ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/IO-AIO/AIO.pm
(Generate patch)

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.55 by root, Sun Oct 22 00:49:29 2006 UTC vs.
Revision 1.88 by root, Sun Oct 29 01:50:29 2006 UTC

15 15
16 aio_read $fh, 30000, 1024, $buffer, 0, sub { 16 aio_read $fh, 30000, 1024, $buffer, 0, sub {
17 $_[0] > 0 or die "read error: $!"; 17 $_[0] > 0 or die "read error: $!";
18 }; 18 };
19 19
20 use IO::AIO 2; # version has aio objects 20 # version 2+ has request and group objects
21 use IO::AIO 2;
21 22
23 aioreq_pri 4; # give next request a very high priority
22 my $req = aio_unlink "/tmp/file", sub { }; 24 my $req = aio_unlink "/tmp/file", sub { };
23 $req->cancel; # cancel request if still in queue 25 $req->cancel; # cancel request if still in queue
24 26
25 # AnyEvent 27 my $grp = aio_group sub { print "all stats done\n" };
28 add $grp aio_stat "..." for ...;
29
30 # AnyEvent integration
26 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; 31 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
27 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); 32 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
28 33
29 # Event 34 # Event integration
30 Event->io (fd => IO::AIO::poll_fileno, 35 Event->io (fd => IO::AIO::poll_fileno,
31 poll => 'r', 36 poll => 'r',
32 cb => \&IO::AIO::poll_cb); 37 cb => \&IO::AIO::poll_cb);
33 38
34 # Glib/Gtk2 39 # Glib/Gtk2 integration
35 add_watch Glib::IO IO::AIO::poll_fileno, 40 add_watch Glib::IO IO::AIO::poll_fileno,
36 in => sub { IO::AIO::poll_cb; 1 }; 41 in => sub { IO::AIO::poll_cb; 1 };
37 42
38 # Tk 43 # Tk integration
39 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", 44 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
40 readable => \&IO::AIO::poll_cb); 45 readable => \&IO::AIO::poll_cb);
41 46
42 # Danga::Socket 47 # Danga::Socket integration
43 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => 48 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
44 \&IO::AIO::poll_cb); 49 \&IO::AIO::poll_cb);
45 50
46
47=head1 DESCRIPTION 51=head1 DESCRIPTION
48 52
49This module implements asynchronous I/O using whatever means your 53This module implements asynchronous I/O using whatever means your
50operating system supports. 54operating system supports.
51 55
56Asynchronous means that operations that can normally block your program
57(e.g. reading from disk) will be done asynchronously: the operation
58will still block, but you can do something else in the meantime. This
59is extremely useful for programs that need to stay interactive even
60when doing heavy I/O (GUI programs, high performance network servers
61etc.), but can also be used to easily do operations in parallel that are
62normally done sequentially, e.g. stat'ing many files, which is much faster
63on a RAID volume or over NFS when you do a number of stat operations
64concurrently.
65
66While this works on all types of file descriptors (for example sockets),
67using these functions on file descriptors that support nonblocking
68operation (again, sockets, pipes etc.) is very inefficient. Use an event
69loop for that (such as the L<Event|Event> module): IO::AIO will naturally
70fit into such an event loop itself.
71
52Currently, a number of threads are started that execute your read/writes 72In this version, a number of threads are started that execute your
53and signal their completion. You don't need thread support in your libc or 73requests and signal their completion. You don't need thread support
54perl, and the threads created by this module will not be visible to the 74in perl, and the threads created by this module will not be visible
55pthreads library. In the future, this module might make use of the native 75to perl. In the future, this module might make use of the native aio
56aio functions available on many operating systems. However, they are often 76functions available on many operating systems. However, they are often
57not well-supported (Linux doesn't allow them on normal files currently, 77not well-supported or restricted (GNU/Linux doesn't allow them on normal
58for example), and they would only support aio_read and aio_write, so the 78files currently, for example), and they would only support aio_read and
59remaining functionality would have to be implemented using threads anyway. 79aio_write, so the remaining functionality would have to be implemented
80using threads anyway.
60 81
61Although the module will work with in the presence of other threads, it is 82Although the module will work with in the presence of other (Perl-)
62currently not reentrant, so use appropriate locking yourself, always call 83threads, it is currently not reentrant in any way, so use appropriate
63C<poll_cb> from within the same thread, or never call C<poll_cb> (or other 84locking yourself, always call C<poll_cb> from within the same thread, or
64C<aio_> functions) recursively. 85never call C<poll_cb> (or other C<aio_> functions) recursively.
86
87=head2 EXAMPLE
88
89This is a simple example that uses the Event module and loads
90F</etc/passwd> asynchronously:
91
92 use Fcntl;
93 use Event;
94 use IO::AIO;
95
96 # register the IO::AIO callback with Event
97 Event->io (fd => IO::AIO::poll_fileno,
98 poll => 'r',
99 cb => \&IO::AIO::poll_cb);
100
101 # queue the request to open /etc/passwd
102 aio_open "/etc/passwd", O_RDONLY, 0, sub {
103 my $fh = $_[0]
104 or die "error while opening: $!";
105
106 # stat'ing filehandles is generally non-blocking
107 my $size = -s $fh;
108
109 # queue a request to read the file
110 my $contents;
111 aio_read $fh, 0, $size, $contents, 0, sub {
112 $_[0] == $size
113 or die "short read: $!";
114
115 close $fh;
116
117 # file contents now in $contents
118 print $contents;
119
120 # exit event loop and program
121 Event::unloop;
122 };
123 };
124
125 # possibly queue up other requests, or open GUI windows,
126 # check for sockets etc. etc.
127
128 # process events as long as there are some:
129 Event::loop;
130
131=head1 REQUEST ANATOMY AND LIFETIME
132
133Every C<aio_*> function creates a request. which is a C data structure not
134directly visible to Perl.
135
136If called in non-void context, every request function returns a Perl
137object representing the request. In void context, nothing is returned,
138which saves a bit of memory.
139
140The perl object is a fairly standard ref-to-hash object. The hash contents
141are not used by IO::AIO so you are free to store anything you like in it.
142
143During their existance, aio requests travel through the following states,
144in order:
145
146=over 4
147
148=item ready
149
150Immediately after a request is created it is put into the ready state,
151waiting for a thread to execute it.
152
153=item execute
154
155A thread has accepted the request for processing and is currently
156executing it (e.g. blocking in read).
157
158=item pending
159
160The request has been executed and is waiting for result processing.
161
162While request submission and execution is fully asynchronous, result
163processing is not and relies on the perl interpreter calling C<poll_cb>
164(or another function with the same effect).
165
166=item result
167
168The request results are processed synchronously by C<poll_cb>.
169
170The C<poll_cb> function will process all outstanding aio requests by
171calling their callbacks, freeing memory associated with them and managing
172any groups they are contained in.
173
174=item done
175
176Request has reached the end of its lifetime and holds no resources anymore
177(except possibly for the Perl object, but its connection to the actual
178aio request is severed and calling its methods will either do nothing or
179result in a runtime error).
180
181=back
65 182
66=cut 183=cut
67 184
68package IO::AIO; 185package IO::AIO;
69 186
71use strict 'vars'; 188use strict 'vars';
72 189
73use base 'Exporter'; 190use base 'Exporter';
74 191
75BEGIN { 192BEGIN {
76 our $VERSION = '2.0'; 193 our $VERSION = '2.1';
77 194
78 our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat 195 our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat
79 aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink 196 aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink
80 aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move 197 aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move
81 aio_group); 198 aio_copy aio_group aio_nop aio_mknod);
82 our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); 199 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
200 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
201 min_parallel max_parallel max_idle
202 nreqs nready npending nthreads
203 max_poll_time max_poll_reqs);
83 204
84 @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; 205 @IO::AIO::GRP::ISA = 'IO::AIO::REQ';
85 206
86 require XSLoader; 207 require XSLoader;
87 XSLoader::load ("IO::AIO", $VERSION); 208 XSLoader::load ("IO::AIO", $VERSION);
88} 209}
89 210
90=head1 FUNCTIONS 211=head1 FUNCTIONS
91 212
92=head2 AIO FUNCTIONS 213=head2 AIO REQUEST FUNCTIONS
93 214
94All the C<aio_*> calls are more or less thin wrappers around the syscall 215All the C<aio_*> calls are more or less thin wrappers around the syscall
95with the same name (sans C<aio_>). The arguments are similar or identical, 216with the same name (sans C<aio_>). The arguments are similar or identical,
96and they all accept an additional (and optional) C<$callback> argument 217and they all accept an additional (and optional) C<$callback> argument
97which must be a code reference. This code reference will get called with 218which must be a code reference. This code reference will get called with
100syscall has been executed asynchronously. 221syscall has been executed asynchronously.
101 222
102All functions expecting a filehandle keep a copy of the filehandle 223All functions expecting a filehandle keep a copy of the filehandle
103internally until the request has finished. 224internally until the request has finished.
104 225
105All requests return objects of type L<IO::AIO::REQ> that allow further 226All functions return request objects of type L<IO::AIO::REQ> that allow
106manipulation of those requests while they are in-flight. 227further manipulation of those requests while they are in-flight.
107 228
108The pathnames you pass to these routines I<must> be absolute and 229The pathnames you pass to these routines I<must> be absolute and
109encoded in byte form. The reason for the former is that at the time the 230encoded as octets. The reason for the former is that at the time the
110request is being executed, the current working directory could have 231request is being executed, the current working directory could have
111changed. Alternatively, you can make sure that you never change the 232changed. Alternatively, you can make sure that you never change the
112current working directory. 233current working directory anywhere in the program and then use relative
234paths.
113 235
114To encode pathnames to byte form, either make sure you either: a) 236To encode pathnames as octets, either make sure you either: a) always pass
115always pass in filenames you got from outside (command line, readdir 237in filenames you got from outside (command line, readdir etc.) without
116etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode 238tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
117your pathnames to the locale (or other) encoding in effect in the user 239your pathnames to the locale (or other) encoding in effect in the user
118environment, d) use Glib::filename_from_unicode on unicode filenames or e) 240environment, d) use Glib::filename_from_unicode on unicode filenames or e)
119use something else. 241use something else to ensure your scalar has the correct contents.
242
243This works, btw. independent of the internal UTF-8 bit, which IO::AIO
244handles correctly wether it is set or not.
120 245
121=over 4 246=over 4
247
248=item $prev_pri = aioreq_pri [$pri]
249
250Returns the priority value that would be used for the next request and, if
251C<$pri> is given, sets the priority for the next aio request.
252
253The default priority is C<0>, the minimum and maximum priorities are C<-4>
254and C<4>, respectively. Requests with higher priority will be serviced
255first.
256
257The priority will be reset to C<0> after each call to one of the C<aio_*>
258functions.
259
260Example: open a file with low priority, then read something from it with
261higher priority so the read request is serviced before other low priority
262open requests (potentially spamming the cache):
263
264 aioreq_pri -3;
265 aio_open ..., sub {
266 return unless $_[0];
267
268 aioreq_pri -2;
269 aio_read $_[0], ..., sub {
270 ...
271 };
272 };
273
274=item aioreq_nice $pri_adjust
275
276Similar to C<aioreq_pri>, but subtracts the given value from the current
277priority, so the effect is cumulative.
122 278
123=item aio_open $pathname, $flags, $mode, $callback->($fh) 279=item aio_open $pathname, $flags, $mode, $callback->($fh)
124 280
125Asynchronously open or create a file and call the callback with a newly 281Asynchronously open or create a file and call the callback with a newly
126created filehandle for the file. 282created filehandle for the file.
176 332
177 aio_read $fh, 7, 15, $buffer, 0, sub { 333 aio_read $fh, 7, 15, $buffer, 0, sub {
178 $_[0] > 0 or die "read error: $!"; 334 $_[0] > 0 or die "read error: $!";
179 print "read $_[0] bytes: <$buffer>\n"; 335 print "read $_[0] bytes: <$buffer>\n";
180 }; 336 };
181
182=item aio_move $srcpath, $dstpath, $callback->($status)
183
184Try to move the I<file> (directories not supported as either source or
185destination) from C<$srcpath> to C<$dstpath> and call the callback with
186the C<0> (error) or C<-1> ok.
187
188This is a composite request that tries to rename(2) the file first. If
189rename files with C<EXDEV>, it creates the destination file with mode 0200
190and copies the contents of the source file into it using C<aio_sendfile>,
191followed by restoring atime, mtime, access mode and uid/gid, in that
192order, and unlinking the C<$srcpath>.
193
194If an error occurs, the partial destination file will be unlinked, if
195possible, except when setting atime, mtime, access mode and uid/gid, where
196errors are being ignored.
197
198=cut
199
200sub aio_move($$$) {
201 my ($src, $dst, $cb) = @_;
202
203 my $grp = aio_group;
204
205 add $grp aio_rename $src, $dst, sub {
206 if ($_[0] && $! == EXDEV) {
207 add $grp aio_open $src, O_RDONLY, 0, sub {
208 if (my $src_fh = $_[0]) {
209 my @stat = stat $src_fh;
210
211 add $grp aio_open $dst, O_WRONLY, 0200, sub {
212 if (my $dst_fh = $_[0]) {
213 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
214 close $src_fh;
215
216 if ($_[0] == $stat[7]) {
217 utime $stat[8], $stat[9], $dst;
218 chmod $stat[2] & 07777, $dst_fh;
219 chown $stat[4], $stat[5], $dst_fh;
220 close $dst_fh;
221
222 add $grp aio_unlink $src, sub {
223 $cb->($_[0]);
224 };
225 } else {
226 my $errno = $!;
227 add $grp aio_unlink $dst, sub {
228 $! = $errno;
229 $cb->(-1);
230 };
231 }
232 };
233 } else {
234 $cb->(-1);
235 }
236 },
237
238 } else {
239 $cb->(-1);
240 }
241 };
242 } else {
243 $cb->($_[0]);
244 }
245 };
246
247 $grp
248}
249 337
250=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) 338=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
251 339
252Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts 340Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
253reading at byte offset C<$in_offset>, and starts writing at the current 341reading at byte offset C<$in_offset>, and starts writing at the current
309=item aio_unlink $pathname, $callback->($status) 397=item aio_unlink $pathname, $callback->($status)
310 398
311Asynchronously unlink (delete) a file and call the callback with the 399Asynchronously unlink (delete) a file and call the callback with the
312result code. 400result code.
313 401
402=item aio_mknod $path, $mode, $dev, $callback->($status)
403
404[EXPERIMENTAL]
405
406Asynchronously create a device node (or fifo). See mknod(2).
407
408The only (POSIX-) portable way of calling this function is:
409
410 aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ...
411
314=item aio_link $srcpath, $dstpath, $callback->($status) 412=item aio_link $srcpath, $dstpath, $callback->($status)
315 413
316Asynchronously create a new link to the existing object at C<$srcpath> at 414Asynchronously create a new link to the existing object at C<$srcpath> at
317the path C<$dstpath> and call the callback with the result code. 415the path C<$dstpath> and call the callback with the result code.
318 416
338sorted, and will B<NOT> include the C<.> and C<..> entries. 436sorted, and will B<NOT> include the C<.> and C<..> entries.
339 437
340The callback a single argument which is either C<undef> or an array-ref 438The callback a single argument which is either C<undef> or an array-ref
341with the filenames. 439with the filenames.
342 440
441=item aio_copy $srcpath, $dstpath, $callback->($status)
442
443Try to copy the I<file> (directories not supported as either source or
444destination) from C<$srcpath> to C<$dstpath> and call the callback with
445the C<0> (error) or C<-1> ok.
446
447This is a composite request that it creates the destination file with
448mode 0200 and copies the contents of the source file into it using
449C<aio_sendfile>, followed by restoring atime, mtime, access mode and
450uid/gid, in that order.
451
452If an error occurs, the partial destination file will be unlinked, if
453possible, except when setting atime, mtime, access mode and uid/gid, where
454errors are being ignored.
455
456=cut
457
458sub aio_copy($$;$) {
459 my ($src, $dst, $cb) = @_;
460
461 my $pri = aioreq_pri;
462 my $grp = aio_group $cb;
463
464 aioreq_pri $pri;
465 add $grp aio_open $src, O_RDONLY, 0, sub {
466 if (my $src_fh = $_[0]) {
467 my @stat = stat $src_fh;
468
469 aioreq_pri $pri;
470 add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
471 if (my $dst_fh = $_[0]) {
472 aioreq_pri $pri;
473 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
474 if ($_[0] == $stat[7]) {
475 $grp->result (0);
476 close $src_fh;
477
478 # those should not normally block. should. should.
479 utime $stat[8], $stat[9], $dst;
480 chmod $stat[2] & 07777, $dst_fh;
481 chown $stat[4], $stat[5], $dst_fh;
482 close $dst_fh;
483 } else {
484 $grp->result (-1);
485 close $src_fh;
486 close $dst_fh;
487
488 aioreq $pri;
489 add $grp aio_unlink $dst;
490 }
491 };
492 } else {
493 $grp->result (-1);
494 }
495 },
496
497 } else {
498 $grp->result (-1);
499 }
500 };
501
502 $grp
503}
504
505=item aio_move $srcpath, $dstpath, $callback->($status)
506
507Try to move the I<file> (directories not supported as either source or
508destination) from C<$srcpath> to C<$dstpath> and call the callback with
509the C<0> (error) or C<-1> ok.
510
511This is a composite request that tries to rename(2) the file first. If
512rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if
513that is successful, unlinking the C<$srcpath>.
514
515=cut
516
517sub aio_move($$;$) {
518 my ($src, $dst, $cb) = @_;
519
520 my $pri = aioreq_pri;
521 my $grp = aio_group $cb;
522
523 aioreq_pri $pri;
524 add $grp aio_rename $src, $dst, sub {
525 if ($_[0] && $! == EXDEV) {
526 aioreq_pri $pri;
527 add $grp aio_copy $src, $dst, sub {
528 $grp->result ($_[0]);
529
530 if (!$_[0]) {
531 aioreq_pri $pri;
532 add $grp aio_unlink $src;
533 }
534 };
535 } else {
536 $grp->result ($_[0]);
537 }
538 };
539
540 $grp
541}
542
343=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 543=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
344 544
345Scans a directory (similar to C<aio_readdir>) but additionally tries to 545Scans a directory (similar to C<aio_readdir>) but additionally tries to
346separate the entries of directory C<$path> into two sets of names, ones 546efficiently separate the entries of directory C<$path> into two sets of
347you can recurse into (directories or links to them), and ones you cannot 547names, directories you can recurse into (directories), and ones you cannot
348recurse into (everything else). 548recurse into (everything else, including symlinks to directories).
349 549
350C<aio_scandir> is a composite request that consists of many sub 550C<aio_scandir> is a composite request that creates of many sub requests_
351requests. C<$maxreq> specifies the maximum number of outstanding aio 551C<$maxreq> specifies the maximum number of outstanding aio requests that
352requests that this function generates. If it is C<< <= 0 >>, then a 552this function generates. If it is C<< <= 0 >>, then a suitable default
353suitable default will be chosen (currently 8). 553will be chosen (currently 4).
354 554
355On error, the callback is called without arguments, otherwise it receives 555On error, the callback is called without arguments, otherwise it receives
356two array-refs with path-relative entry names. 556two array-refs with path-relative entry names.
357 557
358Example: 558Example:
395=cut 595=cut
396 596
397sub aio_scandir($$$) { 597sub aio_scandir($$$) {
398 my ($path, $maxreq, $cb) = @_; 598 my ($path, $maxreq, $cb) = @_;
399 599
600 my $pri = aioreq_pri;
601
400 my $grp = aio_group; 602 my $grp = aio_group $cb;
401 603
402 $maxreq = 8 if $maxreq <= 0; 604 $maxreq = 4 if $maxreq <= 0;
403 605
404 # stat once 606 # stat once
607 aioreq_pri $pri;
405 add $grp aio_stat $path, sub { 608 add $grp aio_stat $path, sub {
406 return $cb->() if $_[0]; 609 return $grp->result () if $_[0];
407 my $now = time; 610 my $now = time;
408 my $hash1 = join ":", (stat _)[0,1,3,7,9]; 611 my $hash1 = join ":", (stat _)[0,1,3,7,9];
409 612
410 # read the directory entries 613 # read the directory entries
614 aioreq_pri $pri;
411 add $grp aio_readdir $path, sub { 615 add $grp aio_readdir $path, sub {
412 my $entries = shift 616 my $entries = shift
413 or return $cb->(); 617 or return $grp->result ();
414 618
415 # stat the dir another time 619 # stat the dir another time
620 aioreq_pri $pri;
416 add $grp aio_stat $path, sub { 621 add $grp aio_stat $path, sub {
417 my $hash2 = join ":", (stat _)[0,1,3,7,9]; 622 my $hash2 = join ":", (stat _)[0,1,3,7,9];
418 623
419 my $ndirs; 624 my $ndirs;
420 625
423 $ndirs = -1; 628 $ndirs = -1;
424 } else { 629 } else {
425 # if nlink == 2, we are finished 630 # if nlink == 2, we are finished
426 # on non-posix-fs's, we rely on nlink < 2 631 # on non-posix-fs's, we rely on nlink < 2
427 $ndirs = (stat _)[3] - 2 632 $ndirs = (stat _)[3] - 2
428 or return $cb->([], $entries); 633 or return $grp->result ([], $entries);
429 } 634 }
430 635
431 # sort into likely dirs and likely nondirs 636 # sort into likely dirs and likely nondirs
432 # dirs == files without ".", short entries first 637 # dirs == files without ".", short entries first
433 $entries = [map $_->[0], 638 $entries = [map $_->[0],
435 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length], 640 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
436 @$entries]; 641 @$entries];
437 642
438 my (@dirs, @nondirs); 643 my (@dirs, @nondirs);
439 644
440 my ($statcb, $schedcb); 645 my $statgrp = add $grp aio_group sub {
441 my $nreq = 0; 646 $grp->result (\@dirs, \@nondirs);
647 };
442 648
443 $schedcb = sub { 649 limit $statgrp $maxreq;
444 if (@$entries) { 650 feed $statgrp sub {
445 if ($nreq < $maxreq) { 651 return unless @$entries;
446 my $ent = pop @$entries; 652 my $entry = pop @$entries;
653
654 aioreq_pri $pri;
655 add $statgrp aio_stat "$path/$entry/.", sub {
656 if ($_[0] < 0) {
657 push @nondirs, $entry;
658 } else {
659 # need to check for real directory
660 aioreq_pri $pri;
661 add $statgrp aio_lstat "$path/$entry", sub {
662 if (-d _) {
663 push @dirs, $entry;
664
665 unless (--$ndirs) {
666 push @nondirs, @$entries;
667 feed $statgrp;
668 }
669 } else {
670 push @nondirs, $entry;
671 }
447 $nreq++; 672 }
448 add $grp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) };
449 } 673 }
450 } elsif (!$nreq) {
451 # finished
452 undef $statcb;
453 undef $schedcb;
454 $cb->(\@dirs, \@nondirs) if $cb;
455 undef $cb;
456 } 674 };
457 }; 675 };
458 $statcb = sub {
459 my ($status, $entry) = @_;
460
461 if ($status < 0) {
462 $nreq--;
463 push @nondirs, $entry;
464 &$schedcb;
465 } else {
466 # need to check for real directory
467 add $grp aio_lstat "$path/$entry", sub {
468 $nreq--;
469
470 if (-d _) {
471 push @dirs, $entry;
472
473 if (!--$ndirs) {
474 push @nondirs, @$entries;
475 $entries = [];
476 }
477 } else {
478 push @nondirs, $entry;
479 }
480
481 &$schedcb;
482 }
483 }
484 };
485
486 &$schedcb while @$entries && $nreq < $maxreq;
487 }; 676 };
488 }; 677 };
489 }; 678 };
490 679
491 $grp 680 $grp
502callback with the fdatasync result code. 691callback with the fdatasync result code.
503 692
504If this call isn't available because your OS lacks it or it couldn't be 693If this call isn't available because your OS lacks it or it couldn't be
505detected, it will be emulated by calling C<fsync> instead. 694detected, it will be emulated by calling C<fsync> instead.
506 695
507=item aio_group $callback->() 696=item aio_group $callback->(...)
508
509[EXPERIMENTAL]
510 697
511This is a very special aio request: Instead of doing something, it is a 698This is a very special aio request: Instead of doing something, it is a
512container for other aio requests, which is useful if you want to bundle 699container for other aio requests, which is useful if you want to bundle
513many requests into a single, composite, request. 700many requests into a single, composite, request with a definite callback
701and the ability to cancel the whole request with its subrequests.
514 702
515Returns an object of class L<IO::AIO::GRP>. See its documentation below 703Returns an object of class L<IO::AIO::GRP>. See its documentation below
516for more info. 704for more info.
517 705
518Example: 706Example:
524 add $grp 712 add $grp
525 (aio_stat ...), 713 (aio_stat ...),
526 (aio_stat ...), 714 (aio_stat ...),
527 ...; 715 ...;
528 716
717=item aio_nop $callback->()
718
719This is a special request - it does nothing in itself and is only used for
720side effects, such as when you want to add a dummy request to a group so
721that finishing the requests in the group depends on executing the given
722code.
723
724While this request does nothing, it still goes through the execution
725phase and still requires a worker thread. Thus, the callback will not
726be executed immediately but only after other requests in the queue have
727entered their execution phase. This can be used to measure request
728latency.
729
529=item aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED* 730=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
530 731
531Mainly used for debugging and benchmarking, this aio request puts one of 732Mainly used for debugging and benchmarking, this aio request puts one of
532the request workers to sleep for the given time. 733the request workers to sleep for the given time.
533 734
735While it is theoretically handy to have simple I/O scheduling requests
736like sleep and file handle readable/writable, the overhead this creates is
737immense (it blocks a thread for a long time) so do not use this function
738except to put your application under artificial I/O pressure.
739
534=back 740=back
535 741
536=head2 IO::AIO::REQ CLASS 742=head2 IO::AIO::REQ CLASS
537 743
538All non-aggregate C<aio_*> functions return an object of this class when 744All non-aggregate C<aio_*> functions return an object of this class when
539called in non-void context. 745called in non-void context.
540 746
541A request always moves through the following five states in its lifetime,
542in order: B<ready> (request has been created, but has not been executed
543yet), B<execute> (request is currently being executed), B<pending>
544(request has been executed but callback has not been called yet),
545B<result> (results are being processed synchronously, includes calling the
546callback) and B<done> (request has reached the end of its lifetime and
547holds no resources anymore).
548
549=over 4 747=over 4
550 748
551=item $req->cancel 749=item cancel $req
552 750
553Cancels the request, if possible. Has the effect of skipping execution 751Cancels the request, if possible. Has the effect of skipping execution
554when entering the B<execute> state and skipping calling the callback when 752when entering the B<execute> state and skipping calling the callback when
555entering the the B<result> state, but will leave the request otherwise 753entering the the B<result> state, but will leave the request otherwise
556untouched. That means that requests that currently execute will not be 754untouched. That means that requests that currently execute will not be
557stopped and resources held by the request will not be freed prematurely. 755stopped and resources held by the request will not be freed prematurely.
558 756
757=item cb $req $callback->(...)
758
759Replace (or simply set) the callback registered to the request.
760
559=back 761=back
560 762
561=head2 IO::AIO::GRP CLASS 763=head2 IO::AIO::GRP CLASS
562 764
563This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to 765This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
577You add requests by calling the C<add> method with one or more 779You add requests by calling the C<add> method with one or more
578C<IO::AIO::REQ> objects: 780C<IO::AIO::REQ> objects:
579 781
580 $grp->add (aio_unlink "..."); 782 $grp->add (aio_unlink "...");
581 783
582 add $grp aio_stat "...", sub { ... }; 784 add $grp aio_stat "...", sub {
785 $_[0] or return $grp->result ("error");
786
787 # add another request dynamically, if first succeeded
788 add $grp aio_open "...", sub {
789 $grp->result ("ok");
790 };
791 };
583 792
584This makes it very easy to create composite requests (see the source of 793This makes it very easy to create composite requests (see the source of
585C<aio_move> for an application) that work and feel like simple requests. 794C<aio_move> for an application) that work and feel like simple requests.
586 795
796=over 4
797
587The IO::AIO::GRP objects will be cleaned up during calls to 798=item * The IO::AIO::GRP objects will be cleaned up during calls to
588C<IO::AIO::poll_cb>, just like any other request. 799C<IO::AIO::poll_cb>, just like any other request.
589 800
590They can be canceled like any other request. Canceling will cancel not 801=item * They can be canceled like any other request. Canceling will cancel not
591just the request itself, but also all requests it contains. 802only the request itself, but also all requests it contains.
592 803
593They can also can also be added to other IO::AIO::GRP objects. 804=item * They can also can also be added to other IO::AIO::GRP objects.
805
806=item * You must not add requests to a group from within the group callback (or
807any later time).
808
809=back
594 810
595Their lifetime, simplified, looks like this: when they are empty, they 811Their lifetime, simplified, looks like this: when they are empty, they
596will finish very quickly. If they contain only requests that are in the 812will finish very quickly. If they contain only requests that are in the
597C<done> state, they will also finish. Otherwise they will continue to 813C<done> state, they will also finish. Otherwise they will continue to
598exist. 814exist.
599 815
816That means after creating a group you have some time to add requests. And
817in the callbacks of those requests, you can add further requests to the
818group. And only when all those requests have finished will the the group
819itself finish.
820
600=over 4 821=over 4
601 822
823=item add $grp ...
824
602=item $grp->add (...) 825=item $grp->add (...)
603 826
604=item add $grp ... 827Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
828be added, including other groups, as long as you do not create circular
829dependencies.
605 830
606Add one or more 831Returns all its arguments.
607Cancels the request, if possible. Has the effect of skipping execution 832
608when entering the B<execute> state and skipping calling the callback when 833=item $grp->cancel_subs
609entering the the B<result> state, but will leave the request otherwise 834
610untouched. That means that requests that currently execute will not be 835Cancel all subrequests and clears any feeder, but not the group request
611stopped and resources held by the request will not be freed prematurely. 836itself. Useful when you queued a lot of events but got a result early.
837
838=item $grp->result (...)
839
840Set the result value(s) that will be passed to the group callback when all
841subrequests have finished and set thre groups errno to the current value
842of errno (just like calling C<errno> without an error number). By default,
843no argument will be passed and errno is zero.
844
845=item $grp->errno ([$errno])
846
847Sets the group errno value to C<$errno>, or the current value of errno
848when the argument is missing.
849
850Every aio request has an associated errno value that is restored when
851the callback is invoked. This method lets you change this value from its
852default (0).
853
854Calling C<result> will also set errno, so make sure you either set C<$!>
855before the call to C<result>, or call c<errno> after it.
856
857=item feed $grp $callback->($grp)
858
859Sets a feeder/generator on this group: every group can have an attached
860generator that generates requests if idle. The idea behind this is that,
861although you could just queue as many requests as you want in a group,
862this might starve other requests for a potentially long time. For
863example, C<aio_scandir> might generate hundreds of thousands C<aio_stat>
864requests, delaying any later requests for a long time.
865
866To avoid this, and allow incremental generation of requests, you can
867instead a group and set a feeder on it that generates those requests. The
868feed callback will be called whenever there are few enough (see C<limit>,
869below) requests active in the group itself and is expected to queue more
870requests.
871
872The feed callback can queue as many requests as it likes (i.e. C<add> does
873not impose any limits).
874
875If the feed does not queue more requests when called, it will be
876automatically removed from the group.
877
878If the feed limit is C<0>, it will be set to C<2> automatically.
879
880Example:
881
882 # stat all files in @files, but only ever use four aio requests concurrently:
883
884 my $grp = aio_group sub { print "finished\n" };
885 limit $grp 4;
886 feed $grp sub {
887 my $file = pop @files
888 or return;
889
890 add $grp aio_stat $file, sub { ... };
891 };
892
893=item limit $grp $num
894
895Sets the feeder limit for the group: The feeder will be called whenever
896the group contains less than this many requests.
897
898Setting the limit to C<0> will pause the feeding process.
612 899
613=back 900=back
614 901
615=head2 SUPPORT FUNCTIONS 902=head2 SUPPORT FUNCTIONS
903
904=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
616 905
617=over 4 906=over 4
618 907
619=item $fileno = IO::AIO::poll_fileno 908=item $fileno = IO::AIO::poll_fileno
620 909
625 914
626See C<poll_cb> for an example. 915See C<poll_cb> for an example.
627 916
628=item IO::AIO::poll_cb 917=item IO::AIO::poll_cb
629 918
630Process all outstanding events on the result pipe. You have to call this 919Process some outstanding events on the result pipe. You have to call this
631regularly. Returns the number of events processed. Returns immediately 920regularly. Returns the number of events processed. Returns immediately
632when no events are outstanding. 921when no events are outstanding. The amount of events processed depends on
922the settings of C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
923
924If not all requests were processed for whatever reason, the filehandle
925will still be ready when C<poll_cb> returns.
633 926
634Example: Install an Event watcher that automatically calls 927Example: Install an Event watcher that automatically calls
635IO::AIO::poll_cb with high priority: 928IO::AIO::poll_cb with high priority:
636 929
637 Event->io (fd => IO::AIO::poll_fileno, 930 Event->io (fd => IO::AIO::poll_fileno,
638 poll => 'r', async => 1, 931 poll => 'r', async => 1,
639 cb => \&IO::AIO::poll_cb); 932 cb => \&IO::AIO::poll_cb);
640 933
934=item IO::AIO::max_poll_reqs $nreqs
935
936=item IO::AIO::max_poll_time $seconds
937
938These set the maximum number of requests (default C<0>, meaning infinity)
939that are being processed by C<IO::AIO::poll_cb> in one call, respectively
940the maximum amount of time (default C<0>, meaning infinity) spent in
941C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
942of time C<poll_cb> is allowed to use).
943
944Setting these is useful if you want to ensure some level of
945interactiveness when perl is not fast enough to process all requests in
946time.
947
948For interactive programs, values such as C<0.01> to C<0.1> should be fine.
949
950Example: Install an Event watcher that automatically calls
951IO::AIO::poll_some with low priority, to ensure that other parts of the
952program get the CPU sometimes even under high AIO load.
953
954 # try not to spend much more than 0.1s in poll_cb
955 IO::AIO::max_poll_time 0.1;
956
957 # use a low priority so other tasks have priority
958 Event->io (fd => IO::AIO::poll_fileno,
959 poll => 'r', nice => 1,
960 cb => &IO::AIO::poll_cb);
961
641=item IO::AIO::poll_wait 962=item IO::AIO::poll_wait
642 963
643Wait till the result filehandle becomes ready for reading (simply does a 964Wait till the result filehandle becomes ready for reading (simply does a
644C<select> on the filehandle. This is useful if you want to synchronously wait 965C<select> on the filehandle. This is useful if you want to synchronously
645for some requests to finish). 966wait for some requests to finish).
646 967
647See C<nreqs> for an example. 968See C<nreqs> for an example.
648 969
970=item IO::AIO::poll
971
972Waits until some requests have been handled.
973
974Strictly equivalent to:
975
976 IO::AIO::poll_wait, IO::AIO::poll_cb
977 if IO::AIO::nreqs;
978
649=item IO::AIO::nreqs 979=item IO::AIO::flush
650 980
651Returns the number of requests currently outstanding (i.e. for which their 981Wait till all outstanding AIO requests have been handled.
652callback has not been invoked yet).
653 982
654Example: wait till there are no outstanding requests anymore: 983Strictly equivalent to:
655 984
656 IO::AIO::poll_wait, IO::AIO::poll_cb 985 IO::AIO::poll_wait, IO::AIO::poll_cb
657 while IO::AIO::nreqs; 986 while IO::AIO::nreqs;
658 987
659=item IO::AIO::flush 988=head3 CONTROLLING THE NUMBER OF THREADS
660
661Wait till all outstanding AIO requests have been handled.
662
663Strictly equivalent to:
664
665 IO::AIO::poll_wait, IO::AIO::poll_cb
666 while IO::AIO::nreqs;
667
668=item IO::AIO::poll
669
670Waits until some requests have been handled.
671
672Strictly equivalent to:
673
674 IO::AIO::poll_wait, IO::AIO::poll_cb
675 if IO::AIO::nreqs;
676 989
677=item IO::AIO::min_parallel $nthreads 990=item IO::AIO::min_parallel $nthreads
678 991
679Set the minimum number of AIO threads to C<$nthreads>. The current default 992Set the minimum number of AIO threads to C<$nthreads>. The current
680is C<4>, which means four asynchronous operations can be done at one time 993default is C<8>, which means eight asynchronous operations can execute
681(the number of outstanding operations, however, is unlimited). 994concurrently at any one time (the number of outstanding requests,
995however, is unlimited).
682 996
683IO::AIO starts threads only on demand, when an AIO request is queued and 997IO::AIO starts threads only on demand, when an AIO request is queued and
684no free thread exists. 998no free thread exists. Please note that queueing up a hundred requests can
999create demand for a hundred threads, even if it turns out that everything
1000is in the cache and could have been processed faster by a single thread.
685 1001
686It is recommended to keep the number of threads low, as some Linux 1002It is recommended to keep the number of threads relatively low, as some
687kernel versions will scale negatively with the number of threads (higher 1003Linux kernel versions will scale negatively with the number of threads
688parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 1004(higher parallelity => MUCH higher latency). With current Linux 2.6
689threads should be fine. 1005versions, 4-32 threads should be fine.
690 1006
691Under most circumstances you don't need to call this function, as the 1007Under most circumstances you don't need to call this function, as the
692module selects a default that is suitable for low to moderate load. 1008module selects a default that is suitable for low to moderate load.
693 1009
694=item IO::AIO::max_parallel $nthreads 1010=item IO::AIO::max_parallel $nthreads
703This module automatically runs C<max_parallel 0> at program end, to ensure 1019This module automatically runs C<max_parallel 0> at program end, to ensure
704that all threads are killed and that there are no outstanding requests. 1020that all threads are killed and that there are no outstanding requests.
705 1021
706Under normal circumstances you don't need to call this function. 1022Under normal circumstances you don't need to call this function.
707 1023
1024=item IO::AIO::max_idle $nthreads
1025
1026Limit the number of threads (default: 4) that are allowed to idle (i.e.,
1027threads that did not get a request to process within 10 seconds). That
1028means if a thread becomes idle while C<$nthreads> other threads are also
1029idle, it will free its resources and exit.
1030
1031This is useful when you allow a large number of threads (e.g. 100 or 1000)
1032to allow for extremely high load situations, but want to free resources
1033under normal circumstances (1000 threads can easily consume 30MB of RAM).
1034
1035The default is probably ok in most situations, especially if thread
1036creation is fast. If thread creation is very slow on your system you might
1037want to use larger values.
1038
708=item $oldnreqs = IO::AIO::max_outstanding $nreqs 1039=item $oldmaxreqs = IO::AIO::max_outstanding $maxreqs
1040
1041This is a very bad function to use in interactive programs because it
1042blocks, and a bad way to reduce concurrency because it is inexact: Better
1043use an C<aio_group> together with a feed callback.
709 1044
710Sets the maximum number of outstanding requests to C<$nreqs>. If you 1045Sets the maximum number of outstanding requests to C<$nreqs>. If you
711try to queue up more than this number of requests, the caller will block until 1046to queue up more than this number of requests, the next call to the
712some requests have been handled. 1047C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
1048function will block until the limit is no longer exceeded.
713 1049
714The default is very large, so normally there is no practical limit. If you 1050The default value is very large, so there is no practical limit on the
715queue up many requests in a loop it often improves speed if you set 1051number of outstanding requests.
716this to a relatively low number, such as C<100>.
717 1052
718Under normal circumstances you don't need to call this function. 1053You can still queue as many requests as you want. Therefore,
1054C<max_oustsanding> is mainly useful in simple scripts (with low values) or
1055as a stop gap to shield against fatal memory overflow (with large values).
1056
1057=head3 STATISTICAL INFORMATION
1058
1059=item IO::AIO::nreqs
1060
1061Returns the number of requests currently in the ready, execute or pending
1062states (i.e. for which their callback has not been invoked yet).
1063
1064Example: wait till there are no outstanding requests anymore:
1065
1066 IO::AIO::poll_wait, IO::AIO::poll_cb
1067 while IO::AIO::nreqs;
1068
1069=item IO::AIO::nready
1070
1071Returns the number of requests currently in the ready state (not yet
1072executed).
1073
1074=item IO::AIO::npending
1075
1076Returns the number of requests currently in the pending state (executed,
1077but not yet processed by poll_cb).
719 1078
720=back 1079=back
721 1080
722=cut 1081=cut
723 1082
735 or return undef; 1094 or return undef;
736 1095
737 *$sym 1096 *$sym
738} 1097}
739 1098
740min_parallel 4; 1099min_parallel 8;
741 1100
742END { 1101END {
743 max_parallel 0; 1102 min_parallel 1;
744} 1103 flush;
1104};
745 1105
7461; 11061;
747 1107
748=head2 FORK BEHAVIOUR 1108=head2 FORK BEHAVIOUR
749 1109
750This module should do "the right thing" when the process using it forks: 1110This module should do "the right thing" when the process using it forks:
751 1111
752Before the fork, IO::AIO enters a quiescent state where no requests 1112Before the fork, IO::AIO enters a quiescent state where no requests
753can be added in other threads and no results will be processed. After 1113can be added in other threads and no results will be processed. After
754the fork the parent simply leaves the quiescent state and continues 1114the fork the parent simply leaves the quiescent state and continues
755request/result processing, while the child clears the request/result 1115request/result processing, while the child frees the request/result queue
756queue (so the requests started before the fork will only be handled in 1116(so that the requests started before the fork will only be handled in the
757the parent). Threads will be started on demand until the limit ste in the 1117parent). Threads will be started on demand until the limit set in the
758parent process has been reached again. 1118parent process has been reached again.
759 1119
760In short: the parent will, after a short pause, continue as if fork had 1120In short: the parent will, after a short pause, continue as if fork had
761not been called, while the child will act as if IO::AIO has not been used 1121not been called, while the child will act as if IO::AIO has not been used
762yet. 1122yet.
763 1123
1124=head2 MEMORY USAGE
1125
1126Per-request usage:
1127
1128Each aio request uses - depending on your architecture - around 100-200
1129bytes of memory. In addition, stat requests need a stat buffer (possibly
1130a few hundred bytes), readdir requires a result buffer and so on. Perl
1131scalars and other data passed into aio requests will also be locked and
1132will consume memory till the request has entered the done state.
1133
1134This is now awfully much, so queuing lots of requests is not usually a
1135problem.
1136
1137Per-thread usage:
1138
1139In the execution phase, some aio requests require more memory for
1140temporary buffers, and each thread requires a stack and other data
1141structures (usually around 16k-128k, depending on the OS).
1142
1143=head1 KNOWN BUGS
1144
1145Known bugs will be fixed in the next release.
1146
764=head1 SEE ALSO 1147=head1 SEE ALSO
765 1148
766L<Coro>, L<Linux::AIO> (obsolete). 1149L<Coro::AIO>.
767 1150
768=head1 AUTHOR 1151=head1 AUTHOR
769 1152
770 Marc Lehmann <schmorp@schmorp.de> 1153 Marc Lehmann <schmorp@schmorp.de>
771 http://home.schmorp.de/ 1154 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines