_timer> when a monotonic clock source is used. The real time
2200 clock advanced as expected, but if it is used as sole clocksource, then a
2201 long suspend would be detected as a time jump by libev, and timers would
2202 be adjusted accordingly.
2204 I would not be surprised to see different behaviour in different between
2205 operating systems, OS versions or even different hardware.
2207 The other form of suspend (job control, or sending a SIGSTOP) will see a
2208 time jump in the monotonic clocks and the realtime clock. If the program
2209 is suspended for a very long time, and monotonic clock sources are in use,
2210 then you can expect C<ev_timer>s to expire as the full suspension time
2211 will be counted towards the timers. When no monotonic clock source is in
2212 use, then libev will again assume a timejump and adjust accordingly.
2214 It might be beneficial for this latter case to call C<ev_suspend>
2215 and C<ev_resume> in code that handles C<SIGTSTP>, to at least get
2216 deterministic behaviour in this case (you can do nothing against
2217 C<SIGSTOP>).
2219 =head3 Watcher-Specific Functions and Data Members
2221 =over 4
2223 =item ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)
2225 =item ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)
2227 Configure the timer to trigger after C<after> seconds (fractional and
2228 negative values are supported). If C<repeat> is C<0.>, then it will
2229 automatically be stopped once the timeout is reached. If it is positive,
2230 then the timer will automatically be configured to trigger again C<repeat>
2231 seconds later, again, and again, until stopped manually.
2233 The timer itself will do a best-effort at avoiding drift, that is, if
2234 you configure a timer to trigger every 10 seconds, then it will normally
2235 trigger at exactly 10 second intervals. If, however, your program cannot
2236 keep up with the timer (because it takes longer than those 10 seconds to
2237 do stuff) the timer will not fire more than once per event loop iteration.
2239 =item ev_timer_again (loop, ev_timer *)
2241 This will act as if the timer timed out, and restarts it again if it is
2242 repeating. It basically works like calling C<ev_timer_stop>, updating the
2243 timeout to the C<repeat> value and calling C<ev_timer_start>.
2245 The exact semantics are as in the following rules, all of which will be
2246 applied to the watcher:
2248 =over 4
2250 =item If the timer is pending, the pending status is always cleared.
2252 =item If the timer is started but non-repeating, stop it (as if it timed
2253 out, without invoking it).
2255 =item If the timer is repeating, make the C<repeat> value the new timeout
2256 and start the timer, if necessary.
2258 =back
2260 This sounds a bit complicated, see L</Be smart about timeouts>, above, for a
2261 usage example.
2263 =item ev_tstamp ev_timer_remaining (loop, ev_timer *)
2265 Returns the remaining time until a timer fires. If the timer is active,
2266 then this time is relative to the current event loop time, otherwise it's
2267 the timeout value currently configured.
2269 That is, after an C<ev_timer_set (w, 5, 7)>, C<ev_timer_remaining> returns
2270 C<5>. When the timer is started and one second passes, C<ev_timer_remaining>
2271 will return C<4>. When the timer expires and is restarted, it will return
2272 roughly C<7> (likely slightly less as callback invocation takes some time,
2273 too), and so on.
2275 =item ev_tstamp repeat [read-write]
2277 The current C<repeat> value. Will be used each time the watcher times out
2278 or C<ev_timer_again> is called, and determines the next timeout (if any),
2279 which is also when any modifications are taken into account.
2281 =back
2283 =head3 Examples
2285 Example: Create a timer that fires after 60 seconds.
2287 static void
2288 one_minute_cb (struct ev_loop *loop, ev_timer *w, int revents)
2289 {
2290 .. one minute over, w is actually stopped right here
2291 }
2293 ev_timer mytimer;
2294 ev_timer_init (&mytimer, one_minute_cb, 60., 0.);
2295 ev_timer_start (loop, &mytimer);
2297 Example: Create a timeout timer that times out after 10 seconds of
2298 inactivity.
2300 static void
2301 timeout_cb (struct ev_loop *loop, ev_timer *w, int revents)
2302 {
2303 .. ten seconds without any activity
2304 }
2306 ev_timer mytimer;
2307 ev_timer_init (&mytimer, timeout_cb, 0., 10.); /* note, only repeat used */
2308 ev_timer_again (&mytimer); /* start timer */
2309 ev_run (loop, 0);
2311 // and in some piece of code that gets executed on any "activity":
2312 // reset the timeout to start ticking again at 10 seconds
2313 ev_timer_again (&mytimer);
2316 =head2 C<ev_periodic> - to cron or not to cron?
2318 Periodic watchers are also timers of a kind, but they are very versatile
2319 (and unfortunately a bit complex).
2321 Unlike C<ev_timer>, periodic watchers are not based on real time (or
2322 relative time, the physical time that passes) but on wall clock time
2323 (absolute time, the thing you can read on your calendar or clock). The
2324 difference is that wall clock time can run faster or slower than real
2325 time, and time jumps are not uncommon (e.g. when you adjust your
2326 wrist-watch).
2328 You can tell a periodic watcher to trigger after some specific point
2329 in time: for example, if you tell a periodic watcher to trigger "in 10
2330 seconds" (by specifying e.g. C<ev_now () + 10.>, that is, an absolute time
2331 not a delay) and then reset your system clock to January of the previous
2332 year, then it will take a year or more to trigger the event (unlike an
2333 C<ev_timer>, which would still trigger roughly 10 seconds after starting
2334 it, as it uses a relative timeout).
2336 C<ev_periodic> watchers can also be used to implement vastly more complex
2337 timers, such as triggering an event on each "midnight, local time", or
2338 other complicated rules. This cannot easily be done with C<ev_timer>
2339 watchers, as those cannot react to time jumps.
2341 As with timers, the callback is guaranteed to be invoked only when the
2342 point in time where it is supposed to trigger has passed. If multiple
2343 timers become ready during the same loop iteration then the ones with
2344 earlier time-out values are invoked before ones with later time-out values
2345 (but this is no longer true when a callback calls C<ev_run> recursively).
2347 =head3 Watcher-Specific Functions and Data Members
2349 =over 4
2351 =item ev_periodic_init (ev_periodic *, callback, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
2353 =item ev_periodic_set (ev_periodic *, ev_tstamp offset, ev_tstamp interval, reschedule_cb)
2355 Lots of arguments, let's sort it out... There are basically three modes of
2356 operation, and we will explain them from simplest to most complex:
2358 =over 4
2360 =item * absolute timer (offset = absolute time, interval = 0, reschedule_cb = 0)
2362 In this configuration the watcher triggers an event after the wall clock
2363 time C<offset> has passed. It will not repeat and will not adjust when a
2364 time jump occurs, that is, if it is to be run at January 1st 2011 then it
2365 will be stopped and invoked when the system clock reaches or surpasses
2366 this point in time.
2368 =item * repeating interval timer (offset = offset within interval, interval > 0, reschedule_cb = 0)
2370 In this mode the watcher will always be scheduled to time out at the next
2371 C<offset + N * interval> time (for some integer N, which can also be
2372 negative) and then repeat, regardless of any time jumps. The C<offset>
2373 argument is merely an offset into the C<interval> periods.
2375 This can be used to create timers that do not drift with respect to the
2376 system clock, for example, here is an C<ev_periodic> that triggers each
2377 hour, on the hour (with respect to UTC):
2379 ev_periodic_set (&periodic, 0., 3600., 0);
2381 This doesn't mean there will always be 3600 seconds in between triggers,
2382 but only that the callback will be called when the system time shows a
2383 full hour (UTC), or more correctly, when the system time is evenly divisible
2384 by 3600.
2386 Another way to think about it (for the mathematically inclined) is that
2387 C<ev_periodic> will try to run the callback in this mode at the next possible
2388 time where C<time = offset (mod interval)>, regardless of any time jumps.
2390 The C<interval> I<MUST> be positive, and for numerical stability, the
2391 interval value should be higher than C<1/8192> (which is around 100
2392 microseconds) and C<offset> should be higher than C<0> and should have
2393 at most a similar magnitude as the current time (say, within a factor of
2394 ten). Typical values for offset are, in fact, C<0> or something between
2395 C<0> and C<interval>, which is also the recommended range.
2397 Note also that there is an upper limit to how often a timer can fire (CPU
2398 speed for example), so if C<interval> is very small then timing stability
2399 will of course deteriorate. Libev itself tries to be exact to be about one
2400 millisecond (if the OS supports it and the machine is fast enough).
2402 =item * manual reschedule mode (offset ignored, interval ignored, reschedule_cb = callback)
2404 In this mode the values for C<interval> and C<offset> are both being
2405 ignored. Instead, each time the periodic watcher gets scheduled, the
2406 reschedule callback will be called with the watcher as first, and the
2407 current time as second argument.
2409 NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, ever,
2410 or make ANY other event loop modifications whatsoever, unless explicitly
2411 allowed by documentation here>.
2413 If you need to stop it, return C<now + 1e30> (or so, fudge fudge) and stop
2414 it afterwards (e.g. by starting an C<ev_prepare> watcher, which is the
2415 only event loop modification you are allowed to do).
2417 The callback prototype is C<ev_tstamp (*reschedule_cb)(ev_periodic
2418 *w, ev_tstamp now)>, e.g.:
2420 static ev_tstamp
2421 my_rescheduler (ev_periodic *w, ev_tstamp now)
2422 {
2423 return now + 60.;
2424 }
2426 It must return the next time to trigger, based on the passed time value
2427 (that is, the lowest time value larger than to the second argument). It
2428 will usually be called just before the callback will be triggered, but
2429 might be called at other times, too.
2431 NOTE: I<< This callback must always return a time that is higher than or
2432 equal to the passed C<now> value >>.
2434 This can be used to create very complex timers, such as a timer that
2435 triggers on "next midnight, local time". To do this, you would calculate
2436 the next midnight after C<now> and return the timestamp value for
2437 this. Here is a (completely untested, no error checking) example on how to
2438 do this:
2440 #include <time.h>
2442 static ev_tstamp
2443 my_rescheduler (ev_periodic *w, ev_tstamp now)
2444 {
2445 time_t tnow = (time_t)now;
2446 struct tm tm;
2447 localtime_r (&tnow, &tm);
2449 tm.tm_sec = tm.tm_min = tm.tm_hour = 0; // midnight current day
2450 ++tm.tm_mday; // midnight next day
2452 return mktime (&tm);
2453 }
2455 Note: this code might run into trouble on days that have more then two
2456 midnights (beginning and end).
2458 =back
2460 =item ev_periodic_again (loop, ev_periodic *)
2462 Simply stops and restarts the periodic watcher again. This is only useful
2463 when you changed some parameters or the reschedule callback would return
2464 a different time than the last time it was called (e.g. in a crond like
2465 program when the crontabs have changed).
2467 =item ev_tstamp ev_periodic_at (ev_periodic *)
2469 When active, returns the absolute time that the watcher is supposed
2470 to trigger next. This is not the same as the C<offset> argument to
2471 C<ev_periodic_set>, but indeed works even in interval and manual
2472 rescheduling modes.
2474 =item ev_tstamp offset [read-write]
2476 When repeating, this contains the offset value, otherwise this is the
2477 absolute point in time (the C<offset> value passed to C<ev_periodic_set>,
2478 although libev might modify this value for better numerical stability).
2480 Can be modified any time, but changes only take effect when the periodic
2481 timer fires or C<ev_periodic_again> is being called.
2483 =item ev_tstamp interval [read-write]
2485 The current interval value. Can be modified any time, but changes only
2486 take effect when the periodic timer fires or C<ev_periodic_again> is being
2487 called.
2489 =item ev_tstamp (*reschedule_cb)(ev_periodic *w, ev_tstamp now) [read-write]
2491 The current reschedule callback, or C<0>, if this functionality is
2492 switched off. Can be changed any time, but changes only take effect when
2493 the periodic timer fires or C<ev_periodic_again> is being called.
2495 =back
2497 =head3 Examples
2499 Example: Call a callback every hour, or, more precisely, whenever the
2500 system time is divisible by 3600. The callback invocation times have
2501 potentially a lot of jitter, but good long-term stability.
2503 static void
2504 clock_cb (struct ev_loop *loop, ev_periodic *w, int revents)
2505 {
2506 ... its now a full hour (UTC, or TAI or whatever your clock follows)
2507 }
2509 ev_periodic hourly_tick;
2510 ev_periodic_init (&hourly_tick, clock_cb, 0., 3600., 0);
2511 ev_periodic_start (loop, &hourly_tick);
2513 Example: The same as above, but use a reschedule callback to do it:
2515 #include <math.h>
2517 static ev_tstamp
2518 my_scheduler_cb (ev_periodic *w, ev_tstamp now)
2519 {
2520 return now + (3600. - fmod (now, 3600.));
2521 }
2523 ev_periodic_init (&hourly_tick, clock_cb, 0., 0., my_scheduler_cb);
2525 Example: Call a callback every hour, starting now:
2527 ev_periodic hourly_tick;
2528 ev_periodic_init (&hourly_tick, clock_cb,
2529 fmod (ev_now (loop), 3600.), 3600., 0);
2530 ev_periodic_start (loop, &hourly_tick);
2533 =head2 C<ev_signal> - signal me when a signal gets signalled!
2535 Signal watchers will trigger an event when the process receives a specific
2536 signal one or more times. Even though signals are very asynchronous, libev
2537 will try its best to deliver signals synchronously, i.e. as part of the
2538 normal event processing, like any other event.
2540 If you want signals to be delivered truly asynchronously, just use
2541 C<sigaction> as you would do without libev and forget about sharing
2542 the signal. You can even use C<ev_async> from a signal handler to
2543 synchronously wake up an event loop.
2545 You can configure as many watchers as you like for the same signal, but
2546 only within the same loop, i.e. you can watch for C<SIGINT> in your
2547 default loop and for C<SIGIO> in another loop, but you cannot watch for
2548 C<SIGINT> in both the default loop and another loop at the same time. At
2549 the moment, C<SIGCHLD> is permanently tied to the default loop.
2551 Only after the first watcher for a signal is started will libev actually
2552 register something with the kernel. It thus coexists with your own signal
2553 handlers as long as you don't register any with libev for the same signal.
2555 If possible and supported, libev will install its handlers with
2556 C<SA_RESTART> (or equivalent) behaviour enabled, so system calls should
2557 not be unduly interrupted. If you have a problem with system calls getting
2558 interrupted by signals you can block all signals in an C<ev_check> watcher
2559 and unblock them in an C<ev_prepare> watcher.
2561 =head3 The special problem of inheritance over fork/execve/pthread_create
2563 Both the signal mask (C<sigprocmask>) and the signal disposition
2564 (C<sigaction>) are unspecified after starting a signal watcher (and after
2565 stopping it again), that is, libev might or might not block the signal,
2566 and might or might not set or restore the installed signal handler (but
2569 While this does not matter for the signal disposition (libev never
2570 sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
2571 C<execve>), this matters for the signal mask: many programs do not expect
2572 certain signals to be blocked.
2574 This means that before calling C<exec> (from the child) you should reset
2575 the signal mask to whatever "default" you expect (all clear is a good
2576 choice usually).
2578 The simplest way to ensure that the signal mask is reset in the child is
2579 to install a fork handler with C<pthread_atfork> that resets it. That will
2580 catch fork calls done by libraries (such as the libc) as well.
2582 In current versions of libev, the signal will not be blocked indefinitely
2583 unless you use the C<signalfd> API (C<EV_SIGNALFD>). While this reduces
2584 the window of opportunity for problems, it will not go away, as libev
2585 I<has> to modify the signal mask, at least temporarily.
2587 So I can't stress this enough: I<If you do not reset your signal mask when
2588 you expect it to be empty, you have a race condition in your code>. This
2589 is not a libev-specific thing, this is true for most event libraries.
2591 =head3 The special problem of threads signal handling
2593 POSIX threads has problematic signal handling semantics, specifically,
2594 a lot of functionality (sigfd, sigwait etc.) only really works if all
2595 threads in a process block signals, which is hard to achieve.
2597 When you want to use sigwait (or mix libev signal handling with your own
2598 for the same signals), you can tackle this problem by globally blocking
2599 all signals before creating any threads (or creating them with a fully set
2600 sigprocmask) and also specifying the C<EVFLAG_NOSIGMASK> when creating
2601 loops. Then designate one thread as "signal receiver thread" which handles
2602 these signals. You can pass on any signals that libev might be interested
2603 in by calling C<ev_feed_signal>.
2605 =head3 Watcher-Specific Functions and Data Members
2607 =over 4
2609 =item ev_signal_init (ev_signal *, callback, int signum)
2611 =item ev_signal_set (ev_signal *, int signum)
2613 Configures the watcher to trigger on the given signal number (usually one
2614 of the C<SIGxxx> constants).
2616 =item int signum [read-only]
2618 The signal the watcher watches out for.
2620 =back
2622 =head3 Examples
2624 Example: Try to exit cleanly on SIGINT.
2626 static void
2627 sigint_cb (struct ev_loop *loop, ev_signal *w, int revents)
2628 {
2629 ev_break (loop, EVBREAK_ALL);
2630 }
2632 ev_signal signal_watcher;
2633 ev_signal_init (&signal_watcher, sigint_cb, SIGINT);
2634 ev_signal_start (loop, &signal_watcher);
2637 =head2 C<ev_child> - watch out for process status changes
2639 Child watchers trigger when your process receives a SIGCHLD in response to
2640 some child status changes (most typically when a child of yours dies or
2641 exits). It is permissible to install a child watcher I<after> the child
2642 has been forked (which implies it might have already exited), as long
2643 as the event loop isn't entered (or is continued from a watcher), i.e.,
2644 forking and then immediately registering a watcher for the child is fine,
2645 but forking and registering a watcher a few event loop iterations later or
2646 in the next callback invocation is not.
2648 Only the default event loop is capable of handling signals, and therefore
2649 you can only register child watchers in the default event loop.
2651 Due to some design glitches inside libev, child watchers will always be
2652 handled at maximum priority (their priority is set to C<EV_MAXPRI> by
2653 libev)
2655 =head3 Process Interaction
2657 Libev grabs C<SIGCHLD> as soon as the default event loop is
2658 initialised. This is necessary to guarantee proper behaviour even if the
2659 first child watcher is started after the child exits. The occurrence
2660 of C<SIGCHLD> is recorded asynchronously, but child reaping is done
2661 synchronously as part of the event loop processing. Libev always reaps all
2662 children, even ones not watched.
2664 =head3 Overriding the Built-In Processing
2666 Libev offers no special support for overriding the built-in child
2667 processing, but if your application collides with libev's default child
2668 handler, you can override it easily by installing your own handler for
2669 C<SIGCHLD> after initialising the default loop, and making sure the
2670 default loop never gets destroyed. You are encouraged, however, to use an
2671 event-based approach to child reaping and thus use libev's support for
2672 that, so other libev users can use C<ev_child> watchers freely.
2674 =head3 Stopping the Child Watcher
2676 Currently, the child watcher never gets stopped, even when the
2677 child terminates, so normally one needs to stop the watcher in the
2678 callback. Future versions of libev might stop the watcher automatically
2679 when a child exit is detected (calling C<ev_child_stop> twice is not a
2680 problem).
2682 =head3 Watcher-Specific Functions and Data Members
2684 =over 4
2686 =item ev_child_init (ev_child *, callback, int pid, int trace)
2688 =item ev_child_set (ev_child *, int pid, int trace)
2690 Configures the watcher to wait for status changes of process C<pid> (or
2691 I<any> process if C<pid> is specified as C<0>). The callback can look
2692 at the C<rstatus> member of the C<ev_child> watcher structure to see
2693 the status word (use the macros from C<sys/wait.h> and see your systems
2694 C<waitpid> documentation). The C<rpid> member contains the pid of the
2695 process causing the status change. C<trace> must be either C<0> (only
2696 activate the watcher when the process terminates) or C<1> (additionally
2697 activate the watcher when the process is stopped or continued).
2699 =item int pid [read-only]
2701 The process id this watcher watches out for, or C<0>, meaning any process id.
2703 =item int rpid [read-write]
2705 The process id that detected a status change.
2707 =item int rstatus [read-write]
2709 The process exit/trace status caused by C<rpid> (see your systems
2710 C<waitpid> and C<sys/wait.h> documentation for details).
2712 =back
2714 =head3 Examples
2716 Example: C<fork()> a new process and install a child handler to wait for
2717 its completion.
2719 ev_child cw;
2721 static void
2722 child_cb (EV_P_ ev_child *w, int revents)
2723 {
2724 ev_child_stop (EV_A_ w);
2725 printf ("process %d exited with status %x\n", w->rpid, w->rstatus);
2726 }
2728 pid_t pid = fork ();
2730 if (pid < 0)
2731 // error
2732 else if (pid == 0)
2733 {
2734 // the forked child executes here
2735 exit (1);
2736 }
2737 else
2738 {
2739 ev_child_init (&cw, child_cb, pid, 0);
2740 ev_child_start (EV_DEFAULT_ &cw);
2741 }
2744 =head2 C<ev_stat> - did the file attributes just change?
2746 This watches a file system path for attribute changes. That is, it calls
2747 C<stat> on that path in regular intervals (or when the OS says it changed)
2748 and sees if it changed compared to the last time, invoking the callback
2749 if it did. Starting the watcher C<stat>'s the file, so only changes that
2750 happen after the watcher has been started will be reported.
2752 The path does not need to exist: changing from "path exists" to "path does
2753 not exist" is a status change like any other. The condition "path does not
2754 exist" (or more correctly "path cannot be stat'ed") is signified by the
2755 C<st_nlink> field being zero (which is otherwise always forced to be at
2756 least one) and all the other fields of the stat buffer having unspecified
2757 contents.
2759 The path I<must not> end in a slash or contain special components such as
2760 C<.> or C<..>. The path I<should> be absolute: If it is relative and
2761 your working directory changes, then the behaviour is undefined.
2763 Since there is no portable change notification interface available, the
2764 portable implementation simply calls C<stat(2)> regularly on the path
2765 to see if it changed somehow. You can specify a recommended polling
2766 interval for this case. If you specify a polling interval of C<0> (highly
2767 recommended!) then a I<suitable, unspecified default> value will be used
2768 (which you can expect to be around five seconds, although this might
2769 change dynamically). Libev will also impose a minimum interval which is
2770 currently around C<0.1>, but that's usually overkill.
2772 This watcher type is not meant for massive numbers of stat watchers,
2773 as even with OS-supported change notifications, this can be
2774 resource-intensive.
2776 At the time of this writing, the only OS-specific interface implemented
2777 is the Linux inotify interface (implementing kqueue support is left as an
2778 exercise for the reader. Note, however, that the author sees no way of
2779 implementing C<ev_stat> semantics with kqueue, except as a hint).
2781 =head3 ABI Issues (Largefile Support)
2783 Libev by default (unless the user overrides this) uses the default
2784 compilation environment, which means that on systems with large file
2785 support disabled by default, you get the 32 bit version of the stat
2786 structure. When using the library from programs that change the ABI to
2787 use 64 bit file offsets the programs will fail. In that case you have to
2788 compile libev with the same flags to get binary compatibility. This is
2789 obviously the case with any flags that change the ABI, but the problem is
2790 most noticeably displayed with ev_stat and large file support.
2792 The solution for this is to lobby your distribution maker to make large
2793 file interfaces available by default (as e.g. FreeBSD does) and not
2794 optional. Libev cannot simply switch on large file support because it has
2795 to exchange stat structures with application programs compiled using the
2796 default compilation environment.
2798 =head3 Inotify and Kqueue
2800 When C<inotify (7)> support has been compiled into libev and present at
2801 runtime, it will be used to speed up change detection where possible. The
2802 inotify descriptor will be created lazily when the first C<ev_stat>
2803 watcher is being started.
2805 Inotify presence does not change the semantics of C<ev_stat> watchers
2806 except that changes might be detected earlier, and in some cases, to avoid
2807 making regular C<stat> calls. Even in the presence of inotify support
2808 there are many cases where libev has to resort to regular C<stat> polling,
2809 but as long as kernel 2.6.25 or newer is used (2.6.24 and older have too
2810 many bugs), the path exists (i.e. stat succeeds), and the path resides on
2811 a local filesystem (libev currently assumes only ext2/3, jfs, reiserfs and
2812 xfs are fully working) libev usually gets away without polling.
2814 There is no support for kqueue, as apparently it cannot be used to
2815 implement this functionality, due to the requirement of having a file
2816 descriptor open on the object at all times, and detecting renames, unlinks
2817 etc. is difficult.
2819 =head3 C<stat ()> is a synchronous operation
2821 Libev doesn't normally do any kind of I/O itself, and so is not blocking
2822 the process. The exception are C<ev_stat> watchers - those call C<stat
2823 ()>, which is a synchronous operation.
2825 For local paths, this usually doesn't matter: unless the system is very
2826 busy or the intervals between stat's are large, a stat call will be fast,
2827 as the path data is usually in memory already (except when starting the
2828 watcher).
2830 For networked file systems, calling C<stat ()> can block an indefinite
2831 time due to network issues, and even under good conditions, a stat call
2832 often takes multiple milliseconds.
2834 Therefore, it is best to avoid using C<ev_stat> watchers on networked
2835 paths, although this is fully supported by libev.
2837 =head3 The special problem of stat time resolution
2839 The C<stat ()> system call only supports full-second resolution portably,
2840 and even on systems where the resolution is higher, most file systems
2841 still only support whole seconds.
2843 That means that, if the time is the only thing that changes, you can
2844 easily miss updates: on the first update, C<ev_stat> detects a change and
2845 calls your callback, which does something. When there is another update
2846 within the same second, C<ev_stat> will be unable to detect unless the
2847 stat data does change in other ways (e.g. file size).
2849 The solution to this is to delay acting on a change for slightly more
2850 than a second (or till slightly after the next full second boundary), using
2851 a roughly one-second-delay C<ev_timer> (e.g. C<ev_timer_set (w, 0., 1.02);
2852 ev_timer_again (loop, w)>).
2854 The C<.02> offset is added to work around small timing inconsistencies
2855 of some operating systems (where the second counter of the current time
2856 might be be delayed. One such system is the Linux kernel, where a call to
2857 C<gettimeofday> might return a timestamp with a full second later than
2858 a subsequent C<time> call - if the equivalent of C<time ()> is used to
2859 update file times then there will be a small window where the kernel uses
2860 the previous second to update file times but libev might already execute
2861 the timer callback).
2863 =head3 Watcher-Specific Functions and Data Members
2865 =over 4
2867 =item ev_stat_init (ev_stat *, callback, const char *path, ev_tstamp interval)
2869 =item ev_stat_set (ev_stat *, const char *path, ev_tstamp interval)
2871 Configures the watcher to wait for status changes of the given
2872 C<path>. The C<interval> is a hint on how quickly a change is expected to
2873 be detected and should normally be specified as C<0> to let libev choose
2874 a suitable value. The memory pointed to by C<path> must point to the same
2875 path for as long as the watcher is active.
2877 The callback will receive an C<EV_STAT> event when a change was detected,
2878 relative to the attributes at the time the watcher was started (or the
2879 last change was detected).
2881 =item ev_stat_stat (loop, ev_stat *)
2883 Updates the stat buffer immediately with new values. If you change the
2884 watched path in your callback, you could call this function to avoid
2885 detecting this change (while introducing a race condition if you are not
2886 the only one changing the path). Can also be useful simply to find out the
2887 new values.
2889 =item ev_statdata attr [read-only]
2891 The most-recently detected attributes of the file. Although the type is
2892 C<ev_statdata>, this is usually the (or one of the) C<struct stat> types
2893 suitable for your system, but you can only rely on the POSIX-standardised
2894 members to be present. If the C<st_nlink> member is C<0>, then there was
2895 some error while C<stat>ing the file.
2897 =item ev_statdata prev [read-only]
2899 The previous attributes of the file. The callback gets invoked whenever
2900 C<prev> != C<attr>, or, more precisely, one or more of these members
2901 differ: C<st_dev>, C<st_ino>, C<st_mode>, C<st_nlink>, C<st_uid>,
2902 C<st_gid>, C<st_rdev>, C<st_size>, C<st_atime>, C<st_mtime>, C<st_ctime>.
2904 =item ev_tstamp interval [read-only]
2906 The specified interval.
2908 =item const char *path [read-only]
2910 The file system path that is being watched.
2912 =back
2914 =head3 Examples
2916 Example: Watch C</etc/passwd> for attribute changes.
2918 static void
2919 passwd_cb (struct ev_loop *loop, ev_stat *w, int revents)
2920 {
2921 /* /etc/passwd changed in some way */
2922 if (w->attr.st_nlink)
2923 {
2924 printf ("passwd current size %ld\n", (long)w->attr.st_size);
2925 printf ("passwd current atime %ld\n", (long)w->attr.st_mtime);
2926 printf ("passwd current mtime %ld\n", (long)w->attr.st_mtime);
2927 }
2928 else
2929 /* you shalt not abuse printf for puts */
2930 puts ("wow, /etc/passwd is not there, expect problems. "
2931 "if this is windows, they already arrived\n");
2932 }
2934 ...
2935 ev_stat passwd;
2937 ev_stat_init (&passwd, passwd_cb, "/etc/passwd", 0.);
2938 ev_stat_start (loop, &passwd);
2940 Example: Like above, but additionally use a one-second delay so we do not
2941 miss updates (however, frequent updates will delay processing, too, so
2942 one might do the work both on C<ev_stat> callback invocation I<and> on
2943 C<ev_timer> callback invocation).
2945 static ev_stat passwd;
2946 static ev_timer timer;
2948 static void
2949 timer_cb (EV_P_ ev_timer *w, int revents)
2950 {
2951 ev_timer_stop (EV_A_ w);
2953 /* now it's one second after the most recent passwd change */
2954 }
2956 static void
2957 stat_cb (EV_P_ ev_stat *w, int revents)
2958 {
2959 /* reset the one-second timer */
2960 ev_timer_again (EV_A_ &timer);
2961 }
2963 ...
2964 ev_stat_init (&passwd, stat_cb, "/etc/passwd", 0.);
2965 ev_stat_start (loop, &passwd);
2966 ev_timer_init (&timer, timer_cb, 0., 1.02);
2969 =head2 C<ev_idle> - when you've got nothing better to do...
2971 Idle watchers trigger events when no other events of the same or higher
2972 priority are pending (prepare, check and other idle watchers do not count
2973 as receiving "events").
2975 That is, as long as your process is busy handling sockets or timeouts
2976 (or even signals, imagine) of the same or higher priority it will not be
2977 triggered. But when your process is idle (or only lower-priority watchers
2978 are pending), the idle watchers are being called once per event loop
2979 iteration - until stopped, that is, or your process receives more events
2980 and becomes busy again with higher priority stuff.
2982 The most noteworthy effect is that as long as any idle watchers are
2983 active, the process will not block when waiting for new events.
2985 Apart from keeping your process non-blocking (which is a useful
2986 effect on its own sometimes), idle watchers are a good place to do
2987 "pseudo-background processing", or delay processing stuff to after the
2988 event loop has handled all outstanding events.
2990 =head3 Abusing an C<ev_idle> watcher for its side-effect
2992 As long as there is at least one active idle watcher, libev will never
2993 sleep unnecessarily. Or in other words, it will loop as fast as possible.
2994 For this to work, the idle watcher doesn't need to be invoked at all - the
2995 lowest priority will do.
2997 This mode of operation can be useful together with an C<ev_check> watcher,
2998 to do something on each event loop iteration - for example to balance load
2999 between different connections.
3001 See L</Abusing an ev_check watcher for its side-effect> for a longer
3002 example.
3004 =head3 Watcher-Specific Functions and Data Members
3006 =over 4
3008 =item ev_idle_init (ev_idle *, callback)
3010 Initialises and configures the idle watcher - it has no parameters of any
3011 kind. There is a C<ev_idle_set> macro, but using it is utterly pointless,
3012 believe me.
3014 =back
3016 =head3 Examples
3018 Example: Dynamically allocate an C<ev_idle> watcher, start it, and in the
3019 callback, free it. Also, use no error checking, as usual.
3021 static void
3022 idle_cb (struct ev_loop *loop, ev_idle *w, int revents)
3023 {
3024 // stop the watcher
3025 ev_idle_stop (loop, w);
3027 // now we can free it
3028 free (w);
3030 // now do something you wanted to do when the program has
3031 // no longer anything immediate to do.
3032 }
3034 ev_idle *idle_watcher = malloc (sizeof (ev_idle));
3035 ev_idle_init (idle_watcher, idle_cb);
3036 ev_idle_start (loop, idle_watcher);
3039 =head2 C<ev_prepare> and C<ev_check> - customise your event loop!
3041 Prepare and check watchers are often (but not always) used in pairs:
3042 prepare watchers get invoked before the process blocks and check watchers
3043 afterwards.
3045 You I<must not> call C<ev_run> (or similar functions that enter the
3046 current event loop) or C<ev_loop_fork> from either C<ev_prepare> or
3047 C<ev_check> watchers. Other loops than the current one are fine,
3048 however. The rationale behind this is that you do not need to check
3049 for recursion in those watchers, i.e. the sequence will always be
3050 C<ev_prepare>, blocking, C<ev_check> so if you have one watcher of each
3051 kind they will always be called in pairs bracketing the blocking call.
3053 Their main purpose is to integrate other event mechanisms into libev and
3054 their use is somewhat advanced. They could be used, for example, to track
3055 variable changes, implement your own watchers, integrate net-snmp or a
3056 coroutine library and lots more. They are also occasionally useful if
3057 you cache some data and want to flush it before blocking (for example,
3058 in X programs you might want to do an C<XFlush ()> in an C<ev_prepare>
3059 watcher).
3061 This is done by examining in each prepare call which file descriptors
3062 need to be watched by the other library, registering C<ev_io> watchers
3063 for them and starting an C<ev_timer> watcher for any timeouts (many
3064 libraries provide exactly this functionality). Then, in the check watcher,
3065 you check for any events that occurred (by checking the pending status
3066 of all watchers and stopping them) and call back into the library. The
3067 I/O and timer callbacks will never actually be called (but must be valid
3068 nevertheless, because you never know, you know?).
3070 As another example, the Perl Coro module uses these hooks to integrate
3071 coroutines into libev programs, by yielding to other active coroutines
3072 during each prepare and only letting the process block if no coroutines
3073 are ready to run (it's actually more complicated: it only runs coroutines
3074 with priority higher than or equal to the event loop and one coroutine
3075 of lower priority, but only once, using idle watchers to keep the event
3076 loop from blocking if lower-priority coroutines are active, thus mapping
3077 low-priority coroutines to idle/background tasks).
3079 When used for this purpose, it is recommended to give C<ev_check> watchers
3080 highest (C<EV_MAXPRI>) priority, to ensure that they are being run before
3081 any other watchers after the poll (this doesn't matter for C<ev_prepare>
3082 watchers).
3084 Also, C<ev_check> watchers (and C<ev_prepare> watchers, too) should not
3085 activate ("feed") events into libev. While libev fully supports this, they
3086 might get executed before other C<ev_check> watchers did their job. As
3087 C<ev_check> watchers are often used to embed other (non-libev) event
3088 loops those other event loops might be in an unusable state until their
3089 C<ev_check> watcher ran (always remind yourself to coexist peacefully with
3090 others).
3092 =head3 Abusing an C<ev_check> watcher for its side-effect
3094 C<ev_check> (and less often also C<ev_prepare>) watchers can also be
3095 useful because they are called once per event loop iteration. For
3096 example, if you want to handle a large number of connections fairly, you
3097 normally only do a bit of work for each active connection, and if there
3098 is more work to do, you wait for the next event loop iteration, so other
3099 connections have a chance of making progress.
3101 Using an C<ev_check> watcher is almost enough: it will be called on the
3102 next event loop iteration. However, that isn't as soon as possible -
3103 without external events, your C<ev_check> watcher will not be invoked.
3105 This is where C<ev_idle> watchers come in handy - all you need is a
3106 single global idle watcher that is active as long as you have one active
3107 C<ev_check> watcher. The C<ev_idle> watcher makes sure the event loop
3108 will not sleep, and the C<ev_check> watcher makes sure a callback gets
3109 invoked. Neither watcher alone can do that.
3111 =head3 Watcher-Specific Functions and Data Members
3113 =over 4
3115 =item ev_prepare_init (ev_prepare *, callback)
3117 =item ev_check_init (ev_check *, callback)
3119 Initialises and configures the prepare or check watcher - they have no
3120 parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set>
3121 macros, but using them is utterly, utterly, utterly and completely
3122 pointless.
3124 =back
3126 =head3 Examples
3128 There are a number of principal ways to embed other event loops or modules
3129 into libev. Here are some ideas on how to include libadns into libev
3130 (there is a Perl module named C<EV::ADNS> that does this, which you could
3131 use as a working example. Another Perl module named C<EV::Glib> embeds a
3132 Glib main context into libev, and finally, C<Glib::EV> embeds EV into the
3133 Glib event loop).
3135 Method 1: Add IO watchers and a timeout watcher in a prepare handler,
3136 and in a check watcher, destroy them and call into libadns. What follows
3137 is pseudo-code only of course. This requires you to either use a low
3138 priority for the check watcher or use C<ev_clear_pending> explicitly, as
3139 the callbacks for the IO/timeout watchers might not have been called yet.
3141 static ev_io iow [nfd];
3142 static ev_timer tw;
3144 static void
3145 io_cb (struct ev_loop *loop, ev_io *w, int revents)
3146 {
3147 }
3149 // create io watchers for each fd and a timer before blocking
3150 static void
3151 adns_prepare_cb (struct ev_loop *loop, ev_prepare *w, int revents)
3152 {
3153 int timeout = 3600000;
3154 struct pollfd fds [nfd];
3155 // actual code will need to loop here and realloc etc.
3156 adns_beforepoll (ads, fds, &nfd, &timeout, timeval_from (ev_time ()));
3158 /* the callback is illegal, but won't be called as we stop during check */
3159 ev_timer_init (&tw, 0, timeout * 1e-3, 0.);
3160 ev_timer_start (loop, &tw);
3162 // create one ev_io per pollfd
3163 for (int i = 0; i < nfd; ++i)
3164 {
3165 ev_io_init (iow + i, io_cb, fds [i].fd,
3166 ((fds [i].events & POLLIN ? EV_READ : 0)
3167 | (fds [i].events & POLLOUT ? EV_WRITE : 0)));
3169 fds [i].revents = 0;
3170 ev_io_start (loop, iow + i);
3171 }
3172 }
3174 // stop all watchers after blocking
3175 static void
3176 adns_check_cb (struct ev_loop *loop, ev_check *w, int revents)
3177 {
3178 ev_timer_stop (loop, &tw);
3180 for (int i = 0; i < nfd; ++i)
3181 {
3182 // set the relevant poll flags
3183 // could also call adns_processreadable etc. here
3184 struct pollfd *fd = fds + i;
3185 int revents = ev_clear_pending (iow + i);
3186 if (revents & EV_READ ) fd->revents |= fd->events & POLLIN;
3187 if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT;
3189 // now stop the watcher
3190 ev_io_stop (loop, iow + i);
3191 }
3193 adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop));
3194 }
3196 Method 2: This would be just like method 1, but you run C<adns_afterpoll>
3197 in the prepare watcher and would dispose of the check watcher.
3199 Method 3: If the module to be embedded supports explicit event
3200 notification (libadns does), you can also make use of the actual watcher
3201 callbacks, and only destroy/create the watchers in the prepare watcher.
3203 static void
3204 timer_cb (EV_P_ ev_timer *w, int revents)
3205 {
3206 adns_state ads = (adns_state)w->data;
3207 update_now (EV_A);
3209 adns_processtimeouts (ads, &tv_now);
3210 }
3212 static void
3213 io_cb (EV_P_ ev_io *w, int revents)
3214 {
3215 adns_state ads = (adns_state)w->data;
3216 update_now (EV_A);
3218 if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now);
3219 if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now);
3220 }
3222 // do not ever call adns_afterpoll
3224 Method 4: Do not use a prepare or check watcher because the module you
3225 want to embed is not flexible enough to support it. Instead, you can
3226 override their poll function. The drawback with this solution is that the
3227 main loop is now no longer controllable by EV. The C<Glib::EV> module uses
3228 this approach, effectively embedding EV as a client into the horrible
3229 libglib event loop.
3231 static gint
3232 event_poll_func (GPollFD *fds, guint nfds, gint timeout)
3233 {
3234 int got_events = 0;
3236 for (n = 0; n < nfds; ++n)
3237 // create/start io watcher that sets the relevant bits in fds[n] and increment got_events
3239 if (timeout >= 0)
3240 // create/start timer
3242 // poll
3243 ev_run (EV_A_ 0);
3245 // stop timer again
3246 if (timeout >= 0)
3247 ev_timer_stop (EV_A_ &to);
3249 // stop io watchers again - their callbacks should have set
3250 for (n = 0; n < nfds; ++n)
3251 ev_io_stop (EV_A_ iow [n]);
3253 return got_events;
3254 }
3257 =head2 C<ev_embed> - when one backend isn't enough...
3259 This is a rather advanced watcher type that lets you embed one event loop
3260 into another (currently only C<ev_io> events are supported in the embedded
3261 loop, other types of watchers might be handled in a delayed or incorrect
3262 fashion and must not be used).
3264 There are primarily two reasons you would want that: work around bugs and
3265 prioritise I/O.
3267 As an example for a bug workaround, the kqueue backend might only support
3268 sockets on some platform, so it is unusable as generic backend, but you
3269 still want to make use of it because you have many sockets and it scales
3270 so nicely. In this case, you would create a kqueue-based loop and embed
3271 it into your default loop (which might use e.g. poll). Overall operation
3272 will be a bit slower because first libev has to call C<poll> and then
3273 C<kevent>, but at least you can use both mechanisms for what they are
3274 best: C<kqueue> for scalable sockets and C<poll> if you want it to work :)
3276 As for prioritising I/O: under rare circumstances you have the case where
3277 some fds have to be watched and handled very quickly (with low latency),
3278 and even priorities and idle watchers might have too much overhead. In
3279 this case you would put all the high priority stuff in one loop and all
3280 the rest in a second one, and embed the second one in the first.
3282 As long as the watcher is active, the callback will be invoked every
3283 time there might be events pending in the embedded loop. The callback
3284 must then call C<ev_embed_sweep (mainloop, watcher)> to make a single
3285 sweep and invoke their callbacks (the callback doesn't need to invoke the
3286 C<ev_embed_sweep> function directly, it could also start an idle watcher
3287 to give the embedded loop strictly lower priority for example).
3289 You can also set the callback to C<0>, in which case the embed watcher
3290 will automatically execute the embedded loop sweep whenever necessary.
3292 Fork detection will be handled transparently while the C<ev_embed> watcher
3293 is active, i.e., the embedded loop will automatically be forked when the
3294 embedding loop forks. In other cases, the user is responsible for calling
3295 C<ev_loop_fork> on the embedded loop.
3297 Unfortunately, not all backends are embeddable: only the ones returned by
3298 C<ev_embeddable_backends> are, which, unfortunately, does not include any
3299 portable one.
3301 So when you want to use this feature you will always have to be prepared
3302 that you cannot get an embeddable loop. The recommended way to get around
3303 this is to have a separate variables for your embeddable loop, try to
3304 create it, and if that fails, use the normal loop for everything.
3306 =head3 C<ev_embed> and fork
3308 While the C<ev_embed> watcher is running, forks in the embedding loop will
3309 automatically be applied to the embedded loop as well, so no special
3310 fork handling is required in that case. When the watcher is not running,
3311 however, it is still the task of the libev user to call C<ev_loop_fork ()>
3312 as applicable.
3314 =head3 Watcher-Specific Functions and Data Members
3316 =over 4
3318 =item ev_embed_init (ev_embed *, callback, struct ev_loop *embedded_loop)
3320 =item ev_embed_set (ev_embed *, struct ev_loop *embedded_loop)
3322 Configures the watcher to embed the given loop, which must be
3323 embeddable. If the callback is C<0>, then C<ev_embed_sweep> will be
3324 invoked automatically, otherwise it is the responsibility of the callback
3325 to invoke it (it will continue to be called until the sweep has been done,
3326 if you do not want that, you need to temporarily stop the embed watcher).
3328 =item ev_embed_sweep (loop, ev_embed *)
3330 Make a single, non-blocking sweep over the embedded loop. This works
3331 similarly to C<ev_run (embedded_loop, EVRUN_NOWAIT)>, but in the most
3332 appropriate way for embedded loops.
3334 =item struct ev_loop *other [read-only]
3336 The embedded event loop.
3338 =back
3340 =head3 Examples
3342 Example: Try to get an embeddable event loop and embed it into the default
3343 event loop. If that is not possible, use the default loop. The default
3344 loop is stored in C<loop_hi>, while the embeddable loop is stored in
3345 C<loop_lo> (which is C<loop_hi> in the case no embeddable loop can be
3346 used).
3348 struct ev_loop *loop_hi = ev_default_init (0);
3349 struct ev_loop *loop_lo = 0;
3350 ev_embed embed;
3352 // see if there is a chance of getting one that works
3353 // (remember that a flags value of 0 means autodetection)
3354 loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
3355 ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
3356 : 0;
3358 // if we got one, then embed it, otherwise default to loop_hi
3359 if (loop_lo)
3360 {
3361 ev_embed_init (&embed, 0, loop_lo);
3362 ev_embed_start (loop_hi, &embed);
3363 }
3364 else
3365 loop_lo = loop_hi;
3367 Example: Check if kqueue is available but not recommended and create
3368 a kqueue backend for use with sockets (which usually work with any
3369 kqueue implementation). Store the kqueue/socket-only event loop in
3370 C<loop_socket>. (One might optionally use C<EVFLAG_NOENV>, too).
3372 struct ev_loop *loop = ev_default_init (0);
3373 struct ev_loop *loop_socket = 0;
3374 ev_embed embed;
3376 if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
3377 if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
3378 {
3379 ev_embed_init (&embed, 0, loop_socket);
3380 ev_embed_start (loop, &embed);
3381 }
3383 if (!loop_socket)
3384 loop_socket = loop;
3386 // now use loop_socket for all sockets, and loop for everything else
3389 =head2 C<ev_fork> - the audacity to resume the event loop after a fork
3391 Fork watchers are called when a C<fork ()> was detected (usually because
3392 whoever is a good citizen cared to tell libev about it by calling
3393 C<ev_loop_fork>). The invocation is done before the event loop blocks next
3394 and before C<ev_check> watchers are being called, and only in the child
3395 after the fork. If whoever good citizen calling C<ev_default_fork> cheats
3396 and calls it in the wrong process, the fork handlers will be invoked, too,
3397 of course.
3399 =head3 The special problem of life after fork - how is it possible?
3401 Most uses of C<fork ()> consist of forking, then some simple calls to set
3402 up/change the process environment, followed by a call to C<exec()>. This
3403 sequence should be handled by libev without any problems.
3405 This changes when the application actually wants to do event handling
3406 in the child, or both parent in child, in effect "continuing" after the
3407 fork.
3409 The default mode of operation (for libev, with application help to detect
3410 forks) is to duplicate all the state in the child, as would be expected
3411 when I<either> the parent I<or> the child process continues.
3413 When both processes want to continue using libev, then this is usually the
3414 wrong result. In that case, usually one process (typically the parent) is
3415 supposed to continue with all watchers in place as before, while the other
3416 process typically wants to start fresh, i.e. without any active watchers.
3418 The cleanest and most efficient way to achieve that with libev is to
3419 simply create a new event loop, which of course will be "empty", and
3420 use that for new watchers. This has the advantage of not touching more
3421 memory than necessary, and thus avoiding the copy-on-write, and the
3422 disadvantage of having to use multiple event loops (which do not support
3423 signal watchers).
3425 When this is not possible, or you want to use the default loop for
3426 other reasons, then in the process that wants to start "fresh", call
3427 C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
3428 Destroying the default loop will "orphan" (not stop) all registered
3429 watchers, so you have to be careful not to execute code that modifies
3430 those watchers. Note also that in that case, you have to re-register any
3431 signal watchers.
3433 =head3 Watcher-Specific Functions and Data Members
3435 =over 4
3437 =item ev_fork_init (ev_fork *, callback)
3439 Initialises and configures the fork watcher - it has no parameters of any
3440 kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
3441 really.
3443 =back
3446 =head2 C<ev_cleanup> - even the best things end
3448 Cleanup watchers are called just before the event loop is being destroyed
3449 by a call to C<ev_loop_destroy>.
3451 While there is no guarantee that the event loop gets destroyed, cleanup
3452 watchers provide a convenient method to install cleanup hooks for your
3453 program, worker threads and so on - you just to make sure to destroy the
3454 loop when you want them to be invoked.
3456 Cleanup watchers are invoked in the same way as any other watcher. Unlike
3457 all other watchers, they do not keep a reference to the event loop (which
3458 makes a lot of sense if you think about it). Like all other watchers, you
3459 can call libev functions in the callback, except C<ev_cleanup_start>.
3461 =head3 Watcher-Specific Functions and Data Members
3463 =over 4
3465 =item ev_cleanup_init (ev_cleanup *, callback)
3467 Initialises and configures the cleanup watcher - it has no parameters of
3468 any kind. There is a C<ev_cleanup_set> macro, but using it is utterly
3469 pointless, I assure you.
3471 =back
3473 Example: Register an atexit handler to destroy the default loop, so any
3474 cleanup functions are called.
3476 static void
3477 program_exits (void)
3478 {
3479 ev_loop_destroy (EV_DEFAULT_UC);
3480 }
3482 ...
3483 atexit (program_exits);
3486 =head2 C<ev_async> - how to wake up an event loop
3488 In general, you cannot use an C<ev_loop> from multiple threads or other
3489 asynchronous sources such as signal handlers (as opposed to multiple event
3490 loops - those are of course safe to use in different threads).
3492 Sometimes, however, you need to wake up an event loop you do not control,
3493 for example because it belongs to another thread. This is what C<ev_async>
3494 watchers do: as long as the C<ev_async> watcher is active, you can signal
3495 it by calling C<ev_async_send>, which is thread- and signal safe.
3497 This functionality is very similar to C<ev_signal> watchers, as signals,
3498 too, are asynchronous in nature, and signals, too, will be compressed
3499 (i.e. the number of callback invocations may be less than the number of
3500 C<ev_async_send> calls). In fact, you could use signal watchers as a kind
3501 of "global async watchers" by using a watcher on an otherwise unused
3502 signal, and C<ev_feed_signal> to signal this watcher from another thread,
3503 even without knowing which loop owns the signal.
3505 =head3 Queueing
3507 C<ev_async> does not support queueing of data in any way. The reason
3508 is that the author does not know of a simple (or any) algorithm for a
3509 multiple-writer-single-reader queue that works in all cases and doesn't
3510 need elaborate support such as pthreads or unportable memory access
3511 semantics.
3513 That means that if you want to queue data, you have to provide your own
3514 queue. But at least I can tell you how to implement locking around your
3515 queue:
3517 =over 4
3519 =item queueing from a signal handler context
3521 To implement race-free queueing, you simply add to the queue in the signal
3522 handler but you block the signal handler in the watcher callback. Here is
3523 an example that does that for some fictitious SIGUSR1 handler:
3525 static ev_async mysig;
3527 static void
3528 sigusr1_handler (void)
3529 {
3530 sometype data;
3532 // no locking etc.
3533 queue_put (data);
3534 ev_async_send (EV_DEFAULT_ &mysig);
3535 }
3537 static void
3538 mysig_cb (EV_P_ ev_async *w, int revents)
3539 {
3540 sometype data;
3541 sigset_t block, prev;
3543 sigemptyset (&block);
3544 sigaddset (&block, SIGUSR1);
3545 sigprocmask (SIG_BLOCK, &block, &prev);
3547 while (queue_get (&data))
3548 process (data);
3550 if (sigismember (&prev, SIGUSR1)
3551 sigprocmask (SIG_UNBLOCK, &block, 0);
3552 }
3554 (Note: pthreads in theory requires you to use C<pthread_setmask>
3555 instead of C<sigprocmask> when you use threads, but libev doesn't do it
3556 either...).
3558 =item queueing from a thread context
3560 The strategy for threads is different, as you cannot (easily) block
3561 threads but you can easily preempt them, so to queue safely you need to
3562 employ a traditional mutex lock, such as in this pthread example:
3564 static ev_async mysig;
3565 static pthread_mutex_t mymutex = PTHREAD_MUTEX_INITIALIZER;
3567 static void
3568 otherthread (void)
3569 {
3570 // only need to lock the actual queueing operation
3571 pthread_mutex_lock (&mymutex);
3572 queue_put (data);
3573 pthread_mutex_unlock (&mymutex);
3575 ev_async_send (EV_DEFAULT_ &mysig);
3576 }
3578 static void
3579 mysig_cb (EV_P_ ev_async *w, int revents)
3580 {
3581 pthread_mutex_lock (&mymutex);
3583 while (queue_get (&data))
3584 process (data);
3586 pthread_mutex_unlock (&mymutex);
3587 }
3589 =back
3592 =head3 Watcher-Specific Functions and Data Members
3594 =over 4
3596 =item ev_async_init (ev_async *, callback)
3598 Initialises and configures the async watcher - it has no parameters of any
3599 kind. There is a C<ev_async_set> macro, but using it is utterly pointless,
3600 trust me.
3602 =item ev_async_send (loop, ev_async *)
3604 Sends/signals/activates the given C<ev_async> watcher, that is, feeds
3605 an C<EV_ASYNC> event on the watcher into the event loop, and instantly
3606 returns.
3608 Unlike C<ev_feed_event>, this call is safe to do from other threads,
3609 signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
3610 embedding section below on what exactly this means).
3612 Note that, as with other watchers in libev, multiple events might get
3613 compressed into a single callback invocation (another way to look at
3614 this is that C<ev_async> watchers are level-triggered: they are set on
3615 C<ev_async_send>, reset when the event loop detects that).
3617 This call incurs the overhead of at most one extra system call per event
3618 loop iteration, if the event loop is blocked, and no syscall at all if
3619 the event loop (or your program) is processing events. That means that
3620 repeated calls are basically free (there is no need to avoid calls for
3621 performance reasons) and that the overhead becomes smaller (typically
3622 zero) under load.
3624 =item bool = ev_async_pending (ev_async *)
3626 Returns a non-zero value when C<ev_async_send> has been called on the
3627 watcher but the event has not yet been processed (or even noted) by the
3628 event loop.
3630 C<ev_async_send> sets a flag in the watcher and wakes up the loop. When
3631 the loop iterates next and checks for the watcher to have become active,
3632 it will reset the flag again. C<ev_async_pending> can be used to very
3633 quickly check whether invoking the loop might be a good idea.
3635 Not that this does I<not> check whether the watcher itself is pending,
3636 only whether it has been requested to make this watcher pending: there
3637 is a time window between the event loop checking and resetting the async
3638 notification, and the callback being invoked.
3640 =back
3645 There are some other functions of possible interest. Described. Here. Now.
3647 =over 4
3649 =item ev_once (loop, int fd, int events, ev_tstamp timeout, callback, arg)
3651 This function combines a simple timer and an I/O watcher, calls your
3652 callback on whichever event happens first and automatically stops both
3653 watchers. This is useful if you want to wait for a single event on an fd
3654 or timeout without having to allocate/configure/start/stop/free one or
3655 more watchers yourself.
3657 If C<fd> is less than 0, then no I/O watcher will be started and the
3658 C<events> argument is being ignored. Otherwise, an C<ev_io> watcher for
3659 the given C<fd> and C<events> set will be created and started.
3661 If C<timeout> is less than 0, then no timeout watcher will be
3662 started. Otherwise an C<ev_timer> watcher with after = C<timeout> (and
3663 repeat = 0) will be started. C<0> is a valid timeout.
3665 The callback has the type C<void (*cb)(int revents, void *arg)> and is
3666 passed an C<revents> set like normal event callbacks (a combination of
3667 C<EV_ERROR>, C<EV_READ>, C<EV_WRITE> or C<EV_TIMER>) and the C<arg>
3668 value passed to C<ev_once>. Note that it is possible to receive I<both>
3669 a timeout and an io event at the same time - you probably should give io
3670 events precedence.
3672 Example: wait up to ten seconds for data to appear on STDIN_FILENO.
3674 static void stdin_ready (int revents, void *arg)
3675 {
3676 if (revents & EV_READ)
3677 /* stdin might have data for us, joy! */;
3678 else if (revents & EV_TIMER)
3679 /* doh, nothing entered */;
3680 }
3682 ev_once (STDIN_FILENO, EV_READ, 10., stdin_ready, 0);
3684 =item ev_feed_fd_event (loop, int fd, int revents)
3686 Feed an event on the given fd, as if a file descriptor backend detected
3687 the given events.
3689 =item ev_feed_signal_event (loop, int signum)
3691 Feed an event as if the given signal occurred. See also C<ev_feed_signal>,
3692 which is async-safe.
3694 =back
3699 This section explains some common idioms that are not immediately
3700 obvious. Note that examples are sprinkled over the whole manual, and this
3701 section only contains stuff that wouldn't fit anywhere else.
3705 Each watcher has, by default, a C<void *data> member that you can read
3706 or modify at any time: libev will completely ignore it. This can be used
3707 to associate arbitrary data with your watcher. If you need more data and
3708 don't want to allocate memory separately and store a pointer to it in that
3709 data member, you can also "subclass" the watcher type and provide your own
3710 data:
3712 struct my_io
3713 {
3714 ev_io io;
3715 int otherfd;
3716 void *somedata;
3717 struct whatever *mostinteresting;
3718 };
3720 ...
3721 struct my_io w;
3722 ev_io_init (&, my_cb, fd, EV_READ);
3724 And since your callback will be called with a pointer to the watcher, you
3725 can cast it back to your own type:
3727 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
3728 {
3729 struct my_io *w = (struct my_io *)w_;
3730 ...
3731 }
3733 More interesting and less C-conformant ways of casting your callback
3734 function type instead have been omitted.
3738 Another common scenario is to use some data structure with multiple
3739 embedded watchers, in effect creating your own watcher that combines
3740 multiple libev event sources into one "super-watcher":
3742 struct my_biggy
3743 {
3744 int some_data;
3745 ev_timer t1;
3746 ev_timer t2;
3747 }
3749 In this case getting the pointer to C<my_biggy> is a bit more
3750 complicated: Either you store the address of your C<my_biggy> struct in
3751 the C<data> member of the watcher (for woozies or C++ coders), or you need
3752 to use some pointer arithmetic using C<offsetof> inside your watchers (for
3753 real programmers):
3755 #include <stddef.h>
3757 static void
3758 t1_cb (EV_P_ ev_timer *w, int revents)
3759 {
3760 struct my_biggy big = (struct my_biggy *)
3761 (((char *)w) - offsetof (struct my_biggy, t1));
3762 }
3764 static void
3765 t2_cb (EV_P_ ev_timer *w, int revents)
3766 {
3767 struct my_biggy big = (struct my_biggy *)
3768 (((char *)w) - offsetof (struct my_biggy, t2));
3769 }
3773 Often you have structures like this in event-based programs:
3775 callback ()
3776 {
3777 free (request);
3778 }
3780 request = start_new_request (..., callback);
3782 The intent is to start some "lengthy" operation. The C<request> could be
3783 used to cancel the operation, or do other things with it.
3785 It's not uncommon to have code paths in C<start_new_request> that
3786 immediately invoke the callback, for example, to report errors. Or you add
3787 some caching layer that finds that it can skip the lengthy aspects of the
3788 operation and simply invoke the callback with the result.
3790 The problem here is that this will happen I<before> C<start_new_request>
3791 has returned, so C<request> is not set.
3793 Even if you pass the request by some safer means to the callback, you
3794 might want to do something to the request after starting it, such as
3795 canceling it, which probably isn't working so well when the callback has
3796 already been invoked.
3798 A common way around all these issues is to make sure that
3799 C<start_new_request> I<always> returns before the callback is invoked. If
3800 C<start_new_request> immediately knows the result, it can artificially
3801 delay invoking the callback by using a C<prepare> or C<idle> watcher for
3802 example, or more sneakily, by reusing an existing (stopped) watcher and
3803 pushing it into the pending queue:
3805 ev_set_cb (watcher, callback);
3806 ev_feed_event (EV_A_ watcher, 0);
3808 This way, C<start_new_request> can safely return before the callback is
3809 invoked, while not delaying callback invocation too much.
3813 Often (especially in GUI toolkits) there are places where you have
3814 I<modal> interaction, which is most easily implemented by recursively
3815 invoking C<ev_run>.
3817 This brings the problem of exiting - a callback might want to finish the
3818 main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
3819 a modal "Are you sure?" dialog is still waiting), or just the nested one
3820 and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
3821 other combination: In these cases, a simple C<ev_break> will not work.
3823 The solution is to maintain "break this loop" variable for each C<ev_run>
3824 invocation, and use a loop around C<ev_run> until the condition is
3825 triggered, using C<EVRUN_ONCE>:
3827 // main loop
3828 int exit_main_loop = 0;
3830 while (!exit_main_loop)
3831 ev_run (EV_DEFAULT_ EVRUN_ONCE);
3833 // in a modal watcher
3834 int exit_nested_loop = 0;
3836 while (!exit_nested_loop)
3837 ev_run (EV_A_ EVRUN_ONCE);
3839 To exit from any of these loops, just set the corresponding exit variable:
3841 // exit modal loop
3842 exit_nested_loop = 1;
3844 // exit main program, after modal loop is finished
3845 exit_main_loop = 1;
3847 // exit both
3848 exit_main_loop = exit_nested_loop = 1;
3852 Here is a fictitious example of how to run an event loop in a different
3853 thread from where callbacks are being invoked and watchers are
3854 created/added/removed.
3856 For a real-world example, see the C<EV::Loop::Async> perl module,
3857 which uses exactly this technique (which is suited for many high-level
3858 languages).
3860 The example uses a pthread mutex to protect the loop data, a condition
3861 variable to wait for callback invocations, an async watcher to notify the
3862 event loop thread and an unspecified mechanism to wake up the main thread.
3864 First, you need to associate some data with the event loop:
3866 typedef struct {
3867 mutex_t lock; /* global loop lock */
3868 ev_async async_w;
3869 thread_t tid;
3870 cond_t invoke_cv;
3871 } userdata;
3873 void prepare_loop (EV_P)
3874 {
3875 // for simplicity, we use a static userdata struct.
3876 static userdata u;
3878 ev_async_init (&u->async_w, async_cb);
3879 ev_async_start (EV_A_ &u->async_w);
3881 pthread_mutex_init (&u->lock, 0);
3882 pthread_cond_init (&u->invoke_cv, 0);
3884 // now associate this with the loop
3885 ev_set_userdata (EV_A_ u);
3886 ev_set_invoke_pending_cb (EV_A_ l_invoke);
3887 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
3889 // then create the thread running ev_run
3890 pthread_create (&u->tid, 0, l_run, EV_A);
3891 }
3893 The callback for the C<ev_async> watcher does nothing: the watcher is used
3894 solely to wake up the event loop so it takes notice of any new watchers
3895 that might have been added:
3897 static void
3898 async_cb (EV_P_ ev_async *w, int revents)
3899 {
3900 // just used for the side effects
3901 }
3903 The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
3904 protecting the loop data, respectively.
3906 static void
3907 l_release (EV_P)
3908 {
3909 userdata *u = ev_userdata (EV_A);
3910 pthread_mutex_unlock (&u->lock);
3911 }
3913 static void
3914 l_acquire (EV_P)
3915 {
3916 userdata *u = ev_userdata (EV_A);
3917 pthread_mutex_lock (&u->lock);
3918 }
3920 The event loop thread first acquires the mutex, and then jumps straight
3921 into C<ev_run>:
3923 void *
3924 l_run (void *thr_arg)
3925 {
3926 struct ev_loop *loop = (struct ev_loop *)thr_arg;
3928 l_acquire (EV_A);
3929 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
3930 ev_run (EV_A_ 0);
3931 l_release (EV_A);
3933 return 0;
3934 }
3936 Instead of invoking all pending watchers, the C<l_invoke> callback will
3937 signal the main thread via some unspecified mechanism (signals? pipe
3938 writes? C<Async::Interrupt>?) and then waits until all pending watchers
3939 have been called (in a while loop because a) spurious wakeups are possible
3940 and b) skipping inter-thread-communication when there are no pending
3941 watchers is very beneficial):
3943 static void
3944 l_invoke (EV_P)
3945 {
3946 userdata *u = ev_userdata (EV_A);
3948 while (ev_pending_count (EV_A))
3949 {
3950 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
3951 pthread_cond_wait (&u->invoke_cv, &u->lock);
3952 }
3953 }
3955 Now, whenever the main thread gets told to invoke pending watchers, it
3956 will grab the lock, call C<ev_invoke_pending> and then signal the loop
3957 thread to continue:
3959 static void
3960 real_invoke_pending (EV_P)
3961 {
3962 userdata *u = ev_userdata (EV_A);
3964 pthread_mutex_lock (&u->lock);
3965 ev_invoke_pending (EV_A);
3966 pthread_cond_signal (&u->invoke_cv);
3967 pthread_mutex_unlock (&u->lock);
3968 }
3970 Whenever you want to start/stop a watcher or do other modifications to an
3971 event loop, you will now have to lock:
3973 ev_timer timeout_watcher;
3974 userdata *u = ev_userdata (EV_A);
3976 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
3978 pthread_mutex_lock (&u->lock);
3979 ev_timer_start (EV_A_ &timeout_watcher);
3980 ev_async_send (EV_A_ &u->async_w);
3981 pthread_mutex_unlock (&u->lock);
3983 Note that sending the C<ev_async> watcher is required because otherwise
3984 an event loop currently blocking in the kernel will have no knowledge
3985 about the newly added timer. By waking up the loop it will pick up any new
3986 watchers in the next event loop iteration.
3990 While the overhead of a callback that e.g. schedules a thread is small, it
3991 is still an overhead. If you embed libev, and your main usage is with some
3992 kind of threads or coroutines, you might want to customise libev so that
3993 doesn't need callbacks anymore.
3995 Imagine you have coroutines that you can switch to using a function
3996 C<switch_to (coro)>, that libev runs in a coroutine called C<libev_coro>
3997 and that due to some magic, the currently active coroutine is stored in a
3998 global called C<current_coro>. Then you can build your own "wait for libev
3999 event" primitive by changing C<EV_CB_DECLARE> and C<EV_CB_INVOKE> (note
4000 the differing C<;> conventions):
4002 #define EV_CB_DECLARE(type) struct my_coro *cb;
4003 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
4005 That means instead of having a C callback function, you store the
4006 coroutine to switch to in each watcher, and instead of having libev call
4007 your callback, you instead have it switch to that coroutine.
4009 A coroutine might now wait for an event with a function called
4010 C<wait_for_event>. (the watcher needs to be started, as always, but it doesn't
4011 matter when, or whether the watcher is active or not when this function is
4012 called):
4014 void
4015 wait_for_event (ev_watcher *w)
4016 {
4017 ev_set_cb (w, current_coro);
4018 switch_to (libev_coro);
4019 }
4021 That basically suspends the coroutine inside C<wait_for_event> and
4022 continues the libev coroutine, which, when appropriate, switches back to
4023 this or any other coroutine.
4025 You can do similar tricks if you have, say, threads with an event queue -
4026 instead of storing a coroutine, you store the queue object and instead of
4027 switching to a coroutine, you push the watcher onto the queue and notify
4028 any waiters.
4030 To embed libev, see L</EMBEDDING>, but in short, it's easiest to create two
4031 files, F<my_ev.h> and F<my_ev.c> that include the respective libev files:
4033 // my_ev.h
4034 #define EV_CB_DECLARE(type) struct my_coro *cb;
4035 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
4036 #include "../libev/ev.h"
4038 // my_ev.c
4039 #define EV_H "my_ev.h"
4040 #include "../libev/ev.c"
4042 And then use F<my_ev.h> when you would normally use F<ev.h>, and compile
4043 F<my_ev.c> into your project. When properly specifying include paths, you
4044 can even use F<ev.h> as header file name directly.
4049 Libev offers a compatibility emulation layer for libevent. It cannot
4050 emulate the internals of libevent, so here are some usage hints:
4052 =over 4
4054 =item * Only the libevent-1.4.1-beta API is being emulated.
4056 This was the newest libevent version available when libev was implemented,
4057 and is still mostly unchanged in 2010.
4059 =item * Use it by including <event.h>, as usual.
4061 =item * The following members are fully supported: ev_base, ev_callback,
4062 ev_arg, ev_fd, ev_res, ev_events.
4064 =item * Avoid using ev_flags and the EVLIST_*-macros, while it is
4065 maintained by libev, it does not work exactly the same way as in libevent (consider
4066 it a private API).
4068 =item * Priorities are not currently supported. Initialising priorities
4069 will fail and all watchers will have the same priority, even though there
4070 is an ev_pri field.
4072 =item * In libevent, the last base created gets the signals, in libev, the
4073 base that registered the signal gets the signals.
4075 =item * Other members are not supported.
4077 =item * The libev emulation is I<not> ABI compatible to libevent, you need
4078 to use the libev header file and library.
4080 =back
4082 =head1 C++ SUPPORT
4084 =head2 C API
4086 The normal C API should work fine when used from C++: both ev.h and the
4087 libev sources can be compiled as C++. Therefore, code that uses the C API
4088 will work fine.
4090 Proper exception specifications might have to be added to callbacks passed
4091 to libev: exceptions may be thrown only from watcher callbacks, all other
4092 callbacks (allocator, syserr, loop acquire/release and periodic reschedule
4093 callbacks) must not throw exceptions, and might need a C<noexcept>
4094 specification. If you have code that needs to be compiled as both C and
4095 C++ you can use the C<EV_NOEXCEPT> macro for this:
4097 static void
4098 fatal_error (const char *msg) EV_NOEXCEPT
4099 {
4100 perror (msg);
4101 abort ();
4102 }
4104 ...
4105 ev_set_syserr_cb (fatal_error);
4107 The only API functions that can currently throw exceptions are C<ev_run>,
4108 C<ev_invoke>, C<ev_invoke_pending> and C<ev_loop_destroy> (the latter
4109 because it runs cleanup watchers).
4111 Throwing exceptions in watcher callbacks is only supported if libev itself
4112 is compiled with a C++ compiler or your C and C++ environments allow
4113 throwing exceptions through C libraries (most do).
4115 =head2 C++ API
4117 Libev comes with some simplistic wrapper classes for C++ that mainly allow
4118 you to use some convenience methods to start/stop watchers and also change
4119 the callback model to a model using method callbacks on objects.
4121 To use it,
4123 #include <ev++.h>
4125 This automatically includes F<ev.h> and puts all of its definitions (many
4126 of them macros) into the global namespace. All C++ specific things are
4127 put into the C<ev> namespace. It should support all the same embedding
4128 options as F<ev.h>, most notably C<EV_MULTIPLICITY>.
4130 Care has been taken to keep the overhead low. The only data member the C++
4131 classes add (compared to plain C-style watchers) is the event loop pointer
4132 that the watcher is associated with (or no additional members at all if
4133 you disable C<EV_MULTIPLICITY> when embedding libev).
4135 Currently, functions, static and non-static member functions and classes
4136 with C<operator ()> can be used as callbacks. Other types should be easy
4137 to add as long as they only need one additional pointer for context. If
4138 you need support for other types of functors please contact the author
4139 (preferably after implementing it).
4141 For all this to work, your C++ compiler either has to use the same calling
4142 conventions as your C compiler (for static member functions), or you have
4143 to embed libev and compile libev itself as C++.
4145 Here is a list of things available in the C<ev> namespace:
4147 =over 4
4149 =item C<ev::READ>, C<ev::WRITE> etc.
4151 These are just enum values with the same values as the C<EV_READ> etc.
4152 macros from F<ev.h>.
4154 =item C<ev::tstamp>, C<ev::now>
4156 Aliases to the same types/functions as with the C<ev_> prefix.
4158 =item C<ev::io>, C<ev::timer>, C<ev::periodic>, C<ev::idle>, C<ev::sig> etc.
4160 For each C<ev_TYPE> watcher in F<ev.h> there is a corresponding class of
4161 the same name in the C<ev> namespace, with the exception of C<ev_signal>
4162 which is called C<ev::sig> to avoid clashes with the C<signal> macro
4163 defined by many implementations.
4165 All of those classes have these methods:
4167 =over 4
4169 =item ev::TYPE::TYPE ()
4171 =item ev::TYPE::TYPE (loop)
4173 =item ev::TYPE::~TYPE
4175 The constructor (optionally) takes an event loop to associate the watcher
4176 with. If it is omitted, it will use C<EV_DEFAULT>.
4178 The constructor calls C<ev_init> for you, which means you have to call the
4179 C<set> method before starting it.
4181 It will not set a callback, however: You have to call the templated C<set>
4182 method to set a callback before you can start the watcher.
4184 (The reason why you have to use a method is a limitation in C++ which does
4185 not allow explicit template arguments for constructors).
4187 The destructor automatically stops the watcher if it is active.
4189 =item w->set<class, &class::method> (object *)
4191 This method sets the callback method to call. The method has to have a
4192 signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as
4193 first argument and the C<revents> as second. The object must be given as
4194 parameter and is stored in the C<data> member of the watcher.
4196 This method synthesizes efficient thunking code to call your method from
4197 the C callback that libev requires. If your compiler can inline your
4198 callback (i.e. it is visible to it at the place of the C<set> call and
4199 your compiler is good :), then the method will be fully inlined into the
4200 thunking function, making it as fast as a direct C callback.
4202 Example: simple class declaration and watcher initialisation
4204 struct myclass
4205 {
4206 void io_cb (ev::io &w, int revents) { }
4207 }
4209 myclass obj;
4210 ev::io iow;
4211 iow.set <myclass, &myclass::io_cb> (&obj);
4213 =item w->set (object *)
4215 This is a variation of a method callback - leaving out the method to call
4216 will default the method to C<operator ()>, which makes it possible to use
4217 functor objects without having to manually specify the C<operator ()> all
4218 the time. Incidentally, you can then also leave out the template argument
4219 list.
4221 The C<operator ()> method prototype must be C<void operator ()(watcher &w,
4222 int revents)>.
4224 See the method-C<set> above for more details.
4226 Example: use a functor object as callback.
4228 struct myfunctor
4229 {
4230 void operator() (ev::io &w, int revents)
4231 {
4232 ...
4233 }
4234 }
4236 myfunctor f;
4238 ev::io w;
4239 w.set (&f);
4241 =item w->set<function> (void *data = 0)
4243 Also sets a callback, but uses a static method or plain function as
4244 callback. The optional C<data> argument will be stored in the watcher's
4245 C<data> member and is free for you to use.
4247 The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>.
4249 See the method-C<set> above for more details.
4251 Example: Use a plain function as callback.
4253 static void io_cb (ev::io &w, int revents) { }
4254 iow.set <io_cb> ();
4256 =item w->set (loop)
4258 Associates a different C<struct ev_loop> with this watcher. You can only
4259 do this when the watcher is inactive (and not pending either).
4261 =item w->set ([arguments])
4263 Basically the same as C<ev_TYPE_set> (except for C<ev::embed> watchers>),
4264 with the same arguments. Either this method or a suitable start method
4265 must be called at least once. Unlike the C counterpart, an active watcher
4266 gets automatically stopped and restarted when reconfiguring it with this
4267 method.
4269 For C<ev::embed> watchers this method is called C<set_embed>, to avoid
4270 clashing with the C<set (loop)> method.
4272 For C<ev::io> watchers there is an additional C<set> method that acepts a
4273 new event mask only, and internally calls C<ev_io_modfify>.
4275 =item w->start ()
4277 Starts the watcher. Note that there is no C<loop> argument, as the
4278 constructor already stores the event loop.
4280 =item w->start ([arguments])
4282 Instead of calling C<set> and C<start> methods separately, it is often
4283 convenient to wrap them in one call. Uses the same type of arguments as
4284 the configure C<set> method of the watcher.
4286 =item w->stop ()
4288 Stops the watcher if it is active. Again, no C<loop> argument.
4290 =item w->again () (C<ev::timer>, C<ev::periodic> only)
4292 For C<ev::timer> and C<ev::periodic>, this invokes the corresponding
4293 C<ev_TYPE_again> function.
4295 =item w->sweep () (C<ev::embed> only)
4297 Invokes C<ev_embed_sweep>.
4299 =item w->update () (C<ev::stat> only)
4301 Invokes C<ev_stat_stat>.
4303 =back
4305 =back
4307 Example: Define a class with two I/O and idle watchers, start the I/O
4308 watchers in the constructor.
4310 class myclass
4311 {
4312 ev::io io ; void io_cb (ev::io &w, int revents);
4313 ev::io io2 ; void io2_cb (ev::io &w, int revents);
4314 ev::idle idle; void idle_cb (ev::idle &w, int revents);
4316 myclass (int fd)
4317 {
4318 io .set <myclass, &myclass::io_cb > (this);
4319 io2 .set <myclass, &myclass::io2_cb > (this);
4320 idle.set <myclass, &myclass::idle_cb> (this);
4322 io.set (fd, ev::WRITE); // configure the watcher
4323 io.start (); // start it whenever convenient
4325 io2.start (fd, ev::READ); // set + start in one call
4326 }
4327 };
4332 Libev does not offer other language bindings itself, but bindings for a
4333 number of languages exist in the form of third-party packages. If you know
4334 any interesting language binding in addition to the ones listed here, drop
4335 me a note.
4337 =over 4
4339 =item Perl
4341 The EV module implements the full libev API and is actually used to test
4342 libev. EV is developed together with libev. Apart from the EV core module,
4343 there are additional modules that implement libev-compatible interfaces
4344 to C<libadns> (C<EV::ADNS>, but C<AnyEvent::DNS> is preferred nowadays),
4345 C<Net::SNMP> (C<Net::SNMP::EV>) and the C<libglib> event core (C<Glib::EV>
4346 and C<EV::Glib>).
4348 It can be found and installed via CPAN, its homepage is at
4349 L<>.
4351 =item Python
4353 Python bindings can be found at L<>. It
4354 seems to be quite complete and well-documented.
4356 =item Ruby
4358 Tony Arcieri has written a ruby extension that offers access to a subset
4359 of the libev API and adds file handle abstractions, asynchronous DNS and
4360 more on top of it. It can be found via gem servers. Its homepage is at
4361 L<>.
4363 Roger Pack reports that using the link order C<-lws2_32 -lmsvcrt-ruby-190>
4364 makes rev work even on mingw.
4366 =item Haskell
4368 A haskell binding to libev is available at
4369 L<>.
4371 =item D
4373 Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
4374 be found at L<>.
4376 =item Ocaml
4378 Erkki Seppala has written Ocaml bindings for libev, to be found at
4379 L<>.
4381 =item Lua
4383 Brian Maher has written a partial interface to libev for lua (at the
4384 time of this writing, only C<ev_io> and C<ev_timer>), to be found at
4385 L<>.
4387 =item Javascript
4389 Node.js (L<>) uses libev as the underlying event library.
4391 =item Others
4393 There are others, and I stopped counting.
4395 =back
4398 =head1 MACRO MAGIC
4400 Libev can be compiled with a variety of options, the most fundamental
4401 of which is C<EV_MULTIPLICITY>. This option determines whether (most)
4402 functions and callbacks have an initial C<struct ev_loop *> argument.
4404 To make it easier to write programs that cope with either variant, the
4405 following macros are defined:
4407 =over 4
4409 =item C<EV_A>, C<EV_A_>
4411 This provides the loop I<argument> for functions, if one is required ("ev
4412 loop argument"). The C<EV_A> form is used when this is the sole argument,
4413 C<EV_A_> is used when other arguments are following. Example:
4415 ev_unref (EV_A);
4416 ev_timer_add (EV_A_ watcher);
4417 ev_run (EV_A_ 0);
4419 It assumes the variable C<loop> of type C<struct ev_loop *> is in scope,
4420 which is often provided by the following macro.
4422 =item C<EV_P>, C<EV_P_>
4424 This provides the loop I<parameter> for functions, if one is required ("ev
4425 loop parameter"). The C<EV_P> form is used when this is the sole parameter,
4426 C<EV_P_> is used when other parameters are following. Example:
4428 // this is how ev_unref is being declared
4429 static void ev_unref (EV_P);
4431 // this is how you can declare your typical callback
4432 static void cb (EV_P_ ev_timer *w, int revents)
4434 It declares a parameter C<loop> of type C<struct ev_loop *>, quite
4435 suitable for use with C<EV_A>.
4437 =item C<EV_DEFAULT>, C<EV_DEFAULT_>
4439 Similar to the other two macros, this gives you the value of the default
4440 loop, if multiple loops are supported ("ev loop default"). The default loop
4441 will be initialised if it isn't already initialised.
4443 For non-multiplicity builds, these macros do nothing, so you always have
4444 to initialise the loop somewhere.
4448 Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
4449 default loop has been initialised (C<UC> == unchecked). Their behaviour
4450 is undefined when the default loop has not been initialised by a previous
4451 execution of C<EV_DEFAULT>, C<EV_DEFAULT_> or C<ev_default_init (...)>.
4453 It is often prudent to use C<EV_DEFAULT> when initialising the first
4454 watcher in a function but use C<EV_DEFAULT_UC> afterwards.
4456 =back
4458 Example: Declare and initialise a check watcher, utilising the above
4459 macros so it will work regardless of whether multiple loops are supported
4460 or not.
4462 static void
4463 check_cb (EV_P_ ev_timer *w, int revents)
4464 {
4465 ev_check_stop (EV_A_ w);
4466 }
4468 ev_check check;
4469 ev_check_init (&check, check_cb);
4470 ev_check_start (EV_DEFAULT_ &check);
4471 ev_run (EV_DEFAULT_ 0);
4473 =head1 EMBEDDING
4475 Libev can (and often is) directly embedded into host
4476 applications. Examples of applications that embed it include the Deliantra
4477 Game Server, the EV perl module, the GNU Virtual Private Ethernet (gvpe)
4478 and rxvt-unicode.
4480 The goal is to enable you to just copy the necessary files into your
4481 source directory without having to change even a single line in them, so
4482 you can easily upgrade by simply copying (or having a checked-out copy of
4483 libev somewhere in your source tree).
4485 =head2 FILESETS
4487 Depending on what features you need you need to include one or more sets of files
4488 in your application.
4490 =head3 CORE EVENT LOOP
4492 To include only the libev core (all the C<ev_*> functions), with manual
4493 configuration (no autoconf):
4495 #define EV_STANDALONE 1
4496 #include "ev.c"
4498 This will automatically include F<ev.h>, too, and should be done in a
4499 single C source file only to provide the function implementations. To use
4500 it, do the same for F<ev.h> in all files wishing to use this API (best
4501 done by writing a wrapper around F<ev.h> that you can include instead and
4502 where you can put other configuration options):
4504 #define EV_STANDALONE 1
4505 #include "ev.h"
4507 Both header files and implementation files can be compiled with a C++
4508 compiler (at least, that's a stated goal, and breakage will be treated
4509 as a bug).
4511 You need the following files in your source tree, or in a directory
4512 in your include path (e.g. in libev/ when using -Ilibev):
4514 ev.h
4515 ev.c
4516 ev_vars.h
4517 ev_wrap.h
4519 ev_win32.c required on win32 platforms only
4521 ev_select.c only when select backend is enabled
4522 ev_poll.c only when poll backend is enabled
4523 ev_epoll.c only when the epoll backend is enabled
4524 ev_linuxaio.c only when the linux aio backend is enabled
4525 ev_iouring.c only when the linux io_uring backend is enabled
4526 ev_kqueue.c only when the kqueue backend is enabled
4527 ev_port.c only when the solaris port backend is enabled
4529 F<ev.c> includes the backend files directly when enabled, so you only need
4530 to compile this single file.
4534 To include the libevent compatibility API, also include:
4536 #include "event.c"
4538 in the file including F<ev.c>, and:
4540 #include "event.h"
4542 in the files that want to use the libevent API. This also includes F<ev.h>.
4544 You need the following additional files for this:
4546 event.h
4547 event.c
4551 Instead of using C<EV_STANDALONE=1> and providing your configuration in
4552 whatever way you want, you can also C<m4_include([libev.m4])> in your
4553 F<> and leave C<EV_STANDALONE> undefined. F<ev.c> will then
4554 include F<config.h> and configure itself accordingly.
4556 For this of course you need the m4 file:
4558 libev.m4
4562 Libev can be configured via a variety of preprocessor symbols you have to
4563 define before including (or compiling) any of its files. The default in
4564 the absence of autoconf is documented for every option.
4566 Symbols marked with "(h)" do not change the ABI, and can have different
4567 values when compiling libev vs. including F<ev.h>, so it is permissible
4568 to redefine them before including F<ev.h> without breaking compatibility
4569 to a compiled library. All other symbols change the ABI, which means all
4570 users of libev and the libev code itself must be compiled with compatible
4571 settings.
4573 =over 4
4575 =item EV_COMPAT3 (h)
4577 Backwards compatibility is a major concern for libev. This is why this
4578 release of libev comes with wrappers for the functions and symbols that
4579 have been renamed between libev version 3 and 4.
4581 You can disable these wrappers (to test compatibility with future
4582 versions) by defining C<EV_COMPAT3> to C<0> when compiling your
4583 sources. This has the additional advantage that you can drop the C<struct>
4584 from C<struct ev_loop> declarations, as libev will provide an C<ev_loop>
4585 typedef in that case.
4587 In some future version, the default for C<EV_COMPAT3> will become C<0>,
4588 and in some even more future version the compatibility code will be
4589 removed completely.
4591 =item EV_STANDALONE (h)
4593 Must always be C<1> if you do not use autoconf configuration, which
4594 keeps libev from including F<config.h>, and it also defines dummy
4595 implementations for some libevent functions (such as logging, which is not
4596 supported). It will also not define any of the structs usually found in
4597 F<event.h> that are not directly supported by the libev core alone.
4599 In standalone mode, libev will still try to automatically deduce the
4600 configuration, but has to be more conservative.
4602 =item EV_USE_FLOOR
4604 If defined to be C<1>, libev will use the C<floor ()> function for its
4605 periodic reschedule calculations, otherwise libev will fall back on a
4606 portable (slower) implementation. If you enable this, you usually have to
4607 link against libm or something equivalent. Enabling this when the C<floor>
4608 function is not available will fail, so the safe default is to not enable
4609 this.
4613 If defined to be C<1>, libev will try to detect the availability of the
4614 monotonic clock option at both compile time and runtime. Otherwise no
4615 use of the monotonic clock option will be attempted. If you enable this,
4616 you usually have to link against librt or something similar. Enabling it
4617 when the functionality isn't available is safe, though, although you have
4618 to make sure you link against any libraries where the C<clock_gettime>
4619 function is hiding in (often F<-lrt>). See also C<EV_USE_CLOCK_SYSCALL>.
4621 =item EV_USE_REALTIME
4623 If defined to be C<1>, libev will try to detect the availability of the
4624 real-time clock option at compile time (and assume its availability
4625 at runtime if successful). Otherwise no use of the real-time clock
4626 option will be attempted. This effectively replaces C<gettimeofday>
4627 by C<clock_get (CLOCK_REALTIME, ...)> and will not normally affect
4628 correctness. See the note about libraries in the description of
4629 C<EV_USE_MONOTONIC>, though. Defaults to the opposite value of
4634 If defined to be C<1>, libev will try to use a direct syscall instead
4635 of calling the system-provided C<clock_gettime> function. This option
4636 exists because on GNU/Linux, C<clock_gettime> is in C<librt>, but C<librt>
4637 unconditionally pulls in C<libpthread>, slowing down single-threaded
4638 programs needlessly. Using a direct syscall is slightly slower (in
4639 theory), because no optimised vdso implementation can be used, but avoids
4640 the pthread dependency. Defaults to C<1> on GNU/Linux with glibc 2.x or
4641 higher, as it simplifies linking (no need for C<-lrt>).
4645 If defined to be C<1>, libev will assume that C<nanosleep ()> is available
4646 and will use it for delays. Otherwise it will use C<select ()>.
4648 =item EV_USE_EVENTFD
4650 If defined to be C<1>, then libev will assume that C<eventfd ()> is
4651 available and will probe for kernel support at runtime. This will improve
4652 C<ev_signal> and C<ev_async> performance and reduce resource consumption.
4653 If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
4654 2.7 or newer, otherwise disabled.
4656 =item EV_USE_SIGNALFD
4658 If defined to be C<1>, then libev will assume that C<signalfd ()> is
4659 available and will probe for kernel support at runtime. This enables
4660 the use of EVFLAG_SIGNALFD for faster and simpler signal handling. If
4661 undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
4662 2.7 or newer, otherwise disabled.
4664 =item EV_USE_TIMERFD
4666 If defined to be C<1>, then libev will assume that C<timerfd ()> is
4667 available and will probe for kernel support at runtime. This allows
4668 libev to detect time jumps accurately. If undefined, it will be enabled
4669 if the headers indicate GNU/Linux + Glibc 2.8 or newer and define
4670 C<TFD_TIMER_CANCEL_ON_SET>, otherwise disabled.
4672 =item EV_USE_EVENTFD
4674 If defined to be C<1>, then libev will assume that C<eventfd ()> is
4675 available and will probe for kernel support at runtime. This will improve
4676 C<ev_signal> and C<ev_async> performance and reduce resource consumption.
4677 If undefined, it will be enabled if the headers indicate GNU/Linux + Glibc
4678 2.7 or newer, otherwise disabled.
4680 =item EV_USE_SELECT
4682 If undefined or defined to be C<1>, libev will compile in support for the
4683 C<select>(2) backend. No attempt at auto-detection will be done: if no
4684 other method takes over, select will be it. Otherwise the select backend
4685 will not be compiled in.
4689 If defined to C<1>, then the select backend will use the system C<fd_set>
4690 structure. This is useful if libev doesn't compile due to a missing
4691 C<NFDBITS> or C<fd_mask> definition or it mis-guesses the bitset layout
4692 on exotic systems. This usually limits the range of file descriptors to
4693 some low limit such as 1024 or might have other limitations (winsocket
4694 only allows 64 sockets). The C<FD_SETSIZE> macro, set before compilation,
4695 configures the maximum size of the C<fd_set>.
4699 When defined to C<1>, the select backend will assume that
4700 select/socket/connect etc. don't understand file descriptors but
4701 wants osf handles on win32 (this is the case when the select to
4702 be used is the winsock select). This means that it will call
4703 C<_get_osfhandle> on the fd to convert it to an OS handle. Otherwise,
4704 it is assumed that all these functions actually work on fds, even
4705 on win32. Should not be defined on non-win32 platforms.
4707 =item EV_FD_TO_WIN32_HANDLE(fd)
4709 If C<EV_SELECT_IS_WINSOCKET> is enabled, then libev needs a way to map
4710 file descriptors to socket handles. When not defining this symbol (the
4711 default), then libev will call C<_get_osfhandle>, which is usually
4712 correct. In some cases, programs use their own file descriptor management,
4713 in which case they can provide this function to map fds to socket handles.
4715 =item EV_WIN32_HANDLE_TO_FD(handle)
4717 If C<EV_SELECT_IS_WINSOCKET> then libev maps handles to file descriptors
4718 using the standard C<_open_osfhandle> function. For programs implementing
4719 their own fd to handle mapping, overwriting this function makes it easier
4720 to do so. This can be done by defining this macro to an appropriate value.
4722 =item EV_WIN32_CLOSE_FD(fd)
4724 If programs implement their own fd to handle mapping on win32, then this
4725 macro can be used to override the C<close> function, useful to unregister
4726 file descriptors again. Note that the replacement function has to close
4727 the underlying OS handle.
4731 If defined to be C<1>, libev will use C<WSASocket> to create its internal
4732 communication socket, which works better in some environments. Otherwise,
4733 the normal C<socket> function will be used, which works better in other
4734 environments.
4736 =item EV_USE_POLL
4738 If defined to be C<1>, libev will compile in support for the C<poll>(2)
4739 backend. Otherwise it will be enabled on non-win32 platforms. It
4740 takes precedence over select.
4742 =item EV_USE_EPOLL
4744 If defined to be C<1>, libev will compile in support for the Linux
4745 C<epoll>(7) backend. Its availability will be detected at runtime,
4746 otherwise another method will be used as fallback. This is the preferred
4747 backend for GNU/Linux systems. If undefined, it will be enabled if the
4748 headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
4750 =item EV_USE_LINUXAIO
4752 If defined to be C<1>, libev will compile in support for the Linux aio
4753 backend (C<EV_USE_EPOLL> must also be enabled). If undefined, it will be
4754 enabled on linux, otherwise disabled.
4756 =item EV_USE_IOURING
4758 If defined to be C<1>, libev will compile in support for the Linux
4759 io_uring backend (C<EV_USE_EPOLL> must also be enabled). Due to it's
4760 current limitations it has to be requested explicitly. If undefined, it
4761 will be enabled on linux, otherwise disabled.
4763 =item EV_USE_KQUEUE
4765 If defined to be C<1>, libev will compile in support for the BSD style
4766 C<kqueue>(2) backend. Its actual availability will be detected at runtime,
4767 otherwise another method will be used as fallback. This is the preferred
4768 backend for BSD and BSD-like systems, although on most BSDs kqueue only
4769 supports some types of fds correctly (the only platform we found that
4770 supports ptys for example was NetBSD), so kqueue might be compiled in, but
4771 not be used unless explicitly requested. The best way to use it is to find
4772 out whether kqueue supports your type of fd properly and use an embedded
4773 kqueue loop.
4775 =item EV_USE_PORT
4777 If defined to be C<1>, libev will compile in support for the Solaris
4778 10 port style backend. Its availability will be detected at runtime,
4779 otherwise another method will be used as fallback. This is the preferred
4780 backend for Solaris 10 systems.
4782 =item EV_USE_DEVPOLL
4784 Reserved for future expansion, works like the USE symbols above.
4786 =item EV_USE_INOTIFY
4788 If defined to be C<1>, libev will compile in support for the Linux inotify
4789 interface to speed up C<ev_stat> watchers. Its actual availability will
4790 be detected at runtime. If undefined, it will be enabled if the headers
4791 indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
4793 =item EV_NO_SMP
4795 If defined to be C<1>, libev will assume that memory is always coherent
4796 between threads, that is, threads can be used, but threads never run on
4797 different cpus (or different cpu cores). This reduces dependencies
4798 and makes libev faster.
4800 =item EV_NO_THREADS
4802 If defined to be C<1>, libev will assume that it will never be called from
4803 different threads (that includes signal handlers), which is a stronger
4804 assumption than C<EV_NO_SMP>, above. This reduces dependencies and makes
4805 libev faster.
4807 =item EV_ATOMIC_T
4809 Libev requires an integer type (suitable for storing C<0> or C<1>) whose
4810 access is atomic with respect to other threads or signal contexts. No
4811 such type is easily found in the C language, so you can provide your own
4812 type that you know is safe for your purposes. It is used both for signal
4813 handler "locking" as well as for signal and thread safety in C<ev_async>
4814 watchers.
4816 In the absence of this define, libev will use C<sig_atomic_t volatile>
4817 (from F<signal.h>), which is usually good enough on most platforms.
4819 =item EV_H (h)
4821 The name of the F<ev.h> header file used to include it. The default if
4822 undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
4823 used to virtually rename the F<ev.h> header file in case of conflicts.
4825 =item EV_CONFIG_H (h)
4827 If C<EV_STANDALONE> isn't C<1>, this variable can be used to override
4828 F<ev.c>'s idea of where to find the F<config.h> file, similarly to
4829 C<EV_H>, above.
4831 =item EV_EVENT_H (h)
4833 Similarly to C<EV_H>, this macro can be used to override F<event.c>'s idea
4834 of how the F<event.h> header can be found, the default is C<"event.h">.
4836 =item EV_PROTOTYPES (h)
4838 If defined to be C<0>, then F<ev.h> will not define any function
4839 prototypes, but still define all the structs and other symbols. This is
4840 occasionally useful if you want to provide your own wrapper functions
4841 around libev functions.
4845 If undefined or defined to C<1>, then all event-loop-specific functions
4846 will have the C<struct ev_loop *> as first argument, and you can create
4847 additional independent event loops. Otherwise there will be no support
4848 for multiple event loops and there is no first event loop pointer
4849 argument. Instead, all functions act on the single default loop.
4851 Note that C<EV_DEFAULT> and C<EV_DEFAULT_> will no longer provide a
4852 default loop when multiplicity is switched off - you always have to
4853 initialise the loop manually in this case.
4855 =item EV_MINPRI
4857 =item EV_MAXPRI
4859 The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to
4860 C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can
4861 provide for more priorities by overriding those symbols (usually defined
4862 to be C<-2> and C<2>, respectively).
4864 When doing priority-based operations, libev usually has to linearly search
4865 all the priorities, so having many of them (hundreds) uses a lot of space
4866 and time, so using the defaults of five priorities (-2 .. +2) is usually
4867 fine.
4869 If your embedding application does not need any priorities, defining these
4870 both to C<0> will save some memory and CPU.
4876 If undefined or defined to be C<1> (and the platform supports it), then
4877 the respective watcher type is supported. If defined to be C<0>, then it
4878 is not. Disabling watcher types mainly saves code size.
4880 =item EV_FEATURES
4882 If you need to shave off some kilobytes of code at the expense of some
4883 speed (but with the full API), you can define this symbol to request
4884 certain subsets of functionality. The default is to enable all features
4885 that can be enabled on the platform.
4887 A typical way to use this symbol is to define it to C<0> (or to a bitset
4888 with some broad features you want) and then selectively re-enable
4889 additional parts you want, for example if you want everything minimal,
4890 but multiple event loop support, async and child watchers and the poll
4891 backend, use this:
4893 #define EV_FEATURES 0
4894 #define EV_MULTIPLICITY 1
4895 #define EV_USE_POLL 1
4896 #define EV_CHILD_ENABLE 1
4897 #define EV_ASYNC_ENABLE 1
4899 The actual value is a bitset, it can be a combination of the following
4900 values (by default, all of these are enabled):
4902 =over 4
4904 =item C<1> - faster/larger code
4906 Use larger code to speed up some operations.
4908 Currently this is used to override some inlining decisions (enlarging the
4909 code size by roughly 30% on amd64).
4911 When optimising for size, use of compiler flags such as C<-Os> with
4912 gcc is recommended, as well as C<-DNDEBUG>, as libev contains a number of
4913 assertions.
4915 The default is off when C<__OPTIMIZE_SIZE__> is defined by your compiler
4916 (e.g. gcc with C<-Os>).
4918 =item C<2> - faster/larger data structures
4920 Replaces the small 2-heap for timer management by a faster 4-heap, larger
4921 hash table sizes and so on. This will usually further increase code size
4922 and can additionally have an effect on the size of data structures at
4923 runtime.
4925 The default is off when C<__OPTIMIZE_SIZE__> is defined by your compiler
4926 (e.g. gcc with C<-Os>).
4928 =item C<4> - full API configuration
4930 This enables priorities (sets C<EV_MAXPRI>=2 and C<EV_MINPRI>=-2), and
4931 enables multiplicity (C<EV_MULTIPLICITY>=1).
4933 =item C<8> - full API
4935 This enables a lot of the "lesser used" API functions. See C<ev.h> for
4936 details on which parts of the API are still available without this
4937 feature, and do not complain if this subset changes over time.
4939 =item C<16> - enable all optional watcher types
4941 Enables all optional watcher types. If you want to selectively enable
4942 only some watcher types other than I/O and timers (e.g. prepare,
4943 embed, async, child...) you can enable them manually by defining
4944 C<EV_watchertype_ENABLE> to C<1> instead.
4946 =item C<32> - enable all backends
4948 This enables all backends - without this feature, you need to enable at
4949 least one backend manually (C<EV_USE_SELECT> is a good choice).
4951 =item C<64> - enable OS-specific "helper" APIs
4953 Enable inotify, eventfd, signalfd and similar OS-specific helper APIs by
4954 default.
4956 =back
4958 Compiling with C<gcc -Os -DEV_STANDALONE -DEV_USE_EPOLL=1 -DEV_FEATURES=0>
4959 reduces the c