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 1.10 by root, Sun Jul 10 23:45:16 2005 UTC vs.
Revision 1.54 by root, Sun Oct 22 00:19:05 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
21
22 my $req = aio_unlink "/tmp/file", sub { };
23 $req->cancel; # cancel request if still in queue
24
25 # AnyEvent
26 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
27 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
28
20 # Event 29 # Event
21 Event->io (fd => IO::AIO::poll_fileno, 30 Event->io (fd => IO::AIO::poll_fileno,
22 poll => 'r', 31 poll => 'r',
23 cb => \&IO::AIO::poll_cb); 32 cb => \&IO::AIO::poll_cb);
24 33
25 # Glib/Gtk2 34 # Glib/Gtk2
26 add_watch Glib::IO IO::AIO::poll_fileno, 35 add_watch Glib::IO IO::AIO::poll_fileno,
27 \&IO::AIO::poll_cb; 36 in => sub { IO::AIO::poll_cb; 1 };
28 37
29 # Tk 38 # Tk
30 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", 39 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
31 readable => \&IO::AIO::poll_cb); 40 readable => \&IO::AIO::poll_cb);
41
42 # Danga::Socket
43 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
44 \&IO::AIO::poll_cb);
45
32 46
33=head1 DESCRIPTION 47=head1 DESCRIPTION
34 48
35This module implements asynchronous I/O using whatever means your 49This module implements asynchronous I/O using whatever means your
36operating system supports. 50operating system supports.
43not well-supported (Linux doesn't allow them on normal files currently, 57not well-supported (Linux doesn't allow them on normal files currently,
44for example), and they would only support aio_read and aio_write, so the 58for example), and they would only support aio_read and aio_write, so the
45remaining functionality would have to be implemented using threads anyway. 59remaining functionality would have to be implemented using threads anyway.
46 60
47Although the module will work with in the presence of other threads, it is 61Although the module will work with in the presence of other threads, it is
48currently not reentrant, so use appropriate locking yourself. 62currently not reentrant, so use appropriate locking yourself, always call
63C<poll_cb> from within the same thread, or never call C<poll_cb> (or other
64C<aio_> functions) recursively.
49 65
50=cut 66=cut
51 67
52package IO::AIO; 68package IO::AIO;
53 69
70no warnings;
71use strict 'vars';
72
54use base 'Exporter'; 73use base 'Exporter';
55 74
56use Fcntl ();
57
58BEGIN { 75BEGIN {
59 $VERSION = 0.3; 76 our $VERSION = '1.99';
60 77
61 @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink 78 our @EXPORT = 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
62 aio_fsync aio_fdatasync aio_readahead); 80 aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move
81 aio_group);
63 @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); 82 our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);
83
84 @IO::AIO::GRP::ISA = 'IO::AIO::REQ';
64 85
65 require XSLoader; 86 require XSLoader;
66 XSLoader::load IO::AIO, $VERSION; 87 XSLoader::load ("IO::AIO", $VERSION);
67} 88}
68 89
69=head1 FUNCTIONS 90=head1 FUNCTIONS
70 91
71=head2 AIO FUNCTIONS 92=head2 AIO FUNCTIONS
72 93
73All the C<aio_*> calls are more or less thin wrappers around the syscall 94All the C<aio_*> calls are more or less thin wrappers around the syscall
74with the same name (sans C<aio_>). The arguments are similar or identical, 95with the same name (sans C<aio_>). The arguments are similar or identical,
75and they all accept an additional C<$callback> argument which must be 96and they all accept an additional (and optional) C<$callback> argument
76a code reference. This code reference will get called with the syscall 97which must be a code reference. This code reference will get called with
77return code (e.g. most syscalls return C<-1> on error, unlike perl, which 98the syscall return code (e.g. most syscalls return C<-1> on error, unlike
78usually delivers "false") as it's sole argument when the given syscall has 99perl, which usually delivers "false") as it's sole argument when the given
79been executed asynchronously. 100syscall has been executed asynchronously.
80 101
81All functions that expect a filehandle will also accept a file descriptor. 102All functions expecting a filehandle keep a copy of the filehandle
103internally until the request has finished.
82 104
105All non-composite requests (requests that are not broken down into
106multiple requests) return objects of type L<IO::AIO::REQ> that allow
107further manipulation of running requests.
108
83The filenames you pass to these routines I<must> be absolute. The reason 109The pathnames you pass to these routines I<must> be absolute and
84is that at the time the request is being executed, the current working 110encoded in byte form. The reason for the former is that at the time the
85directory could have changed. Alternatively, you can make sure that you 111request is being executed, the current working directory could have
112changed. Alternatively, you can make sure that you never change the
86never change the current working directory. 113current working directory.
114
115To encode pathnames to byte form, either make sure you either: a)
116always pass in filenames you got from outside (command line, readdir
117etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode
118your pathnames to the locale (or other) encoding in effect in the user
119environment, d) use Glib::filename_from_unicode on unicode filenames or e)
120use something else.
87 121
88=over 4 122=over 4
89 123
90=item aio_open $pathname, $flags, $mode, $callback 124=item aio_open $pathname, $flags, $mode, $callback->($fh)
91 125
92Asynchronously open or create a file and call the callback with a newly 126Asynchronously open or create a file and call the callback with a newly
93created filehandle for the file. 127created filehandle for the file.
94 128
95The pathname passed to C<aio_open> must be absolute. See API NOTES, above, 129The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
96for an explanation. 130for an explanation.
97 131
98The C<$mode> argument is a bitmask. See the C<Fcntl> module for a 132The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
99list. They are the same as used in C<sysopen>. 133list. They are the same as used by C<sysopen>.
134
135Likewise, C<$mode> specifies the mode of the newly created file, if it
136didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
137except that it is mandatory (i.e. use C<0> if you don't create new files,
138and C<0666> or C<0777> if you do).
100 139
101Example: 140Example:
102 141
103 aio_open "/etc/passwd", O_RDONLY, 0, sub { 142 aio_open "/etc/passwd", O_RDONLY, 0, sub {
104 if ($_[0]) { 143 if ($_[0]) {
107 } else { 146 } else {
108 die "open failed: $!\n"; 147 die "open failed: $!\n";
109 } 148 }
110 }; 149 };
111 150
112=item aio_close $fh, $callback 151=item aio_close $fh, $callback->($status)
113 152
114Asynchronously close a file and call the callback with the result 153Asynchronously close a file and call the callback with the result
115code. I<WARNING:> although accepted, you should not pass in a perl 154code. I<WARNING:> although accepted, you should not pass in a perl
116filehandle here, as perl will likely close the file descriptor itself when 155filehandle here, as perl will likely close the file descriptor another
117the filehandle is destroyed. Normally, you can safely call perls C<close> 156time when the filehandle is destroyed. Normally, you can safely call perls
118or just let filehandles go out of scope. 157C<close> or just let filehandles go out of scope.
119 158
159This is supposed to be a bug in the API, so that might change. It's
160therefore best to avoid this function.
161
120=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback 162=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
121 163
122=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback 164=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
123 165
124Reads or writes C<length> bytes from the specified C<fh> and C<offset> 166Reads or writes C<length> bytes from the specified C<fh> and C<offset>
125into the scalar given by C<data> and offset C<dataoffset> and calls the 167into the scalar given by C<data> and offset C<dataoffset> and calls the
126callback without the actual number of bytes read (or -1 on error, just 168callback without the actual number of bytes read (or -1 on error, just
127like the syscall). 169like the syscall).
128 170
171The C<$data> scalar I<MUST NOT> be modified in any way while the request
172is outstanding. Modifying it can result in segfaults or WW3 (if the
173necessary/optional hardware is installed).
174
129Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at 175Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
130offset C<0> within the scalar: 176offset C<0> within the scalar:
131 177
132 aio_read $fh, 7, 15, $buffer, 0, sub { 178 aio_read $fh, 7, 15, $buffer, 0, sub {
133 $_[0] > 0 or die "read error: $!"; 179 $_[0] > 0 or die "read error: $!";
134 print "read $_[0] bytes: <$buffer>\n"; 180 print "read $_[0] bytes: <$buffer>\n";
135 }; 181 };
136 182
183=item aio_move $srcpath, $dstpath, $callback->($status)
184
185Try to move the I<file> (directories not supported as either source or
186destination) from C<$srcpath> to C<$dstpath> and call the callback with
187the C<0> (error) or C<-1> ok.
188
189This is a composite request that tries to rename(2) the file first. If
190rename files with C<EXDEV>, it creates the destination file with mode 0200
191and copies the contents of the source file into it using C<aio_sendfile>,
192followed by restoring atime, mtime, access mode and uid/gid, in that
193order, and unlinking the C<$srcpath>.
194
195If an error occurs, the partial destination file will be unlinked, if
196possible, except when setting atime, mtime, access mode and uid/gid, where
197errors are being ignored.
198
199=cut
200
201sub aio_move($$$) {
202 my ($src, $dst, $cb) = @_;
203
204 aio_rename $src, $dst, sub {
205 if ($_[0] && $! == EXDEV) {
206 aio_open $src, O_RDONLY, 0, sub {
207 if (my $src_fh = $_[0]) {
208 my @stat = stat $src_fh;
209
210 aio_open $dst, O_WRONLY, 0200, sub {
211 if (my $dst_fh = $_[0]) {
212 aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
213 close $src_fh;
214
215 if ($_[0] == $stat[7]) {
216 utime $stat[8], $stat[9], $dst;
217 chmod $stat[2] & 07777, $dst_fh;
218 chown $stat[4], $stat[5], $dst_fh;
219 close $dst_fh;
220
221 aio_unlink $src, sub {
222 $cb->($_[0]);
223 };
224 } else {
225 my $errno = $!;
226 aio_unlink $dst, sub {
227 $! = $errno;
228 $cb->(-1);
229 };
230 }
231 };
232 } else {
233 $cb->(-1);
234 }
235 },
236
237 } else {
238 $cb->(-1);
239 }
240 };
241 } else {
242 $cb->($_[0]);
243 }
244 };
245}
246
247=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
248
249Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
250reading at byte offset C<$in_offset>, and starts writing at the current
251file offset of C<$out_fh>. Because of that, it is not safe to issue more
252than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
253other.
254
255This call tries to make use of a native C<sendfile> syscall to provide
256zero-copy operation. For this to work, C<$out_fh> should refer to a
257socket, and C<$in_fh> should refer to mmap'able file.
258
259If the native sendfile call fails or is not implemented, it will be
260emulated, so you can call C<aio_sendfile> on any type of filehandle
261regardless of the limitations of the operating system.
262
263Please note, however, that C<aio_sendfile> can read more bytes from
264C<$in_fh> than are written, and there is no way to find out how many
265bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
266provides the number of bytes written to C<$out_fh>. Only if the result
267value equals C<$length> one can assume that C<$length> bytes have been
268read.
269
137=item aio_readahead $fh,$offset,$length, $callback 270=item aio_readahead $fh,$offset,$length, $callback->($retval)
138 271
139Asynchronously reads the specified byte range into the page cache, using
140the C<readahead> syscall. If that syscall doesn't exist the status will be
141C<-1> and C<$!> is set to ENOSYS.
142
143readahead() populates the page cache with data from a file so that 272C<aio_readahead> populates the page cache with data from a file so that
144subsequent reads from that file will not block on disk I/O. The C<$offset> 273subsequent reads from that file will not block on disk I/O. The C<$offset>
145argument specifies the starting point from which data is to be read and 274argument specifies the starting point from which data is to be read and
146C<$length> specifies the number of bytes to be read. I/O is performed in 275C<$length> specifies the number of bytes to be read. I/O is performed in
147whole pages, so that offset is effectively rounded down to a page boundary 276whole pages, so that offset is effectively rounded down to a page boundary
148and bytes are read up to the next page boundary greater than or equal to 277and bytes are read up to the next page boundary greater than or equal to
149(off-set+length). aio_readahead() does not read beyond the end of the 278(off-set+length). C<aio_readahead> does not read beyond the end of the
150file. The current file offset of the file is left unchanged. 279file. The current file offset of the file is left unchanged.
151 280
281If that syscall doesn't exist (likely if your OS isn't Linux) it will be
282emulated by simply reading the data, which would have a similar effect.
283
152=item aio_stat $fh_or_path, $callback 284=item aio_stat $fh_or_path, $callback->($status)
153 285
154=item aio_lstat $fh, $callback 286=item aio_lstat $fh, $callback->($status)
155 287
156Works like perl's C<stat> or C<lstat> in void context. The callback will 288Works like perl's C<stat> or C<lstat> in void context. The callback will
157be called after the stat and the results will be available using C<stat _> 289be called after the stat and the results will be available using C<stat _>
158or C<-s _> etc... 290or C<-s _> etc...
159 291
169 aio_stat "/etc/passwd", sub { 301 aio_stat "/etc/passwd", sub {
170 $_[0] and die "stat failed: $!"; 302 $_[0] and die "stat failed: $!";
171 print "size is ", -s _, "\n"; 303 print "size is ", -s _, "\n";
172 }; 304 };
173 305
174=item aio_unlink $pathname, $callback 306=item aio_unlink $pathname, $callback->($status)
175 307
176Asynchronously unlink (delete) a file and call the callback with the 308Asynchronously unlink (delete) a file and call the callback with the
177result code. 309result code.
178 310
311=item aio_link $srcpath, $dstpath, $callback->($status)
312
313Asynchronously create a new link to the existing object at C<$srcpath> at
314the path C<$dstpath> and call the callback with the result code.
315
316=item aio_symlink $srcpath, $dstpath, $callback->($status)
317
318Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
319the path C<$dstpath> and call the callback with the result code.
320
321=item aio_rename $srcpath, $dstpath, $callback->($status)
322
323Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
324rename(2) and call the callback with the result code.
325
326=item aio_rmdir $pathname, $callback->($status)
327
328Asynchronously rmdir (delete) a directory and call the callback with the
329result code.
330
331=item aio_readdir $pathname, $callback->($entries)
332
333Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
334directory (i.e. opendir + readdir + closedir). The entries will not be
335sorted, and will B<NOT> include the C<.> and C<..> entries.
336
337The callback a single argument which is either C<undef> or an array-ref
338with the filenames.
339
340=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
341
342Scans a directory (similar to C<aio_readdir>) but additionally tries to
343separate the entries of directory C<$path> into two sets of names, ones
344you can recurse into (directories or links to them), and ones you cannot
345recurse into (everything else).
346
347C<aio_scandir> is a composite request that consists of many sub
348requests. C<$maxreq> specifies the maximum number of outstanding aio
349requests that this function generates. If it is C<< <= 0 >>, then a
350suitable default will be chosen (currently 8).
351
352On error, the callback is called without arguments, otherwise it receives
353two array-refs with path-relative entry names.
354
355Example:
356
357 aio_scandir $dir, 0, sub {
358 my ($dirs, $nondirs) = @_;
359 print "real directories: @$dirs\n";
360 print "everything else: @$nondirs\n";
361 };
362
363Implementation notes.
364
365The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
366
367After reading the directory, the modification time, size etc. of the
368directory before and after the readdir is checked, and if they match (and
369isn't the current time), the link count will be used to decide how many
370entries are directories (if >= 2). Otherwise, no knowledge of the number
371of subdirectories will be assumed.
372
373Then entries will be sorted into likely directories (everything without
374a non-initial dot currently) and likely non-directories (everything
375else). Then every entry plus an appended C</.> will be C<stat>'ed,
376likely directories first. If that succeeds, it assumes that the entry
377is a directory or a symlink to directory (which will be checked
378seperately). This is often faster than stat'ing the entry itself because
379filesystems might detect the type of the entry without reading the inode
380data (e.g. ext2fs filetype feature).
381
382If the known number of directories (link count - 2) has been reached, the
383rest of the entries is assumed to be non-directories.
384
385This only works with certainty on POSIX (= UNIX) filesystems, which
386fortunately are the vast majority of filesystems around.
387
388It will also likely work on non-POSIX filesystems with reduced efficiency
389as those tend to return 0 or 1 as link counts, which disables the
390directory counting heuristic.
391
392=cut
393
394sub aio_scandir($$$) {
395 my ($path, $maxreq, $cb) = @_;
396
397 $maxreq = 8 if $maxreq <= 0;
398
399 # stat once
400 aio_stat $path, sub {
401 return $cb->() if $_[0];
402 my $now = time;
403 my $hash1 = join ":", (stat _)[0,1,3,7,9];
404
405 # read the directory entries
406 aio_readdir $path, sub {
407 my $entries = shift
408 or return $cb->();
409
410 # stat the dir another time
411 aio_stat $path, sub {
412 my $hash2 = join ":", (stat _)[0,1,3,7,9];
413
414 my $ndirs;
415
416 # take the slow route if anything looks fishy
417 if ($hash1 ne $hash2 or (stat _)[9] == $now) {
418 $ndirs = -1;
419 } else {
420 # if nlink == 2, we are finished
421 # on non-posix-fs's, we rely on nlink < 2
422 $ndirs = (stat _)[3] - 2
423 or return $cb->([], $entries);
424 }
425
426 # sort into likely dirs and likely nondirs
427 # dirs == files without ".", short entries first
428 $entries = [map $_->[0],
429 sort { $b->[1] cmp $a->[1] }
430 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
431 @$entries];
432
433 my (@dirs, @nondirs);
434
435 my ($statcb, $schedcb);
436 my $nreq = 0;
437
438 $schedcb = sub {
439 if (@$entries) {
440 if ($nreq < $maxreq) {
441 my $ent = pop @$entries;
442 $nreq++;
443 aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) };
444 }
445 } elsif (!$nreq) {
446 # finished
447 undef $statcb;
448 undef $schedcb;
449 $cb->(\@dirs, \@nondirs) if $cb;
450 undef $cb;
451 }
452 };
453 $statcb = sub {
454 my ($status, $entry) = @_;
455
456 if ($status < 0) {
457 $nreq--;
458 push @nondirs, $entry;
459 &$schedcb;
460 } else {
461 # need to check for real directory
462 aio_lstat "$path/$entry", sub {
463 $nreq--;
464
465 if (-d _) {
466 push @dirs, $entry;
467
468 if (!--$ndirs) {
469 push @nondirs, @$entries;
470 $entries = [];
471 }
472 } else {
473 push @nondirs, $entry;
474 }
475
476 &$schedcb;
477 }
478 }
479 };
480
481 &$schedcb while @$entries && $nreq < $maxreq;
482 };
483 };
484 };
485}
486
179=item aio_fsync $fh, $callback 487=item aio_fsync $fh, $callback->($status)
180 488
181Asynchronously call fsync on the given filehandle and call the callback 489Asynchronously call fsync on the given filehandle and call the callback
182with the fsync result code. 490with the fsync result code.
183 491
184=item aio_fdatasync $fh, $callback 492=item aio_fdatasync $fh, $callback->($status)
185 493
186Asynchronously call fdatasync on the given filehandle and call the 494Asynchronously call fdatasync on the given filehandle and call the
187callback with the fdatasync result code. 495callback with the fdatasync result code.
188 496
497If this call isn't available because your OS lacks it or it couldn't be
498detected, it will be emulated by calling C<fsync> instead.
499
500=item aio_group $callback->()
501
502=item aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED*
503
504Mainly used for debugging and benchmarking, this aio request puts one of
505the request workers to sleep for the given time.
506
189=back 507=back
190 508
509=head2 IO::AIO::REQ CLASS
510
511All non-aggregate C<aio_*> functions return an object of this class when
512called in non-void context.
513
514A request always moves through the following five states in its lifetime,
515in order: B<ready> (request has been created, but has not been executed
516yet), B<execute> (request is currently being executed), B<pending>
517(request has been executed but callback has not been called yet),
518B<result> (results are being processed synchronously, includes calling the
519callback) and B<done> (request has reached the end of its lifetime and
520holds no resources anymore).
521
522=over 4
523
524=item $req->cancel
525
526Cancels the request, if possible. Has the effect of skipping execution
527when entering the B<execute> state and skipping calling the callback when
528entering the the B<result> state, but will leave the request otherwise
529untouched. That means that requests that currently execute will not be
530stopped and resources held by the request will not be freed prematurely.
531
532=back
533
191=head2 SUPPORT FUNCTIONS 534=head2 SUPPORT FUNCTIONS
192 535
193=over 4 536=over 4
194 537
195=item $fileno = IO::AIO::poll_fileno 538=item $fileno = IO::AIO::poll_fileno
196 539
197Return the I<request result pipe filehandle>. This filehandle must be 540Return the I<request result pipe file descriptor>. This filehandle must be
198polled for reading by some mechanism outside this module (e.g. Event 541polled for reading by some mechanism outside this module (e.g. Event or
199or select, see below). If the pipe becomes readable you have to call 542select, see below or the SYNOPSIS). If the pipe becomes readable you have
200C<poll_cb> to check the results. 543to call C<poll_cb> to check the results.
201 544
202See C<poll_cb> for an example. 545See C<poll_cb> for an example.
203 546
204=item IO::AIO::poll_cb 547=item IO::AIO::poll_cb
205 548
206Process all outstanding events on the result pipe. You have to call this 549Process all outstanding events on the result pipe. You have to call this
207regularly. Returns the number of events processed. Returns immediately 550regularly. Returns the number of events processed. Returns immediately
208when no events are outstanding. 551when no events are outstanding.
209 552
210You can use Event to multiplex, e.g.: 553Example: Install an Event watcher that automatically calls
554IO::AIO::poll_cb with high priority:
211 555
212 Event->io (fd => IO::AIO::poll_fileno, 556 Event->io (fd => IO::AIO::poll_fileno,
213 poll => 'r', async => 1, 557 poll => 'r', async => 1,
214 cb => \&IO::AIO::poll_cb); 558 cb => \&IO::AIO::poll_cb);
215 559
216=item IO::AIO::poll_wait 560=item IO::AIO::poll_wait
217 561
218Wait till the result filehandle becomes ready for reading (simply does a 562Wait till the result filehandle becomes ready for reading (simply does a
219select on the filehandle. This is useful if you want to synchronously wait 563C<select> on the filehandle. This is useful if you want to synchronously wait
220for some requests to finish). 564for some requests to finish).
221 565
222See C<nreqs> for an example. 566See C<nreqs> for an example.
223 567
224=item IO::AIO::nreqs 568=item IO::AIO::nreqs
225 569
226Returns the number of requests currently outstanding. 570Returns the number of requests currently outstanding (i.e. for which their
571callback has not been invoked yet).
227 572
228Example: wait till there are no outstanding requests anymore: 573Example: wait till there are no outstanding requests anymore:
229 574
230 IO::AIO::poll_wait, IO::AIO::poll_cb 575 IO::AIO::poll_wait, IO::AIO::poll_cb
231 while IO::AIO::nreqs; 576 while IO::AIO::nreqs;
232 577
578=item IO::AIO::flush
579
580Wait till all outstanding AIO requests have been handled.
581
582Strictly equivalent to:
583
584 IO::AIO::poll_wait, IO::AIO::poll_cb
585 while IO::AIO::nreqs;
586
587=item IO::AIO::poll
588
589Waits until some requests have been handled.
590
591Strictly equivalent to:
592
593 IO::AIO::poll_wait, IO::AIO::poll_cb
594 if IO::AIO::nreqs;
595
233=item IO::AIO::min_parallel $nthreads 596=item IO::AIO::min_parallel $nthreads
234 597
235Set the minimum number of AIO threads to C<$nthreads>. The default is 598Set the minimum number of AIO threads to C<$nthreads>. The current default
236C<1>, which means a single asynchronous operation can be done at one time 599is C<4>, which means four asynchronous operations can be done at one time
237(the number of outstanding operations, however, is unlimited). 600(the number of outstanding operations, however, is unlimited).
601
602IO::AIO starts threads only on demand, when an AIO request is queued and
603no free thread exists.
238 604
239It is recommended to keep the number of threads low, as some Linux 605It is recommended to keep the number of threads low, as some Linux
240kernel versions will scale negatively with the number of threads (higher 606kernel versions will scale negatively with the number of threads (higher
241parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 607parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32
242threads should be fine. 608threads should be fine.
243 609
244Under normal circumstances you don't need to call this function, as this 610Under most circumstances you don't need to call this function, as the
245module automatically starts some threads (the exact number might change, 611module selects a default that is suitable for low to moderate load.
246and is currently 4).
247 612
248=item IO::AIO::max_parallel $nthreads 613=item IO::AIO::max_parallel $nthreads
249 614
250Sets the maximum number of AIO threads to C<$nthreads>. If more than 615Sets the maximum number of AIO threads to C<$nthreads>. If more than the
251the specified number of threads are currently running, kill them. This 616specified number of threads are currently running, this function kills
252function blocks until the limit is reached. 617them. This function blocks until the limit is reached.
618
619While C<$nthreads> are zero, aio requests get queued but not executed
620until the number of threads has been increased again.
253 621
254This module automatically runs C<max_parallel 0> at program end, to ensure 622This module automatically runs C<max_parallel 0> at program end, to ensure
255that all threads are killed and that there are no outstanding requests. 623that all threads are killed and that there are no outstanding requests.
256 624
257Under normal circumstances you don't need to call this function. 625Under normal circumstances you don't need to call this function.
261Sets the maximum number of outstanding requests to C<$nreqs>. If you 629Sets the maximum number of outstanding requests to C<$nreqs>. If you
262try to queue up more than this number of requests, the caller will block until 630try to queue up more than this number of requests, the caller will block until
263some requests have been handled. 631some requests have been handled.
264 632
265The default is very large, so normally there is no practical limit. If you 633The default is very large, so normally there is no practical limit. If you
266queue up many requests in a loop it it often improves speed if you set 634queue up many requests in a loop it often improves speed if you set
267this to a relatively low number, such as C<100>. 635this to a relatively low number, such as C<100>.
268 636
269Under normal circumstances you don't need to call this function. 637Under normal circumstances you don't need to call this function.
270 638
271=back 639=back
274 642
275# support function to convert a fd into a perl filehandle 643# support function to convert a fd into a perl filehandle
276sub _fd2fh { 644sub _fd2fh {
277 return undef if $_[0] < 0; 645 return undef if $_[0] < 0;
278 646
279 # try to be perl5.6-compatible 647 # try to generate nice filehandles
280 local *AIO_FH; 648 my $sym = "IO::AIO::fd#$_[0]";
281 open AIO_FH, "+<&=$_[0]" 649 local *$sym;
650
651 open *$sym, "+<&=$_[0]" # usually works under any unix
652 or open *$sym, "<&=$_[0]" # cygwin needs this
653 or open *$sym, ">&=$_[0]" # or this
282 or return undef; 654 or return undef;
283 655
284 *AIO_FH 656 *$sym
285} 657}
286 658
287min_parallel 4; 659min_parallel 4;
288 660
289END { 661END {
290 max_parallel 0; 662 max_parallel 0;
291} 663}
292 664
2931; 6651;
294 666
667=head2 FORK BEHAVIOUR
668
669This module should do "the right thing" when the process using it forks:
670
671Before the fork, IO::AIO enters a quiescent state where no requests
672can be added in other threads and no results will be processed. After
673the fork the parent simply leaves the quiescent state and continues
674request/result processing, while the child clears the request/result
675queue (so the requests started before the fork will only be handled in
676the parent). Threads will be started on demand until the limit ste in the
677parent process has been reached again.
678
679In short: the parent will, after a short pause, continue as if fork had
680not been called, while the child will act as if IO::AIO has not been used
681yet.
682
295=head1 SEE ALSO 683=head1 SEE ALSO
296 684
297L<Coro>, L<Linux::AIO>. 685L<Coro>, L<Linux::AIO> (obsolete).
298 686
299=head1 AUTHOR 687=head1 AUTHOR
300 688
301 Marc Lehmann <schmorp@schmorp.de> 689 Marc Lehmann <schmorp@schmorp.de>
302 http://home.schmorp.de/ 690 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines