[ViewVC] Diff of: cvs/libev/ev.pod

Comparing libev/ev.pod (file contents):
Revision 1.369 by root, Mon May 30 18:34:28 2011 UTC vs.
Revision 1.382 by sf-exg, Mon Aug 15 10:18:07 2011 UTC

 you actually want to know. Also interesting is the combination of
 C<ev_update_now> and C<ev_now>.
 =item ev_sleep (ev_tstamp interval)
-Sleep for the given interval: The current thread will be blocked until
+Sleep for the given interval: The current thread will be blocked
-either it is interrupted or the given time interval has passed. Basically
+until either it is interrupted or the given time interval has
+passed (approximately - it might return a bit earlier even if not
+interrupted). Returns immediately if C<< interval <= 0 >>.
-this is a sub-second-resolution C<sleep ()>.
+Basically this is a sub-second-resolution C<sleep ()>.
+The range of the C<interval> is limited - libev only guarantees to work
+with sleep times of up to one day (C<< interval <= 86400 >>).
 =item int ev_version_major ()
 =item int ev_version_minor ()
 example) that can't properly initialise their signal masks.
 =item C<EVFLAG_NOSIGMASK>
 When this flag is specified, then libev will avoid to modify the signal
-mask. Specifically, this means you ahve to make sure signals are unblocked
+mask. Specifically, this means you have to make sure signals are unblocked
 when you want to receive them.
 This behaviour is useful when you want to do your own signal handling, or
 want to handle signals only in specific threads and want to avoid libev
 unblocking the signals.
 0.1ms) and so on. The biggest issue is fork races, however - if a program
 forks then I<both> parent and child process have to recreate the epoll
 set, which can take considerable time (one syscall per file descriptor)
 and is of course hard to detect.
-Epoll is also notoriously buggy - embedding epoll fds I<should> work, but
+Epoll is also notoriously buggy - embedding epoll fds I<should> work,
-of course I<doesn't>, and epoll just loves to report events for totally
+but of course I<doesn't>, and epoll just loves to report events for
-I<different> file descriptors (even already closed ones, so one cannot
+totally I<different> file descriptors (even already closed ones, so
-even remove them from the set) than registered in the set (especially
+one cannot even remove them from the set) than registered in the set
-on SMP systems). Libev tries to counter these spurious notifications by
+(especially on SMP systems). Libev tries to counter these spurious
-employing an additional generation counter and comparing that against the
+notifications by employing an additional generation counter and comparing
-events to filter out spurious ones, recreating the set when required. Last
+that against the events to filter out spurious ones, recreating the set
+when required. Epoll also erroneously rounds down timeouts, but gives you
+no way to know when and by how much, so sometimes you have to busy-wait
+because epoll returns immediately despite a nonzero timeout. And last
 not least, it also refuses to work with some file descriptors which work
 perfectly fine with C<select> (files, many character devices...).
-Epoll is truly the train wreck analog among event poll mechanisms,
+Epoll is truly the train wreck among event poll mechanisms, a frankenpoll,
-a frankenpoll, cobbled together in a hurry, no thought to design or
+cobbled together in a hurry, no thought to design or interaction with
-interaction with others.
+others. Oh, the pain, will it ever stop...
 While stopping, setting and starting an I/O watcher in the same iteration
 will result in some caching, there is still a system call per such
 incident (because the same I<file descriptor> could point to a different
 I<file description> now), so its best to avoid that. Also, C<dup ()>'ed
 among the OS-specific backends (I vastly prefer correctness over speed
 hacks).
 On the negative side, the interface is I<bizarre> - so bizarre that
 even sun itself gets it wrong in their code examples: The event polling
-function sometimes returning events to the caller even though an error
+function sometimes returns events to the caller even though an error
 occurred, but with no indication whether it has done so or not (yes, it's
-even documented that way) - deadly for edge-triggered interfaces where
+even documented that way) - deadly for edge-triggered interfaces where you
-you absolutely have to know whether an event occurred or not because you
+absolutely have to know whether an event occurred or not because you have
-have to re-arm the watcher.
+to re-arm the watcher.
 Fortunately libev seems to be able to work around these idiocies.
 This backend maps C<EV_READ> and C<EV_WRITE> in the same way as
 C<EVBACKEND_POLL>.
 overhead for the actual polling but can deliver many events at once.
 By setting a higher I<io collect interval> you allow libev to spend more
 time collecting I/O events, so you can handle more events per iteration,
 at the cost of increasing latency. Timeouts (both C<ev_periodic> and
-C<ev_timer>) will be not affected. Setting this to a non-null value will
+C<ev_timer>) will not be affected. Setting this to a non-null value will
 introduce an additional C<ev_sleep ()> call into most loop iterations. The
 sleep time ensures that libev will not poll for I/O events more often then
-once per this interval, on average.
+once per this interval, on average (as long as the host time resolution is
+good enough).
 Likewise, by setting a higher I<timeout collect interval> you allow libev
 to spend more time collecting timeouts, at the expense of increased
 latency/jitter/inexactness (the watcher callback will be called
 later). C<ev_io> watchers will not be affected. Setting this to a non-null
 =over 4
 =item initialiased
-Before a watcher can be registered with the event looop it has to be
+Before a watcher can be registered with the event loop it has to be
 initialised. This can be done with a call to C<ev_TYPE_init>, or calls to
 C<ev_init> followed by the watcher-specific C<ev_TYPE_set> function.
 In this state it is simply some block of memory that is suitable for
 use in an event loop. It can be moved around, freed, reused etc. at
 detecting time jumps is hard, and some inaccuracies are unavoidable (the
 monotonic clock option helps a lot here).
 The callback is guaranteed to be invoked only I<after> its timeout has
 passed (not I<at>, so on systems with very low-resolution clocks this
-might introduce a small delay). If multiple timers become ready during the
+might introduce a small delay, see "the special problem of being too
+early", below). If multiple timers become ready during the same loop
-same loop iteration then the ones with earlier time-out values are invoked
+iteration then the ones with earlier time-out values are invoked before
-before ones of the same priority with later time-out values (but this is
+ones of the same priority with later time-out values (but this is no
-no longer true when a callback calls C<ev_run> recursively).
+longer true when a callback calls C<ev_run> recursively).
 =head3 Be smart about timeouts
 Many real-world problems involve some kind of timeout, usually for error
 recovery. A typical example is an HTTP request - if the other side hangs,
 Method #1 is almost always a bad idea, and buys you nothing. Method #4 is
 rather complicated, but extremely efficient, something that really pays
 off after the first million or so of active timers, i.e. it's usually
 overkill :)
+=head3 The special problem of being too early
+If you ask a timer to call your callback after three seconds, then
+you expect it to be invoked after three seconds - but of course, this
+cannot be guaranteed to infinite precision. Less obviously, it cannot be
+guaranteed to any precision by libev - imagine somebody suspending the
+process a STOP signal for a few hours for example.
+So, libev tries to invoke your callback as soon as possible I<after> the
+delay has occurred, but cannot guarantee this.
+A less obvious failure mode is calling your callback too early: many event
+loops compare timestamps with a "elapsed delay >= requested delay", but
+this can cause your callback to be invoked much earlier than you would
+expect.
+To see why, imagine a system with a clock that only offers full second
+resolution (think windows if you can't come up with a broken enough OS
+yourself). If you schedule a one-second timer at the time 500.9, then the
+event loop will schedule your timeout to elapse at a system time of 500
+(500.9 truncated to the resolution) + 1, or 501.
+If an event library looks at the timeout 0.1s later, it will see "501 >=
+501" and invoke the callback 0.1s after it was started, even though a
+one-second delay was requested - this is being "too early", despite best
+intentions.
+This is the reason why libev will never invoke the callback if the elapsed
+delay equals the requested delay, but only when the elapsed delay is
+larger than the requested delay. In the example above, libev would only invoke
+the callback at system time 502, or 1.1s after the timer was started.
+So, while libev cannot guarantee that your callback will be invoked
+exactly when requested, it I<can> and I<does> guarantee that the requested
+delay has actually elapsed, or in other words, it always errs on the "too
+late" side of things.
 =head3 The special problem of time updates
 Establishing the current time is a costly operation (it usually takes at
 least two system calls): EV therefore updates its idea of the current
 time only before and after C<ev_run> collects new events, which causes a
    ev_timer_set (&timer, after + ev_now () - ev_time (), 0.);
 If the event loop is suspended for a long time, you can also force an
 update of the time returned by C<ev_now ()> by calling C<ev_now_update
 ()>.
+=head3 The special problem of unsychronised clocks
+Modern systems have a variety of clocks - libev itself uses the normal
+"wall clock" clock and, if available, the monotonic clock (to avoid time
+jumps).
+Neither of these clocks is synchronised with each other or any other clock
+on the system, so C<ev_time ()> might return a considerably different time
+than C<gettimeofday ()> or C<time ()>. On a GNU/Linux system, for example,
+a call to C<gettimeofday> might return a second count that is one higher
+than a directly following call to C<time>.
+The moral of this is to only compare libev-related timestamps with
+C<ev_time ()> and C<ev_now ()>, at least if you want better precision than
+a second or so.
+One more problem arises due to this lack of synchronisation: if libev uses
+the system monotonic clock and you compare timestamps from C<ev_time>
+or C<ev_now> from when you started your timer and when your callback is
+invoked, you will find that sometimes the callback is a bit "early".
+This is because C<ev_timer>s work in real time, not wall clock time, so
+libev makes sure your callback is not invoked before the delay happened,
+I<measured according to the real time>, not the system clock.
+If your timeouts are based on a physical timescale (e.g. "time out this
+connection after 100 seconds") then this shouldn't bother you as it is
+exactly the right behaviour.
+If you want to compare wall clock/system timestamps to your timers, then
+you need to use C<ev_periodic>s, as these are based on the wall clock
+time, where your comparisons will always generate correct results.
 =head3 The special problems of suspended animation
 When you leave the server world it is quite customary to hit machines that
 can suspend/hibernate - what happens to the clocks during such a suspend?
 keep up with the timer (because it takes longer than those 10 seconds to
 do stuff) the timer will not fire more than once per event loop iteration.
 =item ev_timer_again (loop, ev_timer *)
-This will act as if the timer timed out and restart it again if it is
+This will act as if the timer timed out and restarts it again if it is
 repeating. The exact semantics are:
 If the timer is pending, its pending status is cleared.
 If the timer is started but non-repeating, stop it (as if it timed out).
 C<ev_async_sent> calls). In fact, you could use signal watchers as a kind
 of "global async watchers" by using a watcher on an otherwise unused
 signal, and C<ev_feed_signal> to signal this watcher from another thread,
 even without knowing which loop owns the signal.
-Unlike C<ev_signal> watchers, C<ev_async> works with any event loop, not
-just the default loop.
 =head3 Queueing
 C<ev_async> does not support queueing of data in any way. The reason
 is that the author does not know of a simple (or any) algorithm for a
 multiple-writer-single-reader queue that works in all cases and doesn't
 Unlike C<ev_feed_event>, this call is safe to do from other threads,
 signal or similar contexts (see the discussion of C<EV_ATOMIC_T> in the
 embedding section below on what exactly this means).
 Note that, as with other watchers in libev, multiple events might get
-compressed into a single callback invocation (another way to look at this
+compressed into a single callback invocation (another way to look at
-is that C<ev_async> watchers are level-triggered, set on C<ev_async_send>,
+this is that C<ev_async> watchers are level-triggered: they are set on
-reset when the event loop detects that).
+C<ev_async_send>, reset when the event loop detects that).
-This call incurs the overhead of a system call only once per event loop
+This call incurs the overhead of at most one extra system call per event
-iteration, so while the overhead might be noticeable, it doesn't apply to
+loop iteration, if the event loop is blocked, and no syscall at all if
-repeated calls to C<ev_async_send> for the same event loop.
+the event loop (or your program) is processing events. That means that
+repeated calls are basically free (there is no need to avoid calls for
+performance reasons) and that the overhead becomes smaller (typically
+zero) under load.
 =item bool = ev_async_pending (ev_async *)
 Returns a non-zero value when C<ev_async_send> has been called on the
 watcher but the event has not yet been processed (or even noted) by the
 watchers in the constructor.
    class myclass
    {
      ev::io   io  ; void io_cb   (ev::io   &w, int revents);
-     ev::io2  io2 ; void io2_cb  (ev::io   &w, int revents);
+     ev::io   io2 ; void io2_cb  (ev::io   &w, int revents);
      ev::idle idle; void idle_cb (ev::idle &w, int revents);
      myclass (int fd)
      {
        io  .set <myclass, &myclass::io_cb  > (this);
 L<http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hlibev>.
 =item D
 Leandro Lucarella has written a D language binding (F<ev.d>) for libev, to
-be found at L<http://proj.llucax.com.ar/wiki/evd>.
+be found at L<http://www.llucax.com.ar/proj/ev.d/index.html>.
 =item Ocaml
 Erkki Seppala has written Ocaml bindings for libev, to be found at
 L<http://modeemi.cs.tut.fi/~flux/software/ocaml-ev/>.
 suitable for use with C<EV_A>.
 =item C<EV_DEFAULT>, C<EV_DEFAULT_>
 Similar to the other two macros, this gives you the value of the default
-loop, if multiple loops are supported ("ev loop default").
+loop, if multiple loops are supported ("ev loop default"). The default loop
+will be initialised if it isn't already initialised.
+For non-multiplicity builds, these macros do nothing, so you always have
+to initialise the loop somewhere.
 =item C<EV_DEFAULT_UC>, C<EV_DEFAULT_UC_>
 Usage identical to C<EV_DEFAULT> and C<EV_DEFAULT_>, but requires that the
 default loop has been initialised (C<UC> == unchecked). Their behaviour
 indicate GNU/Linux + Glibc 2.4 or newer, otherwise disabled.
 =item EV_ATOMIC_T
 Libev requires an integer type (suitable for storing C<0> or C<1>) whose
-access is atomic with respect to other threads or signal contexts. No such
+access is atomic and serialised with respect to other threads or signal
-type is easily found in the C language, so you can provide your own type
+contexts. No such type is easily found in the C language, so you can
-that you know is safe for your purposes. It is used both for signal handler "locking"
+provide your own type that you know is safe for your purposes. It is used
-as well as for signal and thread safety in C<ev_async> watchers.
+both for signal handler "locking" as well as for signal and thread safety
+in C<ev_async> watchers.
 In the absence of this define, libev will use C<sig_atomic_t volatile>
-(from F<signal.h>), which is usually good enough on most platforms.
+(from F<signal.h>), which is usually good enough on most platforms,
+although strictly speaking using a type that also implies a memory fence
+is required.
 =item EV_H (h)
 The name of the F<ev.h> header file used to include it. The default if
 undefined is C<"ev.h"> in F<event.h>, F<ev.c> and F<ev++.h>. This can be
 If undefined or defined to C<1>, then all event-loop-specific functions
 will have the C<struct ev_loop *> as first argument, and you can create
 additional independent event loops. Otherwise there will be no support
 for multiple event loops and there is no first event loop pointer
 argument. Instead, all functions act on the single default loop.
+Note that C<EV_DEFAULT> and C<EV_DEFAULT_> will no longer provide a
+default loop when multiplicity is switched off - you always have to
+initialise the loop manually in this case.
 =item EV_MINPRI
 =item EV_MAXPRI
 requires, and its I/O model is fundamentally incompatible with the POSIX
 model. Libev still offers limited functionality on this platform in
 the form of the C<EVBACKEND_SELECT> backend, and only supports socket
 descriptors. This only applies when using Win32 natively, not when using
 e.g. cygwin. Actually, it only applies to the microsofts own compilers,
-as every compielr comes with a slightly differently broken/incompatible
+as every compiler comes with a slightly differently broken/incompatible
 environment.
 Lifting these limitations would basically require the full
 re-implementation of the I/O system. If you are into this kind of thing,
 then note that glib does exactly that for you in a very portable way (note
 The type C<double> is used to represent timestamps. It is required to
 have at least 51 bits of mantissa (and 9 bits of exponent), which is
 good enough for at least into the year 4000 with millisecond accuracy
 (the design goal for libev). This requirement is overfulfilled by
-implementations using IEEE 754, which is basically all existing ones. With
+implementations using IEEE 754, which is basically all existing ones.
-IEEE 754 doubles, you get microsecond accuracy until at least 2200.
+With IEEE 754 doubles, you get microsecond accuracy until at least the
+year 2255 (and millisecond accuracy till the year 287396 - by then, libev
+is either obsolete or somebody patched it to use C<long double> or
+something like that, just kidding).
 =back
 If you know of other additional requirements drop me a note.
 =item Processing ev_async_send: O(number_of_async_watchers)
 =item Processing signals: O(max_signal_number)
 Sending involves a system call I<iff> there were no other C<ev_async_send>
-calls in the current loop iteration. Checking for async and signal events
+calls in the current loop iteration and the loop is currently
+blocked. Checking for async and signal events involves iterating over all
-involves iterating over all running async watchers or all signal numbers.
+running async watchers or all signal numbers.
 =back
 =head1 PORTING FROM LIBEV 3.X TO 4.X

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing libev/ev.pod (file contents): Revision 1.369 by root, Mon May 30 18:34:28 2011 UTC vs. Revision 1.382 by sf-exg, Mon Aug 15 10:18:07 2011 UTC

Diff Legend

Comparing libev/ev.pod (file contents):
Revision 1.369 by root, Mon May 30 18:34:28 2011 UTC vs.
Revision 1.382 by sf-exg, Mon Aug 15 10:18:07 2011 UTC