… | |
… | |
79 | our $VERSION = '2.0'; |
79 | our $VERSION = '2.0'; |
80 | |
80 | |
81 | our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat |
81 | our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat |
82 | aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink |
82 | aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink |
83 | aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move |
83 | aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move |
84 | aio_group); |
84 | aio_group aio_nop); |
85 | our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); |
85 | our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); |
86 | |
86 | |
87 | @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; |
87 | @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; |
88 | |
88 | |
89 | require XSLoader; |
89 | require XSLoader; |
… | |
… | |
182 | print "read $_[0] bytes: <$buffer>\n"; |
182 | print "read $_[0] bytes: <$buffer>\n"; |
183 | }; |
183 | }; |
184 | |
184 | |
185 | =item aio_move $srcpath, $dstpath, $callback->($status) |
185 | =item aio_move $srcpath, $dstpath, $callback->($status) |
186 | |
186 | |
|
|
187 | [EXPERIMENTAL due to internal aio_group use] |
|
|
188 | |
187 | Try to move the I<file> (directories not supported as either source or |
189 | Try to move the I<file> (directories not supported as either source or |
188 | destination) from C<$srcpath> to C<$dstpath> and call the callback with |
190 | destination) from C<$srcpath> to C<$dstpath> and call the callback with |
189 | the C<0> (error) or C<-1> ok. |
191 | the C<0> (error) or C<-1> ok. |
190 | |
192 | |
191 | This is a composite request that tries to rename(2) the file first. If |
193 | This is a composite request that tries to rename(2) the file first. If |
… | |
… | |
201 | =cut |
203 | =cut |
202 | |
204 | |
203 | sub aio_move($$$) { |
205 | sub aio_move($$$) { |
204 | my ($src, $dst, $cb) = @_; |
206 | my ($src, $dst, $cb) = @_; |
205 | |
207 | |
206 | my $grp = aio_group; |
208 | my $grp = aio_group $cb; |
207 | |
209 | |
208 | add $grp aio_rename $src, $dst, sub { |
210 | add $grp aio_rename $src, $dst, sub { |
209 | if ($_[0] && $! == EXDEV) { |
211 | if ($_[0] && $! == EXDEV) { |
210 | add $grp aio_open $src, O_RDONLY, 0, sub { |
212 | add $grp aio_open $src, O_RDONLY, 0, sub { |
211 | if (my $src_fh = $_[0]) { |
213 | if (my $src_fh = $_[0]) { |
… | |
… | |
221 | chmod $stat[2] & 07777, $dst_fh; |
223 | chmod $stat[2] & 07777, $dst_fh; |
222 | chown $stat[4], $stat[5], $dst_fh; |
224 | chown $stat[4], $stat[5], $dst_fh; |
223 | close $dst_fh; |
225 | close $dst_fh; |
224 | |
226 | |
225 | add $grp aio_unlink $src, sub { |
227 | add $grp aio_unlink $src, sub { |
226 | $cb->($_[0]); |
228 | $grp->result ($_[0]); |
227 | }; |
229 | }; |
228 | } else { |
230 | } else { |
229 | my $errno = $!; |
231 | my $errno = $!; |
230 | add $grp aio_unlink $dst, sub { |
232 | add $grp aio_unlink $dst, sub { |
231 | $! = $errno; |
233 | $! = $errno; |
232 | $cb->(-1); |
234 | $grp->result (-1); |
233 | }; |
235 | }; |
234 | } |
236 | } |
235 | }; |
237 | }; |
236 | } else { |
238 | } else { |
237 | $cb->(-1); |
239 | $grp->result (-1); |
238 | } |
240 | } |
239 | }, |
241 | }, |
240 | |
242 | |
241 | } else { |
243 | } else { |
242 | $cb->(-1); |
244 | $grp->result (-1); |
243 | } |
245 | } |
244 | }; |
246 | }; |
245 | } else { |
247 | } else { |
246 | $cb->($_[0]); |
248 | $grp->result ($_[0]); |
247 | } |
249 | } |
248 | }; |
250 | }; |
249 | |
251 | |
250 | $grp |
252 | $grp |
251 | } |
253 | } |
… | |
… | |
343 | The callback a single argument which is either C<undef> or an array-ref |
345 | The callback a single argument which is either C<undef> or an array-ref |
344 | with the filenames. |
346 | with the filenames. |
345 | |
347 | |
346 | =item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
348 | =item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
347 | |
349 | |
|
|
350 | [EXPERIMENTAL due to internal aio_group use] |
|
|
351 | |
348 | Scans a directory (similar to C<aio_readdir>) but additionally tries to |
352 | Scans a directory (similar to C<aio_readdir>) but additionally tries to |
349 | separate the entries of directory C<$path> into two sets of names, ones |
353 | separate the entries of directory C<$path> into two sets of names, ones |
350 | you can recurse into (directories or links to them), and ones you cannot |
354 | you can recurse into (directories or links to them), and ones you cannot |
351 | recurse into (everything else). |
355 | recurse into (everything else). |
352 | |
356 | |
353 | C<aio_scandir> is a composite request that consists of many sub |
357 | C<aio_scandir> is a composite request that creates of many sub requests_ |
354 | requests. C<$maxreq> specifies the maximum number of outstanding aio |
358 | C<$maxreq> specifies the maximum number of outstanding aio requests that |
355 | requests that this function generates. If it is C<< <= 0 >>, then a |
359 | this function generates. If it is C<< <= 0 >>, then a suitable default |
356 | suitable default will be chosen (currently 8). |
360 | will be chosen (currently 6). |
357 | |
361 | |
358 | On error, the callback is called without arguments, otherwise it receives |
362 | On error, the callback is called without arguments, otherwise it receives |
359 | two array-refs with path-relative entry names. |
363 | two array-refs with path-relative entry names. |
360 | |
364 | |
361 | Example: |
365 | Example: |
… | |
… | |
398 | =cut |
402 | =cut |
399 | |
403 | |
400 | sub aio_scandir($$$) { |
404 | sub aio_scandir($$$) { |
401 | my ($path, $maxreq, $cb) = @_; |
405 | my ($path, $maxreq, $cb) = @_; |
402 | |
406 | |
403 | my $grp = aio_group; |
407 | my $grp = aio_group $cb; |
404 | |
408 | |
405 | $maxreq = 8 if $maxreq <= 0; |
409 | $maxreq = 6 if $maxreq <= 0; |
406 | |
410 | |
407 | # stat once |
411 | # stat once |
408 | add $grp aio_stat $path, sub { |
412 | add $grp aio_stat $path, sub { |
409 | return $cb->() if $_[0]; |
413 | return $grp->result () if $_[0]; |
410 | my $now = time; |
414 | my $now = time; |
411 | my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
415 | my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
412 | |
416 | |
413 | # read the directory entries |
417 | # read the directory entries |
414 | add $grp aio_readdir $path, sub { |
418 | add $grp aio_readdir $path, sub { |
415 | my $entries = shift |
419 | my $entries = shift |
416 | or return $cb->(); |
420 | or return $grp->result (); |
417 | |
421 | |
418 | # stat the dir another time |
422 | # stat the dir another time |
419 | add $grp aio_stat $path, sub { |
423 | add $grp aio_stat $path, sub { |
420 | my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
424 | my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
421 | |
425 | |
… | |
… | |
426 | $ndirs = -1; |
430 | $ndirs = -1; |
427 | } else { |
431 | } else { |
428 | # if nlink == 2, we are finished |
432 | # if nlink == 2, we are finished |
429 | # on non-posix-fs's, we rely on nlink < 2 |
433 | # on non-posix-fs's, we rely on nlink < 2 |
430 | $ndirs = (stat _)[3] - 2 |
434 | $ndirs = (stat _)[3] - 2 |
431 | or return $cb->([], $entries); |
435 | or return $grp->result ([], $entries); |
432 | } |
436 | } |
433 | |
437 | |
434 | # sort into likely dirs and likely nondirs |
438 | # sort into likely dirs and likely nondirs |
435 | # dirs == files without ".", short entries first |
439 | # dirs == files without ".", short entries first |
436 | $entries = [map $_->[0], |
440 | $entries = [map $_->[0], |
… | |
… | |
441 | my (@dirs, @nondirs); |
445 | my (@dirs, @nondirs); |
442 | |
446 | |
443 | my ($statcb, $schedcb); |
447 | my ($statcb, $schedcb); |
444 | my $nreq = 0; |
448 | my $nreq = 0; |
445 | |
449 | |
|
|
450 | my $statgrp = add $grp aio_group; |
|
|
451 | |
446 | $schedcb = sub { |
452 | $schedcb = sub { |
447 | if (@$entries) { |
453 | if (@$entries) { |
448 | if ($nreq < $maxreq) { |
454 | if ($nreq < $maxreq) { |
449 | my $ent = pop @$entries; |
455 | my $ent = pop @$entries; |
450 | $nreq++; |
456 | $nreq++; |
451 | add $grp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; |
457 | add $statgrp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; |
452 | } |
458 | } |
453 | } elsif (!$nreq) { |
459 | } elsif (!$nreq) { |
454 | # finished |
460 | # finished |
|
|
461 | $statgrp->cancel; |
455 | undef $statcb; |
462 | undef $statcb; |
456 | undef $schedcb; |
463 | undef $schedcb; |
457 | $cb->(\@dirs, \@nondirs) if $cb; |
464 | $grp->result (\@dirs, \@nondirs); |
458 | undef $cb; |
|
|
459 | } |
465 | } |
460 | }; |
466 | }; |
461 | $statcb = sub { |
467 | $statcb = sub { |
462 | my ($status, $entry) = @_; |
468 | my ($status, $entry) = @_; |
463 | |
469 | |
… | |
… | |
505 | callback with the fdatasync result code. |
511 | callback with the fdatasync result code. |
506 | |
512 | |
507 | If this call isn't available because your OS lacks it or it couldn't be |
513 | If this call isn't available because your OS lacks it or it couldn't be |
508 | detected, it will be emulated by calling C<fsync> instead. |
514 | detected, it will be emulated by calling C<fsync> instead. |
509 | |
515 | |
510 | =item aio_group $callback->() |
516 | =item aio_group $callback->(...) |
511 | |
517 | |
512 | [EXPERIMENTAL] |
518 | [EXPERIMENTAL] |
513 | |
519 | |
514 | This is a very special aio request: Instead of doing something, it is a |
520 | This is a very special aio request: Instead of doing something, it is a |
515 | container for other aio requests, which is useful if you want to bundle |
521 | container for other aio requests, which is useful if you want to bundle |
… | |
… | |
526 | |
532 | |
527 | add $grp |
533 | add $grp |
528 | (aio_stat ...), |
534 | (aio_stat ...), |
529 | (aio_stat ...), |
535 | (aio_stat ...), |
530 | ...; |
536 | ...; |
|
|
537 | |
|
|
538 | =item aio_nop $callback->() |
|
|
539 | |
|
|
540 | This is a special request - it does nothing in itself and is only used for |
|
|
541 | side effects, such as when you want to add a dummy request to a group so |
|
|
542 | that finishing the requests in the group depends on executing the given |
|
|
543 | code. |
|
|
544 | |
|
|
545 | While this request does nothing, it still goes through the execution |
|
|
546 | phase and still requires a worker thread. Thus, the callback will not |
|
|
547 | be executed immediately but only after other requests in the queue have |
|
|
548 | entered their execution phase. This can be used to measure request |
|
|
549 | latency. |
531 | |
550 | |
532 | =item IO::AIO::aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED* |
551 | =item IO::AIO::aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED* |
533 | |
552 | |
534 | Mainly used for debugging and benchmarking, this aio request puts one of |
553 | Mainly used for debugging and benchmarking, this aio request puts one of |
535 | the request workers to sleep for the given time. |
554 | the request workers to sleep for the given time. |
… | |
… | |
585 | You add requests by calling the C<add> method with one or more |
604 | You add requests by calling the C<add> method with one or more |
586 | C<IO::AIO::REQ> objects: |
605 | C<IO::AIO::REQ> objects: |
587 | |
606 | |
588 | $grp->add (aio_unlink "..."); |
607 | $grp->add (aio_unlink "..."); |
589 | |
608 | |
590 | add $grp aio_stat "...", sub { ... }; |
609 | add $grp aio_stat "...", sub { |
|
|
610 | $_[0] or return $grp->result ("error"); |
|
|
611 | |
|
|
612 | # add another request dynamically, if first succeeded |
|
|
613 | add $grp aio_open "...", sub { |
|
|
614 | $grp->result ("ok"); |
|
|
615 | }; |
|
|
616 | }; |
591 | |
617 | |
592 | This makes it very easy to create composite requests (see the source of |
618 | This makes it very easy to create composite requests (see the source of |
593 | C<aio_move> for an application) that work and feel like simple requests. |
619 | C<aio_move> for an application) that work and feel like simple requests. |
594 | |
620 | |
|
|
621 | =over 4 |
|
|
622 | |
595 | The IO::AIO::GRP objects will be cleaned up during calls to |
623 | =item * The IO::AIO::GRP objects will be cleaned up during calls to |
596 | C<IO::AIO::poll_cb>, just like any other request. |
624 | C<IO::AIO::poll_cb>, just like any other request. |
597 | |
625 | |
598 | They can be canceled like any other request. Canceling will cancel not |
626 | =item * They can be canceled like any other request. Canceling will cancel not |
599 | just the request itself, but also all requests it contains. |
627 | only the request itself, but also all requests it contains. |
600 | |
628 | |
601 | They can also can also be added to other IO::AIO::GRP objects. |
629 | =item * They can also can also be added to other IO::AIO::GRP objects. |
|
|
630 | |
|
|
631 | =item * You must not add requests to a group from within the group callback (or |
|
|
632 | any later time). |
|
|
633 | |
|
|
634 | =item * This does not harmonise well with C<max_outstanding>, so best do |
|
|
635 | not combine C<aio_group> with it. Groups and feeders are recommended for |
|
|
636 | this kind of concurrency-limiting. |
|
|
637 | |
|
|
638 | =back |
602 | |
639 | |
603 | Their lifetime, simplified, looks like this: when they are empty, they |
640 | Their lifetime, simplified, looks like this: when they are empty, they |
604 | will finish very quickly. If they contain only requests that are in the |
641 | will finish very quickly. If they contain only requests that are in the |
605 | C<done> state, they will also finish. Otherwise they will continue to |
642 | C<done> state, they will also finish. Otherwise they will continue to |
606 | exist. |
643 | exist. |
607 | |
644 | |
|
|
645 | That means after creating a group you have some time to add requests. And |
|
|
646 | in the callbacks of those requests, you can add further requests to the |
|
|
647 | group. And only when all those requests have finished will the the group |
|
|
648 | itself finish. |
|
|
649 | |
608 | =over 4 |
650 | =over 4 |
609 | |
651 | |
610 | =item $grp->add (...) |
652 | =item $grp->add (...) |
611 | |
653 | |
612 | =item add $grp ... |
654 | =item add $grp ... |
613 | |
655 | |
614 | Add one or more |
656 | Add one or more requests to the group. Any type of L<IO::AIO::REQ> can |
615 | Cancels the request, if possible. Has the effect of skipping execution |
657 | be added, including other groups, as long as you do not create circular |
616 | when entering the B<execute> state and skipping calling the callback when |
658 | dependencies. |
617 | entering the the B<result> state, but will leave the request otherwise |
659 | |
618 | untouched. That means that requests that currently execute will not be |
660 | Returns all its arguments. |
619 | stopped and resources held by the request will not be freed prematurely. |
661 | |
|
|
662 | =item $grp->result (...) |
|
|
663 | |
|
|
664 | Set the result value(s) that will be passed to the group callback when all |
|
|
665 | subrequests have finished. By default, no argument will be passed. |
|
|
666 | |
|
|
667 | =item $grp->set_feeder ($callback->($grp)) |
|
|
668 | |
|
|
669 | [VERY EXPERIMENTAL] |
|
|
670 | |
|
|
671 | Sets a feeder/generator on this group: every group can have an attached |
|
|
672 | generator that generates requests if idle. The idea behind this is that, |
|
|
673 | although you could just queue as many requests as you want in a group, |
|
|
674 | this might starve other requests for a potentially long time. For |
|
|
675 | example, C<aio_scandir> might generate hundreds of thousands C<aio_stat> |
|
|
676 | requests, delaying any later requests for a long time. |
|
|
677 | |
|
|
678 | To avoid this, and allow incremental generation of requests, you can |
|
|
679 | instead a group and set a feeder on it that generates those requests. The |
|
|
680 | feeder will be called whenever there are few enough (see C<feeder_limit>, |
|
|
681 | below) requests active in the group itself and is expected to queue more |
|
|
682 | requests. |
|
|
683 | |
|
|
684 | The feeder can queue as many requests as it likes (i.e. C<add> does not |
|
|
685 | impose any limits). |
|
|
686 | |
|
|
687 | If the feeder does not queue more requests when called, it will be |
|
|
688 | automatically removed from the group. |
|
|
689 | |
|
|
690 | If the feeder limit is C<0>, it will be set to C<2> automatically. |
|
|
691 | |
|
|
692 | Example: |
|
|
693 | |
|
|
694 | # stat all files in @files, but only ever use four aio requests concurrently: |
|
|
695 | |
|
|
696 | my $grp = aio_group sub { print "finished\n" }; |
|
|
697 | $grp->feeder_limit (4); |
|
|
698 | $grp->set_feeder (sub { |
|
|
699 | my $file = pop @files |
|
|
700 | or return; |
|
|
701 | |
|
|
702 | add $grp aio_stat $file, sub { ... }; |
|
|
703 | }); |
|
|
704 | |
|
|
705 | =item $grp->feeder_limit ($num) |
|
|
706 | |
|
|
707 | Sets the feeder limit for the group: The feeder will be called whenever |
|
|
708 | the group contains less than this many requests. |
|
|
709 | |
|
|
710 | Setting the limit to C<0> will pause the feeding process. |
620 | |
711 | |
621 | =back |
712 | =back |
622 | |
713 | |
623 | =head2 SUPPORT FUNCTIONS |
714 | =head2 SUPPORT FUNCTIONS |
624 | |
715 | |
… | |
… | |
682 | IO::AIO::poll_wait, IO::AIO::poll_cb |
773 | IO::AIO::poll_wait, IO::AIO::poll_cb |
683 | if IO::AIO::nreqs; |
774 | if IO::AIO::nreqs; |
684 | |
775 | |
685 | =item IO::AIO::min_parallel $nthreads |
776 | =item IO::AIO::min_parallel $nthreads |
686 | |
777 | |
687 | Set the minimum number of AIO threads to C<$nthreads>. The current default |
778 | Set the minimum number of AIO threads to C<$nthreads>. The current |
688 | is C<4>, which means four asynchronous operations can be done at one time |
779 | default is C<8>, which means eight asynchronous operations can execute |
689 | (the number of outstanding operations, however, is unlimited). |
780 | concurrently at any one time (the number of outstanding requests, |
|
|
781 | however, is unlimited). |
690 | |
782 | |
691 | IO::AIO starts threads only on demand, when an AIO request is queued and |
783 | IO::AIO starts threads only on demand, when an AIO request is queued and |
692 | no free thread exists. |
784 | no free thread exists. |
693 | |
785 | |
694 | It is recommended to keep the number of threads low, as some Linux |
786 | It is recommended to keep the number of threads relatively low, as some |
695 | kernel versions will scale negatively with the number of threads (higher |
787 | Linux kernel versions will scale negatively with the number of threads |
696 | parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 |
788 | (higher parallelity => MUCH higher latency). With current Linux 2.6 |
697 | threads should be fine. |
789 | versions, 4-32 threads should be fine. |
698 | |
790 | |
699 | Under most circumstances you don't need to call this function, as the |
791 | Under most circumstances you don't need to call this function, as the |
700 | module selects a default that is suitable for low to moderate load. |
792 | module selects a default that is suitable for low to moderate load. |
701 | |
793 | |
702 | =item IO::AIO::max_parallel $nthreads |
794 | =item IO::AIO::max_parallel $nthreads |
… | |
… | |
713 | |
805 | |
714 | Under normal circumstances you don't need to call this function. |
806 | Under normal circumstances you don't need to call this function. |
715 | |
807 | |
716 | =item $oldnreqs = IO::AIO::max_outstanding $nreqs |
808 | =item $oldnreqs = IO::AIO::max_outstanding $nreqs |
717 | |
809 | |
|
|
810 | [DEPRECATED] |
|
|
811 | |
718 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
812 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
719 | try to queue up more than this number of requests, the caller will block until |
813 | try to queue up more than this number of requests, the caller will block until |
720 | some requests have been handled. |
814 | some requests have been handled. |
721 | |
815 | |
722 | The default is very large, so normally there is no practical limit. If you |
816 | The default is very large, so normally there is no practical limit. If you |
723 | queue up many requests in a loop it often improves speed if you set |
817 | queue up many requests in a loop it often improves speed if you set |
724 | this to a relatively low number, such as C<100>. |
818 | this to a relatively low number, such as C<100>. |
|
|
819 | |
|
|
820 | This function does not work well together with C<aio_group>'s, and their |
|
|
821 | feeder interface is better suited to limiting concurrency, so do not use |
|
|
822 | this function. |
725 | |
823 | |
726 | Under normal circumstances you don't need to call this function. |
824 | Under normal circumstances you don't need to call this function. |
727 | |
825 | |
728 | =back |
826 | =back |
729 | |
827 | |
… | |
… | |
743 | or return undef; |
841 | or return undef; |
744 | |
842 | |
745 | *$sym |
843 | *$sym |
746 | } |
844 | } |
747 | |
845 | |
748 | min_parallel 4; |
846 | min_parallel 8; |
749 | |
847 | |
750 | END { |
848 | END { |
751 | max_parallel 0; |
849 | max_parallel 0; |
752 | } |
850 | } |
753 | |
851 | |
… | |
… | |
767 | |
865 | |
768 | In short: the parent will, after a short pause, continue as if fork had |
866 | In short: the parent will, after a short pause, continue as if fork had |
769 | not been called, while the child will act as if IO::AIO has not been used |
867 | not been called, while the child will act as if IO::AIO has not been used |
770 | yet. |
868 | yet. |
771 | |
869 | |
|
|
870 | =head2 MEMORY USAGE |
|
|
871 | |
|
|
872 | Each aio request uses - depending on your architecture - around 128 bytes |
|
|
873 | of memory. In addition, stat requests need a stat buffer (possibly a few |
|
|
874 | hundred bytes). Perl scalars and other data passed into aio requests will |
|
|
875 | also be locked. |
|
|
876 | |
|
|
877 | This is now awfully much, so queuing lots of requests is not usually a |
|
|
878 | problem. |
|
|
879 | |
|
|
880 | Each thread needs a stack area which is usually around 16k, sometimes much |
|
|
881 | larger, depending on the OS. |
|
|
882 | |
772 | =head1 SEE ALSO |
883 | =head1 SEE ALSO |
773 | |
884 | |
774 | L<Coro>, L<Linux::AIO> (obsolete). |
885 | L<Coro>, L<Linux::AIO> (obsolete). |
775 | |
886 | |
776 | =head1 AUTHOR |
887 | =head1 AUTHOR |