ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libev/ev.3
(Generate patch)

Comparing libev/ev.3 (file contents):
Revision 1.100 by root, Tue Oct 29 12:13:37 2013 UTC vs.
Revision 1.114 by root, Tue Jun 25 06:36:59 2019 UTC

1.\" Automatically generated by Pod::Man 2.27 (Pod::Simple 3.28) 1.\" Automatically generated by Pod::Man 4.11 (Pod::Simple 3.35)
2.\" 2.\"
3.\" Standard preamble: 3.\" Standard preamble:
4.\" ======================================================================== 4.\" ========================================================================
5.de Sp \" Vertical space (when we can't use .PP) 5.de Sp \" Vertical space (when we can't use .PP)
6.if t .sp .5v 6.if t .sp .5v
44.\" 44.\"
45.\" Escape single quotes in literal strings from groff's Unicode transform. 45.\" Escape single quotes in literal strings from groff's Unicode transform.
46.ie \n(.g .ds Aq \(aq 46.ie \n(.g .ds Aq \(aq
47.el .ds Aq ' 47.el .ds Aq '
48.\" 48.\"
49.\" If the F register is turned on, we'll generate index entries on stderr for 49.\" If the F register is >0, we'll generate index entries on stderr for
50.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index 50.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
51.\" entries marked with X<> in POD. Of course, you'll have to process the 51.\" entries marked with X<> in POD. Of course, you'll have to process the
52.\" output yourself in some meaningful fashion. 52.\" output yourself in some meaningful fashion.
53.\" 53.\"
54.\" Avoid warning from groff about undefined register 'F'. 54.\" Avoid warning from groff about undefined register 'F'.
55.de IX 55.de IX
56.. 56..
57.nr rF 0 57.nr rF 0
58.if \n(.g .if rF .nr rF 1 58.if \n(.g .if rF .nr rF 1
59.if (\n(rF:(\n(.g==0)) \{ 59.if (\n(rF:(\n(.g==0)) \{\
60. if \nF \{ 60. if \nF \{\
61. de IX 61. de IX
62. tm Index:\\$1\t\\n%\t"\\$2" 62. tm Index:\\$1\t\\n%\t"\\$2"
63.. 63..
64. if !\nF==2 \{ 64. if !\nF==2 \{\
65. nr % 0 65. nr % 0
66. nr F 2 66. nr F 2
67. \} 67. \}
68. \} 68. \}
69.\} 69.\}
131.\} 131.\}
132.rm #[ #] #H #V #F C 132.rm #[ #] #H #V #F C
133.\" ======================================================================== 133.\" ========================================================================
134.\" 134.\"
135.IX Title "LIBEV 3" 135.IX Title "LIBEV 3"
136.TH LIBEV 3 "2013-10-29" "libev-4.15" "libev - high performance full featured event loop" 136.TH LIBEV 3 "2019-06-25" "libev-4.25" "libev - high performance full featured event loop"
137.\" For nroff, turn off justification. Always turn off hyphenation; it makes 137.\" For nroff, turn off justification. Always turn off hyphenation; it makes
138.\" way too many mistakes in technical documents. 138.\" way too many mistakes in technical documents.
139.if n .ad l 139.if n .ad l
140.nh 140.nh
141.SH "NAME" 141.SH "NAME"
240watchers\fR, which are relatively small C structures you initialise with the 240watchers\fR, which are relatively small C structures you initialise with the
241details of the event, and then hand it over to libev by \fIstarting\fR the 241details of the event, and then hand it over to libev by \fIstarting\fR the
242watcher. 242watcher.
243.SS "\s-1FEATURES\s0" 243.SS "\s-1FEATURES\s0"
244.IX Subsection "FEATURES" 244.IX Subsection "FEATURES"
245Libev supports \f(CW\*(C`select\*(C'\fR, \f(CW\*(C`poll\*(C'\fR, the Linux-specific \f(CW\*(C`epoll\*(C'\fR, the 245Libev supports \f(CW\*(C`select\*(C'\fR, \f(CW\*(C`poll\*(C'\fR, the Linux-specific aio and \f(CW\*(C`epoll\*(C'\fR
246BSD-specific \f(CW\*(C`kqueue\*(C'\fR and the Solaris-specific event port mechanisms 246interfaces, the BSD-specific \f(CW\*(C`kqueue\*(C'\fR and the Solaris-specific event port
247for file descriptor events (\f(CW\*(C`ev_io\*(C'\fR), the Linux \f(CW\*(C`inotify\*(C'\fR interface 247mechanisms for file descriptor events (\f(CW\*(C`ev_io\*(C'\fR), the Linux \f(CW\*(C`inotify\*(C'\fR
248(for \f(CW\*(C`ev_stat\*(C'\fR), Linux eventfd/signalfd (for faster and cleaner 248interface (for \f(CW\*(C`ev_stat\*(C'\fR), Linux eventfd/signalfd (for faster and cleaner
249inter-thread wakeup (\f(CW\*(C`ev_async\*(C'\fR)/signal handling (\f(CW\*(C`ev_signal\*(C'\fR)) relative 249inter-thread wakeup (\f(CW\*(C`ev_async\*(C'\fR)/signal handling (\f(CW\*(C`ev_signal\*(C'\fR)) relative
250timers (\f(CW\*(C`ev_timer\*(C'\fR), absolute timers with customised rescheduling 250timers (\f(CW\*(C`ev_timer\*(C'\fR), absolute timers with customised rescheduling
251(\f(CW\*(C`ev_periodic\*(C'\fR), synchronous signals (\f(CW\*(C`ev_signal\*(C'\fR), process status 251(\f(CW\*(C`ev_periodic\*(C'\fR), synchronous signals (\f(CW\*(C`ev_signal\*(C'\fR), process status
252change events (\f(CW\*(C`ev_child\*(C'\fR), and event watchers dealing with the event 252change events (\f(CW\*(C`ev_child\*(C'\fR), and event watchers dealing with the event
253loop mechanism itself (\f(CW\*(C`ev_idle\*(C'\fR, \f(CW\*(C`ev_embed\*(C'\fR, \f(CW\*(C`ev_prepare\*(C'\fR and 253loop mechanism itself (\f(CW\*(C`ev_idle\*(C'\fR, \f(CW\*(C`ev_embed\*(C'\fR, \f(CW\*(C`ev_prepare\*(C'\fR and
392.Sp 392.Sp
393You could override this function in high-availability programs to, say, 393You could override this function in high-availability programs to, say,
394free some memory if it cannot allocate memory, to use a special allocator, 394free some memory if it cannot allocate memory, to use a special allocator,
395or even to sleep a while and retry until some memory is available. 395or even to sleep a while and retry until some memory is available.
396.Sp 396.Sp
397Example: The following is the \f(CW\*(C`realloc\*(C'\fR function that libev itself uses
398which should work with \f(CW\*(C`realloc\*(C'\fR and \f(CW\*(C`free\*(C'\fR functions of all kinds and
399is probably a good basis for your own implementation.
400.Sp
401.Vb 5
402\& static void *
403\& ev_realloc_emul (void *ptr, long size) EV_NOEXCEPT
404\& {
405\& if (size)
406\& return realloc (ptr, size);
407\&
408\& free (ptr);
409\& return 0;
410\& }
411.Ve
412.Sp
397Example: Replace the libev allocator with one that waits a bit and then 413Example: Replace the libev allocator with one that waits a bit and then
398retries (example requires a standards-compliant \f(CW\*(C`realloc\*(C'\fR). 414retries.
399.Sp 415.Sp
400.Vb 6 416.Vb 8
401\& static void * 417\& static void *
402\& persistent_realloc (void *ptr, size_t size) 418\& persistent_realloc (void *ptr, size_t size)
403\& { 419\& {
420\& if (!size)
421\& {
422\& free (ptr);
423\& return 0;
424\& }
425\&
404\& for (;;) 426\& for (;;)
405\& { 427\& {
406\& void *newptr = realloc (ptr, size); 428\& void *newptr = realloc (ptr, size);
407\& 429\&
408\& if (newptr) 430\& if (newptr)
536make libev check for a fork in each iteration by enabling this flag. 558make libev check for a fork in each iteration by enabling this flag.
537.Sp 559.Sp
538This works by calling \f(CW\*(C`getpid ()\*(C'\fR on every iteration of the loop, 560This works by calling \f(CW\*(C`getpid ()\*(C'\fR on every iteration of the loop,
539and thus this might slow down your event loop if you do a lot of loop 561and thus this might slow down your event loop if you do a lot of loop
540iterations and little real work, but is usually not noticeable (on my 562iterations and little real work, but is usually not noticeable (on my
541GNU/Linux system for example, \f(CW\*(C`getpid\*(C'\fR is actually a simple 5\-insn sequence 563GNU/Linux system for example, \f(CW\*(C`getpid\*(C'\fR is actually a simple 5\-insn
542without a system call and thus \fIvery\fR fast, but my GNU/Linux system also has 564sequence without a system call and thus \fIvery\fR fast, but my GNU/Linux
543\&\f(CW\*(C`pthread_atfork\*(C'\fR which is even faster). 565system also has \f(CW\*(C`pthread_atfork\*(C'\fR which is even faster). (Update: glibc
566versions 2.25 apparently removed the \f(CW\*(C`getpid\*(C'\fR optimisation again).
544.Sp 567.Sp
545The big advantage of this flag is that you can forget about fork (and 568The big advantage of this flag is that you can forget about fork (and
546forget about forgetting to tell libev about forking) when you use this 569forget about forgetting to tell libev about forking, although you still
547flag. 570have to ignore \f(CW\*(C`SIGPIPE\*(C'\fR) when you use this flag.
548.Sp 571.Sp
549This flag setting cannot be overridden or specified in the \f(CW\*(C`LIBEV_FLAGS\*(C'\fR 572This flag setting cannot be overridden or specified in the \f(CW\*(C`LIBEV_FLAGS\*(C'\fR
550environment variable. 573environment variable.
551.ie n .IP """EVFLAG_NOINOTIFY""" 4 574.ie n .IP """EVFLAG_NOINOTIFY""" 4
552.el .IP "\f(CWEVFLAG_NOINOTIFY\fR" 4 575.el .IP "\f(CWEVFLAG_NOINOTIFY\fR" 4
584.Sp 607.Sp
585This flag's behaviour will become the default in future versions of libev. 608This flag's behaviour will become the default in future versions of libev.
586.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4 609.ie n .IP """EVBACKEND_SELECT"" (value 1, portable select backend)" 4
587.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4 610.el .IP "\f(CWEVBACKEND_SELECT\fR (value 1, portable select backend)" 4
588.IX Item "EVBACKEND_SELECT (value 1, portable select backend)" 611.IX Item "EVBACKEND_SELECT (value 1, portable select backend)"
589This is your standard \fIselect\fR\|(2) backend. Not \fIcompletely\fR standard, as 612This is your standard \fBselect\fR\|(2) backend. Not \fIcompletely\fR standard, as
590libev tries to roll its own fd_set with no limits on the number of fds, 613libev tries to roll its own fd_set with no limits on the number of fds,
591but if that fails, expect a fairly low limit on the number of fds when 614but if that fails, expect a fairly low limit on the number of fds when
592using this backend. It doesn't scale too well (O(highest_fd)), but its 615using this backend. It doesn't scale too well (O(highest_fd)), but its
593usually the fastest backend for a low number of (low-numbered :) fds. 616usually the fastest backend for a low number of (low-numbered :) fds.
594.Sp 617.Sp
603\&\f(CW\*(C`writefds\*(C'\fR set (and to work around Microsoft Windows bugs, also onto the 626\&\f(CW\*(C`writefds\*(C'\fR set (and to work around Microsoft Windows bugs, also onto the
604\&\f(CW\*(C`exceptfds\*(C'\fR set on that platform). 627\&\f(CW\*(C`exceptfds\*(C'\fR set on that platform).
605.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4 628.ie n .IP """EVBACKEND_POLL"" (value 2, poll backend, available everywhere except on windows)" 4
606.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4 629.el .IP "\f(CWEVBACKEND_POLL\fR (value 2, poll backend, available everywhere except on windows)" 4
607.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)" 630.IX Item "EVBACKEND_POLL (value 2, poll backend, available everywhere except on windows)"
608And this is your standard \fIpoll\fR\|(2) backend. It's more complicated 631And this is your standard \fBpoll\fR\|(2) backend. It's more complicated
609than select, but handles sparse fds better and has no artificial 632than select, but handles sparse fds better and has no artificial
610limit on the number of fds you can use (except it will slow down 633limit on the number of fds you can use (except it will slow down
611considerably with a lot of inactive fds). It scales similarly to select, 634considerably with a lot of inactive fds). It scales similarly to select,
612i.e. O(total_fds). See the entry for \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR, above, for 635i.e. O(total_fds). See the entry for \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR, above, for
613performance tips. 636performance tips.
615This backend maps \f(CW\*(C`EV_READ\*(C'\fR to \f(CW\*(C`POLLIN | POLLERR | POLLHUP\*(C'\fR, and 638This backend maps \f(CW\*(C`EV_READ\*(C'\fR to \f(CW\*(C`POLLIN | POLLERR | POLLHUP\*(C'\fR, and
616\&\f(CW\*(C`EV_WRITE\*(C'\fR to \f(CW\*(C`POLLOUT | POLLERR | POLLHUP\*(C'\fR. 639\&\f(CW\*(C`EV_WRITE\*(C'\fR to \f(CW\*(C`POLLOUT | POLLERR | POLLHUP\*(C'\fR.
617.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4 640.ie n .IP """EVBACKEND_EPOLL"" (value 4, Linux)" 4
618.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4 641.el .IP "\f(CWEVBACKEND_EPOLL\fR (value 4, Linux)" 4
619.IX Item "EVBACKEND_EPOLL (value 4, Linux)" 642.IX Item "EVBACKEND_EPOLL (value 4, Linux)"
620Use the linux-specific \fIepoll\fR\|(7) interface (for both pre\- and post\-2.6.9 643Use the Linux-specific \fBepoll\fR\|(7) interface (for both pre\- and post\-2.6.9
621kernels). 644kernels).
622.Sp 645.Sp
623For few fds, this backend is a bit little slower than poll and select, but 646For few fds, this backend is a bit little slower than poll and select, but
624it scales phenomenally better. While poll and select usually scale like 647it scales phenomenally better. While poll and select usually scale like
625O(total_fds) where total_fds is the total number of fds (or the highest 648O(total_fds) where total_fds is the total number of fds (or the highest
671All this means that, in practice, \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR can be as fast or 694All this means that, in practice, \f(CW\*(C`EVBACKEND_SELECT\*(C'\fR can be as fast or
672faster than epoll for maybe up to a hundred file descriptors, depending on 695faster than epoll for maybe up to a hundred file descriptors, depending on
673the usage. So sad. 696the usage. So sad.
674.Sp 697.Sp
675While nominally embeddable in other event loops, this feature is broken in 698While nominally embeddable in other event loops, this feature is broken in
676all kernel versions tested so far. 699a lot of kernel revisions, but probably(!) works in current versions.
700.Sp
701This backend maps \f(CW\*(C`EV_READ\*(C'\fR and \f(CW\*(C`EV_WRITE\*(C'\fR in the same way as
702\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR.
703.ie n .IP """EVBACKEND_LINUXAIO"" (value 64, Linux)" 4
704.el .IP "\f(CWEVBACKEND_LINUXAIO\fR (value 64, Linux)" 4
705.IX Item "EVBACKEND_LINUXAIO (value 64, Linux)"
706Use the Linux-specific Linux \s-1AIO\s0 (\fInot\fR \f(CWaio(7)\fR but \f(CWio_submit(2)\fR) event interface available in post\-4.18 kernels (but libev
707only tries to use it in 4.19+).
708.Sp
709This is another Linux train wreck of an event interface.
710.Sp
711If this backend works for you (as of this writing, it was very
712experimental), it is the best event interface available on Linux and might
713be well worth enabling it \- if it isn't available in your kernel this will
714be detected and this backend will be skipped.
715.Sp
716This backend can batch oneshot requests and supports a user-space ring
717buffer to receive events. It also doesn't suffer from most of the design
718problems of epoll (such as not being able to remove event sources from
719the epoll set), and generally sounds too good to be true. Because, this
720being the Linux kernel, of course it suffers from a whole new set of
721limitations, forcing you to fall back to epoll, inheriting all its design
722issues.
723.Sp
724For one, it is not easily embeddable (but probably could be done using
725an event fd at some extra overhead). It also is subject to a system wide
726limit that can be configured in \fI/proc/sys/fs/aio\-max\-nr\fR. If no \s-1AIO\s0
727requests are left, this backend will be skipped during initialisation, and
728will switch to epoll when the loop is active.
729.Sp
730Most problematic in practice, however, is that not all file descriptors
731work with it. For example, in Linux 5.1, \s-1TCP\s0 sockets, pipes, event fds,
732files, \fI/dev/null\fR and many others are supported, but ttys do not work
733properly (a known bug that the kernel developers don't care about, see
734<https://lore.kernel.org/patchwork/patch/1047453/>), so this is not
735(yet?) a generic event polling interface.
736.Sp
737Overall, it seems the Linux developers just don't want it to have a
738generic event handling mechanism other than \f(CW\*(C`select\*(C'\fR or \f(CW\*(C`poll\*(C'\fR.
739.Sp
740To work around all these problem, the current version of libev uses its
741epoll backend as a fallback for file descriptor types that do not work. Or
742falls back completely to epoll if the kernel acts up.
677.Sp 743.Sp
678This backend maps \f(CW\*(C`EV_READ\*(C'\fR and \f(CW\*(C`EV_WRITE\*(C'\fR in the same way as 744This backend maps \f(CW\*(C`EV_READ\*(C'\fR and \f(CW\*(C`EV_WRITE\*(C'\fR in the same way as
679\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR. 745\&\f(CW\*(C`EVBACKEND_POLL\*(C'\fR.
680.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4 746.ie n .IP """EVBACKEND_KQUEUE"" (value 8, most \s-1BSD\s0 clones)" 4
681.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4 747.el .IP "\f(CWEVBACKEND_KQUEUE\fR (value 8, most \s-1BSD\s0 clones)" 4
682.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)" 748.IX Item "EVBACKEND_KQUEUE (value 8, most BSD clones)"
683Kqueue deserves special mention, as at the time of this writing, it 749Kqueue deserves special mention, as at the time this backend was
684was broken on all BSDs except NetBSD (usually it doesn't work reliably 750implemented, it was broken on all BSDs except NetBSD (usually it doesn't
685with anything but sockets and pipes, except on Darwin, where of course 751work reliably with anything but sockets and pipes, except on Darwin,
686it's completely useless). Unlike epoll, however, whose brokenness 752where of course it's completely useless). Unlike epoll, however, whose
687is by design, these kqueue bugs can (and eventually will) be fixed 753brokenness is by design, these kqueue bugs can be (and mostly have been)
688without \s-1API\s0 changes to existing programs. For this reason it's not being 754fixed without \s-1API\s0 changes to existing programs. For this reason it's not
689\&\*(L"auto-detected\*(R" unless you explicitly specify it in the flags (i.e. using 755being \*(L"auto-detected\*(R" on all platforms unless you explicitly specify it
690\&\f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR) or libev was compiled on a known-to-be-good (\-enough) 756in the flags (i.e. using \f(CW\*(C`EVBACKEND_KQUEUE\*(C'\fR) or libev was compiled on a
691system like NetBSD. 757known-to-be-good (\-enough) system like NetBSD.
692.Sp 758.Sp
693You still can embed kqueue into a normal poll or select backend and use it 759You still can embed kqueue into a normal poll or select backend and use it
694only for sockets (after having made sure that sockets work with kqueue on 760only for sockets (after having made sure that sockets work with kqueue on
695the target platform). See \f(CW\*(C`ev_embed\*(C'\fR watchers for more info. 761the target platform). See \f(CW\*(C`ev_embed\*(C'\fR watchers for more info.
696.Sp 762.Sp
697It scales in the same way as the epoll backend, but the interface to the 763It scales in the same way as the epoll backend, but the interface to the
698kernel is more efficient (which says nothing about its actual speed, of 764kernel is more efficient (which says nothing about its actual speed, of
699course). While stopping, setting and starting an I/O watcher does never 765course). While stopping, setting and starting an I/O watcher does never
700cause an extra system call as with \f(CW\*(C`EVBACKEND_EPOLL\*(C'\fR, it still adds up to 766cause an extra system call as with \f(CW\*(C`EVBACKEND_EPOLL\*(C'\fR, it still adds up to
701two event changes per incident. Support for \f(CW\*(C`fork ()\*(C'\fR is very bad (you 767two event changes per incident. Support for \f(CW\*(C`fork ()\*(C'\fR is very bad (you
702might have to leak fd's on fork, but it's more sane than epoll) and it 768might have to leak fds on fork, but it's more sane than epoll) and it
703drops fds silently in similarly hard-to-detect cases. 769drops fds silently in similarly hard-to-detect cases.
704.Sp 770.Sp
705This backend usually performs well under most conditions. 771This backend usually performs well under most conditions.
706.Sp 772.Sp
707While nominally embeddable in other event loops, this doesn't work 773While nominally embeddable in other event loops, this doesn't work
785used if available. 851used if available.
786.Sp 852.Sp
787.Vb 1 853.Vb 1
788\& struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE); 854\& struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_KQUEUE);
789.Ve 855.Ve
856.Sp
857Example: Similarly, on linux, you mgiht want to take advantage of the
858linux aio backend if possible, but fall back to something else if that
859isn't available.
860.Sp
861.Vb 1
862\& struct ev_loop *loop = ev_loop_new (ev_recommended_backends () | EVBACKEND_LINUXAIO);
863.Ve
790.RE 864.RE
791.IP "ev_loop_destroy (loop)" 4 865.IP "ev_loop_destroy (loop)" 4
792.IX Item "ev_loop_destroy (loop)" 866.IX Item "ev_loop_destroy (loop)"
793Destroys an event loop object (frees all memory and kernel state 867Destroys an event loop object (frees all memory and kernel state
794etc.). None of the active event watchers will be stopped in the normal 868etc.). None of the active event watchers will be stopped in the normal
810except in the rare occasion where you really need to free its resources. 884except in the rare occasion where you really need to free its resources.
811If you need dynamically allocated loops it is better to use \f(CW\*(C`ev_loop_new\*(C'\fR 885If you need dynamically allocated loops it is better to use \f(CW\*(C`ev_loop_new\*(C'\fR
812and \f(CW\*(C`ev_loop_destroy\*(C'\fR. 886and \f(CW\*(C`ev_loop_destroy\*(C'\fR.
813.IP "ev_loop_fork (loop)" 4 887.IP "ev_loop_fork (loop)" 4
814.IX Item "ev_loop_fork (loop)" 888.IX Item "ev_loop_fork (loop)"
815This function sets a flag that causes subsequent \f(CW\*(C`ev_run\*(C'\fR iterations to 889This function sets a flag that causes subsequent \f(CW\*(C`ev_run\*(C'\fR iterations
816reinitialise the kernel state for backends that have one. Despite the 890to reinitialise the kernel state for backends that have one. Despite
817name, you can call it anytime, but it makes most sense after forking, in 891the name, you can call it anytime you are allowed to start or stop
818the child process. You \fImust\fR call it (or use \f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR) in the 892watchers (except inside an \f(CW\*(C`ev_prepare\*(C'\fR callback), but it makes most
819child before resuming or calling \f(CW\*(C`ev_run\*(C'\fR. 893sense after forking, in the child process. You \fImust\fR call it (or use
894\&\f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR) in the child before resuming or calling \f(CW\*(C`ev_run\*(C'\fR.
895.Sp
896In addition, if you want to reuse a loop (via this function or
897\&\f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR), you \fIalso\fR have to ignore \f(CW\*(C`SIGPIPE\*(C'\fR.
820.Sp 898.Sp
821Again, you \fIhave\fR to call it on \fIany\fR loop that you want to re-use after 899Again, you \fIhave\fR to call it on \fIany\fR loop that you want to re-use after
822a fork, \fIeven if you do not plan to use the loop in the parent\fR. This is 900a fork, \fIeven if you do not plan to use the loop in the parent\fR. This is
823because some kernel interfaces *cough* \fIkqueue\fR *cough* do funny things 901because some kernel interfaces *cough* \fIkqueue\fR *cough* do funny things
824during fork. 902during fork.
1360bug in your program. 1438bug in your program.
1361.Sp 1439.Sp
1362Libev will usually signal a few \*(L"dummy\*(R" events together with an error, for 1440Libev will usually signal a few \*(L"dummy\*(R" events together with an error, for
1363example it might indicate that a fd is readable or writable, and if your 1441example it might indicate that a fd is readable or writable, and if your
1364callbacks is well-written it can just attempt the operation and cope with 1442callbacks is well-written it can just attempt the operation and cope with
1365the error from \fIread()\fR or \fIwrite()\fR. This will not work in multi-threaded 1443the error from \fBread()\fR or \fBwrite()\fR. This will not work in multi-threaded
1366programs, though, as the fd could already be closed and reused for another 1444programs, though, as the fd could already be closed and reused for another
1367thing, so beware. 1445thing, so beware.
1368.SS "\s-1GENERIC WATCHER FUNCTIONS\s0" 1446.SS "\s-1GENERIC WATCHER FUNCTIONS\s0"
1369.IX Subsection "GENERIC WATCHER FUNCTIONS" 1447.IX Subsection "GENERIC WATCHER FUNCTIONS"
1370.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4 1448.ie n .IP """ev_init"" (ev_TYPE *watcher, callback)" 4
1721But really, best use non-blocking mode. 1799But really, best use non-blocking mode.
1722.PP 1800.PP
1723\fIThe special problem of disappearing file descriptors\fR 1801\fIThe special problem of disappearing file descriptors\fR
1724.IX Subsection "The special problem of disappearing file descriptors" 1802.IX Subsection "The special problem of disappearing file descriptors"
1725.PP 1803.PP
1726Some backends (e.g. kqueue, epoll) need to be told about closing a file 1804Some backends (e.g. kqueue, epoll, linuxaio) need to be told about closing
1727descriptor (either due to calling \f(CW\*(C`close\*(C'\fR explicitly or any other means, 1805a file descriptor (either due to calling \f(CW\*(C`close\*(C'\fR explicitly or any other
1728such as \f(CW\*(C`dup2\*(C'\fR). The reason is that you register interest in some file 1806means, such as \f(CW\*(C`dup2\*(C'\fR). The reason is that you register interest in some
1729descriptor, but when it goes away, the operating system will silently drop 1807file descriptor, but when it goes away, the operating system will silently
1730this interest. If another file descriptor with the same number then is 1808drop this interest. If another file descriptor with the same number then
1731registered with libev, there is no efficient way to see that this is, in 1809is registered with libev, there is no efficient way to see that this is,
1732fact, a different file descriptor. 1810in fact, a different file descriptor.
1733.PP 1811.PP
1734To avoid having to explicitly tell libev about such cases, libev follows 1812To avoid having to explicitly tell libev about such cases, libev follows
1735the following policy: Each time \f(CW\*(C`ev_io_set\*(C'\fR is being called, libev 1813the following policy: Each time \f(CW\*(C`ev_io_set\*(C'\fR is being called, libev
1736will assume that this is potentially a new file descriptor, otherwise 1814will assume that this is potentially a new file descriptor, otherwise
1737it is assumed that the file descriptor stays the same. That means that 1815it is assumed that the file descriptor stays the same. That means that
1789reuse the same code path. 1867reuse the same code path.
1790.PP 1868.PP
1791\fIThe special problem of fork\fR 1869\fIThe special problem of fork\fR
1792.IX Subsection "The special problem of fork" 1870.IX Subsection "The special problem of fork"
1793.PP 1871.PP
1794Some backends (epoll, kqueue) do not support \f(CW\*(C`fork ()\*(C'\fR at all or exhibit 1872Some backends (epoll, kqueue, probably linuxaio) do not support \f(CW\*(C`fork ()\*(C'\fR
1795useless behaviour. Libev fully supports fork, but needs to be told about 1873at all or exhibit useless behaviour. Libev fully supports fork, but needs
1796it in the child if you want to continue to use it in the child. 1874to be told about it in the child if you want to continue to use it in the
1875child.
1797.PP 1876.PP
1798To support fork in your child processes, you have to call \f(CW\*(C`ev_loop_fork 1877To support fork in your child processes, you have to call \f(CW\*(C`ev_loop_fork
1799()\*(C'\fR after a fork in the child, enable \f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR, or resort to 1878()\*(C'\fR after a fork in the child, enable \f(CW\*(C`EVFLAG_FORKCHECK\*(C'\fR, or resort to
1800\&\f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR. 1879\&\f(CW\*(C`EVBACKEND_SELECT\*(C'\fR or \f(CW\*(C`EVBACKEND_POLL\*(C'\fR.
1801.PP 1880.PP
1806when writing to a pipe whose other end has been closed, your program gets 1885when writing to a pipe whose other end has been closed, your program gets
1807sent a \s-1SIGPIPE,\s0 which, by default, aborts your program. For most programs 1886sent a \s-1SIGPIPE,\s0 which, by default, aborts your program. For most programs
1808this is sensible behaviour, for daemons, this is usually undesirable. 1887this is sensible behaviour, for daemons, this is usually undesirable.
1809.PP 1888.PP
1810So when you encounter spurious, unexplained daemon exits, make sure you 1889So when you encounter spurious, unexplained daemon exits, make sure you
1811ignore \s-1SIGPIPE \s0(and maybe make sure you log the exit status of your daemon 1890ignore \s-1SIGPIPE\s0 (and maybe make sure you log the exit status of your daemon
1812somewhere, as that would have given you a big clue). 1891somewhere, as that would have given you a big clue).
1813.PP 1892.PP
1814\fIThe special problem of \fIaccept()\fIing when you can't\fR 1893\fIThe special problem of \f(BIaccept()\fIing when you can't\fR
1815.IX Subsection "The special problem of accept()ing when you can't" 1894.IX Subsection "The special problem of accept()ing when you can't"
1816.PP 1895.PP
1817Many implementations of the \s-1POSIX \s0\f(CW\*(C`accept\*(C'\fR function (for example, 1896Many implementations of the \s-1POSIX\s0 \f(CW\*(C`accept\*(C'\fR function (for example,
1818found in post\-2004 Linux) have the peculiar behaviour of not removing a 1897found in post\-2004 Linux) have the peculiar behaviour of not removing a
1819connection from the pending queue in all error cases. 1898connection from the pending queue in all error cases.
1820.PP 1899.PP
1821For example, larger servers often run out of file descriptors (because 1900For example, larger servers often run out of file descriptors (because
1822of resource limits), causing \f(CW\*(C`accept\*(C'\fR to fail with \f(CW\*(C`ENFILE\*(C'\fR but not 1901of resource limits), causing \f(CW\*(C`accept\*(C'\fR to fail with \f(CW\*(C`ENFILE\*(C'\fR but not
2161.PP 2240.PP
2162The relative timeouts are calculated relative to the \f(CW\*(C`ev_now ()\*(C'\fR 2241The relative timeouts are calculated relative to the \f(CW\*(C`ev_now ()\*(C'\fR
2163time. This is usually the right thing as this timestamp refers to the time 2242time. This is usually the right thing as this timestamp refers to the time
2164of the event triggering whatever timeout you are modifying/starting. If 2243of the event triggering whatever timeout you are modifying/starting. If
2165you suspect event processing to be delayed and you \fIneed\fR to base the 2244you suspect event processing to be delayed and you \fIneed\fR to base the
2166timeout on the current time, use something like this to adjust for this: 2245timeout on the current time, use something like the following to adjust
2246for it:
2167.PP 2247.PP
2168.Vb 1 2248.Vb 1
2169\& ev_timer_set (&timer, after + ev_now () \- ev_time (), 0.); 2249\& ev_timer_set (&timer, after + (ev_time () \- ev_now ()), 0.);
2170.Ve 2250.Ve
2171.PP 2251.PP
2172If the event loop is suspended for a long time, you can also force an 2252If the event loop is suspended for a long time, you can also force an
2173update of the time returned by \f(CW\*(C`ev_now ()\*(C'\fR by calling \f(CW\*(C`ev_now_update 2253update of the time returned by \f(CW\*(C`ev_now ()\*(C'\fR by calling \f(CW\*(C`ev_now_update
2174()\*(C'\fR. 2254()\*(C'\fR, although that will push the event time of all outstanding events
2255further into the future.
2175.PP 2256.PP
2176\fIThe special problem of unsynchronised clocks\fR 2257\fIThe special problem of unsynchronised clocks\fR
2177.IX Subsection "The special problem of unsynchronised clocks" 2258.IX Subsection "The special problem of unsynchronised clocks"
2178.PP 2259.PP
2179Modern systems have a variety of clocks \- libev itself uses the normal 2260Modern systems have a variety of clocks \- libev itself uses the normal
2244.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)" 2325.IX Item "ev_timer_init (ev_timer *, callback, ev_tstamp after, ev_tstamp repeat)"
2245.PD 0 2326.PD 0
2246.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4 2327.IP "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 4
2247.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)" 2328.IX Item "ev_timer_set (ev_timer *, ev_tstamp after, ev_tstamp repeat)"
2248.PD 2329.PD
2249Configure the timer to trigger after \f(CW\*(C`after\*(C'\fR seconds. If \f(CW\*(C`repeat\*(C'\fR 2330Configure the timer to trigger after \f(CW\*(C`after\*(C'\fR seconds (fractional and
2250is \f(CW0.\fR, then it will automatically be stopped once the timeout is 2331negative values are supported). If \f(CW\*(C`repeat\*(C'\fR is \f(CW0.\fR, then it will
2251reached. If it is positive, then the timer will automatically be 2332automatically be stopped once the timeout is reached. If it is positive,
2252configured to trigger again \f(CW\*(C`repeat\*(C'\fR seconds later, again, and again, 2333then the timer will automatically be configured to trigger again \f(CW\*(C`repeat\*(C'\fR
2253until stopped manually. 2334seconds later, again, and again, until stopped manually.
2254.Sp 2335.Sp
2255The timer itself will do a best-effort at avoiding drift, that is, if 2336The timer itself will do a best-effort at avoiding drift, that is, if
2256you configure a timer to trigger every 10 seconds, then it will normally 2337you configure a timer to trigger every 10 seconds, then it will normally
2257trigger at exactly 10 second intervals. If, however, your program cannot 2338trigger at exactly 10 second intervals. If, however, your program cannot
2258keep up with the timer (because it takes longer than those 10 seconds to 2339keep up with the timer (because it takes longer than those 10 seconds to
2340Periodic watchers are also timers of a kind, but they are very versatile 2421Periodic watchers are also timers of a kind, but they are very versatile
2341(and unfortunately a bit complex). 2422(and unfortunately a bit complex).
2342.PP 2423.PP
2343Unlike \f(CW\*(C`ev_timer\*(C'\fR, periodic watchers are not based on real time (or 2424Unlike \f(CW\*(C`ev_timer\*(C'\fR, periodic watchers are not based on real time (or
2344relative time, the physical time that passes) but on wall clock time 2425relative time, the physical time that passes) but on wall clock time
2345(absolute time, the thing you can read on your calender or clock). The 2426(absolute time, the thing you can read on your calendar or clock). The
2346difference is that wall clock time can run faster or slower than real 2427difference is that wall clock time can run faster or slower than real
2347time, and time jumps are not uncommon (e.g. when you adjust your 2428time, and time jumps are not uncommon (e.g. when you adjust your
2348wrist-watch). 2429wrist-watch).
2349.PP 2430.PP
2350You can tell a periodic watcher to trigger after some specific point 2431You can tell a periodic watcher to trigger after some specific point
2355\&\f(CW\*(C`ev_timer\*(C'\fR, which would still trigger roughly 10 seconds after starting 2436\&\f(CW\*(C`ev_timer\*(C'\fR, which would still trigger roughly 10 seconds after starting
2356it, as it uses a relative timeout). 2437it, as it uses a relative timeout).
2357.PP 2438.PP
2358\&\f(CW\*(C`ev_periodic\*(C'\fR watchers can also be used to implement vastly more complex 2439\&\f(CW\*(C`ev_periodic\*(C'\fR watchers can also be used to implement vastly more complex
2359timers, such as triggering an event on each \*(L"midnight, local time\*(R", or 2440timers, such as triggering an event on each \*(L"midnight, local time\*(R", or
2360other complicated rules. This cannot be done with \f(CW\*(C`ev_timer\*(C'\fR watchers, as 2441other complicated rules. This cannot easily be done with \f(CW\*(C`ev_timer\*(C'\fR
2361those cannot react to time jumps. 2442watchers, as those cannot react to time jumps.
2362.PP 2443.PP
2363As with timers, the callback is guaranteed to be invoked only when the 2444As with timers, the callback is guaranteed to be invoked only when the
2364point in time where it is supposed to trigger has passed. If multiple 2445point in time where it is supposed to trigger has passed. If multiple
2365timers become ready during the same loop iteration then the ones with 2446timers become ready during the same loop iteration then the ones with
2366earlier time-out values are invoked before ones with later time-out values 2447earlier time-out values are invoked before ones with later time-out values
2427In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`offset\*(C'\fR are both being 2508In this mode the values for \f(CW\*(C`interval\*(C'\fR and \f(CW\*(C`offset\*(C'\fR are both being
2428ignored. Instead, each time the periodic watcher gets scheduled, the 2509ignored. Instead, each time the periodic watcher gets scheduled, the
2429reschedule callback will be called with the watcher as first, and the 2510reschedule callback will be called with the watcher as first, and the
2430current time as second argument. 2511current time as second argument.
2431.Sp 2512.Sp
2432\&\s-1NOTE: \s0\fIThis callback \s-1MUST NOT\s0 stop or destroy any periodic watcher, ever, 2513\&\s-1NOTE:\s0 \fIThis callback \s-1MUST NOT\s0 stop or destroy any periodic watcher, ever,
2433or make \s-1ANY\s0 other event loop modifications whatsoever, unless explicitly 2514or make \s-1ANY\s0 other event loop modifications whatsoever, unless explicitly
2434allowed by documentation here\fR. 2515allowed by documentation here\fR.
2435.Sp 2516.Sp
2436If you need to stop it, return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop 2517If you need to stop it, return \f(CW\*(C`now + 1e30\*(C'\fR (or so, fudge fudge) and stop
2437it afterwards (e.g. by starting an \f(CW\*(C`ev_prepare\*(C'\fR watcher, which is the 2518it afterwards (e.g. by starting an \f(CW\*(C`ev_prepare\*(C'\fR watcher, which is the
2451It must return the next time to trigger, based on the passed time value 2532It must return the next time to trigger, based on the passed time value
2452(that is, the lowest time value larger than to the second argument). It 2533(that is, the lowest time value larger than to the second argument). It
2453will usually be called just before the callback will be triggered, but 2534will usually be called just before the callback will be triggered, but
2454might be called at other times, too. 2535might be called at other times, too.
2455.Sp 2536.Sp
2456\&\s-1NOTE: \s0\fIThis callback must always return a time that is higher than or 2537\&\s-1NOTE:\s0 \fIThis callback must always return a time that is higher than or
2457equal to the passed \f(CI\*(C`now\*(C'\fI value\fR. 2538equal to the passed \f(CI\*(C`now\*(C'\fI value\fR.
2458.Sp 2539.Sp
2459This can be used to create very complex timers, such as a timer that 2540This can be used to create very complex timers, such as a timer that
2460triggers on \*(L"next midnight, local time\*(R". To do this, you would calculate the 2541triggers on \*(L"next midnight, local time\*(R". To do this, you would calculate
2461next midnight after \f(CW\*(C`now\*(C'\fR and return the timestamp value for this. How 2542the next midnight after \f(CW\*(C`now\*(C'\fR and return the timestamp value for
2462you do this is, again, up to you (but it is not trivial, which is the main 2543this. Here is a (completely untested, no error checking) example on how to
2463reason I omitted it as an example). 2544do this:
2545.Sp
2546.Vb 1
2547\& #include <time.h>
2548\&
2549\& static ev_tstamp
2550\& my_rescheduler (ev_periodic *w, ev_tstamp now)
2551\& {
2552\& time_t tnow = (time_t)now;
2553\& struct tm tm;
2554\& localtime_r (&tnow, &tm);
2555\&
2556\& tm.tm_sec = tm.tm_min = tm.tm_hour = 0; // midnight current day
2557\& ++tm.tm_mday; // midnight next day
2558\&
2559\& return mktime (&tm);
2560\& }
2561.Ve
2562.Sp
2563Note: this code might run into trouble on days that have more then two
2564midnights (beginning and end).
2464.RE 2565.RE
2465.RS 4 2566.RS 4
2466.RE 2567.RE
2467.IP "ev_periodic_again (loop, ev_periodic *)" 4 2568.IP "ev_periodic_again (loop, ev_periodic *)" 4
2468.IX Item "ev_periodic_again (loop, ev_periodic *)" 2569.IX Item "ev_periodic_again (loop, ev_periodic *)"
2553only within the same loop, i.e. you can watch for \f(CW\*(C`SIGINT\*(C'\fR in your 2654only within the same loop, i.e. you can watch for \f(CW\*(C`SIGINT\*(C'\fR in your
2554default loop and for \f(CW\*(C`SIGIO\*(C'\fR in another loop, but you cannot watch for 2655default loop and for \f(CW\*(C`SIGIO\*(C'\fR in another loop, but you cannot watch for
2555\&\f(CW\*(C`SIGINT\*(C'\fR in both the default loop and another loop at the same time. At 2656\&\f(CW\*(C`SIGINT\*(C'\fR in both the default loop and another loop at the same time. At
2556the moment, \f(CW\*(C`SIGCHLD\*(C'\fR is permanently tied to the default loop. 2657the moment, \f(CW\*(C`SIGCHLD\*(C'\fR is permanently tied to the default loop.
2557.PP 2658.PP
2558When the first watcher gets started will libev actually register something 2659Only after the first watcher for a signal is started will libev actually
2559with the kernel (thus it coexists with your own signal handlers as long as 2660register something with the kernel. It thus coexists with your own signal
2560you don't register any with libev for the same signal). 2661handlers as long as you don't register any with libev for the same signal.
2561.PP 2662.PP
2562If possible and supported, libev will install its handlers with 2663If possible and supported, libev will install its handlers with
2563\&\f(CW\*(C`SA_RESTART\*(C'\fR (or equivalent) behaviour enabled, so system calls should 2664\&\f(CW\*(C`SA_RESTART\*(C'\fR (or equivalent) behaviour enabled, so system calls should
2564not be unduly interrupted. If you have a problem with system calls getting 2665not be unduly interrupted. If you have a problem with system calls getting
2565interrupted by signals you can block all signals in an \f(CW\*(C`ev_check\*(C'\fR watcher 2666interrupted by signals you can block all signals in an \f(CW\*(C`ev_check\*(C'\fR watcher
2586The simplest way to ensure that the signal mask is reset in the child is 2687The simplest way to ensure that the signal mask is reset in the child is
2587to install a fork handler with \f(CW\*(C`pthread_atfork\*(C'\fR that resets it. That will 2688to install a fork handler with \f(CW\*(C`pthread_atfork\*(C'\fR that resets it. That will
2588catch fork calls done by libraries (such as the libc) as well. 2689catch fork calls done by libraries (such as the libc) as well.
2589.PP 2690.PP
2590In current versions of libev, the signal will not be blocked indefinitely 2691In current versions of libev, the signal will not be blocked indefinitely
2591unless you use the \f(CW\*(C`signalfd\*(C'\fR \s-1API \s0(\f(CW\*(C`EV_SIGNALFD\*(C'\fR). While this reduces 2692unless you use the \f(CW\*(C`signalfd\*(C'\fR \s-1API\s0 (\f(CW\*(C`EV_SIGNALFD\*(C'\fR). While this reduces
2592the window of opportunity for problems, it will not go away, as libev 2693the window of opportunity for problems, it will not go away, as libev
2593\&\fIhas\fR to modify the signal mask, at least temporarily. 2694\&\fIhas\fR to modify the signal mask, at least temporarily.
2594.PP 2695.PP
2595So I can't stress this enough: \fIIf you do not reset your signal mask when 2696So I can't stress this enough: \fIIf you do not reset your signal mask when
2596you expect it to be empty, you have a race condition in your code\fR. This 2697you expect it to be empty, you have a race condition in your code\fR. This
3048.IX Subsection "ev_prepare and ev_check - customise your event loop!" 3149.IX Subsection "ev_prepare and ev_check - customise your event loop!"
3049Prepare and check watchers are often (but not always) used in pairs: 3150Prepare and check watchers are often (but not always) used in pairs:
3050prepare watchers get invoked before the process blocks and check watchers 3151prepare watchers get invoked before the process blocks and check watchers
3051afterwards. 3152afterwards.
3052.PP 3153.PP
3053You \fImust not\fR call \f(CW\*(C`ev_run\*(C'\fR or similar functions that enter 3154You \fImust not\fR call \f(CW\*(C`ev_run\*(C'\fR (or similar functions that enter the
3054the current event loop from either \f(CW\*(C`ev_prepare\*(C'\fR or \f(CW\*(C`ev_check\*(C'\fR 3155current event loop) or \f(CW\*(C`ev_loop_fork\*(C'\fR from either \f(CW\*(C`ev_prepare\*(C'\fR or
3055watchers. Other loops than the current one are fine, however. The 3156\&\f(CW\*(C`ev_check\*(C'\fR watchers. Other loops than the current one are fine,
3056rationale behind this is that you do not need to check for recursion in 3157however. The rationale behind this is that you do not need to check
3057those watchers, i.e. the sequence will always be \f(CW\*(C`ev_prepare\*(C'\fR, blocking, 3158for recursion in those watchers, i.e. the sequence will always be
3058\&\f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each kind they will always be 3159\&\f(CW\*(C`ev_prepare\*(C'\fR, blocking, \f(CW\*(C`ev_check\*(C'\fR so if you have one watcher of each
3059called in pairs bracketing the blocking call. 3160kind they will always be called in pairs bracketing the blocking call.
3060.PP 3161.PP
3061Their main purpose is to integrate other event mechanisms into libev and 3162Their main purpose is to integrate other event mechanisms into libev and
3062their use is somewhat advanced. They could be used, for example, to track 3163their use is somewhat advanced. They could be used, for example, to track
3063variable changes, implement your own watchers, integrate net-snmp or a 3164variable changes, implement your own watchers, integrate net-snmp or a
3064coroutine library and lots more. They are also occasionally useful if 3165coroutine library and lots more. They are also occasionally useful if
3358.PP 3459.PP
3359.Vb 3 3460.Vb 3
3360\& struct ev_loop *loop_hi = ev_default_init (0); 3461\& struct ev_loop *loop_hi = ev_default_init (0);
3361\& struct ev_loop *loop_lo = 0; 3462\& struct ev_loop *loop_lo = 0;
3362\& ev_embed embed; 3463\& ev_embed embed;
3363\& 3464\&
3364\& // see if there is a chance of getting one that works 3465\& // see if there is a chance of getting one that works
3365\& // (remember that a flags value of 0 means autodetection) 3466\& // (remember that a flags value of 0 means autodetection)
3366\& loop_lo = ev_embeddable_backends () & ev_recommended_backends () 3467\& loop_lo = ev_embeddable_backends () & ev_recommended_backends ()
3367\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ()) 3468\& ? ev_loop_new (ev_embeddable_backends () & ev_recommended_backends ())
3368\& : 0; 3469\& : 0;
3384.PP 3485.PP
3385.Vb 3 3486.Vb 3
3386\& struct ev_loop *loop = ev_default_init (0); 3487\& struct ev_loop *loop = ev_default_init (0);
3387\& struct ev_loop *loop_socket = 0; 3488\& struct ev_loop *loop_socket = 0;
3388\& ev_embed embed; 3489\& ev_embed embed;
3389\& 3490\&
3390\& if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE) 3491\& if (ev_supported_backends () & ~ev_recommended_backends () & EVBACKEND_KQUEUE)
3391\& if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE)) 3492\& if ((loop_socket = ev_loop_new (EVBACKEND_KQUEUE))
3392\& { 3493\& {
3393\& ev_embed_init (&embed, 0, loop_socket); 3494\& ev_embed_init (&embed, 0, loop_socket);
3394\& ev_embed_start (loop, &embed); 3495\& ev_embed_start (loop, &embed);
3411of course. 3512of course.
3412.PP 3513.PP
3413\fIThe special problem of life after fork \- how is it possible?\fR 3514\fIThe special problem of life after fork \- how is it possible?\fR
3414.IX Subsection "The special problem of life after fork - how is it possible?" 3515.IX Subsection "The special problem of life after fork - how is it possible?"
3415.PP 3516.PP
3416Most uses of \f(CW\*(C`fork()\*(C'\fR consist of forking, then some simple calls to set 3517Most uses of \f(CW\*(C`fork ()\*(C'\fR consist of forking, then some simple calls to set
3417up/change the process environment, followed by a call to \f(CW\*(C`exec()\*(C'\fR. This 3518up/change the process environment, followed by a call to \f(CW\*(C`exec()\*(C'\fR. This
3418sequence should be handled by libev without any problems. 3519sequence should be handled by libev without any problems.
3419.PP 3520.PP
3420This changes when the application actually wants to do event handling 3521This changes when the application actually wants to do event handling
3421in the child, or both parent in child, in effect \*(L"continuing\*(R" after the 3522in the child, or both parent in child, in effect \*(L"continuing\*(R" after the
3638is a time window between the event loop checking and resetting the async 3739is a time window between the event loop checking and resetting the async
3639notification, and the callback being invoked. 3740notification, and the callback being invoked.
3640.SH "OTHER FUNCTIONS" 3741.SH "OTHER FUNCTIONS"
3641.IX Header "OTHER FUNCTIONS" 3742.IX Header "OTHER FUNCTIONS"
3642There are some other functions of possible interest. Described. Here. Now. 3743There are some other functions of possible interest. Described. Here. Now.
3643.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 4 3744.IP "ev_once (loop, int fd, int events, ev_tstamp timeout, callback, arg)" 4
3644.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback)" 3745.IX Item "ev_once (loop, int fd, int events, ev_tstamp timeout, callback, arg)"
3645This function combines a simple timer and an I/O watcher, calls your 3746This function combines a simple timer and an I/O watcher, calls your
3646callback on whichever event happens first and automatically stops both 3747callback on whichever event happens first and automatically stops both
3647watchers. This is useful if you want to wait for a single event on an fd 3748watchers. This is useful if you want to wait for a single event on an fd
3648or timeout without having to allocate/configure/start/stop/free one or 3749or timeout without having to allocate/configure/start/stop/free one or
3649more watchers yourself. 3750more watchers yourself.
4049files, \fImy_ev.h\fR and \fImy_ev.c\fR that include the respective libev files: 4150files, \fImy_ev.h\fR and \fImy_ev.c\fR that include the respective libev files:
4050.PP 4151.PP
4051.Vb 4 4152.Vb 4
4052\& // my_ev.h 4153\& // my_ev.h
4053\& #define EV_CB_DECLARE(type) struct my_coro *cb; 4154\& #define EV_CB_DECLARE(type) struct my_coro *cb;
4054\& #define EV_CB_INVOKE(watcher) switch_to ((watcher)\->cb); 4155\& #define EV_CB_INVOKE(watcher) switch_to ((watcher)\->cb)
4055\& #include "../libev/ev.h" 4156\& #include "../libev/ev.h"
4056\& 4157\&
4057\& // my_ev.c 4158\& // my_ev.c
4058\& #define EV_H "my_ev.h" 4159\& #define EV_H "my_ev.h"
4059\& #include "../libev/ev.c" 4160\& #include "../libev/ev.c"
4099The normal C \s-1API\s0 should work fine when used from \*(C+: both ev.h and the 4200The normal C \s-1API\s0 should work fine when used from \*(C+: both ev.h and the
4100libev sources can be compiled as \*(C+. Therefore, code that uses the C \s-1API\s0 4201libev sources can be compiled as \*(C+. Therefore, code that uses the C \s-1API\s0
4101will work fine. 4202will work fine.
4102.PP 4203.PP
4103Proper exception specifications might have to be added to callbacks passed 4204Proper exception specifications might have to be added to callbacks passed
4104to libev: exceptions may be thrown only from watcher callbacks, all 4205to libev: exceptions may be thrown only from watcher callbacks, all other
4105other callbacks (allocator, syserr, loop acquire/release and periodic 4206callbacks (allocator, syserr, loop acquire/release and periodic reschedule
4106reschedule callbacks) must not throw exceptions, and might need a \f(CW\*(C`throw 4207callbacks) must not throw exceptions, and might need a \f(CW\*(C`noexcept\*(C'\fR
4107()\*(C'\fR specification. If you have code that needs to be compiled as both C 4208specification. If you have code that needs to be compiled as both C and
4108and \*(C+ you can use the \f(CW\*(C`EV_THROW\*(C'\fR macro for this: 4209\&\*(C+ you can use the \f(CW\*(C`EV_NOEXCEPT\*(C'\fR macro for this:
4109.PP 4210.PP
4110.Vb 6 4211.Vb 6
4111\& static void 4212\& static void
4112\& fatal_error (const char *msg) EV_THROW 4213\& fatal_error (const char *msg) EV_NOEXCEPT
4113\& { 4214\& {
4114\& perror (msg); 4215\& perror (msg);
4115\& abort (); 4216\& abort ();
4116\& } 4217\& }
4117\& 4218\&
4245\& void operator() (ev::io &w, int revents) 4346\& void operator() (ev::io &w, int revents)
4246\& { 4347\& {
4247\& ... 4348\& ...
4248\& } 4349\& }
4249\& } 4350\& }
4250\& 4351\&
4251\& myfunctor f; 4352\& myfunctor f;
4252\& 4353\&
4253\& ev::io w; 4354\& ev::io w;
4254\& w.set (&f); 4355\& w.set (&f);
4255.Ve 4356.Ve
4491\& #include "ev.c" 4592\& #include "ev.c"
4492.Ve 4593.Ve
4493.PP 4594.PP
4494This will automatically include \fIev.h\fR, too, and should be done in a 4595This will automatically include \fIev.h\fR, too, and should be done in a
4495single C source file only to provide the function implementations. To use 4596single C source file only to provide the function implementations. To use
4496it, do the same for \fIev.h\fR in all files wishing to use this \s-1API \s0(best 4597it, do the same for \fIev.h\fR in all files wishing to use this \s-1API\s0 (best
4497done by writing a wrapper around \fIev.h\fR that you can include instead and 4598done by writing a wrapper around \fIev.h\fR that you can include instead and
4498where you can put other configuration options): 4599where you can put other configuration options):
4499.PP 4600.PP
4500.Vb 2 4601.Vb 2
4501\& #define EV_STANDALONE 1 4602\& #define EV_STANDALONE 1
4515\& ev_vars.h 4616\& ev_vars.h
4516\& ev_wrap.h 4617\& ev_wrap.h
4517\& 4618\&
4518\& ev_win32.c required on win32 platforms only 4619\& ev_win32.c required on win32 platforms only
4519\& 4620\&
4520\& ev_select.c only when select backend is enabled (which is enabled by default) 4621\& ev_select.c only when select backend is enabled
4521\& ev_poll.c only when poll backend is enabled (disabled by default) 4622\& ev_poll.c only when poll backend is enabled
4522\& ev_epoll.c only when the epoll backend is enabled (disabled by default) 4623\& ev_epoll.c only when the epoll backend is enabled
4624\& ev_linuxaio.c only when the linux aio backend is enabled
4523\& ev_kqueue.c only when the kqueue backend is enabled (disabled by default) 4625\& ev_kqueue.c only when the kqueue backend is enabled
4524\& ev_port.c only when the solaris port backend is enabled (disabled by default) 4626\& ev_port.c only when the solaris port backend is enabled
4525.Ve 4627.Ve
4526.PP 4628.PP
4527\&\fIev.c\fR includes the backend files directly when enabled, so you only need 4629\&\fIev.c\fR includes the backend files directly when enabled, so you only need
4528to compile this single file. 4630to compile this single file.
4529.PP 4631.PP
4574values when compiling libev vs. including \fIev.h\fR, so it is permissible 4676values when compiling libev vs. including \fIev.h\fR, so it is permissible
4575to redefine them before including \fIev.h\fR without breaking compatibility 4677to redefine them before including \fIev.h\fR without breaking compatibility
4576to a compiled library. All other symbols change the \s-1ABI,\s0 which means all 4678to a compiled library. All other symbols change the \s-1ABI,\s0 which means all
4577users of libev and the libev code itself must be compiled with compatible 4679users of libev and the libev code itself must be compiled with compatible
4578settings. 4680settings.
4579.IP "\s-1EV_COMPAT3 \s0(h)" 4 4681.IP "\s-1EV_COMPAT3\s0 (h)" 4
4580.IX Item "EV_COMPAT3 (h)" 4682.IX Item "EV_COMPAT3 (h)"
4581Backwards compatibility is a major concern for libev. This is why this 4683Backwards compatibility is a major concern for libev. This is why this
4582release of libev comes with wrappers for the functions and symbols that 4684release of libev comes with wrappers for the functions and symbols that
4583have been renamed between libev version 3 and 4. 4685have been renamed between libev version 3 and 4.
4584.Sp 4686.Sp
4589typedef in that case. 4691typedef in that case.
4590.Sp 4692.Sp
4591In some future version, the default for \f(CW\*(C`EV_COMPAT3\*(C'\fR will become \f(CW0\fR, 4693In some future version, the default for \f(CW\*(C`EV_COMPAT3\*(C'\fR will become \f(CW0\fR,
4592and in some even more future version the compatibility code will be 4694and in some even more future version the compatibility code will be
4593removed completely. 4695removed completely.
4594.IP "\s-1EV_STANDALONE \s0(h)" 4 4696.IP "\s-1EV_STANDALONE\s0 (h)" 4
4595.IX Item "EV_STANDALONE (h)" 4697.IX Item "EV_STANDALONE (h)"
4596Must always be \f(CW1\fR if you do not use autoconf configuration, which 4698Must always be \f(CW1\fR if you do not use autoconf configuration, which
4597keeps libev from including \fIconfig.h\fR, and it also defines dummy 4699keeps libev from including \fIconfig.h\fR, and it also defines dummy
4598implementations for some libevent functions (such as logging, which is not 4700implementations for some libevent functions (such as logging, which is not
4599supported). It will also not define any of the structs usually found in 4701supported). It will also not define any of the structs usually found in
4708If defined to be \f(CW1\fR, libev will compile in support for the Linux 4810If defined to be \f(CW1\fR, libev will compile in support for the Linux
4709\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime, 4811\&\f(CW\*(C`epoll\*(C'\fR(7) backend. Its availability will be detected at runtime,
4710otherwise another method will be used as fallback. This is the preferred 4812otherwise another method will be used as fallback. This is the preferred
4711backend for GNU/Linux systems. If undefined, it will be enabled if the 4813backend for GNU/Linux systems. If undefined, it will be enabled if the
4712headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled. 4814headers indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
4815.IP "\s-1EV_USE_LINUXAIO\s0" 4
4816.IX Item "EV_USE_LINUXAIO"
4817If defined to be \f(CW1\fR, libev will compile in support for the Linux
4818aio backend. Due to it's currenbt limitations it has to be requested
4819explicitly. If undefined, it will be enabled on linux, otherwise
4820disabled.
4713.IP "\s-1EV_USE_KQUEUE\s0" 4 4821.IP "\s-1EV_USE_KQUEUE\s0" 4
4714.IX Item "EV_USE_KQUEUE" 4822.IX Item "EV_USE_KQUEUE"
4715If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style 4823If defined to be \f(CW1\fR, libev will compile in support for the \s-1BSD\s0 style
4716\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime, 4824\&\f(CW\*(C`kqueue\*(C'\fR(2) backend. Its actual availability will be detected at runtime,
4717otherwise another method will be used as fallback. This is the preferred 4825otherwise another method will be used as fallback. This is the preferred
4757handler \*(L"locking\*(R" as well as for signal and thread safety in \f(CW\*(C`ev_async\*(C'\fR 4865handler \*(L"locking\*(R" as well as for signal and thread safety in \f(CW\*(C`ev_async\*(C'\fR
4758watchers. 4866watchers.
4759.Sp 4867.Sp
4760In the absence of this define, libev will use \f(CW\*(C`sig_atomic_t volatile\*(C'\fR 4868In the absence of this define, libev will use \f(CW\*(C`sig_atomic_t volatile\*(C'\fR
4761(from \fIsignal.h\fR), which is usually good enough on most platforms. 4869(from \fIsignal.h\fR), which is usually good enough on most platforms.
4762.IP "\s-1EV_H \s0(h)" 4 4870.IP "\s-1EV_H\s0 (h)" 4
4763.IX Item "EV_H (h)" 4871.IX Item "EV_H (h)"
4764The name of the \fIev.h\fR header file used to include it. The default if 4872The name of the \fIev.h\fR header file used to include it. The default if
4765undefined is \f(CW"ev.h"\fR in \fIevent.h\fR, \fIev.c\fR and \fIev++.h\fR. This can be 4873undefined is \f(CW"ev.h"\fR in \fIevent.h\fR, \fIev.c\fR and \fIev++.h\fR. This can be
4766used to virtually rename the \fIev.h\fR header file in case of conflicts. 4874used to virtually rename the \fIev.h\fR header file in case of conflicts.
4767.IP "\s-1EV_CONFIG_H \s0(h)" 4 4875.IP "\s-1EV_CONFIG_H\s0 (h)" 4
4768.IX Item "EV_CONFIG_H (h)" 4876.IX Item "EV_CONFIG_H (h)"
4769If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override 4877If \f(CW\*(C`EV_STANDALONE\*(C'\fR isn't \f(CW1\fR, this variable can be used to override
4770\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to 4878\&\fIev.c\fR's idea of where to find the \fIconfig.h\fR file, similarly to
4771\&\f(CW\*(C`EV_H\*(C'\fR, above. 4879\&\f(CW\*(C`EV_H\*(C'\fR, above.
4772.IP "\s-1EV_EVENT_H \s0(h)" 4 4880.IP "\s-1EV_EVENT_H\s0 (h)" 4
4773.IX Item "EV_EVENT_H (h)" 4881.IX Item "EV_EVENT_H (h)"
4774Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea 4882Similarly to \f(CW\*(C`EV_H\*(C'\fR, this macro can be used to override \fIevent.c\fR's idea
4775of how the \fIevent.h\fR header can be found, the default is \f(CW"event.h"\fR. 4883of how the \fIevent.h\fR header can be found, the default is \f(CW"event.h"\fR.
4776.IP "\s-1EV_PROTOTYPES \s0(h)" 4 4884.IP "\s-1EV_PROTOTYPES\s0 (h)" 4
4777.IX Item "EV_PROTOTYPES (h)" 4885.IX Item "EV_PROTOTYPES (h)"
4778If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function 4886If defined to be \f(CW0\fR, then \fIev.h\fR will not define any function
4779prototypes, but still define all the structs and other symbols. This is 4887prototypes, but still define all the structs and other symbols. This is
4780occasionally useful if you want to provide your own wrapper functions 4888occasionally useful if you want to provide your own wrapper functions
4781around libev functions. 4889around libev functions.
4990.Vb 3 5098.Vb 3
4991\& #define EV_COMMON \e 5099\& #define EV_COMMON \e
4992\& SV *self; /* contains this struct */ \e 5100\& SV *self; /* contains this struct */ \e
4993\& SV *cb_sv, *fh /* note no trailing ";" */ 5101\& SV *cb_sv, *fh /* note no trailing ";" */
4994.Ve 5102.Ve
4995.IP "\s-1EV_CB_DECLARE \s0(type)" 4 5103.IP "\s-1EV_CB_DECLARE\s0 (type)" 4
4996.IX Item "EV_CB_DECLARE (type)" 5104.IX Item "EV_CB_DECLARE (type)"
4997.PD 0 5105.PD 0
4998.IP "\s-1EV_CB_INVOKE \s0(watcher, revents)" 4 5106.IP "\s-1EV_CB_INVOKE\s0 (watcher, revents)" 4
4999.IX Item "EV_CB_INVOKE (watcher, revents)" 5107.IX Item "EV_CB_INVOKE (watcher, revents)"
5000.IP "ev_set_cb (ev, cb)" 4 5108.IP "ev_set_cb (ev, cb)" 4
5001.IX Item "ev_set_cb (ev, cb)" 5109.IX Item "ev_set_cb (ev, cb)"
5002.PD 5110.PD
5003Can be used to change the callback member declaration in each watcher, 5111Can be used to change the callback member declaration in each watcher,
5006their default definitions. One possible use for overriding these is to 5114their default definitions. One possible use for overriding these is to
5007avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use 5115avoid the \f(CW\*(C`struct ev_loop *\*(C'\fR as first argument in all cases, or to use
5008method calls instead of plain function calls in \*(C+. 5116method calls instead of plain function calls in \*(C+.
5009.SS "\s-1EXPORTED API SYMBOLS\s0" 5117.SS "\s-1EXPORTED API SYMBOLS\s0"
5010.IX Subsection "EXPORTED API SYMBOLS" 5118.IX Subsection "EXPORTED API SYMBOLS"
5011If you need to re-export the \s-1API \s0(e.g. via a \s-1DLL\s0) and you need a list of 5119If you need to re-export the \s-1API\s0 (e.g. via a \s-1DLL\s0) and you need a list of
5012exported symbols, you can use the provided \fISymbol.*\fR files which list 5120exported symbols, you can use the provided \fISymbol.*\fR files which list
5013all public symbols, one per line: 5121all public symbols, one per line:
5014.PP 5122.PP
5015.Vb 2 5123.Vb 2
5016\& Symbols.ev for libev proper 5124\& Symbols.ev for libev proper
5248.PP 5356.PP
5249\fI\f(CI\*(C`select\*(C'\fI is buggy\fR 5357\fI\f(CI\*(C`select\*(C'\fI is buggy\fR
5250.IX Subsection "select is buggy" 5358.IX Subsection "select is buggy"
5251.PP 5359.PP
5252All that's left is \f(CW\*(C`select\*(C'\fR, and of course Apple found a way to fuck this 5360All that's left is \f(CW\*(C`select\*(C'\fR, and of course Apple found a way to fuck this
5253one up as well: On \s-1OS/X, \s0\f(CW\*(C`select\*(C'\fR actively limits the number of file 5361one up as well: On \s-1OS/X,\s0 \f(CW\*(C`select\*(C'\fR actively limits the number of file
5254descriptors you can pass in to 1024 \- your program suddenly crashes when 5362descriptors you can pass in to 1024 \- your program suddenly crashes when
5255you use more. 5363you use more.
5256.PP 5364.PP
5257There is an undocumented \*(L"workaround\*(R" for this \- defining 5365There is an undocumented \*(L"workaround\*(R" for this \- defining
5258\&\f(CW\*(C`_DARWIN_UNLIMITED_SELECT\*(C'\fR, which libev tries to use, so select \fIshould\fR 5366\&\f(CW\*(C`_DARWIN_UNLIMITED_SELECT\*(C'\fR, which libev tries to use, so select \fIshould\fR
5406Libev assumes not only that all watcher pointers have the same internal 5514Libev assumes not only that all watcher pointers have the same internal
5407structure (guaranteed by \s-1POSIX\s0 but not by \s-1ISO C\s0 for example), but it also 5515structure (guaranteed by \s-1POSIX\s0 but not by \s-1ISO C\s0 for example), but it also
5408assumes that the same (machine) code can be used to call any watcher 5516assumes that the same (machine) code can be used to call any watcher
5409callback: The watcher callbacks have different type signatures, but libev 5517callback: The watcher callbacks have different type signatures, but libev
5410calls them using an \f(CW\*(C`ev_watcher *\*(C'\fR internally. 5518calls them using an \f(CW\*(C`ev_watcher *\*(C'\fR internally.
5519.IP "null pointers and integer zero are represented by 0 bytes" 4
5520.IX Item "null pointers and integer zero are represented by 0 bytes"
5521Libev uses \f(CW\*(C`memset\*(C'\fR to initialise structs and arrays to \f(CW0\fR bytes, and
5522relies on this setting pointers and integers to null.
5411.IP "pointer accesses must be thread-atomic" 4 5523.IP "pointer accesses must be thread-atomic" 4
5412.IX Item "pointer accesses must be thread-atomic" 5524.IX Item "pointer accesses must be thread-atomic"
5413Accessing a pointer value must be atomic, it must both be readable and 5525Accessing a pointer value must be atomic, it must both be readable and
5414writable in one piece \- this is the case on all current architectures. 5526writable in one piece \- this is the case on all current architectures.
5415.ie n .IP """sig_atomic_t volatile"" must be thread-atomic as well" 4 5527.ie n .IP """sig_atomic_t volatile"" must be thread-atomic as well" 4

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines