--- IO-AIO/AIO.pm 2006/10/22 10:33:26 1.59 +++ IO-AIO/AIO.pm 2006/10/23 00:50:10 1.64 @@ -81,7 +81,7 @@ our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move - aio_group); + aio_group aio_nop); our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; @@ -354,10 +354,10 @@ you can recurse into (directories or links to them), and ones you cannot recurse into (everything else). -C is a composite request that consists of many sub -requests. C<$maxreq> specifies the maximum number of outstanding aio -requests that this function generates. If it is C<< <= 0 >>, then a -suitable default will be chosen (currently 8). +C is a composite request that creates of many sub requests_ +C<$maxreq> specifies the maximum number of outstanding aio requests that +this function generates. If it is C<< <= 0 >>, then a suitable default +will be chosen (currently 6). On error, the callback is called without arguments, otherwise it receives two array-refs with path-relative entry names. @@ -406,7 +406,7 @@ my $grp = aio_group $cb; - $maxreq = 8 if $maxreq <= 0; + $maxreq = 6 if $maxreq <= 0; # stat once add $grp aio_stat $path, sub { @@ -447,19 +447,21 @@ my ($statcb, $schedcb); my $nreq = 0; + my $statgrp = add $grp aio_group; + $schedcb = sub { if (@$entries) { if ($nreq < $maxreq) { my $ent = pop @$entries; $nreq++; - add $grp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; + add $statgrp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; } } elsif (!$nreq) { # finished + $statgrp->cancel; undef $statcb; undef $schedcb; - $grp->result (\@dirs, \@nondirs) if $cb; - undef $cb; + $grp->result (\@dirs, \@nondirs); } }; $statcb = sub { @@ -533,6 +535,19 @@ (aio_stat ...), ...; +=item aio_nop $callback->() + +This is a special request - it does nothing in itself and is only used for +side effects, such as when you want to add a dummy request to a group so +that finishing the requests in the group depends on executing the given +code. + +While this request does nothing, it still goes through the execution +phase and still requires a worker thread. Thus, the callback will not +be executed immediately but only after other requests in the queue have +entered their execution phase. This can be used to measure request +latency. + =item IO::AIO::aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED* Mainly used for debugging and benchmarking, this aio request puts one of @@ -603,13 +618,24 @@ This makes it very easy to create composite requests (see the source of C for an application) that work and feel like simple requests. -The IO::AIO::GRP objects will be cleaned up during calls to +=over 4 + +=item * The IO::AIO::GRP objects will be cleaned up during calls to C, just like any other request. -They can be canceled like any other request. Canceling will cancel not +=item * They can be canceled like any other request. Canceling will cancel not only the request itself, but also all requests it contains. -They can also can also be added to other IO::AIO::GRP objects. +=item * They can also can also be added to other IO::AIO::GRP objects. + +=item * You must not add requests to a group from within the group callback (or +any later time). + +=item * This does not harmonise well with C, so best do +not combine C with it. Groups and feeders are recommended for +this kind of concurrency-limiting. + +=back Their lifetime, simplified, looks like this: when they are empty, they will finish very quickly. If they contain only requests that are in the @@ -638,6 +664,51 @@ Set the result value(s) that will be passed to the group callback when all subrequests have finished. By default, no argument will be passed. +=item $grp->set_feeder ($callback->($grp)) + +[VERY EXPERIMENTAL] + +Sets a feeder/generator on this group: every group can have an attached +generator that generates requests if idle. The idea behind this is that, +although you could just queue as many requests as you want in a group, +this might starve other requests for a potentially long time. For +example, C might generate hundreds of thousands C +requests, delaying any later requests for a long time. + +To avoid this, and allow incremental generation of requests, you can +instead a group and set a feeder on it that generates those requests. The +feeder will be called whenever there are few enough (see C, +below) requests active in the group itself and is expected to queue more +requests. + +The feeder can queue as many requests as it likes (i.e. C does not +impose any limits). + +If the feeder does not queue more requests when called, it will be +automatically removed from the group. + +If the feeder limit is C<0>, it will be set to C<2> automatically. + +Example: + + # stat all files in @files, but only ever use four aio requests concurrently: + + my $grp = aio_group sub { print "finished\n" }; + $grp->feeder_limit (4); + $grp->set_feeder (sub { + my $file = pop @files + or return; + + add $grp aio_stat $file, sub { ... }; + }); + +=item $grp->feeder_limit ($num) + +Sets the feeder limit for the group: The feeder will be called whenever +the group contains less than this many requests. + +Setting the limit to C<0> will pause the feeding process. + =back =head2 SUPPORT FUNCTIONS @@ -704,17 +775,18 @@ =item IO::AIO::min_parallel $nthreads -Set the minimum number of AIO threads to C<$nthreads>. The current default -is C<4>, which means four asynchronous operations can be done at one time -(the number of outstanding operations, however, is unlimited). +Set the minimum number of AIO threads to C<$nthreads>. The current +default is C<8>, which means eight asynchronous operations can execute +concurrently at any one time (the number of outstanding requests, +however, is unlimited). IO::AIO starts threads only on demand, when an AIO request is queued and no free thread exists. -It is recommended to keep the number of threads low, as some Linux -kernel versions will scale negatively with the number of threads (higher -parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 -threads should be fine. +It is recommended to keep the number of threads relatively low, as some +Linux kernel versions will scale negatively with the number of threads +(higher parallelity => MUCH higher latency). With current Linux 2.6 +versions, 4-32 threads should be fine. Under most circumstances you don't need to call this function, as the module selects a default that is suitable for low to moderate load. @@ -735,6 +807,8 @@ =item $oldnreqs = IO::AIO::max_outstanding $nreqs +[DEPRECATED] + Sets the maximum number of outstanding requests to C<$nreqs>. If you try to queue up more than this number of requests, the caller will block until some requests have been handled. @@ -743,6 +817,10 @@ queue up many requests in a loop it often improves speed if you set this to a relatively low number, such as C<100>. +This function does not work well together with C's, and their +feeder interface is better suited to limiting concurrency, so do not use +this function. + Under normal circumstances you don't need to call this function. =back @@ -765,7 +843,7 @@ *$sym } -min_parallel 4; +min_parallel 8; END { max_parallel 0; @@ -789,6 +867,19 @@ not been called, while the child will act as if IO::AIO has not been used yet. +=head2 MEMORY USAGE + +Each aio request uses - depending on your architecture - around 128 bytes +of memory. In addition, stat requests need a stat buffer (possibly a few +hundred bytes). Perl scalars and other data passed into aio requests will +also be locked. + +This is now awfully much, so queuing lots of requests is not usually a +problem. + +Each thread needs a stack area which is usually around 16k, sometimes much +larger, depending on the OS. + =head1 SEE ALSO L, L (obsolete).