… | |
… | |
47 | |
47 | |
48 | return 0; |
48 | return 0; |
49 | } |
49 | } |
50 | |
50 | |
51 | =head1 DESCRIPTION |
51 | =head1 DESCRIPTION |
|
|
52 | |
|
|
53 | The newest version of this document is also available as a html-formatted |
|
|
54 | web page you might find easier to navigate when reading it for the first |
|
|
55 | time: L<http://cvs.schmorp.de/libev/ev.html>. |
52 | |
56 | |
53 | Libev is an event loop: you register interest in certain events (such as a |
57 | Libev is an event loop: you register interest in certain events (such as a |
54 | file descriptor being readable or a timeout occuring), and it will manage |
58 | file descriptor being readable or a timeout occuring), and it will manage |
55 | these event sources and provide your program with events. |
59 | these event sources and provide your program with events. |
56 | |
60 | |
… | |
… | |
113 | |
117 | |
114 | =item int ev_version_major () |
118 | =item int ev_version_major () |
115 | |
119 | |
116 | =item int ev_version_minor () |
120 | =item int ev_version_minor () |
117 | |
121 | |
118 | You can find out the major and minor version numbers of the library |
122 | You can find out the major and minor ABI version numbers of the library |
119 | you linked against by calling the functions C<ev_version_major> and |
123 | you linked against by calling the functions C<ev_version_major> and |
120 | C<ev_version_minor>. If you want, you can compare against the global |
124 | C<ev_version_minor>. If you want, you can compare against the global |
121 | symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the |
125 | symbols C<EV_VERSION_MAJOR> and C<EV_VERSION_MINOR>, which specify the |
122 | version of the library your program was compiled against. |
126 | version of the library your program was compiled against. |
123 | |
127 | |
|
|
128 | These version numbers refer to the ABI version of the library, not the |
|
|
129 | release version. |
|
|
130 | |
124 | Usually, it's a good idea to terminate if the major versions mismatch, |
131 | Usually, it's a good idea to terminate if the major versions mismatch, |
125 | as this indicates an incompatible change. Minor versions are usually |
132 | as this indicates an incompatible change. Minor versions are usually |
126 | compatible to older versions, so a larger minor version alone is usually |
133 | compatible to older versions, so a larger minor version alone is usually |
127 | not a problem. |
134 | not a problem. |
128 | |
135 | |
129 | Example: Make sure we haven't accidentally been linked against the wrong |
136 | Example: Make sure we haven't accidentally been linked against the wrong |
130 | version. |
137 | version. |
… | |
… | |
274 | a fork, you can also make libev check for a fork in each iteration by |
281 | a fork, you can also make libev check for a fork in each iteration by |
275 | enabling this flag. |
282 | enabling this flag. |
276 | |
283 | |
277 | This works by calling C<getpid ()> on every iteration of the loop, |
284 | This works by calling C<getpid ()> on every iteration of the loop, |
278 | and thus this might slow down your event loop if you do a lot of loop |
285 | and thus this might slow down your event loop if you do a lot of loop |
279 | iterations and little real work, but is usually not noticable (on my |
286 | iterations and little real work, but is usually not noticeable (on my |
280 | Linux system for example, C<getpid> is actually a simple 5-insn sequence |
287 | Linux system for example, C<getpid> is actually a simple 5-insn sequence |
281 | without a syscall and thus I<very> fast, but my Linux system also has |
288 | without a syscall and thus I<very> fast, but my Linux system also has |
282 | C<pthread_atfork> which is even faster). |
289 | C<pthread_atfork> which is even faster). |
283 | |
290 | |
284 | The big advantage of this flag is that you can forget about fork (and |
291 | The big advantage of this flag is that you can forget about fork (and |
… | |
… | |
430 | |
437 | |
431 | Like C<ev_default_fork>, but acts on an event loop created by |
438 | Like C<ev_default_fork>, but acts on an event loop created by |
432 | C<ev_loop_new>. Yes, you have to call this on every allocated event loop |
439 | C<ev_loop_new>. Yes, you have to call this on every allocated event loop |
433 | after fork, and how you do this is entirely your own problem. |
440 | after fork, and how you do this is entirely your own problem. |
434 | |
441 | |
|
|
442 | =item unsigned int ev_loop_count (loop) |
|
|
443 | |
|
|
444 | Returns the count of loop iterations for the loop, which is identical to |
|
|
445 | the number of times libev did poll for new events. It starts at C<0> and |
|
|
446 | happily wraps around with enough iterations. |
|
|
447 | |
|
|
448 | This value can sometimes be useful as a generation counter of sorts (it |
|
|
449 | "ticks" the number of loop iterations), as it roughly corresponds with |
|
|
450 | C<ev_prepare> and C<ev_check> calls. |
|
|
451 | |
435 | =item unsigned int ev_backend (loop) |
452 | =item unsigned int ev_backend (loop) |
436 | |
453 | |
437 | Returns one of the C<EVBACKEND_*> flags indicating the event backend in |
454 | Returns one of the C<EVBACKEND_*> flags indicating the event backend in |
438 | use. |
455 | use. |
439 | |
456 | |
… | |
… | |
472 | libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is |
489 | libev watchers. However, a pair of C<ev_prepare>/C<ev_check> watchers is |
473 | usually a better approach for this kind of thing. |
490 | usually a better approach for this kind of thing. |
474 | |
491 | |
475 | Here are the gory details of what C<ev_loop> does: |
492 | Here are the gory details of what C<ev_loop> does: |
476 | |
493 | |
|
|
494 | - Before the first iteration, call any pending watchers. |
477 | * If there are no active watchers (reference count is zero), return. |
495 | * If there are no active watchers (reference count is zero), return. |
478 | - Queue prepare watchers and then call all outstanding watchers. |
496 | - Queue all prepare watchers and then call all outstanding watchers. |
479 | - If we have been forked, recreate the kernel state. |
497 | - If we have been forked, recreate the kernel state. |
480 | - Update the kernel state with all outstanding changes. |
498 | - Update the kernel state with all outstanding changes. |
481 | - Update the "event loop time". |
499 | - Update the "event loop time". |
482 | - Calculate for how long to block. |
500 | - Calculate for how long to block. |
483 | - Block the process, waiting for any events. |
501 | - Block the process, waiting for any events. |
… | |
… | |
722 | =item bool ev_is_pending (ev_TYPE *watcher) |
740 | =item bool ev_is_pending (ev_TYPE *watcher) |
723 | |
741 | |
724 | Returns a true value iff the watcher is pending, (i.e. it has outstanding |
742 | Returns a true value iff the watcher is pending, (i.e. it has outstanding |
725 | events but its callback has not yet been invoked). As long as a watcher |
743 | events but its callback has not yet been invoked). As long as a watcher |
726 | is pending (but not active) you must not call an init function on it (but |
744 | is pending (but not active) you must not call an init function on it (but |
727 | C<ev_TYPE_set> is safe) and you must make sure the watcher is available to |
745 | C<ev_TYPE_set> is safe), you must not change its priority, and you must |
728 | libev (e.g. you cnanot C<free ()> it). |
746 | make sure the watcher is available to libev (e.g. you cannot C<free ()> |
|
|
747 | it). |
729 | |
748 | |
730 | =item callback ev_cb (ev_TYPE *watcher) |
749 | =item callback ev_cb (ev_TYPE *watcher) |
731 | |
750 | |
732 | Returns the callback currently set on the watcher. |
751 | Returns the callback currently set on the watcher. |
733 | |
752 | |
734 | =item ev_cb_set (ev_TYPE *watcher, callback) |
753 | =item ev_cb_set (ev_TYPE *watcher, callback) |
735 | |
754 | |
736 | Change the callback. You can change the callback at virtually any time |
755 | Change the callback. You can change the callback at virtually any time |
737 | (modulo threads). |
756 | (modulo threads). |
|
|
757 | |
|
|
758 | =item ev_set_priority (ev_TYPE *watcher, priority) |
|
|
759 | |
|
|
760 | =item int ev_priority (ev_TYPE *watcher) |
|
|
761 | |
|
|
762 | Set and query the priority of the watcher. The priority is a small |
|
|
763 | integer between C<EV_MAXPRI> (default: C<2>) and C<EV_MINPRI> |
|
|
764 | (default: C<-2>). Pending watchers with higher priority will be invoked |
|
|
765 | before watchers with lower priority, but priority will not keep watchers |
|
|
766 | from being executed (except for C<ev_idle> watchers). |
|
|
767 | |
|
|
768 | This means that priorities are I<only> used for ordering callback |
|
|
769 | invocation after new events have been received. This is useful, for |
|
|
770 | example, to reduce latency after idling, or more often, to bind two |
|
|
771 | watchers on the same event and make sure one is called first. |
|
|
772 | |
|
|
773 | If you need to suppress invocation when higher priority events are pending |
|
|
774 | you need to look at C<ev_idle> watchers, which provide this functionality. |
|
|
775 | |
|
|
776 | You I<must not> change the priority of a watcher as long as it is active or |
|
|
777 | pending. |
|
|
778 | |
|
|
779 | The default priority used by watchers when no priority has been set is |
|
|
780 | always C<0>, which is supposed to not be too high and not be too low :). |
|
|
781 | |
|
|
782 | Setting a priority outside the range of C<EV_MINPRI> to C<EV_MAXPRI> is |
|
|
783 | fine, as long as you do not mind that the priority value you query might |
|
|
784 | or might not have been adjusted to be within valid range. |
|
|
785 | |
|
|
786 | =item ev_invoke (loop, ev_TYPE *watcher, int revents) |
|
|
787 | |
|
|
788 | Invoke the C<watcher> with the given C<loop> and C<revents>. Neither |
|
|
789 | C<loop> nor C<revents> need to be valid as long as the watcher callback |
|
|
790 | can deal with that fact. |
|
|
791 | |
|
|
792 | =item int ev_clear_pending (loop, ev_TYPE *watcher) |
|
|
793 | |
|
|
794 | If the watcher is pending, this function returns clears its pending status |
|
|
795 | and returns its C<revents> bitset (as if its callback was invoked). If the |
|
|
796 | watcher isn't pending it does nothing and returns C<0>. |
738 | |
797 | |
739 | =back |
798 | =back |
740 | |
799 | |
741 | |
800 | |
742 | =head2 ASSOCIATING CUSTOM DATA WITH A WATCHER |
801 | =head2 ASSOCIATING CUSTOM DATA WITH A WATCHER |
… | |
… | |
848 | it is best to always use non-blocking I/O: An extra C<read>(2) returning |
907 | it is best to always use non-blocking I/O: An extra C<read>(2) returning |
849 | C<EAGAIN> is far preferable to a program hanging until some data arrives. |
908 | C<EAGAIN> is far preferable to a program hanging until some data arrives. |
850 | |
909 | |
851 | If you cannot run the fd in non-blocking mode (for example you should not |
910 | If you cannot run the fd in non-blocking mode (for example you should not |
852 | play around with an Xlib connection), then you have to seperately re-test |
911 | play around with an Xlib connection), then you have to seperately re-test |
853 | wether a file descriptor is really ready with a known-to-be good interface |
912 | whether a file descriptor is really ready with a known-to-be good interface |
854 | such as poll (fortunately in our Xlib example, Xlib already does this on |
913 | such as poll (fortunately in our Xlib example, Xlib already does this on |
855 | its own, so its quite safe to use). |
914 | its own, so its quite safe to use). |
|
|
915 | |
|
|
916 | =head3 The special problem of disappearing file descriptors |
|
|
917 | |
|
|
918 | Some backends (e.g kqueue, epoll) need to be told about closing a file |
|
|
919 | descriptor (either by calling C<close> explicitly or by any other means, |
|
|
920 | such as C<dup>). The reason is that you register interest in some file |
|
|
921 | descriptor, but when it goes away, the operating system will silently drop |
|
|
922 | this interest. If another file descriptor with the same number then is |
|
|
923 | registered with libev, there is no efficient way to see that this is, in |
|
|
924 | fact, a different file descriptor. |
|
|
925 | |
|
|
926 | To avoid having to explicitly tell libev about such cases, libev follows |
|
|
927 | the following policy: Each time C<ev_io_set> is being called, libev |
|
|
928 | will assume that this is potentially a new file descriptor, otherwise |
|
|
929 | it is assumed that the file descriptor stays the same. That means that |
|
|
930 | you I<have> to call C<ev_io_set> (or C<ev_io_init>) when you change the |
|
|
931 | descriptor even if the file descriptor number itself did not change. |
|
|
932 | |
|
|
933 | This is how one would do it normally anyway, the important point is that |
|
|
934 | the libev application should not optimise around libev but should leave |
|
|
935 | optimisations to libev. |
|
|
936 | |
856 | |
937 | |
857 | =over 4 |
938 | =over 4 |
858 | |
939 | |
859 | =item ev_io_init (ev_io *, callback, int fd, int events) |
940 | =item ev_io_init (ev_io *, callback, int fd, int events) |
860 | |
941 | |
… | |
… | |
1018 | but on wallclock time (absolute time). You can tell a periodic watcher |
1099 | but on wallclock time (absolute time). You can tell a periodic watcher |
1019 | to trigger "at" some specific point in time. For example, if you tell a |
1100 | to trigger "at" some specific point in time. For example, if you tell a |
1020 | periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now () |
1101 | periodic watcher to trigger in 10 seconds (by specifiying e.g. C<ev_now () |
1021 | + 10.>) and then reset your system clock to the last year, then it will |
1102 | + 10.>) and then reset your system clock to the last year, then it will |
1022 | take a year to trigger the event (unlike an C<ev_timer>, which would trigger |
1103 | take a year to trigger the event (unlike an C<ev_timer>, which would trigger |
1023 | roughly 10 seconds later and of course not if you reset your system time |
1104 | roughly 10 seconds later). |
1024 | again). |
|
|
1025 | |
1105 | |
1026 | They can also be used to implement vastly more complex timers, such as |
1106 | They can also be used to implement vastly more complex timers, such as |
1027 | triggering an event on eahc midnight, local time. |
1107 | triggering an event on each midnight, local time or other, complicated, |
|
|
1108 | rules. |
1028 | |
1109 | |
1029 | As with timers, the callback is guarenteed to be invoked only when the |
1110 | As with timers, the callback is guarenteed to be invoked only when the |
1030 | time (C<at>) has been passed, but if multiple periodic timers become ready |
1111 | time (C<at>) has been passed, but if multiple periodic timers become ready |
1031 | during the same loop iteration then order of execution is undefined. |
1112 | during the same loop iteration then order of execution is undefined. |
1032 | |
1113 | |
… | |
… | |
1039 | Lots of arguments, lets sort it out... There are basically three modes of |
1120 | Lots of arguments, lets sort it out... There are basically three modes of |
1040 | operation, and we will explain them from simplest to complex: |
1121 | operation, and we will explain them from simplest to complex: |
1041 | |
1122 | |
1042 | =over 4 |
1123 | =over 4 |
1043 | |
1124 | |
1044 | =item * absolute timer (interval = reschedule_cb = 0) |
1125 | =item * absolute timer (at = time, interval = reschedule_cb = 0) |
1045 | |
1126 | |
1046 | In this configuration the watcher triggers an event at the wallclock time |
1127 | In this configuration the watcher triggers an event at the wallclock time |
1047 | C<at> and doesn't repeat. It will not adjust when a time jump occurs, |
1128 | C<at> and doesn't repeat. It will not adjust when a time jump occurs, |
1048 | that is, if it is to be run at January 1st 2011 then it will run when the |
1129 | that is, if it is to be run at January 1st 2011 then it will run when the |
1049 | system time reaches or surpasses this time. |
1130 | system time reaches or surpasses this time. |
1050 | |
1131 | |
1051 | =item * non-repeating interval timer (interval > 0, reschedule_cb = 0) |
1132 | =item * non-repeating interval timer (at = offset, interval > 0, reschedule_cb = 0) |
1052 | |
1133 | |
1053 | In this mode the watcher will always be scheduled to time out at the next |
1134 | In this mode the watcher will always be scheduled to time out at the next |
1054 | C<at + N * interval> time (for some integer N) and then repeat, regardless |
1135 | C<at + N * interval> time (for some integer N, which can also be negative) |
1055 | of any time jumps. |
1136 | and then repeat, regardless of any time jumps. |
1056 | |
1137 | |
1057 | This can be used to create timers that do not drift with respect to system |
1138 | This can be used to create timers that do not drift with respect to system |
1058 | time: |
1139 | time: |
1059 | |
1140 | |
1060 | ev_periodic_set (&periodic, 0., 3600., 0); |
1141 | ev_periodic_set (&periodic, 0., 3600., 0); |
… | |
… | |
1066 | |
1147 | |
1067 | Another way to think about it (for the mathematically inclined) is that |
1148 | Another way to think about it (for the mathematically inclined) is that |
1068 | C<ev_periodic> will try to run the callback in this mode at the next possible |
1149 | C<ev_periodic> will try to run the callback in this mode at the next possible |
1069 | time where C<time = at (mod interval)>, regardless of any time jumps. |
1150 | time where C<time = at (mod interval)>, regardless of any time jumps. |
1070 | |
1151 | |
|
|
1152 | For numerical stability it is preferable that the C<at> value is near |
|
|
1153 | C<ev_now ()> (the current time), but there is no range requirement for |
|
|
1154 | this value. |
|
|
1155 | |
1071 | =item * manual reschedule mode (reschedule_cb = callback) |
1156 | =item * manual reschedule mode (at and interval ignored, reschedule_cb = callback) |
1072 | |
1157 | |
1073 | In this mode the values for C<interval> and C<at> are both being |
1158 | In this mode the values for C<interval> and C<at> are both being |
1074 | ignored. Instead, each time the periodic watcher gets scheduled, the |
1159 | ignored. Instead, each time the periodic watcher gets scheduled, the |
1075 | reschedule callback will be called with the watcher as first, and the |
1160 | reschedule callback will be called with the watcher as first, and the |
1076 | current time as second argument. |
1161 | current time as second argument. |
1077 | |
1162 | |
1078 | NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, |
1163 | NOTE: I<This callback MUST NOT stop or destroy any periodic watcher, |
1079 | ever, or make any event loop modifications>. If you need to stop it, |
1164 | ever, or make any event loop modifications>. If you need to stop it, |
1080 | return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by |
1165 | return C<now + 1e30> (or so, fudge fudge) and stop it afterwards (e.g. by |
1081 | starting a prepare watcher). |
1166 | starting an C<ev_prepare> watcher, which is legal). |
1082 | |
1167 | |
1083 | Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w, |
1168 | Its prototype is C<ev_tstamp (*reschedule_cb)(struct ev_periodic *w, |
1084 | ev_tstamp now)>, e.g.: |
1169 | ev_tstamp now)>, e.g.: |
1085 | |
1170 | |
1086 | static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now) |
1171 | static ev_tstamp my_rescheduler (struct ev_periodic *w, ev_tstamp now) |
… | |
… | |
1108 | |
1193 | |
1109 | Simply stops and restarts the periodic watcher again. This is only useful |
1194 | Simply stops and restarts the periodic watcher again. This is only useful |
1110 | when you changed some parameters or the reschedule callback would return |
1195 | when you changed some parameters or the reschedule callback would return |
1111 | a different time than the last time it was called (e.g. in a crond like |
1196 | a different time than the last time it was called (e.g. in a crond like |
1112 | program when the crontabs have changed). |
1197 | program when the crontabs have changed). |
|
|
1198 | |
|
|
1199 | =item ev_tstamp offset [read-write] |
|
|
1200 | |
|
|
1201 | When repeating, this contains the offset value, otherwise this is the |
|
|
1202 | absolute point in time (the C<at> value passed to C<ev_periodic_set>). |
|
|
1203 | |
|
|
1204 | Can be modified any time, but changes only take effect when the periodic |
|
|
1205 | timer fires or C<ev_periodic_again> is being called. |
1113 | |
1206 | |
1114 | =item ev_tstamp interval [read-write] |
1207 | =item ev_tstamp interval [read-write] |
1115 | |
1208 | |
1116 | The current interval value. Can be modified any time, but changes only |
1209 | The current interval value. Can be modified any time, but changes only |
1117 | take effect when the periodic timer fires or C<ev_periodic_again> is being |
1210 | take effect when the periodic timer fires or C<ev_periodic_again> is being |
… | |
… | |
1341 | ev_stat_start (loop, &passwd); |
1434 | ev_stat_start (loop, &passwd); |
1342 | |
1435 | |
1343 | |
1436 | |
1344 | =head2 C<ev_idle> - when you've got nothing better to do... |
1437 | =head2 C<ev_idle> - when you've got nothing better to do... |
1345 | |
1438 | |
1346 | Idle watchers trigger events when there are no other events are pending |
1439 | Idle watchers trigger events when no other events of the same or higher |
1347 | (prepare, check and other idle watchers do not count). That is, as long |
1440 | priority are pending (prepare, check and other idle watchers do not |
1348 | as your process is busy handling sockets or timeouts (or even signals, |
1441 | count). |
1349 | imagine) it will not be triggered. But when your process is idle all idle |
1442 | |
1350 | watchers are being called again and again, once per event loop iteration - |
1443 | That is, as long as your process is busy handling sockets or timeouts |
|
|
1444 | (or even signals, imagine) of the same or higher priority it will not be |
|
|
1445 | triggered. But when your process is idle (or only lower-priority watchers |
|
|
1446 | are pending), the idle watchers are being called once per event loop |
1351 | until stopped, that is, or your process receives more events and becomes |
1447 | iteration - until stopped, that is, or your process receives more events |
1352 | busy. |
1448 | and becomes busy again with higher priority stuff. |
1353 | |
1449 | |
1354 | The most noteworthy effect is that as long as any idle watchers are |
1450 | The most noteworthy effect is that as long as any idle watchers are |
1355 | active, the process will not block when waiting for new events. |
1451 | active, the process will not block when waiting for new events. |
1356 | |
1452 | |
1357 | Apart from keeping your process non-blocking (which is a useful |
1453 | Apart from keeping your process non-blocking (which is a useful |
… | |
… | |
1423 | with priority higher than or equal to the event loop and one coroutine |
1519 | with priority higher than or equal to the event loop and one coroutine |
1424 | of lower priority, but only once, using idle watchers to keep the event |
1520 | of lower priority, but only once, using idle watchers to keep the event |
1425 | loop from blocking if lower-priority coroutines are active, thus mapping |
1521 | loop from blocking if lower-priority coroutines are active, thus mapping |
1426 | low-priority coroutines to idle/background tasks). |
1522 | low-priority coroutines to idle/background tasks). |
1427 | |
1523 | |
|
|
1524 | It is recommended to give C<ev_check> watchers highest (C<EV_MAXPRI>) |
|
|
1525 | priority, to ensure that they are being run before any other watchers |
|
|
1526 | after the poll. Also, C<ev_check> watchers (and C<ev_prepare> watchers, |
|
|
1527 | too) should not activate ("feed") events into libev. While libev fully |
|
|
1528 | supports this, they will be called before other C<ev_check> watchers did |
|
|
1529 | their job. As C<ev_check> watchers are often used to embed other event |
|
|
1530 | loops those other event loops might be in an unusable state until their |
|
|
1531 | C<ev_check> watcher ran (always remind yourself to coexist peacefully with |
|
|
1532 | others). |
|
|
1533 | |
1428 | =over 4 |
1534 | =over 4 |
1429 | |
1535 | |
1430 | =item ev_prepare_init (ev_prepare *, callback) |
1536 | =item ev_prepare_init (ev_prepare *, callback) |
1431 | |
1537 | |
1432 | =item ev_check_init (ev_check *, callback) |
1538 | =item ev_check_init (ev_check *, callback) |
… | |
… | |
1435 | parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set> |
1541 | parameters of any kind. There are C<ev_prepare_set> and C<ev_check_set> |
1436 | macros, but using them is utterly, utterly and completely pointless. |
1542 | macros, but using them is utterly, utterly and completely pointless. |
1437 | |
1543 | |
1438 | =back |
1544 | =back |
1439 | |
1545 | |
1440 | Example: To include a library such as adns, you would add IO watchers |
1546 | There are a number of principal ways to embed other event loops or modules |
1441 | and a timeout watcher in a prepare handler, as required by libadns, and |
1547 | into libev. Here are some ideas on how to include libadns into libev |
|
|
1548 | (there is a Perl module named C<EV::ADNS> that does this, which you could |
|
|
1549 | use for an actually working example. Another Perl module named C<EV::Glib> |
|
|
1550 | embeds a Glib main context into libev, and finally, C<Glib::EV> embeds EV |
|
|
1551 | into the Glib event loop). |
|
|
1552 | |
|
|
1553 | Method 1: Add IO watchers and a timeout watcher in a prepare handler, |
1442 | in a check watcher, destroy them and call into libadns. What follows is |
1554 | and in a check watcher, destroy them and call into libadns. What follows |
1443 | pseudo-code only of course: |
1555 | is pseudo-code only of course. This requires you to either use a low |
|
|
1556 | priority for the check watcher or use C<ev_clear_pending> explicitly, as |
|
|
1557 | the callbacks for the IO/timeout watchers might not have been called yet. |
1444 | |
1558 | |
1445 | static ev_io iow [nfd]; |
1559 | static ev_io iow [nfd]; |
1446 | static ev_timer tw; |
1560 | static ev_timer tw; |
1447 | |
1561 | |
1448 | static void |
1562 | static void |
1449 | io_cb (ev_loop *loop, ev_io *w, int revents) |
1563 | io_cb (ev_loop *loop, ev_io *w, int revents) |
1450 | { |
1564 | { |
1451 | // set the relevant poll flags |
|
|
1452 | // could also call adns_processreadable etc. here |
|
|
1453 | struct pollfd *fd = (struct pollfd *)w->data; |
|
|
1454 | if (revents & EV_READ ) fd->revents |= fd->events & POLLIN; |
|
|
1455 | if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT; |
|
|
1456 | } |
1565 | } |
1457 | |
1566 | |
1458 | // create io watchers for each fd and a timer before blocking |
1567 | // create io watchers for each fd and a timer before blocking |
1459 | static void |
1568 | static void |
1460 | adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents) |
1569 | adns_prepare_cb (ev_loop *loop, ev_prepare *w, int revents) |
… | |
… | |
1466 | |
1575 | |
1467 | /* the callback is illegal, but won't be called as we stop during check */ |
1576 | /* the callback is illegal, but won't be called as we stop during check */ |
1468 | ev_timer_init (&tw, 0, timeout * 1e-3); |
1577 | ev_timer_init (&tw, 0, timeout * 1e-3); |
1469 | ev_timer_start (loop, &tw); |
1578 | ev_timer_start (loop, &tw); |
1470 | |
1579 | |
1471 | // create on ev_io per pollfd |
1580 | // create one ev_io per pollfd |
1472 | for (int i = 0; i < nfd; ++i) |
1581 | for (int i = 0; i < nfd; ++i) |
1473 | { |
1582 | { |
1474 | ev_io_init (iow + i, io_cb, fds [i].fd, |
1583 | ev_io_init (iow + i, io_cb, fds [i].fd, |
1475 | ((fds [i].events & POLLIN ? EV_READ : 0) |
1584 | ((fds [i].events & POLLIN ? EV_READ : 0) |
1476 | | (fds [i].events & POLLOUT ? EV_WRITE : 0))); |
1585 | | (fds [i].events & POLLOUT ? EV_WRITE : 0))); |
1477 | |
1586 | |
1478 | fds [i].revents = 0; |
1587 | fds [i].revents = 0; |
1479 | iow [i].data = fds + i; |
|
|
1480 | ev_io_start (loop, iow + i); |
1588 | ev_io_start (loop, iow + i); |
1481 | } |
1589 | } |
1482 | } |
1590 | } |
1483 | |
1591 | |
1484 | // stop all watchers after blocking |
1592 | // stop all watchers after blocking |
… | |
… | |
1486 | adns_check_cb (ev_loop *loop, ev_check *w, int revents) |
1594 | adns_check_cb (ev_loop *loop, ev_check *w, int revents) |
1487 | { |
1595 | { |
1488 | ev_timer_stop (loop, &tw); |
1596 | ev_timer_stop (loop, &tw); |
1489 | |
1597 | |
1490 | for (int i = 0; i < nfd; ++i) |
1598 | for (int i = 0; i < nfd; ++i) |
|
|
1599 | { |
|
|
1600 | // set the relevant poll flags |
|
|
1601 | // could also call adns_processreadable etc. here |
|
|
1602 | struct pollfd *fd = fds + i; |
|
|
1603 | int revents = ev_clear_pending (iow + i); |
|
|
1604 | if (revents & EV_READ ) fd->revents |= fd->events & POLLIN; |
|
|
1605 | if (revents & EV_WRITE) fd->revents |= fd->events & POLLOUT; |
|
|
1606 | |
|
|
1607 | // now stop the watcher |
1491 | ev_io_stop (loop, iow + i); |
1608 | ev_io_stop (loop, iow + i); |
|
|
1609 | } |
1492 | |
1610 | |
1493 | adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop)); |
1611 | adns_afterpoll (adns, fds, nfd, timeval_from (ev_now (loop)); |
|
|
1612 | } |
|
|
1613 | |
|
|
1614 | Method 2: This would be just like method 1, but you run C<adns_afterpoll> |
|
|
1615 | in the prepare watcher and would dispose of the check watcher. |
|
|
1616 | |
|
|
1617 | Method 3: If the module to be embedded supports explicit event |
|
|
1618 | notification (adns does), you can also make use of the actual watcher |
|
|
1619 | callbacks, and only destroy/create the watchers in the prepare watcher. |
|
|
1620 | |
|
|
1621 | static void |
|
|
1622 | timer_cb (EV_P_ ev_timer *w, int revents) |
|
|
1623 | { |
|
|
1624 | adns_state ads = (adns_state)w->data; |
|
|
1625 | update_now (EV_A); |
|
|
1626 | |
|
|
1627 | adns_processtimeouts (ads, &tv_now); |
|
|
1628 | } |
|
|
1629 | |
|
|
1630 | static void |
|
|
1631 | io_cb (EV_P_ ev_io *w, int revents) |
|
|
1632 | { |
|
|
1633 | adns_state ads = (adns_state)w->data; |
|
|
1634 | update_now (EV_A); |
|
|
1635 | |
|
|
1636 | if (revents & EV_READ ) adns_processreadable (ads, w->fd, &tv_now); |
|
|
1637 | if (revents & EV_WRITE) adns_processwriteable (ads, w->fd, &tv_now); |
|
|
1638 | } |
|
|
1639 | |
|
|
1640 | // do not ever call adns_afterpoll |
|
|
1641 | |
|
|
1642 | Method 4: Do not use a prepare or check watcher because the module you |
|
|
1643 | want to embed is too inflexible to support it. Instead, youc na override |
|
|
1644 | their poll function. The drawback with this solution is that the main |
|
|
1645 | loop is now no longer controllable by EV. The C<Glib::EV> module does |
|
|
1646 | this. |
|
|
1647 | |
|
|
1648 | static gint |
|
|
1649 | event_poll_func (GPollFD *fds, guint nfds, gint timeout) |
|
|
1650 | { |
|
|
1651 | int got_events = 0; |
|
|
1652 | |
|
|
1653 | for (n = 0; n < nfds; ++n) |
|
|
1654 | // create/start io watcher that sets the relevant bits in fds[n] and increment got_events |
|
|
1655 | |
|
|
1656 | if (timeout >= 0) |
|
|
1657 | // create/start timer |
|
|
1658 | |
|
|
1659 | // poll |
|
|
1660 | ev_loop (EV_A_ 0); |
|
|
1661 | |
|
|
1662 | // stop timer again |
|
|
1663 | if (timeout >= 0) |
|
|
1664 | ev_timer_stop (EV_A_ &to); |
|
|
1665 | |
|
|
1666 | // stop io watchers again - their callbacks should have set |
|
|
1667 | for (n = 0; n < nfds; ++n) |
|
|
1668 | ev_io_stop (EV_A_ iow [n]); |
|
|
1669 | |
|
|
1670 | return got_events; |
1494 | } |
1671 | } |
1495 | |
1672 | |
1496 | |
1673 | |
1497 | =head2 C<ev_embed> - when one backend isn't enough... |
1674 | =head2 C<ev_embed> - when one backend isn't enough... |
1498 | |
1675 | |
… | |
… | |
1702 | |
1879 | |
1703 | To use it, |
1880 | To use it, |
1704 | |
1881 | |
1705 | #include <ev++.h> |
1882 | #include <ev++.h> |
1706 | |
1883 | |
1707 | (it is not installed by default). This automatically includes F<ev.h> |
1884 | This automatically includes F<ev.h> and puts all of its definitions (many |
1708 | and puts all of its definitions (many of them macros) into the global |
1885 | of them macros) into the global namespace. All C++ specific things are |
1709 | namespace. All C++ specific things are put into the C<ev> namespace. |
1886 | put into the C<ev> namespace. It should support all the same embedding |
|
|
1887 | options as F<ev.h>, most notably C<EV_MULTIPLICITY>. |
1710 | |
1888 | |
1711 | It should support all the same embedding options as F<ev.h>, most notably |
1889 | Care has been taken to keep the overhead low. The only data member the C++ |
1712 | C<EV_MULTIPLICITY>. |
1890 | classes add (compared to plain C-style watchers) is the event loop pointer |
|
|
1891 | that the watcher is associated with (or no additional members at all if |
|
|
1892 | you disable C<EV_MULTIPLICITY> when embedding libev). |
|
|
1893 | |
|
|
1894 | Currently, functions, and static and non-static member functions can be |
|
|
1895 | used as callbacks. Other types should be easy to add as long as they only |
|
|
1896 | need one additional pointer for context. If you need support for other |
|
|
1897 | types of functors please contact the author (preferably after implementing |
|
|
1898 | it). |
1713 | |
1899 | |
1714 | Here is a list of things available in the C<ev> namespace: |
1900 | Here is a list of things available in the C<ev> namespace: |
1715 | |
1901 | |
1716 | =over 4 |
1902 | =over 4 |
1717 | |
1903 | |
… | |
… | |
1733 | |
1919 | |
1734 | All of those classes have these methods: |
1920 | All of those classes have these methods: |
1735 | |
1921 | |
1736 | =over 4 |
1922 | =over 4 |
1737 | |
1923 | |
1738 | =item ev::TYPE::TYPE (object *, object::method *) |
1924 | =item ev::TYPE::TYPE () |
1739 | |
1925 | |
1740 | =item ev::TYPE::TYPE (object *, object::method *, struct ev_loop *) |
1926 | =item ev::TYPE::TYPE (struct ev_loop *) |
1741 | |
1927 | |
1742 | =item ev::TYPE::~TYPE |
1928 | =item ev::TYPE::~TYPE |
1743 | |
1929 | |
1744 | The constructor takes a pointer to an object and a method pointer to |
1930 | The constructor (optionally) takes an event loop to associate the watcher |
1745 | the event handler callback to call in this class. The constructor calls |
1931 | with. If it is omitted, it will use C<EV_DEFAULT>. |
1746 | C<ev_init> for you, which means you have to call the C<set> method |
1932 | |
1747 | before starting it. If you do not specify a loop then the constructor |
1933 | The constructor calls C<ev_init> for you, which means you have to call the |
1748 | automatically associates the default loop with this watcher. |
1934 | C<set> method before starting it. |
|
|
1935 | |
|
|
1936 | It will not set a callback, however: You have to call the templated C<set> |
|
|
1937 | method to set a callback before you can start the watcher. |
|
|
1938 | |
|
|
1939 | (The reason why you have to use a method is a limitation in C++ which does |
|
|
1940 | not allow explicit template arguments for constructors). |
1749 | |
1941 | |
1750 | The destructor automatically stops the watcher if it is active. |
1942 | The destructor automatically stops the watcher if it is active. |
|
|
1943 | |
|
|
1944 | =item w->set<class, &class::method> (object *) |
|
|
1945 | |
|
|
1946 | This method sets the callback method to call. The method has to have a |
|
|
1947 | signature of C<void (*)(ev_TYPE &, int)>, it receives the watcher as |
|
|
1948 | first argument and the C<revents> as second. The object must be given as |
|
|
1949 | parameter and is stored in the C<data> member of the watcher. |
|
|
1950 | |
|
|
1951 | This method synthesizes efficient thunking code to call your method from |
|
|
1952 | the C callback that libev requires. If your compiler can inline your |
|
|
1953 | callback (i.e. it is visible to it at the place of the C<set> call and |
|
|
1954 | your compiler is good :), then the method will be fully inlined into the |
|
|
1955 | thunking function, making it as fast as a direct C callback. |
|
|
1956 | |
|
|
1957 | Example: simple class declaration and watcher initialisation |
|
|
1958 | |
|
|
1959 | struct myclass |
|
|
1960 | { |
|
|
1961 | void io_cb (ev::io &w, int revents) { } |
|
|
1962 | } |
|
|
1963 | |
|
|
1964 | myclass obj; |
|
|
1965 | ev::io iow; |
|
|
1966 | iow.set <myclass, &myclass::io_cb> (&obj); |
|
|
1967 | |
|
|
1968 | =item w->set<function> (void *data = 0) |
|
|
1969 | |
|
|
1970 | Also sets a callback, but uses a static method or plain function as |
|
|
1971 | callback. The optional C<data> argument will be stored in the watcher's |
|
|
1972 | C<data> member and is free for you to use. |
|
|
1973 | |
|
|
1974 | The prototype of the C<function> must be C<void (*)(ev::TYPE &w, int)>. |
|
|
1975 | |
|
|
1976 | See the method-C<set> above for more details. |
|
|
1977 | |
|
|
1978 | Example: |
|
|
1979 | |
|
|
1980 | static void io_cb (ev::io &w, int revents) { } |
|
|
1981 | iow.set <io_cb> (); |
1751 | |
1982 | |
1752 | =item w->set (struct ev_loop *) |
1983 | =item w->set (struct ev_loop *) |
1753 | |
1984 | |
1754 | Associates a different C<struct ev_loop> with this watcher. You can only |
1985 | Associates a different C<struct ev_loop> with this watcher. You can only |
1755 | do this when the watcher is inactive (and not pending either). |
1986 | do this when the watcher is inactive (and not pending either). |
1756 | |
1987 | |
1757 | =item w->set ([args]) |
1988 | =item w->set ([args]) |
1758 | |
1989 | |
1759 | Basically the same as C<ev_TYPE_set>, with the same args. Must be |
1990 | Basically the same as C<ev_TYPE_set>, with the same args. Must be |
1760 | called at least once. Unlike the C counterpart, an active watcher gets |
1991 | called at least once. Unlike the C counterpart, an active watcher gets |
1761 | automatically stopped and restarted. |
1992 | automatically stopped and restarted when reconfiguring it with this |
|
|
1993 | method. |
1762 | |
1994 | |
1763 | =item w->start () |
1995 | =item w->start () |
1764 | |
1996 | |
1765 | Starts the watcher. Note that there is no C<loop> argument as the |
1997 | Starts the watcher. Note that there is no C<loop> argument, as the |
1766 | constructor already takes the loop. |
1998 | constructor already stores the event loop. |
1767 | |
1999 | |
1768 | =item w->stop () |
2000 | =item w->stop () |
1769 | |
2001 | |
1770 | Stops the watcher if it is active. Again, no C<loop> argument. |
2002 | Stops the watcher if it is active. Again, no C<loop> argument. |
1771 | |
2003 | |
… | |
… | |
1796 | |
2028 | |
1797 | myclass (); |
2029 | myclass (); |
1798 | } |
2030 | } |
1799 | |
2031 | |
1800 | myclass::myclass (int fd) |
2032 | myclass::myclass (int fd) |
1801 | : io (this, &myclass::io_cb), |
|
|
1802 | idle (this, &myclass::idle_cb) |
|
|
1803 | { |
2033 | { |
|
|
2034 | io .set <myclass, &myclass::io_cb > (this); |
|
|
2035 | idle.set <myclass, &myclass::idle_cb> (this); |
|
|
2036 | |
1804 | io.start (fd, ev::READ); |
2037 | io.start (fd, ev::READ); |
1805 | } |
2038 | } |
1806 | |
2039 | |
1807 | |
2040 | |
1808 | =head1 MACRO MAGIC |
2041 | =head1 MACRO MAGIC |
1809 | |
2042 | |
1810 | Libev can be compiled with a variety of options, the most fundemantal is |
2043 | Libev can be compiled with a variety of options, the most fundemantal is |
1811 | C<EV_MULTIPLICITY>. This option determines wether (most) functions and |
2044 | C<EV_MULTIPLICITY>. This option determines whether (most) functions and |
1812 | callbacks have an initial C<struct ev_loop *> argument. |
2045 | callbacks have an initial C<struct ev_loop *> argument. |
1813 | |
2046 | |
1814 | To make it easier to write programs that cope with either variant, the |
2047 | To make it easier to write programs that cope with either variant, the |
1815 | following macros are defined: |
2048 | following macros are defined: |
1816 | |
2049 | |
… | |
… | |
1850 | loop, if multiple loops are supported ("ev loop default"). |
2083 | loop, if multiple loops are supported ("ev loop default"). |
1851 | |
2084 | |
1852 | =back |
2085 | =back |
1853 | |
2086 | |
1854 | Example: Declare and initialise a check watcher, utilising the above |
2087 | Example: Declare and initialise a check watcher, utilising the above |
1855 | macros so it will work regardless of wether multiple loops are supported |
2088 | macros so it will work regardless of whether multiple loops are supported |
1856 | or not. |
2089 | or not. |
1857 | |
2090 | |
1858 | static void |
2091 | static void |
1859 | check_cb (EV_P_ ev_timer *w, int revents) |
2092 | check_cb (EV_P_ ev_timer *w, int revents) |
1860 | { |
2093 | { |
… | |
… | |
2085 | will have the C<struct ev_loop *> as first argument, and you can create |
2318 | will have the C<struct ev_loop *> as first argument, and you can create |
2086 | additional independent event loops. Otherwise there will be no support |
2319 | additional independent event loops. Otherwise there will be no support |
2087 | for multiple event loops and there is no first event loop pointer |
2320 | for multiple event loops and there is no first event loop pointer |
2088 | argument. Instead, all functions act on the single default loop. |
2321 | argument. Instead, all functions act on the single default loop. |
2089 | |
2322 | |
|
|
2323 | =item EV_MINPRI |
|
|
2324 | |
|
|
2325 | =item EV_MAXPRI |
|
|
2326 | |
|
|
2327 | The range of allowed priorities. C<EV_MINPRI> must be smaller or equal to |
|
|
2328 | C<EV_MAXPRI>, but otherwise there are no non-obvious limitations. You can |
|
|
2329 | provide for more priorities by overriding those symbols (usually defined |
|
|
2330 | to be C<-2> and C<2>, respectively). |
|
|
2331 | |
|
|
2332 | When doing priority-based operations, libev usually has to linearly search |
|
|
2333 | all the priorities, so having many of them (hundreds) uses a lot of space |
|
|
2334 | and time, so using the defaults of five priorities (-2 .. +2) is usually |
|
|
2335 | fine. |
|
|
2336 | |
|
|
2337 | If your embedding app does not need any priorities, defining these both to |
|
|
2338 | C<0> will save some memory and cpu. |
|
|
2339 | |
2090 | =item EV_PERIODIC_ENABLE |
2340 | =item EV_PERIODIC_ENABLE |
2091 | |
2341 | |
2092 | If undefined or defined to be C<1>, then periodic timers are supported. If |
2342 | If undefined or defined to be C<1>, then periodic timers are supported. If |
|
|
2343 | defined to be C<0>, then they are not. Disabling them saves a few kB of |
|
|
2344 | code. |
|
|
2345 | |
|
|
2346 | =item EV_IDLE_ENABLE |
|
|
2347 | |
|
|
2348 | If undefined or defined to be C<1>, then idle watchers are supported. If |
2093 | defined to be C<0>, then they are not. Disabling them saves a few kB of |
2349 | defined to be C<0>, then they are not. Disabling them saves a few kB of |
2094 | code. |
2350 | code. |
2095 | |
2351 | |
2096 | =item EV_EMBED_ENABLE |
2352 | =item EV_EMBED_ENABLE |
2097 | |
2353 | |
… | |
… | |
2190 | |
2446 | |
2191 | In this section the complexities of (many of) the algorithms used inside |
2447 | In this section the complexities of (many of) the algorithms used inside |
2192 | libev will be explained. For complexity discussions about backends see the |
2448 | libev will be explained. For complexity discussions about backends see the |
2193 | documentation for C<ev_default_init>. |
2449 | documentation for C<ev_default_init>. |
2194 | |
2450 | |
|
|
2451 | All of the following are about amortised time: If an array needs to be |
|
|
2452 | extended, libev needs to realloc and move the whole array, but this |
|
|
2453 | happens asymptotically never with higher number of elements, so O(1) might |
|
|
2454 | mean it might do a lengthy realloc operation in rare cases, but on average |
|
|
2455 | it is much faster and asymptotically approaches constant time. |
|
|
2456 | |
2195 | =over 4 |
2457 | =over 4 |
2196 | |
2458 | |
2197 | =item Starting and stopping timer/periodic watchers: O(log skipped_other_timers) |
2459 | =item Starting and stopping timer/periodic watchers: O(log skipped_other_timers) |
2198 | |
2460 | |
|
|
2461 | This means that, when you have a watcher that triggers in one hour and |
|
|
2462 | there are 100 watchers that would trigger before that then inserting will |
|
|
2463 | have to skip those 100 watchers. |
|
|
2464 | |
2199 | =item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers) |
2465 | =item Changing timer/periodic watchers (by autorepeat, again): O(log skipped_other_timers) |
2200 | |
2466 | |
|
|
2467 | That means that for changing a timer costs less than removing/adding them |
|
|
2468 | as only the relative motion in the event queue has to be paid for. |
|
|
2469 | |
2201 | =item Starting io/check/prepare/idle/signal/child watchers: O(1) |
2470 | =item Starting io/check/prepare/idle/signal/child watchers: O(1) |
2202 | |
2471 | |
|
|
2472 | These just add the watcher into an array or at the head of a list. |
2203 | =item Stopping check/prepare/idle watchers: O(1) |
2473 | =item Stopping check/prepare/idle watchers: O(1) |
2204 | |
2474 | |
2205 | =item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE)) |
2475 | =item Stopping an io/signal/child watcher: O(number_of_watchers_for_this_(fd/signal/pid % EV_PID_HASHSIZE)) |
2206 | |
2476 | |
|
|
2477 | These watchers are stored in lists then need to be walked to find the |
|
|
2478 | correct watcher to remove. The lists are usually short (you don't usually |
|
|
2479 | have many watchers waiting for the same fd or signal). |
|
|
2480 | |
2207 | =item Finding the next timer per loop iteration: O(1) |
2481 | =item Finding the next timer per loop iteration: O(1) |
2208 | |
2482 | |
2209 | =item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd) |
2483 | =item Each change on a file descriptor per loop iteration: O(number_of_watchers_for_this_fd) |
2210 | |
2484 | |
|
|
2485 | A change means an I/O watcher gets started or stopped, which requires |
|
|
2486 | libev to recalculate its status (and possibly tell the kernel). |
|
|
2487 | |
2211 | =item Activating one watcher: O(1) |
2488 | =item Activating one watcher: O(1) |
2212 | |
2489 | |
|
|
2490 | =item Priority handling: O(number_of_priorities) |
|
|
2491 | |
|
|
2492 | Priorities are implemented by allocating some space for each |
|
|
2493 | priority. When doing priority-based operations, libev usually has to |
|
|
2494 | linearly search all the priorities. |
|
|
2495 | |
2213 | =back |
2496 | =back |
2214 | |
2497 | |
2215 | |
2498 | |
2216 | =head1 AUTHOR |
2499 | =head1 AUTHOR |
2217 | |
2500 | |