… | |
… | |
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 | # AnyEvent |
20 | # version 2+ has request and group objects |
|
|
21 | use IO::AIO 2; |
|
|
22 | |
|
|
23 | my $req = aio_unlink "/tmp/file", sub { }; |
|
|
24 | $req->cancel; # cancel request if still in queue |
|
|
25 | |
|
|
26 | my $grp = aio_group sub { print "all stats done\n" }; |
|
|
27 | add $grp aio_stat "..." for ...; |
|
|
28 | |
|
|
29 | # AnyEvent integration |
21 | open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; |
30 | open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; |
22 | my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); |
31 | my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); |
23 | |
32 | |
24 | # Event |
33 | # Event integration |
25 | Event->io (fd => IO::AIO::poll_fileno, |
34 | Event->io (fd => IO::AIO::poll_fileno, |
26 | poll => 'r', |
35 | poll => 'r', |
27 | cb => \&IO::AIO::poll_cb); |
36 | cb => \&IO::AIO::poll_cb); |
28 | |
37 | |
29 | # Glib/Gtk2 |
38 | # Glib/Gtk2 integration |
30 | add_watch Glib::IO IO::AIO::poll_fileno, |
39 | add_watch Glib::IO IO::AIO::poll_fileno, |
31 | in => sub { IO::AIO::poll_cb; 1 }; |
40 | in => sub { IO::AIO::poll_cb; 1 }; |
32 | |
41 | |
33 | # Tk |
42 | # Tk integration |
34 | Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", |
43 | Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", |
35 | readable => \&IO::AIO::poll_cb); |
44 | readable => \&IO::AIO::poll_cb); |
36 | |
45 | |
37 | # Danga::Socket |
46 | # Danga::Socket integration |
38 | Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
47 | Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
39 | \&IO::AIO::poll_cb); |
48 | \&IO::AIO::poll_cb); |
40 | |
|
|
41 | |
49 | |
42 | =head1 DESCRIPTION |
50 | =head1 DESCRIPTION |
43 | |
51 | |
44 | This module implements asynchronous I/O using whatever means your |
52 | This module implements asynchronous I/O using whatever means your |
45 | operating system supports. |
53 | operating system supports. |
… | |
… | |
61 | =cut |
69 | =cut |
62 | |
70 | |
63 | package IO::AIO; |
71 | package IO::AIO; |
64 | |
72 | |
65 | no warnings; |
73 | no warnings; |
|
|
74 | use strict 'vars'; |
66 | |
75 | |
67 | use base 'Exporter'; |
76 | use base 'Exporter'; |
68 | |
77 | |
69 | use Fcntl (); |
|
|
70 | |
|
|
71 | BEGIN { |
78 | BEGIN { |
72 | $VERSION = '1.72'; |
79 | our $VERSION = '2.0'; |
73 | |
80 | |
74 | @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 |
75 | 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 |
76 | aio_fsync aio_fdatasync aio_readahead); |
83 | aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move |
|
|
84 | aio_group aio_nop); |
77 | @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel |
85 | our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); |
78 | max_outstanding nreqs); |
86 | |
|
|
87 | @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; |
79 | |
88 | |
80 | require XSLoader; |
89 | require XSLoader; |
81 | XSLoader::load IO::AIO, $VERSION; |
90 | XSLoader::load ("IO::AIO", $VERSION); |
82 | } |
91 | } |
83 | |
92 | |
84 | =head1 FUNCTIONS |
93 | =head1 FUNCTIONS |
85 | |
94 | |
86 | =head2 AIO FUNCTIONS |
95 | =head2 AIO FUNCTIONS |
… | |
… | |
93 | perl, which usually delivers "false") as it's sole argument when the given |
102 | perl, which usually delivers "false") as it's sole argument when the given |
94 | syscall has been executed asynchronously. |
103 | syscall has been executed asynchronously. |
95 | |
104 | |
96 | All functions expecting a filehandle keep a copy of the filehandle |
105 | All functions expecting a filehandle keep a copy of the filehandle |
97 | internally until the request has finished. |
106 | internally until the request has finished. |
|
|
107 | |
|
|
108 | All requests return objects of type L<IO::AIO::REQ> that allow further |
|
|
109 | manipulation of those requests while they are in-flight. |
98 | |
110 | |
99 | The pathnames you pass to these routines I<must> be absolute and |
111 | The pathnames you pass to these routines I<must> be absolute and |
100 | encoded in byte form. The reason for the former is that at the time the |
112 | encoded in byte form. The reason for the former is that at the time the |
101 | request is being executed, the current working directory could have |
113 | request is being executed, the current working directory could have |
102 | changed. Alternatively, you can make sure that you never change the |
114 | changed. Alternatively, you can make sure that you never change the |
… | |
… | |
168 | aio_read $fh, 7, 15, $buffer, 0, sub { |
180 | aio_read $fh, 7, 15, $buffer, 0, sub { |
169 | $_[0] > 0 or die "read error: $!"; |
181 | $_[0] > 0 or die "read error: $!"; |
170 | print "read $_[0] bytes: <$buffer>\n"; |
182 | print "read $_[0] bytes: <$buffer>\n"; |
171 | }; |
183 | }; |
172 | |
184 | |
|
|
185 | =item aio_move $srcpath, $dstpath, $callback->($status) |
|
|
186 | |
|
|
187 | [EXPERIMENTAL due to internal aio_group use] |
|
|
188 | |
|
|
189 | Try to move the I<file> (directories not supported as either source or |
|
|
190 | destination) from C<$srcpath> to C<$dstpath> and call the callback with |
|
|
191 | the C<0> (error) or C<-1> ok. |
|
|
192 | |
|
|
193 | This is a composite request that tries to rename(2) the file first. If |
|
|
194 | rename files with C<EXDEV>, it creates the destination file with mode 0200 |
|
|
195 | and copies the contents of the source file into it using C<aio_sendfile>, |
|
|
196 | followed by restoring atime, mtime, access mode and uid/gid, in that |
|
|
197 | order, and unlinking the C<$srcpath>. |
|
|
198 | |
|
|
199 | If an error occurs, the partial destination file will be unlinked, if |
|
|
200 | possible, except when setting atime, mtime, access mode and uid/gid, where |
|
|
201 | errors are being ignored. |
|
|
202 | |
|
|
203 | =cut |
|
|
204 | |
|
|
205 | sub aio_move($$$) { |
|
|
206 | my ($src, $dst, $cb) = @_; |
|
|
207 | |
|
|
208 | my $grp = aio_group $cb; |
|
|
209 | |
|
|
210 | add $grp aio_rename $src, $dst, sub { |
|
|
211 | if ($_[0] && $! == EXDEV) { |
|
|
212 | add $grp aio_open $src, O_RDONLY, 0, sub { |
|
|
213 | if (my $src_fh = $_[0]) { |
|
|
214 | my @stat = stat $src_fh; |
|
|
215 | |
|
|
216 | add $grp aio_open $dst, O_WRONLY, 0200, sub { |
|
|
217 | if (my $dst_fh = $_[0]) { |
|
|
218 | add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub { |
|
|
219 | close $src_fh; |
|
|
220 | |
|
|
221 | if ($_[0] == $stat[7]) { |
|
|
222 | utime $stat[8], $stat[9], $dst; |
|
|
223 | chmod $stat[2] & 07777, $dst_fh; |
|
|
224 | chown $stat[4], $stat[5], $dst_fh; |
|
|
225 | close $dst_fh; |
|
|
226 | |
|
|
227 | add $grp aio_unlink $src, sub { |
|
|
228 | $grp->result ($_[0]); |
|
|
229 | }; |
|
|
230 | } else { |
|
|
231 | my $errno = $!; |
|
|
232 | add $grp aio_unlink $dst, sub { |
|
|
233 | $! = $errno; |
|
|
234 | $grp->result (-1); |
|
|
235 | }; |
|
|
236 | } |
|
|
237 | }; |
|
|
238 | } else { |
|
|
239 | $grp->result (-1); |
|
|
240 | } |
|
|
241 | }, |
|
|
242 | |
|
|
243 | } else { |
|
|
244 | $grp->result (-1); |
|
|
245 | } |
|
|
246 | }; |
|
|
247 | } else { |
|
|
248 | $grp->result ($_[0]); |
|
|
249 | } |
|
|
250 | }; |
|
|
251 | |
|
|
252 | $grp |
|
|
253 | } |
|
|
254 | |
173 | =item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
255 | =item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
174 | |
256 | |
175 | Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts |
257 | Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts |
176 | reading at byte offset C<$in_offset>, and starts writing at the current |
258 | reading at byte offset C<$in_offset>, and starts writing at the current |
177 | file offset of C<$out_fh>. Because of that, it is not safe to issue more |
259 | file offset of C<$out_fh>. Because of that, it is not safe to issue more |
… | |
… | |
232 | =item aio_unlink $pathname, $callback->($status) |
314 | =item aio_unlink $pathname, $callback->($status) |
233 | |
315 | |
234 | Asynchronously unlink (delete) a file and call the callback with the |
316 | Asynchronously unlink (delete) a file and call the callback with the |
235 | result code. |
317 | result code. |
236 | |
318 | |
|
|
319 | =item aio_link $srcpath, $dstpath, $callback->($status) |
|
|
320 | |
|
|
321 | Asynchronously create a new link to the existing object at C<$srcpath> at |
|
|
322 | the path C<$dstpath> and call the callback with the result code. |
|
|
323 | |
|
|
324 | =item aio_symlink $srcpath, $dstpath, $callback->($status) |
|
|
325 | |
|
|
326 | Asynchronously create a new symbolic link to the existing object at C<$srcpath> at |
|
|
327 | the path C<$dstpath> and call the callback with the result code. |
|
|
328 | |
|
|
329 | =item aio_rename $srcpath, $dstpath, $callback->($status) |
|
|
330 | |
|
|
331 | Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as |
|
|
332 | rename(2) and call the callback with the result code. |
|
|
333 | |
237 | =item aio_rmdir $pathname, $callback->($status) |
334 | =item aio_rmdir $pathname, $callback->($status) |
238 | |
335 | |
239 | Asynchronously rmdir (delete) a directory and call the callback with the |
336 | Asynchronously rmdir (delete) a directory and call the callback with the |
240 | result code. |
337 | result code. |
241 | |
338 | |
… | |
… | |
248 | 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 |
249 | with the filenames. |
346 | with the filenames. |
250 | |
347 | |
251 | =item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
348 | =item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
252 | |
349 | |
|
|
350 | [EXPERIMENTAL due to internal aio_group use] |
|
|
351 | |
253 | Scans a directory (similar to C<aio_readdir>) and tries to separate the |
352 | Scans a directory (similar to C<aio_readdir>) but additionally tries to |
254 | entries of directory C<$path> into two sets of names, ones you can recurse |
353 | separate the entries of directory C<$path> into two sets of names, ones |
255 | into (directories), and ones you cannot recurse into (everything else). |
354 | you can recurse into (directories or links to them), and ones you cannot |
|
|
355 | recurse into (everything else). |
256 | |
356 | |
257 | C<aio_scandir> is a composite request that consists of many |
357 | C<aio_scandir> is a composite request that creates of many sub requests_ |
258 | aio-primitives. C<$maxreq> specifies the maximum number of outstanding |
358 | C<$maxreq> specifies the maximum number of outstanding aio requests that |
259 | aio 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 |
260 | suitable default will be chosen (currently 8). |
360 | will be chosen (currently 6). |
261 | |
361 | |
262 | On error, the callback is called without arguments, otherwise it receives |
362 | On error, the callback is called without arguments, otherwise it receives |
263 | two array-refs with path-relative entry names. |
363 | two array-refs with path-relative entry names. |
264 | |
364 | |
265 | Example: |
365 | Example: |
… | |
… | |
273 | Implementation notes. |
373 | Implementation notes. |
274 | |
374 | |
275 | The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. |
375 | The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. |
276 | |
376 | |
277 | After reading the directory, the modification time, size etc. of the |
377 | After reading the directory, the modification time, size etc. of the |
278 | directory before and after the readdir is checked, and if they match, the |
378 | directory before and after the readdir is checked, and if they match (and |
279 | link count will be used to decide how many entries are directories (if |
379 | isn't the current time), the link count will be used to decide how many |
280 | >= 2). Otherwise, no knowledge of the number of subdirectories will be |
380 | entries are directories (if >= 2). Otherwise, no knowledge of the number |
281 | assumed. |
381 | of subdirectories will be assumed. |
282 | |
382 | |
283 | Then entires will be sorted into likely directories (everything without a |
383 | Then entries will be sorted into likely directories (everything without |
284 | non-initial dot) and likely non-directories (everything else). Then every |
384 | a non-initial dot currently) and likely non-directories (everything |
285 | entry + C</.> will be C<stat>'ed, likely directories first. This is often |
385 | else). Then every entry plus an appended C</.> will be C<stat>'ed, |
|
|
386 | likely directories first. If that succeeds, it assumes that the entry |
|
|
387 | is a directory or a symlink to directory (which will be checked |
|
|
388 | seperately). This is often faster than stat'ing the entry itself because |
286 | faster because filesystems might detect the type of the entry without |
389 | filesystems might detect the type of the entry without reading the inode |
287 | reading the inode data (e.g. ext2fs filetype feature). If that succeeds, |
390 | data (e.g. ext2fs filetype feature). |
288 | it assumes that the entry is a directory or a symlink to directory (which |
|
|
289 | will be checked seperately). |
|
|
290 | |
391 | |
291 | If the known number of directories has been reached, the rest of the |
392 | If the known number of directories (link count - 2) has been reached, the |
292 | entries is assumed to be non-directories. |
393 | rest of the entries is assumed to be non-directories. |
|
|
394 | |
|
|
395 | This only works with certainty on POSIX (= UNIX) filesystems, which |
|
|
396 | fortunately are the vast majority of filesystems around. |
|
|
397 | |
|
|
398 | It will also likely work on non-POSIX filesystems with reduced efficiency |
|
|
399 | as those tend to return 0 or 1 as link counts, which disables the |
|
|
400 | directory counting heuristic. |
293 | |
401 | |
294 | =cut |
402 | =cut |
295 | |
403 | |
296 | sub aio_scandir($$$) { |
404 | sub aio_scandir($$$) { |
297 | my ($path, $maxreq, $cb) = @_; |
405 | my ($path, $maxreq, $cb) = @_; |
298 | |
406 | |
|
|
407 | my $grp = aio_group $cb; |
|
|
408 | |
299 | $maxreq = 8 if $maxreq <= 0; |
409 | $maxreq = 6 if $maxreq <= 0; |
300 | |
410 | |
301 | # stat once |
411 | # stat once |
302 | aio_stat $path, sub { |
412 | add $grp aio_stat $path, sub { |
303 | return $cb->() if $_[0]; |
413 | return $grp->result () if $_[0]; |
|
|
414 | my $now = time; |
304 | my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
415 | my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
305 | |
416 | |
306 | # read the directory entries |
417 | # read the directory entries |
307 | aio_readdir $path, sub { |
418 | add $grp aio_readdir $path, sub { |
308 | my $entries = shift |
419 | my $entries = shift |
309 | or return $cb->(); |
420 | or return $grp->result (); |
310 | |
421 | |
311 | # stat the dir another time |
422 | # stat the dir another time |
312 | aio_stat $path, sub { |
423 | add $grp aio_stat $path, sub { |
313 | my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
424 | my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
314 | |
425 | |
315 | my $ndirs; |
426 | my $ndirs; |
316 | |
427 | |
317 | # take the slow route if anything looks fishy |
428 | # take the slow route if anything looks fishy |
318 | if ($hash1 ne $hash2) { |
429 | if ($hash1 ne $hash2 or (stat _)[9] == $now) { |
319 | $ndirs = -1; |
430 | $ndirs = -1; |
320 | } else { |
431 | } else { |
321 | # if nlink == 2, we are finished |
432 | # if nlink == 2, we are finished |
322 | # on non-posix-fs's, we rely on nlink < 2 |
433 | # on non-posix-fs's, we rely on nlink < 2 |
323 | $ndirs = (stat _)[3] - 2 |
434 | $ndirs = (stat _)[3] - 2 |
324 | or return $cb->([], $entries); |
435 | or return $grp->result ([], $entries); |
325 | } |
436 | } |
326 | |
437 | |
327 | # sort into likely dirs and likely nondirs |
438 | # sort into likely dirs and likely nondirs |
328 | # dirs == files without ".", short entries first |
439 | # dirs == files without ".", short entries first |
329 | $entries = [map $_->[0], |
440 | $entries = [map $_->[0], |
… | |
… | |
334 | my (@dirs, @nondirs); |
445 | my (@dirs, @nondirs); |
335 | |
446 | |
336 | my ($statcb, $schedcb); |
447 | my ($statcb, $schedcb); |
337 | my $nreq = 0; |
448 | my $nreq = 0; |
338 | |
449 | |
|
|
450 | my $statgrp = add $grp aio_group; |
|
|
451 | |
339 | $schedcb = sub { |
452 | $schedcb = sub { |
340 | if (@$entries) { |
453 | if (@$entries) { |
341 | if ($nreq < $maxreq) { |
454 | if ($nreq < $maxreq) { |
342 | my $ent = pop @$entries; |
455 | my $ent = pop @$entries; |
343 | $nreq++; |
456 | $nreq++; |
344 | aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; |
457 | add $statgrp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; |
345 | } |
458 | } |
346 | } elsif (!$nreq) { |
459 | } elsif (!$nreq) { |
347 | # finished |
460 | # finished |
|
|
461 | $statgrp->cancel; |
348 | undef $statcb; |
462 | undef $statcb; |
349 | undef $schedcb; |
463 | undef $schedcb; |
350 | $cb->(\@dirs, \@nondirs) if $cb; |
464 | $grp->result (\@dirs, \@nondirs); |
351 | undef $cb; |
|
|
352 | } |
465 | } |
353 | }; |
466 | }; |
354 | $statcb = sub { |
467 | $statcb = sub { |
355 | my ($status, $entry) = @_; |
468 | my ($status, $entry) = @_; |
356 | |
469 | |
… | |
… | |
358 | $nreq--; |
471 | $nreq--; |
359 | push @nondirs, $entry; |
472 | push @nondirs, $entry; |
360 | &$schedcb; |
473 | &$schedcb; |
361 | } else { |
474 | } else { |
362 | # need to check for real directory |
475 | # need to check for real directory |
363 | aio_lstat "$path/$entry", sub { |
476 | add $grp aio_lstat "$path/$entry", sub { |
364 | $nreq--; |
477 | $nreq--; |
365 | |
478 | |
366 | if (-d _) { |
479 | if (-d _) { |
367 | push @dirs, $entry; |
480 | push @dirs, $entry; |
368 | |
481 | |
… | |
… | |
381 | |
494 | |
382 | &$schedcb while @$entries && $nreq < $maxreq; |
495 | &$schedcb while @$entries && $nreq < $maxreq; |
383 | }; |
496 | }; |
384 | }; |
497 | }; |
385 | }; |
498 | }; |
|
|
499 | |
|
|
500 | $grp |
386 | } |
501 | } |
387 | |
502 | |
388 | =item aio_fsync $fh, $callback->($status) |
503 | =item aio_fsync $fh, $callback->($status) |
389 | |
504 | |
390 | Asynchronously call fsync on the given filehandle and call the callback |
505 | Asynchronously call fsync on the given filehandle and call the callback |
… | |
… | |
395 | Asynchronously call fdatasync on the given filehandle and call the |
510 | Asynchronously call fdatasync on the given filehandle and call the |
396 | callback with the fdatasync result code. |
511 | callback with the fdatasync result code. |
397 | |
512 | |
398 | 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 |
399 | detected, it will be emulated by calling C<fsync> instead. |
514 | detected, it will be emulated by calling C<fsync> instead. |
|
|
515 | |
|
|
516 | =item aio_group $callback->(...) |
|
|
517 | |
|
|
518 | [EXPERIMENTAL] |
|
|
519 | |
|
|
520 | This is a very special aio request: Instead of doing something, it is a |
|
|
521 | container for other aio requests, which is useful if you want to bundle |
|
|
522 | many requests into a single, composite, request. |
|
|
523 | |
|
|
524 | Returns an object of class L<IO::AIO::GRP>. See its documentation below |
|
|
525 | for more info. |
|
|
526 | |
|
|
527 | Example: |
|
|
528 | |
|
|
529 | my $grp = aio_group sub { |
|
|
530 | print "all stats done\n"; |
|
|
531 | }; |
|
|
532 | |
|
|
533 | add $grp |
|
|
534 | (aio_stat ...), |
|
|
535 | (aio_stat ...), |
|
|
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. |
|
|
550 | |
|
|
551 | =item IO::AIO::aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED* |
|
|
552 | |
|
|
553 | Mainly used for debugging and benchmarking, this aio request puts one of |
|
|
554 | the request workers to sleep for the given time. |
|
|
555 | |
|
|
556 | While it is theoretically handy to have simple I/O scheduling requests |
|
|
557 | like sleep and file handle readable/writable, the overhead this creates |
|
|
558 | is immense, so do not use this function except to put your application |
|
|
559 | under artificial I/O pressure. |
|
|
560 | |
|
|
561 | =back |
|
|
562 | |
|
|
563 | =head2 IO::AIO::REQ CLASS |
|
|
564 | |
|
|
565 | All non-aggregate C<aio_*> functions return an object of this class when |
|
|
566 | called in non-void context. |
|
|
567 | |
|
|
568 | A request always moves through the following five states in its lifetime, |
|
|
569 | in order: B<ready> (request has been created, but has not been executed |
|
|
570 | yet), B<execute> (request is currently being executed), B<pending> |
|
|
571 | (request has been executed but callback has not been called yet), |
|
|
572 | B<result> (results are being processed synchronously, includes calling the |
|
|
573 | callback) and B<done> (request has reached the end of its lifetime and |
|
|
574 | holds no resources anymore). |
|
|
575 | |
|
|
576 | =over 4 |
|
|
577 | |
|
|
578 | =item $req->cancel |
|
|
579 | |
|
|
580 | Cancels the request, if possible. Has the effect of skipping execution |
|
|
581 | when entering the B<execute> state and skipping calling the callback when |
|
|
582 | entering the the B<result> state, but will leave the request otherwise |
|
|
583 | untouched. That means that requests that currently execute will not be |
|
|
584 | stopped and resources held by the request will not be freed prematurely. |
|
|
585 | |
|
|
586 | =back |
|
|
587 | |
|
|
588 | =head2 IO::AIO::GRP CLASS |
|
|
589 | |
|
|
590 | This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to |
|
|
591 | objects of this class, too. |
|
|
592 | |
|
|
593 | A IO::AIO::GRP object is a special request that can contain multiple other |
|
|
594 | aio requests. |
|
|
595 | |
|
|
596 | You create one by calling the C<aio_group> constructing function with a |
|
|
597 | callback that will be called when all contained requests have entered the |
|
|
598 | C<done> state: |
|
|
599 | |
|
|
600 | my $grp = aio_group sub { |
|
|
601 | print "all requests are done\n"; |
|
|
602 | }; |
|
|
603 | |
|
|
604 | You add requests by calling the C<add> method with one or more |
|
|
605 | C<IO::AIO::REQ> objects: |
|
|
606 | |
|
|
607 | $grp->add (aio_unlink "..."); |
|
|
608 | |
|
|
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 | }; |
|
|
617 | |
|
|
618 | This makes it very easy to create composite requests (see the source of |
|
|
619 | C<aio_move> for an application) that work and feel like simple requests. |
|
|
620 | |
|
|
621 | =over 4 |
|
|
622 | |
|
|
623 | =item * The IO::AIO::GRP objects will be cleaned up during calls to |
|
|
624 | C<IO::AIO::poll_cb>, just like any other request. |
|
|
625 | |
|
|
626 | =item * They can be canceled like any other request. Canceling will cancel not |
|
|
627 | only the request itself, but also all requests it contains. |
|
|
628 | |
|
|
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 |
|
|
639 | |
|
|
640 | Their lifetime, simplified, looks like this: when they are empty, they |
|
|
641 | will finish very quickly. If they contain only requests that are in the |
|
|
642 | C<done> state, they will also finish. Otherwise they will continue to |
|
|
643 | exist. |
|
|
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 | |
|
|
650 | =over 4 |
|
|
651 | |
|
|
652 | =item $grp->add (...) |
|
|
653 | |
|
|
654 | =item add $grp ... |
|
|
655 | |
|
|
656 | Add one or more requests to the group. Any type of L<IO::AIO::REQ> can |
|
|
657 | be added, including other groups, as long as you do not create circular |
|
|
658 | dependencies. |
|
|
659 | |
|
|
660 | Returns all its arguments. |
|
|
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. |
400 | |
711 | |
401 | =back |
712 | =back |
402 | |
713 | |
403 | =head2 SUPPORT FUNCTIONS |
714 | =head2 SUPPORT FUNCTIONS |
404 | |
715 | |
… | |
… | |
462 | IO::AIO::poll_wait, IO::AIO::poll_cb |
773 | IO::AIO::poll_wait, IO::AIO::poll_cb |
463 | if IO::AIO::nreqs; |
774 | if IO::AIO::nreqs; |
464 | |
775 | |
465 | =item IO::AIO::min_parallel $nthreads |
776 | =item IO::AIO::min_parallel $nthreads |
466 | |
777 | |
467 | Set the minimum number of AIO threads to C<$nthreads>. The current default |
778 | Set the minimum number of AIO threads to C<$nthreads>. The current |
468 | 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 |
469 | (the number of outstanding operations, however, is unlimited). |
780 | concurrently at any one time (the number of outstanding requests, |
|
|
781 | however, is unlimited). |
470 | |
782 | |
471 | 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 |
472 | no free thread exists. |
784 | no free thread exists. |
473 | |
785 | |
474 | 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 |
475 | kernel versions will scale negatively with the number of threads (higher |
787 | Linux kernel versions will scale negatively with the number of threads |
476 | parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 |
788 | (higher parallelity => MUCH higher latency). With current Linux 2.6 |
477 | threads should be fine. |
789 | versions, 4-32 threads should be fine. |
478 | |
790 | |
479 | 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 |
480 | 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. |
481 | |
793 | |
482 | =item IO::AIO::max_parallel $nthreads |
794 | =item IO::AIO::max_parallel $nthreads |
… | |
… | |
493 | |
805 | |
494 | Under normal circumstances you don't need to call this function. |
806 | Under normal circumstances you don't need to call this function. |
495 | |
807 | |
496 | =item $oldnreqs = IO::AIO::max_outstanding $nreqs |
808 | =item $oldnreqs = IO::AIO::max_outstanding $nreqs |
497 | |
809 | |
|
|
810 | [DEPRECATED] |
|
|
811 | |
498 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
812 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
499 | try to queue up more than this number of requests, the caller will block until |
813 | try to queue up more than this number of requests, the caller will block until |
500 | some requests have been handled. |
814 | some requests have been handled. |
501 | |
815 | |
502 | The default is very large, so normally there is no practical limit. If you |
816 | The default is very large, so normally there is no practical limit. If you |
503 | queue up many requests in a loop it often improves speed if you set |
817 | queue up many requests in a loop it often improves speed if you set |
504 | this to a relatively low number, such as C<100>. |
818 | this to a relatively low number, such as C<100>. |
|
|
819 | |
|
|
820 | This function does not work well together with C<aio_group>'s, and their |
|
|
821 | feeder interface is better suited to limiting concurrency, so do not use |
|
|
822 | this function. |
505 | |
823 | |
506 | Under normal circumstances you don't need to call this function. |
824 | Under normal circumstances you don't need to call this function. |
507 | |
825 | |
508 | =back |
826 | =back |
509 | |
827 | |
… | |
… | |
523 | or return undef; |
841 | or return undef; |
524 | |
842 | |
525 | *$sym |
843 | *$sym |
526 | } |
844 | } |
527 | |
845 | |
528 | min_parallel 4; |
846 | min_parallel 8; |
529 | |
847 | |
530 | END { |
848 | END { |
531 | max_parallel 0; |
849 | max_parallel 0; |
532 | } |
850 | } |
533 | |
851 | |
534 | 1; |
852 | 1; |
535 | |
853 | |
536 | =head2 FORK BEHAVIOUR |
854 | =head2 FORK BEHAVIOUR |
|
|
855 | |
|
|
856 | This module should do "the right thing" when the process using it forks: |
537 | |
857 | |
538 | Before the fork, IO::AIO enters a quiescent state where no requests |
858 | Before the fork, IO::AIO enters a quiescent state where no requests |
539 | can be added in other threads and no results will be processed. After |
859 | can be added in other threads and no results will be processed. After |
540 | the fork the parent simply leaves the quiescent state and continues |
860 | the fork the parent simply leaves the quiescent state and continues |
541 | request/result processing, while the child clears the request/result |
861 | request/result processing, while the child clears the request/result |
542 | queue (so the requests started before the fork will only be handled in |
862 | queue (so the requests started before the fork will only be handled in |
543 | the parent). Threats will be started on demand until the limit ste in the |
863 | the parent). Threads will be started on demand until the limit ste in the |
544 | parent process has been reached again. |
864 | parent process has been reached again. |
545 | |
865 | |
|
|
866 | In short: the parent will, after a short pause, continue as if fork had |
|
|
867 | not been called, while the child will act as if IO::AIO has not been used |
|
|
868 | yet. |
|
|
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 | |
546 | =head1 SEE ALSO |
883 | =head1 SEE ALSO |
547 | |
884 | |
548 | L<Coro>, L<Linux::AIO>. |
885 | L<Coro>, L<Linux::AIO> (obsolete). |
549 | |
886 | |
550 | =head1 AUTHOR |
887 | =head1 AUTHOR |
551 | |
888 | |
552 | Marc Lehmann <schmorp@schmorp.de> |
889 | Marc Lehmann <schmorp@schmorp.de> |
553 | http://home.schmorp.de/ |
890 | http://home.schmorp.de/ |