ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.pod
(Generate patch)

Comparing libev/ev.pod (file contents):
Revision 1.310 by root, Thu Oct 21 12:32:47 2010 UTC vs.
Revision 1.366 by sf-exg, Thu Feb 3 16:21:08 2011 UTC

43 43
44 int 44 int
45 main (void) 45 main (void)
46 { 46 {
47 // use the default event loop unless you have special needs 47 // use the default event loop unless you have special needs
48 struct ev_loop *loop = ev_default_loop (0); 48 struct ev_loop *loop = EV_DEFAULT;
49 49
50 // initialise an io watcher, then start it 50 // initialise an io watcher, then start it
51 // this one will watch for stdin to become readable 51 // this one will watch for stdin to become readable
52 ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ); 52 ev_io_init (&stdin_watcher, stdin_cb, /*STDIN_FILENO*/ 0, EV_READ);
53 ev_io_start (loop, &stdin_watcher); 53 ev_io_start (loop, &stdin_watcher);
58 ev_timer_start (loop, &timeout_watcher); 58 ev_timer_start (loop, &timeout_watcher);
59 59
60 // now wait for events to arrive 60 // now wait for events to arrive
61 ev_run (loop, 0); 61 ev_run (loop, 0);
62 62
63 // unloop was called, so exit 63 // break was called, so exit
64 return 0; 64 return 0;
65 } 65 }
66 66
67=head1 ABOUT THIS DOCUMENT 67=head1 ABOUT THIS DOCUMENT
68 68
77on event-based programming, nor will it introduce event-based programming 77on event-based programming, nor will it introduce event-based programming
78with libev. 78with libev.
79 79
80Familiarity with event based programming techniques in general is assumed 80Familiarity with event based programming techniques in general is assumed
81throughout this document. 81throughout this document.
82
83=head1 WHAT TO READ WHEN IN A HURRY
84
85This manual tries to be very detailed, but unfortunately, this also makes
86it very long. If you just want to know the basics of libev, I suggest
87reading L<ANATOMY OF A WATCHER>, then the L<EXAMPLE PROGRAM> above and
88look up the missing functions in L<GLOBAL FUNCTIONS> and the C<ev_io> and
89C<ev_timer> sections in L<WATCHER TYPES>.
82 90
83=head1 ABOUT LIBEV 91=head1 ABOUT LIBEV
84 92
85Libev is an event loop: you register interest in certain events (such as a 93Libev is an event loop: you register interest in certain events (such as a
86file descriptor being readable or a timeout occurring), and it will manage 94file descriptor being readable or a timeout occurring), and it will manage
124this argument. 132this argument.
125 133
126=head2 TIME REPRESENTATION 134=head2 TIME REPRESENTATION
127 135
128Libev represents time as a single floating point number, representing 136Libev represents time as a single floating point number, representing
129the (fractional) number of seconds since the (POSIX) epoch (in practise 137the (fractional) number of seconds since the (POSIX) epoch (in practice
130somewhere near the beginning of 1970, details are complicated, don't 138somewhere near the beginning of 1970, details are complicated, don't
131ask). This type is called C<ev_tstamp>, which is what you should use 139ask). This type is called C<ev_tstamp>, which is what you should use
132too. It usually aliases to the C<double> type in C. When you need to do 140too. It usually aliases to the C<double> type in C. When you need to do
133any calculations on it, you should treat it as some floating point value. 141any calculations on it, you should treat it as some floating point value.
134 142
165 173
166=item ev_tstamp ev_time () 174=item ev_tstamp ev_time ()
167 175
168Returns the current time as libev would use it. Please note that the 176Returns the current time as libev would use it. Please note that the
169C<ev_now> function is usually faster and also often returns the timestamp 177C<ev_now> function is usually faster and also often returns the timestamp
170you actually want to know. 178you actually want to know. Also interesting is the combination of
179C<ev_update_now> and C<ev_now>.
171 180
172=item ev_sleep (ev_tstamp interval) 181=item ev_sleep (ev_tstamp interval)
173 182
174Sleep for the given interval: The current thread will be blocked until 183Sleep for the given interval: The current thread will be blocked until
175either it is interrupted or the given time interval has passed. Basically 184either it is interrupted or the given time interval has passed. Basically
192as this indicates an incompatible change. Minor versions are usually 201as this indicates an incompatible change. Minor versions are usually
193compatible to older versions, so a larger minor version alone is usually 202compatible to older versions, so a larger minor version alone is usually
194not a problem. 203not a problem.
195 204
196Example: Make sure we haven't accidentally been linked against the wrong 205Example: Make sure we haven't accidentally been linked against the wrong
197version (note, however, that this will not detect ABI mismatches :). 206version (note, however, that this will not detect other ABI mismatches,
207such as LFS or reentrancy).
198 208
199 assert (("libev version mismatch", 209 assert (("libev version mismatch",
200 ev_version_major () == EV_VERSION_MAJOR 210 ev_version_major () == EV_VERSION_MAJOR
201 && ev_version_minor () >= EV_VERSION_MINOR)); 211 && ev_version_minor () >= EV_VERSION_MINOR));
202 212
213 assert (("sorry, no epoll, no sex", 223 assert (("sorry, no epoll, no sex",
214 ev_supported_backends () & EVBACKEND_EPOLL)); 224 ev_supported_backends () & EVBACKEND_EPOLL));
215 225
216=item unsigned int ev_recommended_backends () 226=item unsigned int ev_recommended_backends ()
217 227
218Return the set of all backends compiled into this binary of libev and also 228Return the set of all backends compiled into this binary of libev and
219recommended for this platform. This set is often smaller than the one 229also recommended for this platform, meaning it will work for most file
230descriptor types. This set is often smaller than the one returned by
220returned by C<ev_supported_backends>, as for example kqueue is broken on 231C<ev_supported_backends>, as for example kqueue is broken on most BSDs
221most BSDs and will not be auto-detected unless you explicitly request it 232and will not be auto-detected unless you explicitly request it (assuming
222(assuming you know what you are doing). This is the set of backends that 233you know what you are doing). This is the set of backends that libev will
223libev will probe for if you specify no backends explicitly. 234probe for if you specify no backends explicitly.
224 235
225=item unsigned int ev_embeddable_backends () 236=item unsigned int ev_embeddable_backends ()
226 237
227Returns the set of backends that are embeddable in other event loops. This 238Returns the set of backends that are embeddable in other event loops. This
228is the theoretical, all-platform, value. To find which backends 239value is platform-specific but can include backends not available on the
229might be supported on the current system, you would need to look at 240current system. To find which embeddable backends might be supported on
230C<ev_embeddable_backends () & ev_supported_backends ()>, likewise for 241the current system, you would need to look at C<ev_embeddable_backends ()
231recommended ones. 242& ev_supported_backends ()>, likewise for recommended ones.
232 243
233See the description of C<ev_embed> watchers for more info. 244See the description of C<ev_embed> watchers for more info.
234 245
235=item ev_set_allocator (void *(*cb)(void *ptr, long size)) [NOT REENTRANT] 246=item ev_set_allocator (void *(*cb)(void *ptr, long size))
236 247
237Sets the allocation function to use (the prototype is similar - the 248Sets the allocation function to use (the prototype is similar - the
238semantics are identical to the C<realloc> C89/SuS/POSIX function). It is 249semantics are identical to the C<realloc> C89/SuS/POSIX function). It is
239used to allocate and free memory (no surprises here). If it returns zero 250used to allocate and free memory (no surprises here). If it returns zero
240when memory needs to be allocated (C<size != 0>), the library might abort 251when memory needs to be allocated (C<size != 0>), the library might abort
266 } 277 }
267 278
268 ... 279 ...
269 ev_set_allocator (persistent_realloc); 280 ev_set_allocator (persistent_realloc);
270 281
271=item ev_set_syserr_cb (void (*cb)(const char *msg)); [NOT REENTRANT] 282=item ev_set_syserr_cb (void (*cb)(const char *msg))
272 283
273Set the callback function to call on a retryable system call error (such 284Set the callback function to call on a retryable system call error (such
274as failed select, poll, epoll_wait). The message is a printable string 285as failed select, poll, epoll_wait). The message is a printable string
275indicating the system call or subsystem causing the problem. If this 286indicating the system call or subsystem causing the problem. If this
276callback is set, then libev will expect it to remedy the situation, no 287callback is set, then libev will expect it to remedy the situation, no
288 } 299 }
289 300
290 ... 301 ...
291 ev_set_syserr_cb (fatal_error); 302 ev_set_syserr_cb (fatal_error);
292 303
304=item ev_feed_signal (int signum)
305
306This function can be used to "simulate" a signal receive. It is completely
307safe to call this function at any time, from any context, including signal
308handlers or random threads.
309
310Its main use is to customise signal handling in your process, especially
311in the presence of threads. For example, you could block signals
312by default in all threads (and specifying C<EVFLAG_NOSIGMASK> when
313creating any loops), and in one thread, use C<sigwait> or any other
314mechanism to wait for signals, then "deliver" them to libev by calling
315C<ev_feed_signal>.
316
293=back 317=back
294 318
295=head1 FUNCTIONS CONTROLLING THE EVENT LOOP 319=head1 FUNCTIONS CONTROLLING EVENT LOOPS
296 320
297An event loop is described by a C<struct ev_loop *> (the C<struct> is 321An event loop is described by a C<struct ev_loop *> (the C<struct> is
298I<not> optional in case unless libev 3 compatibility is disabled, as libev 322I<not> optional in this case unless libev 3 compatibility is disabled, as
2993 had an C<ev_loop> function colliding with the struct name). 323libev 3 had an C<ev_loop> function colliding with the struct name).
300 324
301The library knows two types of such loops, the I<default> loop, which 325The library knows two types of such loops, the I<default> loop, which
302supports signals and child events, and dynamically created event loops 326supports child process events, and dynamically created event loops which
303which do not. 327do not.
304 328
305=over 4 329=over 4
306 330
307=item struct ev_loop *ev_default_loop (unsigned int flags) 331=item struct ev_loop *ev_default_loop (unsigned int flags)
308 332
309This will initialise the default event loop if it hasn't been initialised 333This returns the "default" event loop object, which is what you should
310yet and return it. If the default loop could not be initialised, returns 334normally use when you just need "the event loop". Event loop objects and
311false. If it already was initialised it simply returns it (and ignores the 335the C<flags> parameter are described in more detail in the entry for
312flags. If that is troubling you, check C<ev_backend ()> afterwards). 336C<ev_loop_new>.
337
338If the default loop is already initialised then this function simply
339returns it (and ignores the flags. If that is troubling you, check
340C<ev_backend ()> afterwards). Otherwise it will create it with the given
341flags, which should almost always be C<0>, unless the caller is also the
342one calling C<ev_run> or otherwise qualifies as "the main program".
313 343
314If you don't know what event loop to use, use the one returned from this 344If you don't know what event loop to use, use the one returned from this
315function. 345function (or via the C<EV_DEFAULT> macro).
316 346
317Note that this function is I<not> thread-safe, so if you want to use it 347Note that this function is I<not> thread-safe, so if you want to use it
318from multiple threads, you have to lock (note also that this is unlikely, 348from multiple threads, you have to employ some kind of mutex (note also
319as loops cannot be shared easily between threads anyway). 349that this case is unlikely, as loops cannot be shared easily between
350threads anyway).
320 351
321The default loop is the only loop that can handle C<ev_signal> and 352The default loop is the only loop that can handle C<ev_child> watchers,
322C<ev_child> watchers, and to do this, it always registers a handler 353and to do this, it always registers a handler for C<SIGCHLD>. If this is
323for C<SIGCHLD>. If this is a problem for your application you can either 354a problem for your application you can either create a dynamic loop with
324create a dynamic loop with C<ev_loop_new> that doesn't do that, or you 355C<ev_loop_new> which doesn't do that, or you can simply overwrite the
325can simply overwrite the C<SIGCHLD> signal handler I<after> calling 356C<SIGCHLD> signal handler I<after> calling C<ev_default_init>.
326C<ev_default_init>. 357
358Example: This is the most typical usage.
359
360 if (!ev_default_loop (0))
361 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
362
363Example: Restrict libev to the select and poll backends, and do not allow
364environment settings to be taken into account:
365
366 ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
367
368=item struct ev_loop *ev_loop_new (unsigned int flags)
369
370This will create and initialise a new event loop object. If the loop
371could not be initialised, returns false.
372
373This function is thread-safe, and one common way to use libev with
374threads is indeed to create one loop per thread, and using the default
375loop in the "main" or "initial" thread.
327 376
328The flags argument can be used to specify special behaviour or specific 377The flags argument can be used to specify special behaviour or specific
329backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>). 378backends to use, and is usually specified as C<0> (or C<EVFLAG_AUTO>).
330 379
331The following flags are supported: 380The following flags are supported:
366environment variable. 415environment variable.
367 416
368=item C<EVFLAG_NOINOTIFY> 417=item C<EVFLAG_NOINOTIFY>
369 418
370When this flag is specified, then libev will not attempt to use the 419When this flag is specified, then libev will not attempt to use the
371I<inotify> API for it's C<ev_stat> watchers. Apart from debugging and 420I<inotify> API for its C<ev_stat> watchers. Apart from debugging and
372testing, this flag can be useful to conserve inotify file descriptors, as 421testing, this flag can be useful to conserve inotify file descriptors, as
373otherwise each loop using C<ev_stat> watchers consumes one inotify handle. 422otherwise each loop using C<ev_stat> watchers consumes one inotify handle.
374 423
375=item C<EVFLAG_SIGNALFD> 424=item C<EVFLAG_SIGNALFD>
376 425
377When this flag is specified, then libev will attempt to use the 426When this flag is specified, then libev will attempt to use the
378I<signalfd> API for it's C<ev_signal> (and C<ev_child>) watchers. This API 427I<signalfd> API for its C<ev_signal> (and C<ev_child>) watchers. This API
379delivers signals synchronously, which makes it both faster and might make 428delivers signals synchronously, which makes it both faster and might make
380it possible to get the queued signal data. It can also simplify signal 429it possible to get the queued signal data. It can also simplify signal
381handling with threads, as long as you properly block signals in your 430handling with threads, as long as you properly block signals in your
382threads that are not interested in handling them. 431threads that are not interested in handling them.
383 432
384Signalfd will not be used by default as this changes your signal mask, and 433Signalfd will not be used by default as this changes your signal mask, and
385there are a lot of shoddy libraries and programs (glib's threadpool for 434there are a lot of shoddy libraries and programs (glib's threadpool for
386example) that can't properly initialise their signal masks. 435example) that can't properly initialise their signal masks.
436
437=item C<EVFLAG_NOSIGMASK>
438
439When this flag is specified, then libev will avoid to modify the signal
440mask. Specifically, this means you ahve to make sure signals are unblocked
441when you want to receive them.
442
443This behaviour is useful when you want to do your own signal handling, or
444want to handle signals only in specific threads and want to avoid libev
445unblocking the signals.
446
447It's also required by POSIX in a threaded program, as libev calls
448C<sigprocmask>, whose behaviour is officially unspecified.
449
450This flag's behaviour will become the default in future versions of libev.
387 451
388=item C<EVBACKEND_SELECT> (value 1, portable select backend) 452=item C<EVBACKEND_SELECT> (value 1, portable select backend)
389 453
390This is your standard select(2) backend. Not I<completely> standard, as 454This is your standard select(2) backend. Not I<completely> standard, as
391libev tries to roll its own fd_set with no limits on the number of fds, 455libev tries to roll its own fd_set with no limits on the number of fds,
427epoll scales either O(1) or O(active_fds). 491epoll scales either O(1) or O(active_fds).
428 492
429The epoll mechanism deserves honorable mention as the most misdesigned 493The epoll mechanism deserves honorable mention as the most misdesigned
430of the more advanced event mechanisms: mere annoyances include silently 494of the more advanced event mechanisms: mere annoyances include silently
431dropping file descriptors, requiring a system call per change per file 495dropping file descriptors, requiring a system call per change per file
432descriptor (and unnecessary guessing of parameters), problems with dup and 496descriptor (and unnecessary guessing of parameters), problems with dup,
497returning before the timeout value, resulting in additional iterations
498(and only giving 5ms accuracy while select on the same platform gives
433so on. The biggest issue is fork races, however - if a program forks then 4990.1ms) and so on. The biggest issue is fork races, however - if a program
434I<both> parent and child process have to recreate the epoll set, which can 500forks then I<both> parent and child process have to recreate the epoll
435take considerable time (one syscall per file descriptor) and is of course 501set, which can take considerable time (one syscall per file descriptor)
436hard to detect. 502and is of course hard to detect.
437 503
438Epoll is also notoriously buggy - embedding epoll fds I<should> work, but 504Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
439of course I<doesn't>, and epoll just loves to report events for totally 505of course I<doesn't>, and epoll just loves to report events for totally
440I<different> file descriptors (even already closed ones, so one cannot 506I<different> file descriptors (even already closed ones, so one cannot
441even remove them from the set) than registered in the set (especially 507even remove them from the set) than registered in the set (especially
443employing an additional generation counter and comparing that against the 509employing an additional generation counter and comparing that against the
444events to filter out spurious ones, recreating the set when required. Last 510events to filter out spurious ones, recreating the set when required. Last
445not least, it also refuses to work with some file descriptors which work 511not least, it also refuses to work with some file descriptors which work
446perfectly fine with C<select> (files, many character devices...). 512perfectly fine with C<select> (files, many character devices...).
447 513
514Epoll is truly the train wreck analog among event poll mechanisms,
515a frankenpoll, cobbled together in a hurry, no thought to design or
516interaction with others.
517
448While stopping, setting and starting an I/O watcher in the same iteration 518While stopping, setting and starting an I/O watcher in the same iteration
449will result in some caching, there is still a system call per such 519will result in some caching, there is still a system call per such
450incident (because the same I<file descriptor> could point to a different 520incident (because the same I<file descriptor> could point to a different
451I<file description> now), so its best to avoid that. Also, C<dup ()>'ed 521I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
452file descriptors might not work very well if you register events for both 522file descriptors might not work very well if you register events for both
517=item C<EVBACKEND_PORT> (value 32, Solaris 10) 587=item C<EVBACKEND_PORT> (value 32, Solaris 10)
518 588
519This uses the Solaris 10 event port mechanism. As with everything on Solaris, 589This uses the Solaris 10 event port mechanism. As with everything on Solaris,
520it's really slow, but it still scales very well (O(active_fds)). 590it's really slow, but it still scales very well (O(active_fds)).
521 591
522Please note that Solaris event ports can deliver a lot of spurious
523notifications, so you need to use non-blocking I/O or other means to avoid
524blocking when no data (or space) is available.
525
526While this backend scales well, it requires one system call per active 592While this backend scales well, it requires one system call per active
527file descriptor per loop iteration. For small and medium numbers of file 593file descriptor per loop iteration. For small and medium numbers of file
528descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend 594descriptors a "slow" C<EVBACKEND_SELECT> or C<EVBACKEND_POLL> backend
529might perform better. 595might perform better.
530 596
531On the positive side, with the exception of the spurious readiness 597On the positive side, this backend actually performed fully to
532notifications, this backend actually performed fully to specification
533in all tests and is fully embeddable, which is a rare feat among the 598specification in all tests and is fully embeddable, which is a rare feat
534OS-specific backends (I vastly prefer correctness over speed hacks). 599among the OS-specific backends (I vastly prefer correctness over speed
600hacks).
601
602On the negative side, the interface is I<bizarre> - so bizarre that
603even sun itself gets it wrong in their code examples: The event polling
604function sometimes returning events to the caller even though an error
605occurred, but with no indication whether it has done so or not (yes, it's
606even documented that way) - deadly for edge-triggered interfaces where
607you absolutely have to know whether an event occurred or not because you
608have to re-arm the watcher.
609
610Fortunately libev seems to be able to work around these idiocies.
535 611
536This backend maps C<EV_READ> and C<EV_WRITE> in the same way as 612This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
537C<EVBACKEND_POLL>. 613C<EVBACKEND_POLL>.
538 614
539=item C<EVBACKEND_ALL> 615=item C<EVBACKEND_ALL>
540 616
541Try all backends (even potentially broken ones that wouldn't be tried 617Try all backends (even potentially broken ones that wouldn't be tried
542with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as 618with C<EVFLAG_AUTO>). Since this is a mask, you can do stuff such as
543C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>. 619C<EVBACKEND_ALL & ~EVBACKEND_KQUEUE>.
544 620
545It is definitely not recommended to use this flag. 621It is definitely not recommended to use this flag, use whatever
622C<ev_recommended_backends ()> returns, or simply do not specify a backend
623at all.
624
625=item C<EVBACKEND_MASK>
626
627Not a backend at all, but a mask to select all backend bits from a
628C<flags> value, in case you want to mask out any backends from a flags
629value (e.g. when modifying the C<LIBEV_FLAGS> environment variable).
546 630
547=back 631=back
548 632
549If one or more of the backend flags are or'ed into the flags value, 633If one or more of the backend flags are or'ed into the flags value,
550then only these backends will be tried (in the reverse order as listed 634then only these backends will be tried (in the reverse order as listed
551here). If none are specified, all backends in C<ev_recommended_backends 635here). If none are specified, all backends in C<ev_recommended_backends
552()> will be tried. 636()> will be tried.
553 637
554Example: This is the most typical usage.
555
556 if (!ev_default_loop (0))
557 fatal ("could not initialise libev, bad $LIBEV_FLAGS in environment?");
558
559Example: Restrict libev to the select and poll backends, and do not allow
560environment settings to be taken into account:
561
562 ev_default_loop (EVBACKEND_POLL | EVBACKEND_SELECT | EVFLAG_NOENV);
563
564Example: Use whatever libev has to offer, but make sure that kqueue is
565used if available (warning, breaks stuff, best use only with your own
566private event loop and only if you know the OS supports your types of
567fds):
568
569 ev_default_loop (ev_recommended_backends () | EVBACKEND_KQUEUE);
570
571=item struct ev_loop *ev_loop_new (unsigned int flags)
572
573Similar to C<ev_default_loop>, but always creates a new event loop that is
574always distinct from the default loop.
575
576Note that this function I<is> thread-safe, and one common way to use
577libev with threads is indeed to create one loop per thread, and using the
578default loop in the "main" or "initial" thread.
579
580Example: Try to create a event loop that uses epoll and nothing else. 638Example: Try to create a event loop that uses epoll and nothing else.
581 639
582 struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV); 640 struct ev_loop *epoller = ev_loop_new (EVBACKEND_EPOLL | EVFLAG_NOENV);
583 if (!epoller) 641 if (!epoller)
584 fatal ("no epoll found here, maybe it hides under your chair"); 642 fatal ("no epoll found here, maybe it hides under your chair");
585 643
644Example: Use whatever libev has to offer, but make sure that kqueue is
645used if available.
646
647 struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
648
586=item ev_default_destroy () 649=item ev_loop_destroy (loop)
587 650
588Destroys the default loop (frees all memory and kernel state etc.). None 651Destroys an event loop object (frees all memory and kernel state
589of the active event watchers will be stopped in the normal sense, so 652etc.). None of the active event watchers will be stopped in the normal
590e.g. C<ev_is_active> might still return true. It is your responsibility to 653sense, so e.g. C<ev_is_active> might still return true. It is your
591either stop all watchers cleanly yourself I<before> calling this function, 654responsibility to either stop all watchers cleanly yourself I<before>
592or cope with the fact afterwards (which is usually the easiest thing, you 655calling this function, or cope with the fact afterwards (which is usually
593can just ignore the watchers and/or C<free ()> them for example). 656the easiest thing, you can just ignore the watchers and/or C<free ()> them
657for example).
594 658
595Note that certain global state, such as signal state (and installed signal 659Note that certain global state, such as signal state (and installed signal
596handlers), will not be freed by this function, and related watchers (such 660handlers), will not be freed by this function, and related watchers (such
597as signal and child watchers) would need to be stopped manually. 661as signal and child watchers) would need to be stopped manually.
598 662
599In general it is not advisable to call this function except in the 663This function is normally used on loop objects allocated by
600rare occasion where you really need to free e.g. the signal handling 664C<ev_loop_new>, but it can also be used on the default loop returned by
665C<ev_default_loop>, in which case it is not thread-safe.
666
667Note that it is not advisable to call this function on the default loop
668except in the rare occasion where you really need to free its resources.
601pipe fds. If you need dynamically allocated loops it is better to use 669If you need dynamically allocated loops it is better to use C<ev_loop_new>
602C<ev_loop_new> and C<ev_loop_destroy>. 670and C<ev_loop_destroy>.
603 671
604=item ev_loop_destroy (loop) 672=item ev_loop_fork (loop)
605 673
606Like C<ev_default_destroy>, but destroys an event loop created by an
607earlier call to C<ev_loop_new>.
608
609=item ev_default_fork ()
610
611This function sets a flag that causes subsequent C<ev_run> iterations 674This function sets a flag that causes subsequent C<ev_run> iterations to
612to reinitialise the kernel state for backends that have one. Despite the 675reinitialise the kernel state for backends that have one. Despite the
613name, you can call it anytime, but it makes most sense after forking, in 676name, you can call it anytime, but it makes most sense after forking, in
614the child process (or both child and parent, but that again makes little 677the child process. You I<must> call it (or use C<EVFLAG_FORKCHECK>) in the
615sense). You I<must> call it in the child before using any of the libev 678child before resuming or calling C<ev_run>.
616functions, and it will only take effect at the next C<ev_run> iteration.
617 679
618Again, you I<have> to call it on I<any> loop that you want to re-use after 680Again, you I<have> to call it on I<any> loop that you want to re-use after
619a fork, I<even if you do not plan to use the loop in the parent>. This is 681a fork, I<even if you do not plan to use the loop in the parent>. This is
620because some kernel interfaces *cough* I<kqueue> *cough* do funny things 682because some kernel interfaces *cough* I<kqueue> *cough* do funny things
621during fork. 683during fork.
626call it at all (in fact, C<epoll> is so badly broken that it makes a 688call it at all (in fact, C<epoll> is so badly broken that it makes a
627difference, but libev will usually detect this case on its own and do a 689difference, but libev will usually detect this case on its own and do a
628costly reset of the backend). 690costly reset of the backend).
629 691
630The function itself is quite fast and it's usually not a problem to call 692The function itself is quite fast and it's usually not a problem to call
631it just in case after a fork. To make this easy, the function will fit in 693it just in case after a fork.
632quite nicely into a call to C<pthread_atfork>:
633 694
695Example: Automate calling C<ev_loop_fork> on the default loop when
696using pthreads.
697
698 static void
699 post_fork_child (void)
700 {
701 ev_loop_fork (EV_DEFAULT);
702 }
703
704 ...
634 pthread_atfork (0, 0, ev_default_fork); 705 pthread_atfork (0, 0, post_fork_child);
635
636=item ev_loop_fork (loop)
637
638Like C<ev_default_fork>, but acts on an event loop created by
639C<ev_loop_new>. Yes, you have to call this on every allocated event loop
640after fork that you want to re-use in the child, and how you keep track of
641them is entirely your own problem.
642 706
643=item int ev_is_default_loop (loop) 707=item int ev_is_default_loop (loop)
644 708
645Returns true when the given loop is, in fact, the default loop, and false 709Returns true when the given loop is, in fact, the default loop, and false
646otherwise. 710otherwise.
657prepare and check phases. 721prepare and check phases.
658 722
659=item unsigned int ev_depth (loop) 723=item unsigned int ev_depth (loop)
660 724
661Returns the number of times C<ev_run> was entered minus the number of 725Returns the number of times C<ev_run> was entered minus the number of
662times C<ev_run> was exited, in other words, the recursion depth. 726times C<ev_run> was exited normally, in other words, the recursion depth.
663 727
664Outside C<ev_run>, this number is zero. In a callback, this number is 728Outside C<ev_run>, this number is zero. In a callback, this number is
665C<1>, unless C<ev_run> was invoked recursively (or from another thread), 729C<1>, unless C<ev_run> was invoked recursively (or from another thread),
666in which case it is higher. 730in which case it is higher.
667 731
668Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread 732Leaving C<ev_run> abnormally (setjmp/longjmp, cancelling the thread,
669etc.), doesn't count as "exit" - consider this as a hint to avoid such 733throwing an exception etc.), doesn't count as "exit" - consider this
670ungentleman-like behaviour unless it's really convenient. 734as a hint to avoid such ungentleman-like behaviour unless it's really
735convenient, in which case it is fully supported.
671 736
672=item unsigned int ev_backend (loop) 737=item unsigned int ev_backend (loop)
673 738
674Returns one of the C<EVBACKEND_*> flags indicating the event backend in 739Returns one of the C<EVBACKEND_*> flags indicating the event backend in
675use. 740use.
736relying on all watchers to be stopped when deciding when a program has 801relying on all watchers to be stopped when deciding when a program has
737finished (especially in interactive programs), but having a program 802finished (especially in interactive programs), but having a program
738that automatically loops as long as it has to and no longer by virtue 803that automatically loops as long as it has to and no longer by virtue
739of relying on its watchers stopping correctly, that is truly a thing of 804of relying on its watchers stopping correctly, that is truly a thing of
740beauty. 805beauty.
806
807This function is also I<mostly> exception-safe - you can break out of
808a C<ev_run> call by calling C<longjmp> in a callback, throwing a C++
809exception and so on. This does not decrement the C<ev_depth> value, nor
810will it clear any outstanding C<EVBREAK_ONE> breaks.
741 811
742A flags value of C<EVRUN_NOWAIT> will look for new events, will handle 812A flags value of C<EVRUN_NOWAIT> will look for new events, will handle
743those events and any already outstanding ones, but will not wait and 813those events and any already outstanding ones, but will not wait and
744block your process in case there are no events and will return after one 814block your process in case there are no events and will return after one
745iteration of the loop. This is sometimes useful to poll and handle new 815iteration of the loop. This is sometimes useful to poll and handle new
798anymore. 868anymore.
799 869
800 ... queue jobs here, make sure they register event watchers as long 870 ... queue jobs here, make sure they register event watchers as long
801 ... as they still have work to do (even an idle watcher will do..) 871 ... as they still have work to do (even an idle watcher will do..)
802 ev_run (my_loop, 0); 872 ev_run (my_loop, 0);
803 ... jobs done or somebody called unloop. yeah! 873 ... jobs done or somebody called break. yeah!
804 874
805=item ev_break (loop, how) 875=item ev_break (loop, how)
806 876
807Can be used to make a call to C<ev_run> return early (but only after it 877Can be used to make a call to C<ev_run> return early (but only after it
808has processed all outstanding events). The C<how> argument must be either 878has processed all outstanding events). The C<how> argument must be either
809C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or 879C<EVBREAK_ONE>, which will make the innermost C<ev_run> call return, or
810C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return. 880C<EVBREAK_ALL>, which will make all nested C<ev_run> calls return.
811 881
812This "unloop state" will be cleared when entering C<ev_run> again. 882This "break state" will be cleared on the next call to C<ev_run>.
813 883
814It is safe to call C<ev_break> from outside any C<ev_run> calls. ##TODO## 884It is safe to call C<ev_break> from outside any C<ev_run> calls, too, in
885which case it will have no effect.
815 886
816=item ev_ref (loop) 887=item ev_ref (loop)
817 888
818=item ev_unref (loop) 889=item ev_unref (loop)
819 890
840running when nothing else is active. 911running when nothing else is active.
841 912
842 ev_signal exitsig; 913 ev_signal exitsig;
843 ev_signal_init (&exitsig, sig_cb, SIGINT); 914 ev_signal_init (&exitsig, sig_cb, SIGINT);
844 ev_signal_start (loop, &exitsig); 915 ev_signal_start (loop, &exitsig);
845 evf_unref (loop); 916 ev_unref (loop);
846 917
847Example: For some weird reason, unregister the above signal handler again. 918Example: For some weird reason, unregister the above signal handler again.
848 919
849 ev_ref (loop); 920 ev_ref (loop);
850 ev_signal_stop (loop, &exitsig); 921 ev_signal_stop (loop, &exitsig);
908 979
909=item ev_invoke_pending (loop) 980=item ev_invoke_pending (loop)
910 981
911This call will simply invoke all pending watchers while resetting their 982This call will simply invoke all pending watchers while resetting their
912pending state. Normally, C<ev_run> does this automatically when required, 983pending state. Normally, C<ev_run> does this automatically when required,
913but when overriding the invoke callback this call comes handy. 984but when overriding the invoke callback this call comes handy. This
985function can be invoked from a watcher - this can be useful for example
986when you want to do some lengthy calculation and want to pass further
987event handling to another thread (you still have to make sure only one
988thread executes within C<ev_invoke_pending> or C<ev_run> of course).
914 989
915=item int ev_pending_count (loop) 990=item int ev_pending_count (loop)
916 991
917Returns the number of pending watchers - zero indicates that no watchers 992Returns the number of pending watchers - zero indicates that no watchers
918are pending. 993are pending.
958See also the locking example in the C<THREADS> section later in this 1033See also the locking example in the C<THREADS> section later in this
959document. 1034document.
960 1035
961=item ev_set_userdata (loop, void *data) 1036=item ev_set_userdata (loop, void *data)
962 1037
963=item ev_userdata (loop) 1038=item void *ev_userdata (loop)
964 1039
965Set and retrieve a single C<void *> associated with a loop. When 1040Set and retrieve a single C<void *> associated with a loop. When
966C<ev_set_userdata> has never been called, then C<ev_userdata> returns 1041C<ev_set_userdata> has never been called, then C<ev_userdata> returns
967C<0.> 1042C<0>.
968 1043
969These two functions can be used to associate arbitrary data with a loop, 1044These two functions can be used to associate arbitrary data with a loop,
970and are intended solely for the C<invoke_pending_cb>, C<release> and 1045and are intended solely for the C<invoke_pending_cb>, C<release> and
971C<acquire> callbacks described above, but of course can be (ab-)used for 1046C<acquire> callbacks described above, but of course can be (ab-)used for
972any other purpose as well. 1047any other purpose as well.
990 1065
991In the following description, uppercase C<TYPE> in names stands for the 1066In the following description, uppercase C<TYPE> in names stands for the
992watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer 1067watcher type, e.g. C<ev_TYPE_start> can mean C<ev_timer_start> for timer
993watchers and C<ev_io_start> for I/O watchers. 1068watchers and C<ev_io_start> for I/O watchers.
994 1069
995A watcher is a structure that you create and register to record your 1070A watcher is an opaque structure that you allocate and register to record
996interest in some event. For instance, if you want to wait for STDIN to 1071your interest in some event. To make a concrete example, imagine you want
997become readable, you would create an C<ev_io> watcher for that: 1072to wait for STDIN to become readable, you would create an C<ev_io> watcher
1073for that:
998 1074
999 static void my_cb (struct ev_loop *loop, ev_io *w, int revents) 1075 static void my_cb (struct ev_loop *loop, ev_io *w, int revents)
1000 { 1076 {
1001 ev_io_stop (w); 1077 ev_io_stop (w);
1002 ev_break (loop, EVBREAK_ALL); 1078 ev_break (loop, EVBREAK_ALL);
1017stack). 1093stack).
1018 1094
1019Each watcher has an associated watcher structure (called C<struct ev_TYPE> 1095Each watcher has an associated watcher structure (called C<struct ev_TYPE>
1020or simply C<ev_TYPE>, as typedefs are provided for all watcher structs). 1096or simply C<ev_TYPE>, as typedefs are provided for all watcher structs).
1021 1097
1022Each watcher structure must be initialised by a call to C<ev_init 1098Each watcher structure must be initialised by a call to C<ev_init (watcher
1023(watcher *, callback)>, which expects a callback to be provided. This 1099*, callback)>, which expects a callback to be provided. This callback is
1024callback gets invoked each time the event occurs (or, in the case of I/O 1100invoked each time the event occurs (or, in the case of I/O watchers, each
1025watchers, each time the event loop detects that the file descriptor given 1101time the event loop detects that the file descriptor given is readable
1026is readable and/or writable). 1102and/or writable).
1027 1103
1028Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >> 1104Each watcher type further has its own C<< ev_TYPE_set (watcher *, ...) >>
1029macro to configure it, with arguments specific to the watcher type. There 1105macro to configure it, with arguments specific to the watcher type. There
1030is also a macro to combine initialisation and setting in one call: C<< 1106is also a macro to combine initialisation and setting in one call: C<<
1031ev_TYPE_init (watcher *, callback, ...) >>. 1107ev_TYPE_init (watcher *, callback, ...) >>.
1099=item C<EV_FORK> 1175=item C<EV_FORK>
1100 1176
1101The event loop has been resumed in the child process after fork (see 1177The event loop has been resumed in the child process after fork (see
1102C<ev_fork>). 1178C<ev_fork>).
1103 1179
1180=item C<EV_CLEANUP>
1181
1182The event loop is about to be destroyed (see C<ev_cleanup>).
1183
1104=item C<EV_ASYNC> 1184=item C<EV_ASYNC>
1105 1185
1106The given async watcher has been asynchronously notified (see C<ev_async>). 1186The given async watcher has been asynchronously notified (see C<ev_async>).
1107 1187
1108=item C<EV_CUSTOM> 1188=item C<EV_CUSTOM>
1280See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related 1360See also C<ev_feed_fd_event> and C<ev_feed_signal_event> for related
1281functions that do not need a watcher. 1361functions that do not need a watcher.
1282 1362
1283=back 1363=back
1284 1364
1365See also the L<ASSOCIATING CUSTOM DATA WITH A WATCHER> and L<BUILDING YOUR
1366OWN COMPOSITE WATCHERS> idioms.
1285 1367
1286=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER 1368=head2 WATCHER STATES
1287 1369
1288Each watcher has, by default, a member C<void *data> that you can change 1370There are various watcher states mentioned throughout this manual -
1289and read at any time: libev will completely ignore it. This can be used 1371active, pending and so on. In this section these states and the rules to
1290to associate arbitrary data with your watcher. If you need more data and 1372transition between them will be described in more detail - and while these
1291don't want to allocate memory and store a pointer to it in that data 1373rules might look complicated, they usually do "the right thing".
1292member, you can also "subclass" the watcher type and provide your own
1293data:
1294 1374
1295 struct my_io 1375=over 4
1296 {
1297 ev_io io;
1298 int otherfd;
1299 void *somedata;
1300 struct whatever *mostinteresting;
1301 };
1302 1376
1303 ... 1377=item initialiased
1304 struct my_io w;
1305 ev_io_init (&w.io, my_cb, fd, EV_READ);
1306 1378
1307And since your callback will be called with a pointer to the watcher, you 1379Before a watcher can be registered with the event looop it has to be
1308can cast it back to your own type: 1380initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
1381C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
1309 1382
1310 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents) 1383In this state it is simply some block of memory that is suitable for
1311 { 1384use in an event loop. It can be moved around, freed, reused etc. at
1312 struct my_io *w = (struct my_io *)w_; 1385will - as long as you either keep the memory contents intact, or call
1313 ... 1386C<ev_TYPE_init> again.
1314 }
1315 1387
1316More interesting and less C-conformant ways of casting your callback type 1388=item started/running/active
1317instead have been omitted.
1318 1389
1319Another common scenario is to use some data structure with multiple 1390Once a watcher has been started with a call to C<ev_TYPE_start> it becomes
1320embedded watchers: 1391property of the event loop, and is actively waiting for events. While in
1392this state it cannot be accessed (except in a few documented ways), moved,
1393freed or anything else - the only legal thing is to keep a pointer to it,
1394and call libev functions on it that are documented to work on active watchers.
1321 1395
1322 struct my_biggy 1396=item pending
1323 {
1324 int some_data;
1325 ev_timer t1;
1326 ev_timer t2;
1327 }
1328 1397
1329In this case getting the pointer to C<my_biggy> is a bit more 1398If a watcher is active and libev determines that an event it is interested
1330complicated: Either you store the address of your C<my_biggy> struct 1399in has occurred (such as a timer expiring), it will become pending. It will
1331in the C<data> member of the watcher (for woozies), or you need to use 1400stay in this pending state until either it is stopped or its callback is
1332some pointer arithmetic using C<offsetof> inside your watchers (for real 1401about to be invoked, so it is not normally pending inside the watcher
1333programmers): 1402callback.
1334 1403
1335 #include <stddef.h> 1404The watcher might or might not be active while it is pending (for example,
1405an expired non-repeating timer can be pending but no longer active). If it
1406is stopped, it can be freely accessed (e.g. by calling C<ev_TYPE_set>),
1407but it is still property of the event loop at this time, so cannot be
1408moved, freed or reused. And if it is active the rules described in the
1409previous item still apply.
1336 1410
1337 static void 1411It is also possible to feed an event on a watcher that is not active (e.g.
1338 t1_cb (EV_P_ ev_timer *w, int revents) 1412via C<ev_feed_event>), in which case it becomes pending without being
1339 { 1413active.
1340 struct my_biggy big = (struct my_biggy *)
1341 (((char *)w) - offsetof (struct my_biggy, t1));
1342 }
1343 1414
1344 static void 1415=item stopped
1345 t2_cb (EV_P_ ev_timer *w, int revents) 1416
1346 { 1417A watcher can be stopped implicitly by libev (in which case it might still
1347 struct my_biggy big = (struct my_biggy *) 1418be pending), or explicitly by calling its C<ev_TYPE_stop> function. The
1348 (((char *)w) - offsetof (struct my_biggy, t2)); 1419latter will clear any pending state the watcher might be in, regardless
1349 } 1420of whether it was active or not, so stopping a watcher explicitly before
1421freeing it is often a good idea.
1422
1423While stopped (and not pending) the watcher is essentially in the
1424initialised state, that is, it can be reused, moved, modified in any way
1425you wish (but when you trash the memory block, you need to C<ev_TYPE_init>
1426it again).
1427
1428=back
1350 1429
1351=head2 WATCHER PRIORITY MODELS 1430=head2 WATCHER PRIORITY MODELS
1352 1431
1353Many event loops support I<watcher priorities>, which are usually small 1432Many event loops support I<watcher priorities>, which are usually small
1354integers that influence the ordering of event callback invocation 1433integers that influence the ordering of event callback invocation
1481In general you can register as many read and/or write event watchers per 1560In general you can register as many read and/or write event watchers per
1482fd as you want (as long as you don't confuse yourself). Setting all file 1561fd as you want (as long as you don't confuse yourself). Setting all file
1483descriptors to non-blocking mode is also usually a good idea (but not 1562descriptors to non-blocking mode is also usually a good idea (but not
1484required if you know what you are doing). 1563required if you know what you are doing).
1485 1564
1486If you cannot use non-blocking mode, then force the use of a
1487known-to-be-good backend (at the time of this writing, this includes only
1488C<EVBACKEND_SELECT> and C<EVBACKEND_POLL>). The same applies to file
1489descriptors for which non-blocking operation makes no sense (such as
1490files) - libev doesn't guarantee any specific behaviour in that case.
1491
1492Another thing you have to watch out for is that it is quite easy to 1565Another thing you have to watch out for is that it is quite easy to
1493receive "spurious" readiness notifications, that is your callback might 1566receive "spurious" readiness notifications, that is, your callback might
1494be called with C<EV_READ> but a subsequent C<read>(2) will actually block 1567be called with C<EV_READ> but a subsequent C<read>(2) will actually block
1495because there is no data. Not only are some backends known to create a 1568because there is no data. It is very easy to get into this situation even
1496lot of those (for example Solaris ports), it is very easy to get into 1569with a relatively standard program structure. Thus it is best to always
1497this situation even with a relatively standard program structure. Thus 1570use non-blocking I/O: An extra C<read>(2) returning C<EAGAIN> is far
1498it is best to always use non-blocking I/O: An extra C<read>(2) returning
1499C<EAGAIN> is far preferable to a program hanging until some data arrives. 1571preferable to a program hanging until some data arrives.
1500 1572
1501If you cannot run the fd in non-blocking mode (for example you should 1573If you cannot run the fd in non-blocking mode (for example you should
1502not play around with an Xlib connection), then you have to separately 1574not play around with an Xlib connection), then you have to separately
1503re-test whether a file descriptor is really ready with a known-to-be good 1575re-test whether a file descriptor is really ready with a known-to-be good
1504interface such as poll (fortunately in our Xlib example, Xlib already 1576interface such as poll (fortunately in the case of Xlib, it already does
1505does this on its own, so its quite safe to use). Some people additionally 1577this on its own, so its quite safe to use). Some people additionally
1506use C<SIGALRM> and an interval timer, just to be sure you won't block 1578use C<SIGALRM> and an interval timer, just to be sure you won't block
1507indefinitely. 1579indefinitely.
1508 1580
1509But really, best use non-blocking mode. 1581But really, best use non-blocking mode.
1510 1582
1538 1610
1539There is no workaround possible except not registering events 1611There is no workaround possible except not registering events
1540for potentially C<dup ()>'ed file descriptors, or to resort to 1612for potentially C<dup ()>'ed file descriptors, or to resort to
1541C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>. 1613C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1542 1614
1615=head3 The special problem of files
1616
1617Many people try to use C<select> (or libev) on file descriptors
1618representing files, and expect it to become ready when their program
1619doesn't block on disk accesses (which can take a long time on their own).
1620
1621However, this cannot ever work in the "expected" way - you get a readiness
1622notification as soon as the kernel knows whether and how much data is
1623there, and in the case of open files, that's always the case, so you
1624always get a readiness notification instantly, and your read (or possibly
1625write) will still block on the disk I/O.
1626
1627Another way to view it is that in the case of sockets, pipes, character
1628devices and so on, there is another party (the sender) that delivers data
1629on its own, but in the case of files, there is no such thing: the disk
1630will not send data on its own, simply because it doesn't know what you
1631wish to read - you would first have to request some data.
1632
1633Since files are typically not-so-well supported by advanced notification
1634mechanism, libev tries hard to emulate POSIX behaviour with respect
1635to files, even though you should not use it. The reason for this is
1636convenience: sometimes you want to watch STDIN or STDOUT, which is
1637usually a tty, often a pipe, but also sometimes files or special devices
1638(for example, C<epoll> on Linux works with F</dev/random> but not with
1639F</dev/urandom>), and even though the file might better be served with
1640asynchronous I/O instead of with non-blocking I/O, it is still useful when
1641it "just works" instead of freezing.
1642
1643So avoid file descriptors pointing to files when you know it (e.g. use
1644libeio), but use them when it is convenient, e.g. for STDIN/STDOUT, or
1645when you rarely read from a file instead of from a socket, and want to
1646reuse the same code path.
1647
1543=head3 The special problem of fork 1648=head3 The special problem of fork
1544 1649
1545Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit 1650Some backends (epoll, kqueue) do not support C<fork ()> at all or exhibit
1546useless behaviour. Libev fully supports fork, but needs to be told about 1651useless behaviour. Libev fully supports fork, but needs to be told about
1547it in the child. 1652it in the child if you want to continue to use it in the child.
1548 1653
1549To support fork in your programs, you either have to call 1654To support fork in your child processes, you have to call C<ev_loop_fork
1550C<ev_default_fork ()> or C<ev_loop_fork ()> after a fork in the child, 1655()> after a fork in the child, enable C<EVFLAG_FORKCHECK>, or resort to
1551enable C<EVFLAG_FORKCHECK>, or resort to C<EVBACKEND_SELECT> or 1656C<EVBACKEND_SELECT> or C<EVBACKEND_POLL>.
1552C<EVBACKEND_POLL>.
1553 1657
1554=head3 The special problem of SIGPIPE 1658=head3 The special problem of SIGPIPE
1555 1659
1556While not really specific to libev, it is easy to forget about C<SIGPIPE>: 1660While not really specific to libev, it is easy to forget about C<SIGPIPE>:
1557when writing to a pipe whose other end has been closed, your program gets 1661when writing to a pipe whose other end has been closed, your program gets
2173 2277
2174=head2 C<ev_signal> - signal me when a signal gets signalled! 2278=head2 C<ev_signal> - signal me when a signal gets signalled!
2175 2279
2176Signal watchers will trigger an event when the process receives a specific 2280Signal watchers will trigger an event when the process receives a specific
2177signal one or more times. Even though signals are very asynchronous, libev 2281signal one or more times. Even though signals are very asynchronous, libev
2178will try it's best to deliver signals synchronously, i.e. as part of the 2282will try its best to deliver signals synchronously, i.e. as part of the
2179normal event processing, like any other event. 2283normal event processing, like any other event.
2180 2284
2181If you want signals to be delivered truly asynchronously, just use 2285If you want signals to be delivered truly asynchronously, just use
2182C<sigaction> as you would do without libev and forget about sharing 2286C<sigaction> as you would do without libev and forget about sharing
2183the signal. You can even use C<ev_async> from a signal handler to 2287the signal. You can even use C<ev_async> from a signal handler to
2202=head3 The special problem of inheritance over fork/execve/pthread_create 2306=head3 The special problem of inheritance over fork/execve/pthread_create
2203 2307
2204Both the signal mask (C<sigprocmask>) and the signal disposition 2308Both the signal mask (C<sigprocmask>) and the signal disposition
2205(C<sigaction>) are unspecified after starting a signal watcher (and after 2309(C<sigaction>) are unspecified after starting a signal watcher (and after
2206stopping it again), that is, libev might or might not block the signal, 2310stopping it again), that is, libev might or might not block the signal,
2207and might or might not set or restore the installed signal handler. 2311and might or might not set or restore the installed signal handler (but
2312see C<EVFLAG_NOSIGMASK>).
2208 2313
2209While this does not matter for the signal disposition (libev never 2314While this does not matter for the signal disposition (libev never
2210sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on 2315sets signals to C<SIG_IGN>, so handlers will be reset to C<SIG_DFL> on
2211C<execve>), this matters for the signal mask: many programs do not expect 2316C<execve>), this matters for the signal mask: many programs do not expect
2212certain signals to be blocked. 2317certain signals to be blocked.
2225I<has> to modify the signal mask, at least temporarily. 2330I<has> to modify the signal mask, at least temporarily.
2226 2331
2227So I can't stress this enough: I<If you do not reset your signal mask when 2332So I can't stress this enough: I<If you do not reset your signal mask when
2228you expect it to be empty, you have a race condition in your code>. This 2333you expect it to be empty, you have a race condition in your code>. This
2229is not a libev-specific thing, this is true for most event libraries. 2334is not a libev-specific thing, this is true for most event libraries.
2335
2336=head3 The special problem of threads signal handling
2337
2338POSIX threads has problematic signal handling semantics, specifically,
2339a lot of functionality (sigfd, sigwait etc.) only really works if all
2340threads in a process block signals, which is hard to achieve.
2341
2342When you want to use sigwait (or mix libev signal handling with your own
2343for the same signals), you can tackle this problem by globally blocking
2344all signals before creating any threads (or creating them with a fully set
2345sigprocmask) and also specifying the C<EVFLAG_NOSIGMASK> when creating
2346loops. Then designate one thread as "signal receiver thread" which handles
2347these signals. You can pass on any signals that libev might be interested
2348in by calling C<ev_feed_signal>.
2230 2349
2231=head3 Watcher-Specific Functions and Data Members 2350=head3 Watcher-Specific Functions and Data Members
2232 2351
2233=over 4 2352=over 4
2234 2353
3008disadvantage of having to use multiple event loops (which do not support 3127disadvantage of having to use multiple event loops (which do not support
3009signal watchers). 3128signal watchers).
3010 3129
3011When this is not possible, or you want to use the default loop for 3130When this is not possible, or you want to use the default loop for
3012other reasons, then in the process that wants to start "fresh", call 3131other reasons, then in the process that wants to start "fresh", call
3013C<ev_default_destroy ()> followed by C<ev_default_loop (...)>. Destroying 3132C<ev_loop_destroy (EV_DEFAULT)> followed by C<ev_default_loop (...)>.
3014the default loop will "orphan" (not stop) all registered watchers, so you 3133Destroying the default loop will "orphan" (not stop) all registered
3015have to be careful not to execute code that modifies those watchers. Note 3134watchers, so you have to be careful not to execute code that modifies
3016also that in that case, you have to re-register any signal watchers. 3135those watchers. Note also that in that case, you have to re-register any
3136signal watchers.
3017 3137
3018=head3 Watcher-Specific Functions and Data Members 3138=head3 Watcher-Specific Functions and Data Members
3019 3139
3020=over 4 3140=over 4
3021 3141
3022=item ev_fork_init (ev_signal *, callback) 3142=item ev_fork_init (ev_fork *, callback)
3023 3143
3024Initialises and configures the fork watcher - it has no parameters of any 3144Initialises and configures the fork watcher - it has no parameters of any
3025kind. There is a C<ev_fork_set> macro, but using it is utterly pointless, 3145kind. There is a C<ev_fork_set> macro, but using it is utterly pointless,
3026believe me. 3146really.
3027 3147
3028=back 3148=back
3029 3149
3030 3150
3151=head2 C<ev_cleanup> - even the best things end
3152
3153Cleanup watchers are called just before the event loop is being destroyed
3154by a call to C<ev_loop_destroy>.
3155
3156While there is no guarantee that the event loop gets destroyed, cleanup
3157watchers provide a convenient method to install cleanup hooks for your
3158program, worker threads and so on - you just to make sure to destroy the
3159loop when you want them to be invoked.
3160
3161Cleanup watchers are invoked in the same way as any other watcher. Unlike
3162all other watchers, they do not keep a reference to the event loop (which
3163makes a lot of sense if you think about it). Like all other watchers, you
3164can call libev functions in the callback, except C<ev_cleanup_start>.
3165
3166=head3 Watcher-Specific Functions and Data Members
3167
3168=over 4
3169
3170=item ev_cleanup_init (ev_cleanup *, callback)
3171
3172Initialises and configures the cleanup watcher - it has no parameters of
3173any kind. There is a C<ev_cleanup_set> macro, but using it is utterly
3174pointless, I assure you.
3175
3176=back
3177
3178Example: Register an atexit handler to destroy the default loop, so any
3179cleanup functions are called.
3180
3181 static void
3182 program_exits (void)
3183 {
3184 ev_loop_destroy (EV_DEFAULT_UC);
3185 }
3186
3187 ...
3188 atexit (program_exits);
3189
3190
3031=head2 C<ev_async> - how to wake up an event loop 3191=head2 C<ev_async> - how to wake up an event loop
3032 3192
3033In general, you cannot use an C<ev_run> from multiple threads or other 3193In general, you cannot use an C<ev_loop> from multiple threads or other
3034asynchronous sources such as signal handlers (as opposed to multiple event 3194asynchronous sources such as signal handlers (as opposed to multiple event
3035loops - those are of course safe to use in different threads). 3195loops - those are of course safe to use in different threads).
3036 3196
3037Sometimes, however, you need to wake up an event loop you do not control, 3197Sometimes, however, you need to wake up an event loop you do not control,
3038for example because it belongs to another thread. This is what C<ev_async> 3198for example because it belongs to another thread. This is what C<ev_async>
3040it by calling C<ev_async_send>, which is thread- and signal safe. 3200it by calling C<ev_async_send>, which is thread- and signal safe.
3041 3201
3042This functionality is very similar to C<ev_signal> watchers, as signals, 3202This functionality is very similar to C<ev_signal> watchers, as signals,
3043too, are asynchronous in nature, and signals, too, will be compressed 3203too, are asynchronous in nature, and signals, too, will be compressed
3044(i.e. the number of callback invocations may be less than the number of 3204(i.e. the number of callback invocations may be less than the number of
3045C<ev_async_sent> calls). 3205C<ev_async_sent> calls). In fact, you could use signal watchers as a kind
3206of "global async watchers" by using a watcher on an otherwise unused
3207signal, and C<ev_feed_signal> to signal this watcher from another thread,
3208even without knowing which loop owns the signal.
3046 3209
3047Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not 3210Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
3048just the default loop. 3211just the default loop.
3049 3212
3050=head3 Queueing 3213=head3 Queueing
3145trust me. 3308trust me.
3146 3309
3147=item ev_async_send (loop, ev_async *) 3310=item ev_async_send (loop, ev_async *)
3148 3311
3149Sends/signals/activates the given C<ev_async> watcher, that is, feeds 3312Sends/signals/activates the given C<ev_async> watcher, that is, feeds
3150an C<EV_ASYNC> event on the watcher into the event loop. Unlike 3313an C<EV_ASYNC> event on the watcher into the event loop, and instantly
3314returns.
3315
3151C<ev_feed_event>, this call is safe to do from other threads, signal or 3316Unlike C<ev_feed_event>, this call is safe to do from other threads,
3152similar contexts (see the discussion of C<EV_ATOMIC_T> in the embedding 3317signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
3153section below on what exactly this means). 3318embedding section below on what exactly this means).
3154 3319
3155Note that, as with other watchers in libev, multiple events might get 3320Note that, as with other watchers in libev, multiple events might get
3156compressed into a single callback invocation (another way to look at this 3321compressed into a single callback invocation (another way to look at this
3157is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>, 3322is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
3158reset when the event loop detects that). 3323reset when the event loop detects that).
3226Feed an event on the given fd, as if a file descriptor backend detected 3391Feed an event on the given fd, as if a file descriptor backend detected
3227the given events it. 3392the given events it.
3228 3393
3229=item ev_feed_signal_event (loop, int signum) 3394=item ev_feed_signal_event (loop, int signum)
3230 3395
3231Feed an event as if the given signal occurred (C<loop> must be the default 3396Feed an event as if the given signal occurred. See also C<ev_feed_signal>,
3232loop!). 3397which is async-safe.
3233 3398
3234=back 3399=back
3400
3401
3402=head1 COMMON OR USEFUL IDIOMS (OR BOTH)
3403
3404This section explains some common idioms that are not immediately
3405obvious. Note that examples are sprinkled over the whole manual, and this
3406section only contains stuff that wouldn't fit anywhere else.
3407
3408=head2 ASSOCIATING CUSTOM DATA WITH A WATCHER
3409
3410Each watcher has, by default, a C<void *data> member that you can read
3411or modify at any time: libev will completely ignore it. This can be used
3412to associate arbitrary data with your watcher. If you need more data and
3413don't want to allocate memory separately and store a pointer to it in that
3414data member, you can also "subclass" the watcher type and provide your own
3415data:
3416
3417 struct my_io
3418 {
3419 ev_io io;
3420 int otherfd;
3421 void *somedata;
3422 struct whatever *mostinteresting;
3423 };
3424
3425 ...
3426 struct my_io w;
3427 ev_io_init (&w.io, my_cb, fd, EV_READ);
3428
3429And since your callback will be called with a pointer to the watcher, you
3430can cast it back to your own type:
3431
3432 static void my_cb (struct ev_loop *loop, ev_io *w_, int revents)
3433 {
3434 struct my_io *w = (struct my_io *)w_;
3435 ...
3436 }
3437
3438More interesting and less C-conformant ways of casting your callback
3439function type instead have been omitted.
3440
3441=head2 BUILDING YOUR OWN COMPOSITE WATCHERS
3442
3443Another common scenario is to use some data structure with multiple
3444embedded watchers, in effect creating your own watcher that combines
3445multiple libev event sources into one "super-watcher":
3446
3447 struct my_biggy
3448 {
3449 int some_data;
3450 ev_timer t1;
3451 ev_timer t2;
3452 }
3453
3454In this case getting the pointer to C<my_biggy> is a bit more
3455complicated: Either you store the address of your C<my_biggy> struct in
3456the C<data> member of the watcher (for woozies or C++ coders), or you need
3457to use some pointer arithmetic using C<offsetof> inside your watchers (for
3458real programmers):
3459
3460 #include <stddef.h>
3461
3462 static void
3463 t1_cb (EV_P_ ev_timer *w, int revents)
3464 {
3465 struct my_biggy big = (struct my_biggy *)
3466 (((char *)w) - offsetof (struct my_biggy, t1));
3467 }
3468
3469 static void
3470 t2_cb (EV_P_ ev_timer *w, int revents)
3471 {
3472 struct my_biggy big = (struct my_biggy *)
3473 (((char *)w) - offsetof (struct my_biggy, t2));
3474 }
3475
3476=head2 MODEL/NESTED EVENT LOOP INVOCATIONS AND EXIT CONDITIONS
3477
3478Often (especially in GUI toolkits) there are places where you have
3479I<modal> interaction, which is most easily implemented by recursively
3480invoking C<ev_run>.
3481
3482This brings the problem of exiting - a callback might want to finish the
3483main C<ev_run> call, but not the nested one (e.g. user clicked "Quit", but
3484a modal "Are you sure?" dialog is still waiting), or just the nested one
3485and not the main one (e.g. user clocked "Ok" in a modal dialog), or some
3486other combination: In these cases, C<ev_break> will not work alone.
3487
3488The solution is to maintain "break this loop" variable for each C<ev_run>
3489invocation, and use a loop around C<ev_run> until the condition is
3490triggered, using C<EVRUN_ONCE>:
3491
3492 // main loop
3493 int exit_main_loop = 0;
3494
3495 while (!exit_main_loop)
3496 ev_run (EV_DEFAULT_ EVRUN_ONCE);
3497
3498 // in a model watcher
3499 int exit_nested_loop = 0;
3500
3501 while (!exit_nested_loop)
3502 ev_run (EV_A_ EVRUN_ONCE);
3503
3504To exit from any of these loops, just set the corresponding exit variable:
3505
3506 // exit modal loop
3507 exit_nested_loop = 1;
3508
3509 // exit main program, after modal loop is finished
3510 exit_main_loop = 1;
3511
3512 // exit both
3513 exit_main_loop = exit_nested_loop = 1;
3514
3515=head2 THREAD LOCKING EXAMPLE
3516
3517Here is a fictitious example of how to run an event loop in a different
3518thread from where callbacks are being invoked and watchers are
3519created/added/removed.
3520
3521For a real-world example, see the C<EV::Loop::Async> perl module,
3522which uses exactly this technique (which is suited for many high-level
3523languages).
3524
3525The example uses a pthread mutex to protect the loop data, a condition
3526variable to wait for callback invocations, an async watcher to notify the
3527event loop thread and an unspecified mechanism to wake up the main thread.
3528
3529First, you need to associate some data with the event loop:
3530
3531 typedef struct {
3532 mutex_t lock; /* global loop lock */
3533 ev_async async_w;
3534 thread_t tid;
3535 cond_t invoke_cv;
3536 } userdata;
3537
3538 void prepare_loop (EV_P)
3539 {
3540 // for simplicity, we use a static userdata struct.
3541 static userdata u;
3542
3543 ev_async_init (&u->async_w, async_cb);
3544 ev_async_start (EV_A_ &u->async_w);
3545
3546 pthread_mutex_init (&u->lock, 0);
3547 pthread_cond_init (&u->invoke_cv, 0);
3548
3549 // now associate this with the loop
3550 ev_set_userdata (EV_A_ u);
3551 ev_set_invoke_pending_cb (EV_A_ l_invoke);
3552 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
3553
3554 // then create the thread running ev_run
3555 pthread_create (&u->tid, 0, l_run, EV_A);
3556 }
3557
3558The callback for the C<ev_async> watcher does nothing: the watcher is used
3559solely to wake up the event loop so it takes notice of any new watchers
3560that might have been added:
3561
3562 static void
3563 async_cb (EV_P_ ev_async *w, int revents)
3564 {
3565 // just used for the side effects
3566 }
3567
3568The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
3569protecting the loop data, respectively.
3570
3571 static void
3572 l_release (EV_P)
3573 {
3574 userdata *u = ev_userdata (EV_A);
3575 pthread_mutex_unlock (&u->lock);
3576 }
3577
3578 static void
3579 l_acquire (EV_P)
3580 {
3581 userdata *u = ev_userdata (EV_A);
3582 pthread_mutex_lock (&u->lock);
3583 }
3584
3585The event loop thread first acquires the mutex, and then jumps straight
3586into C<ev_run>:
3587
3588 void *
3589 l_run (void *thr_arg)
3590 {
3591 struct ev_loop *loop = (struct ev_loop *)thr_arg;
3592
3593 l_acquire (EV_A);
3594 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
3595 ev_run (EV_A_ 0);
3596 l_release (EV_A);
3597
3598 return 0;
3599 }
3600
3601Instead of invoking all pending watchers, the C<l_invoke> callback will
3602signal the main thread via some unspecified mechanism (signals? pipe
3603writes? C<Async::Interrupt>?) and then waits until all pending watchers
3604have been called (in a while loop because a) spurious wakeups are possible
3605and b) skipping inter-thread-communication when there are no pending
3606watchers is very beneficial):
3607
3608 static void
3609 l_invoke (EV_P)
3610 {
3611 userdata *u = ev_userdata (EV_A);
3612
3613 while (ev_pending_count (EV_A))
3614 {
3615 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
3616 pthread_cond_wait (&u->invoke_cv, &u->lock);
3617 }
3618 }
3619
3620Now, whenever the main thread gets told to invoke pending watchers, it
3621will grab the lock, call C<ev_invoke_pending> and then signal the loop
3622thread to continue:
3623
3624 static void
3625 real_invoke_pending (EV_P)
3626 {
3627 userdata *u = ev_userdata (EV_A);
3628
3629 pthread_mutex_lock (&u->lock);
3630 ev_invoke_pending (EV_A);
3631 pthread_cond_signal (&u->invoke_cv);
3632 pthread_mutex_unlock (&u->lock);
3633 }
3634
3635Whenever you want to start/stop a watcher or do other modifications to an
3636event loop, you will now have to lock:
3637
3638 ev_timer timeout_watcher;
3639 userdata *u = ev_userdata (EV_A);
3640
3641 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
3642
3643 pthread_mutex_lock (&u->lock);
3644 ev_timer_start (EV_A_ &timeout_watcher);
3645 ev_async_send (EV_A_ &u->async_w);
3646 pthread_mutex_unlock (&u->lock);
3647
3648Note that sending the C<ev_async> watcher is required because otherwise
3649an event loop currently blocking in the kernel will have no knowledge
3650about the newly added timer. By waking up the loop it will pick up any new
3651watchers in the next event loop iteration.
3652
3653=head2 THREADS, COROUTINES, CONTINUATIONS, QUEUES... INSTEAD OF CALLBACKS
3654
3655While the overhead of a callback that e.g. schedules a thread is small, it
3656is still an overhead. If you embed libev, and your main usage is with some
3657kind of threads or coroutines, you might want to customise libev so that
3658doesn't need callbacks anymore.
3659
3660Imagine you have coroutines that you can switch to using a function
3661C<switch_to (coro)>, that libev runs in a coroutine called C<libev_coro>
3662and that due to some magic, the currently active coroutine is stored in a
3663global called C<current_coro>. Then you can build your own "wait for libev
3664event" primitive by changing C<EV_CB_DECLARE> and C<EV_CB_INVOKE> (note
3665the differing C<;> conventions):
3666
3667 #define EV_CB_DECLARE(type) struct my_coro *cb;
3668 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb)
3669
3670That means instead of having a C callback function, you store the
3671coroutine to switch to in each watcher, and instead of having libev call
3672your callback, you instead have it switch to that coroutine.
3673
3674A coroutine might now wait for an event with a function called
3675C<wait_for_event>. (the watcher needs to be started, as always, but it doesn't
3676matter when, or whether the watcher is active or not when this function is
3677called):
3678
3679 void
3680 wait_for_event (ev_watcher *w)
3681 {
3682 ev_cb_set (w) = current_coro;
3683 switch_to (libev_coro);
3684 }
3685
3686That basically suspends the coroutine inside C<wait_for_event> and
3687continues the libev coroutine, which, when appropriate, switches back to
3688this or any other coroutine. I am sure if you sue this your own :)
3689
3690You can do similar tricks if you have, say, threads with an event queue -
3691instead of storing a coroutine, you store the queue object and instead of
3692switching to a coroutine, you push the watcher onto the queue and notify
3693any waiters.
3694
3695To embed libev, see L<EMBEDDING>, but in short, it's easiest to create two
3696files, F<my_ev.h> and F<my_ev.c> that include the respective libev files:
3697
3698 // my_ev.h
3699 #define EV_CB_DECLARE(type) struct my_coro *cb;
3700 #define EV_CB_INVOKE(watcher) switch_to ((watcher)->cb);
3701 #include "../libev/ev.h"
3702
3703 // my_ev.c
3704 #define EV_H "my_ev.h"
3705 #include "../libev/ev.c"
3706
3707And then use F<my_ev.h> when you would normally use F<ev.h>, and compile
3708F<my_ev.c> into your project. When properly specifying include paths, you
3709can even use F<ev.h> as header file name directly.
3235 3710
3236 3711
3237=head1 LIBEVENT EMULATION 3712=head1 LIBEVENT EMULATION
3238 3713
3239Libev offers a compatibility emulation layer for libevent. It cannot 3714Libev offers a compatibility emulation layer for libevent. It cannot
3240emulate the internals of libevent, so here are some usage hints: 3715emulate the internals of libevent, so here are some usage hints:
3241 3716
3242=over 4 3717=over 4
3718
3719=item * Only the libevent-1.4.1-beta API is being emulated.
3720
3721This was the newest libevent version available when libev was implemented,
3722and is still mostly unchanged in 2010.
3243 3723
3244=item * Use it by including <event.h>, as usual. 3724=item * Use it by including <event.h>, as usual.
3245 3725
3246=item * The following members are fully supported: ev_base, ev_callback, 3726=item * The following members are fully supported: ev_base, ev_callback,
3247ev_arg, ev_fd, ev_res, ev_events. 3727ev_arg, ev_fd, ev_res, ev_events.
3253=item * Priorities are not currently supported. Initialising priorities 3733=item * Priorities are not currently supported. Initialising priorities
3254will fail and all watchers will have the same priority, even though there 3734will fail and all watchers will have the same priority, even though there
3255is an ev_pri field. 3735is an ev_pri field.
3256 3736
3257=item * In libevent, the last base created gets the signals, in libev, the 3737=item * In libevent, the last base created gets the signals, in libev, the
3258first base created (== the default loop) gets the signals. 3738base that registered the signal gets the signals.
3259 3739
3260=item * Other members are not supported. 3740=item * Other members are not supported.
3261 3741
3262=item * The libev emulation is I<not> ABI compatible to libevent, you need 3742=item * The libev emulation is I<not> ABI compatible to libevent, you need
3263to use the libev header file and library. 3743to use the libev header file and library.
3282Care has been taken to keep the overhead low. The only data member the C++ 3762Care has been taken to keep the overhead low. The only data member the C++
3283classes add (compared to plain C-style watchers) is the event loop pointer 3763classes add (compared to plain C-style watchers) is the event loop pointer
3284that the watcher is associated with (or no additional members at all if 3764that the watcher is associated with (or no additional members at all if
3285you disable C<EV_MULTIPLICITY> when embedding libev). 3765you disable C<EV_MULTIPLICITY> when embedding libev).
3286 3766
3287Currently, functions, and static and non-static member functions can be 3767Currently, functions, static and non-static member functions and classes
3288used as callbacks. Other types should be easy to add as long as they only 3768with C<operator ()> can be used as callbacks. Other types should be easy
3289need one additional pointer for context. If you need support for other 3769to add as long as they only need one additional pointer for context. If
3290types of functors please contact the author (preferably after implementing 3770you need support for other types of functors please contact the author
3291it). 3771(preferably after implementing it).
3292 3772
3293Here is a list of things available in the C<ev> namespace: 3773Here is a list of things available in the C<ev> namespace:
3294 3774
3295=over 4 3775=over 4
3296 3776
4164And a F<ev_cpp.C> implementation file that contains libev proper and is compiled: 4644And a F<ev_cpp.C> implementation file that contains libev proper and is compiled:
4165 4645
4166 #include "ev_cpp.h" 4646 #include "ev_cpp.h"
4167 #include "ev.c" 4647 #include "ev.c"
4168 4648
4169=head1 INTERACTION WITH OTHER PROGRAMS OR LIBRARIES 4649=head1 INTERACTION WITH OTHER PROGRAMS, LIBRARIES OR THE ENVIRONMENT
4170 4650
4171=head2 THREADS AND COROUTINES 4651=head2 THREADS AND COROUTINES
4172 4652
4173=head3 THREADS 4653=head3 THREADS
4174 4654
4225default loop and triggering an C<ev_async> watcher from the default loop 4705default loop and triggering an C<ev_async> watcher from the default loop
4226watcher callback into the event loop interested in the signal. 4706watcher callback into the event loop interested in the signal.
4227 4707
4228=back 4708=back
4229 4709
4230=head4 THREAD LOCKING EXAMPLE 4710See also L<THREAD LOCKING EXAMPLE>.
4231
4232Here is a fictitious example of how to run an event loop in a different
4233thread than where callbacks are being invoked and watchers are
4234created/added/removed.
4235
4236For a real-world example, see the C<EV::Loop::Async> perl module,
4237which uses exactly this technique (which is suited for many high-level
4238languages).
4239
4240The example uses a pthread mutex to protect the loop data, a condition
4241variable to wait for callback invocations, an async watcher to notify the
4242event loop thread and an unspecified mechanism to wake up the main thread.
4243
4244First, you need to associate some data with the event loop:
4245
4246 typedef struct {
4247 mutex_t lock; /* global loop lock */
4248 ev_async async_w;
4249 thread_t tid;
4250 cond_t invoke_cv;
4251 } userdata;
4252
4253 void prepare_loop (EV_P)
4254 {
4255 // for simplicity, we use a static userdata struct.
4256 static userdata u;
4257
4258 ev_async_init (&u->async_w, async_cb);
4259 ev_async_start (EV_A_ &u->async_w);
4260
4261 pthread_mutex_init (&u->lock, 0);
4262 pthread_cond_init (&u->invoke_cv, 0);
4263
4264 // now associate this with the loop
4265 ev_set_userdata (EV_A_ u);
4266 ev_set_invoke_pending_cb (EV_A_ l_invoke);
4267 ev_set_loop_release_cb (EV_A_ l_release, l_acquire);
4268
4269 // then create the thread running ev_loop
4270 pthread_create (&u->tid, 0, l_run, EV_A);
4271 }
4272
4273The callback for the C<ev_async> watcher does nothing: the watcher is used
4274solely to wake up the event loop so it takes notice of any new watchers
4275that might have been added:
4276
4277 static void
4278 async_cb (EV_P_ ev_async *w, int revents)
4279 {
4280 // just used for the side effects
4281 }
4282
4283The C<l_release> and C<l_acquire> callbacks simply unlock/lock the mutex
4284protecting the loop data, respectively.
4285
4286 static void
4287 l_release (EV_P)
4288 {
4289 userdata *u = ev_userdata (EV_A);
4290 pthread_mutex_unlock (&u->lock);
4291 }
4292
4293 static void
4294 l_acquire (EV_P)
4295 {
4296 userdata *u = ev_userdata (EV_A);
4297 pthread_mutex_lock (&u->lock);
4298 }
4299
4300The event loop thread first acquires the mutex, and then jumps straight
4301into C<ev_run>:
4302
4303 void *
4304 l_run (void *thr_arg)
4305 {
4306 struct ev_loop *loop = (struct ev_loop *)thr_arg;
4307
4308 l_acquire (EV_A);
4309 pthread_setcanceltype (PTHREAD_CANCEL_ASYNCHRONOUS, 0);
4310 ev_run (EV_A_ 0);
4311 l_release (EV_A);
4312
4313 return 0;
4314 }
4315
4316Instead of invoking all pending watchers, the C<l_invoke> callback will
4317signal the main thread via some unspecified mechanism (signals? pipe
4318writes? C<Async::Interrupt>?) and then waits until all pending watchers
4319have been called (in a while loop because a) spurious wakeups are possible
4320and b) skipping inter-thread-communication when there are no pending
4321watchers is very beneficial):
4322
4323 static void
4324 l_invoke (EV_P)
4325 {
4326 userdata *u = ev_userdata (EV_A);
4327
4328 while (ev_pending_count (EV_A))
4329 {
4330 wake_up_other_thread_in_some_magic_or_not_so_magic_way ();
4331 pthread_cond_wait (&u->invoke_cv, &u->lock);
4332 }
4333 }
4334
4335Now, whenever the main thread gets told to invoke pending watchers, it
4336will grab the lock, call C<ev_invoke_pending> and then signal the loop
4337thread to continue:
4338
4339 static void
4340 real_invoke_pending (EV_P)
4341 {
4342 userdata *u = ev_userdata (EV_A);
4343
4344 pthread_mutex_lock (&u->lock);
4345 ev_invoke_pending (EV_A);
4346 pthread_cond_signal (&u->invoke_cv);
4347 pthread_mutex_unlock (&u->lock);
4348 }
4349
4350Whenever you want to start/stop a watcher or do other modifications to an
4351event loop, you will now have to lock:
4352
4353 ev_timer timeout_watcher;
4354 userdata *u = ev_userdata (EV_A);
4355
4356 ev_timer_init (&timeout_watcher, timeout_cb, 5.5, 0.);
4357
4358 pthread_mutex_lock (&u->lock);
4359 ev_timer_start (EV_A_ &timeout_watcher);
4360 ev_async_send (EV_A_ &u->async_w);
4361 pthread_mutex_unlock (&u->lock);
4362
4363Note that sending the C<ev_async> watcher is required because otherwise
4364an event loop currently blocking in the kernel will have no knowledge
4365about the newly added timer. By waking up the loop it will pick up any new
4366watchers in the next event loop iteration.
4367 4711
4368=head3 COROUTINES 4712=head3 COROUTINES
4369 4713
4370Libev is very accommodating to coroutines ("cooperative threads"): 4714Libev is very accommodating to coroutines ("cooperative threads"):
4371libev fully supports nesting calls to its functions from different 4715libev fully supports nesting calls to its functions from different
4467=head3 C<kqueue> is buggy 4811=head3 C<kqueue> is buggy
4468 4812
4469The kqueue syscall is broken in all known versions - most versions support 4813The kqueue syscall is broken in all known versions - most versions support
4470only sockets, many support pipes. 4814only sockets, many support pipes.
4471 4815
4472Libev tries to work around this by not using C<kqueue> by default on 4816Libev tries to work around this by not using C<kqueue> by default on this
4473this rotten platform, but of course you can still ask for it when creating 4817rotten platform, but of course you can still ask for it when creating a
4474a loop. 4818loop - embedding a socket-only kqueue loop into a select-based one is
4819probably going to work well.
4475 4820
4476=head3 C<poll> is buggy 4821=head3 C<poll> is buggy
4477 4822
4478Instead of fixing C<kqueue>, Apple replaced their (working) C<poll> 4823Instead of fixing C<kqueue>, Apple replaced their (working) C<poll>
4479implementation by something calling C<kqueue> internally around the 10.5.6 4824implementation by something calling C<kqueue> internally around the 10.5.6
4498 4843
4499=head3 C<errno> reentrancy 4844=head3 C<errno> reentrancy
4500 4845
4501The default compile environment on Solaris is unfortunately so 4846The default compile environment on Solaris is unfortunately so
4502thread-unsafe that you can't even use components/libraries compiled 4847thread-unsafe that you can't even use components/libraries compiled
4503without C<-D_REENTRANT> (as long as they use C<errno>), which, of course, 4848without C<-D_REENTRANT> in a threaded program, which, of course, isn't
4504isn't defined by default. 4849defined by default. A valid, if stupid, implementation choice.
4505 4850
4506If you want to use libev in threaded environments you have to make sure 4851If you want to use libev in threaded environments you have to make sure
4507it's compiled with C<_REENTRANT> defined. 4852it's compiled with C<_REENTRANT> defined.
4508 4853
4509=head3 Event port backend 4854=head3 Event port backend
4510 4855
4511The scalable event interface for Solaris is called "event ports". Unfortunately, 4856The scalable event interface for Solaris is called "event
4512this mechanism is very buggy. If you run into high CPU usage, your program 4857ports". Unfortunately, this mechanism is very buggy in all major
4858releases. If you run into high CPU usage, your program freezes or you get
4513freezes or you get a large number of spurious wakeups, make sure you have 4859a large number of spurious wakeups, make sure you have all the relevant
4514all the relevant and latest kernel patches applied. No, I don't know which 4860and latest kernel patches applied. No, I don't know which ones, but there
4515ones, but there are multiple ones. 4861are multiple ones to apply, and afterwards, event ports actually work
4862great.
4516 4863
4517If you can't get it to work, you can try running the program by setting 4864If you can't get it to work, you can try running the program by setting
4518the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and 4865the environment variable C<LIBEV_FLAGS=3> to only allow C<poll> and
4519C<select> backends. 4866C<select> backends.
4520 4867
4521=head2 AIX POLL BUG 4868=head2 AIX POLL BUG
4522 4869
4523AIX unfortunately has a broken C<poll.h> header. Libev works around 4870AIX unfortunately has a broken C<poll.h> header. Libev works around
4524this by trying to avoid the poll backend altogether (i.e. it's not even 4871this by trying to avoid the poll backend altogether (i.e. it's not even
4525compiled in), which normally isn't a big problem as C<select> works fine 4872compiled in), which normally isn't a big problem as C<select> works fine
4526with large bitsets, and AIX is dead anyway. 4873with large bitsets on AIX, and AIX is dead anyway.
4527 4874
4528=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS 4875=head2 WIN32 PLATFORM LIMITATIONS AND WORKAROUNDS
4529 4876
4530=head3 General issues 4877=head3 General issues
4531 4878
4637structure (guaranteed by POSIX but not by ISO C for example), but it also 4984structure (guaranteed by POSIX but not by ISO C for example), but it also
4638assumes that the same (machine) code can be used to call any watcher 4985assumes that the same (machine) code can be used to call any watcher
4639callback: The watcher callbacks have different type signatures, but libev 4986callback: The watcher callbacks have different type signatures, but libev
4640calls them using an C<ev_watcher *> internally. 4987calls them using an C<ev_watcher *> internally.
4641 4988
4989=item pointer accesses must be thread-atomic
4990
4991Accessing a pointer value must be atomic, it must both be readable and
4992writable in one piece - this is the case on all current architectures.
4993
4642=item C<sig_atomic_t volatile> must be thread-atomic as well 4994=item C<sig_atomic_t volatile> must be thread-atomic as well
4643 4995
4644The type C<sig_atomic_t volatile> (or whatever is defined as 4996The type C<sig_atomic_t volatile> (or whatever is defined as
4645C<EV_ATOMIC_T>) must be atomic with respect to accesses from different 4997C<EV_ATOMIC_T>) must be atomic with respect to accesses from different
4646threads. This is not part of the specification for C<sig_atomic_t>, but is 4998threads. This is not part of the specification for C<sig_atomic_t>, but is
4752=back 5104=back
4753 5105
4754 5106
4755=head1 PORTING FROM LIBEV 3.X TO 4.X 5107=head1 PORTING FROM LIBEV 3.X TO 4.X
4756 5108
4757The major version 4 introduced some minor incompatible changes to the API. 5109The major version 4 introduced some incompatible changes to the API.
4758 5110
4759At the moment, the C<ev.h> header file tries to implement superficial 5111At the moment, the C<ev.h> header file provides compatibility definitions
4760compatibility, so most programs should still compile. Those might be 5112for all changes, so most programs should still compile. The compatibility
4761removed in later versions of libev, so better update early than late. 5113layer might be removed in later versions of libev, so better update to the
5114new API early than late.
4762 5115
4763=over 4 5116=over 4
5117
5118=item C<EV_COMPAT3> backwards compatibility mechanism
5119
5120The backward compatibility mechanism can be controlled by
5121C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
5122section.
5123
5124=item C<ev_default_destroy> and C<ev_default_fork> have been removed
5125
5126These calls can be replaced easily by their C<ev_loop_xxx> counterparts:
5127
5128 ev_loop_destroy (EV_DEFAULT_UC);
5129 ev_loop_fork (EV_DEFAULT);
4764 5130
4765=item function/symbol renames 5131=item function/symbol renames
4766 5132
4767A number of functions and symbols have been renamed: 5133A number of functions and symbols have been renamed:
4768 5134
4787ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme 5153ev_loop> anymore and C<EV_TIMER> now follows the same naming scheme
4788as all other watcher types. Note that C<ev_loop_fork> is still called 5154as all other watcher types. Note that C<ev_loop_fork> is still called
4789C<ev_loop_fork> because it would otherwise clash with the C<ev_fork> 5155C<ev_loop_fork> because it would otherwise clash with the C<ev_fork>
4790typedef. 5156typedef.
4791 5157
4792=item C<EV_COMPAT3> backwards compatibility mechanism
4793
4794The backward compatibility mechanism can be controlled by
4795C<EV_COMPAT3>. See L<PREPROCESSOR SYMBOLS/MACROS> in the L<EMBEDDING>
4796section.
4797
4798=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES> 5158=item C<EV_MINIMAL> mechanism replaced by C<EV_FEATURES>
4799 5159
4800The preprocessor symbol C<EV_MINIMAL> has been replaced by a different 5160The preprocessor symbol C<EV_MINIMAL> has been replaced by a different
4801mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile 5161mechanism, C<EV_FEATURES>. Programs using C<EV_MINIMAL> usually compile
4802and work, but the library code will of course be larger. 5162and work, but the library code will of course be larger.
4808 5168
4809=over 4 5169=over 4
4810 5170
4811=item active 5171=item active
4812 5172
4813A watcher is active as long as it has been started (has been attached to 5173A watcher is active as long as it has been started and not yet stopped.
4814an event loop) but not yet stopped (disassociated from the event loop). 5174See L<WATCHER STATES> for details.
4815 5175
4816=item application 5176=item application
4817 5177
4818In this document, an application is whatever is using libev. 5178In this document, an application is whatever is using libev.
5179
5180=item backend
5181
5182The part of the code dealing with the operating system interfaces.
4819 5183
4820=item callback 5184=item callback
4821 5185
4822The address of a function that is called when some event has been 5186The address of a function that is called when some event has been
4823detected. Callbacks are being passed the event loop, the watcher that 5187detected. Callbacks are being passed the event loop, the watcher that
4824received the event, and the actual event bitset. 5188received the event, and the actual event bitset.
4825 5189
4826=item callback invocation 5190=item callback/watcher invocation
4827 5191
4828The act of calling the callback associated with a watcher. 5192The act of calling the callback associated with a watcher.
4829 5193
4830=item event 5194=item event
4831 5195
4850The model used to describe how an event loop handles and processes 5214The model used to describe how an event loop handles and processes
4851watchers and events. 5215watchers and events.
4852 5216
4853=item pending 5217=item pending
4854 5218
4855A watcher is pending as soon as the corresponding event has been detected, 5219A watcher is pending as soon as the corresponding event has been
4856and stops being pending as soon as the watcher will be invoked or its 5220detected. See L<WATCHER STATES> for details.
4857pending status is explicitly cleared by the application.
4858
4859A watcher can be pending, but not active. Stopping a watcher also clears
4860its pending status.
4861 5221
4862=item real time 5222=item real time
4863 5223
4864The physical time that is observed. It is apparently strictly monotonic :) 5224The physical time that is observed. It is apparently strictly monotonic :)
4865 5225
4866=item wall-clock time 5226=item wall-clock time
4867 5227
4868The time and date as shown on clocks. Unlike real time, it can actually 5228The time and date as shown on clocks. Unlike real time, it can actually
4869be wrong and jump forwards and backwards, e.g. when the you adjust your 5229be wrong and jump forwards and backwards, e.g. when you adjust your
4870clock. 5230clock.
4871 5231
4872=item watcher 5232=item watcher
4873 5233
4874A data structure that describes interest in certain events. Watchers need 5234A data structure that describes interest in certain events. Watchers need
4875to be started (attached to an event loop) before they can receive events. 5235to be started (attached to an event loop) before they can receive events.
4876 5236
4877=item watcher invocation
4878
4879The act of calling the callback associated with a watcher.
4880
4881=back 5237=back
4882 5238
4883=head1 AUTHOR 5239=head1 AUTHOR
4884 5240
4885Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael Magnusson. 5241Marc Lehmann <libev@schmorp.de>, with repeated corrections by Mikael
5242Magnusson and Emanuele Giaquinta, and minor corrections by many others.
4886 5243

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines