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 1.76 by root, Wed Oct 25 08:20:14 2006 UTC vs.
Revision 1.120 by root, Sun Dec 2 21:51:36 2007 UTC

… …
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use IO::AIO; 7 use IO::AIO;
8 8
9 aio_open "/etc/passwd", O_RDONLY, 0, sub { 9 aio_open "/etc/passwd", O_RDONLY, 0, sub {
10 my ($fh) = @_; 10 my $fh = shift
11 or die "/etc/passwd: $!";
11 ... 12 ...
12 }; 13 };
13 14
14 aio_unlink "/tmp/file", sub { }; 15 aio_unlink "/tmp/file", sub { };
15 16
… …
25 $req->cancel; # cancel request if still in queue 26 $req->cancel; # cancel request if still in queue
26 27
27 my $grp = aio_group sub { print "all stats done\n" }; 28 my $grp = aio_group sub { print "all stats done\n" };
28 add $grp aio_stat "..." for ...; 29 add $grp aio_stat "..." for ...;
29 30
30 # AnyEvent integration 31 # AnyEvent integration (EV, Event, Glib, Tk, urxvt, pureperl...)
31 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; 32 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
32 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); 33 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
34
35 # EV integration
36 my $w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
33 37
34 # Event integration 38 # Event integration
35 Event->io (fd => IO::AIO::poll_fileno, 39 Event->io (fd => IO::AIO::poll_fileno,
36 poll => 'r', 40 poll => 'r',
37 cb => \&IO::AIO::poll_cb); 41 cb => \&IO::AIO::poll_cb);
… …
50 54
51=head1 DESCRIPTION 55=head1 DESCRIPTION
52 56
53This module implements asynchronous I/O using whatever means your 57This module implements asynchronous I/O using whatever means your
54operating system supports. 58operating system supports.
59
60Asynchronous means that operations that can normally block your program
61(e.g. reading from disk) will be done asynchronously: the operation
62will still block, but you can do something else in the meantime. This
63is extremely useful for programs that need to stay interactive even
64when doing heavy I/O (GUI programs, high performance network servers
65etc.), but can also be used to easily do operations in parallel that are
66normally done sequentially, e.g. stat'ing many files, which is much faster
67on a RAID volume or over NFS when you do a number of stat operations
68concurrently.
69
70While most of this works on all types of file descriptors (for
71example sockets), using these functions on file descriptors that
72support nonblocking operation (again, sockets, pipes etc.) is very
73inefficient. Use an event loop for that (such as the L<Event|Event>
74module): IO::AIO will naturally fit into such an event loop itself.
55 75
56In this version, a number of threads are started that execute your 76In this version, a number of threads are started that execute your
57requests and signal their completion. You don't need thread support 77requests and signal their completion. You don't need thread support
58in perl, and the threads created by this module will not be visible 78in perl, and the threads created by this module will not be visible
59to perl. In the future, this module might make use of the native aio 79to perl. In the future, this module might make use of the native aio
60functions available on many operating systems. However, they are often 80functions available on many operating systems. However, they are often
61not well-supported or restricted (Linux doesn't allow them on normal 81not well-supported or restricted (GNU/Linux doesn't allow them on normal
62files currently, for example), and they would only support aio_read and 82files currently, for example), and they would only support aio_read and
63aio_write, so the remaining functionality would have to be implemented 83aio_write, so the remaining functionality would have to be implemented
64using threads anyway. 84using threads anyway.
65 85
66Although the module will work with in the presence of other (Perl-) 86Although the module will work in the presence of other (Perl-) threads,
67threads, it is currently not reentrant in any way, so use appropriate 87it is currently not reentrant in any way, so use appropriate locking
68locking yourself, always call C<poll_cb> from within the same thread, or 88yourself, always call C<poll_cb> from within the same thread, or never
69never call C<poll_cb> (or other C<aio_> functions) recursively. 89call C<poll_cb> (or other C<aio_> functions) recursively.
90
91=head2 EXAMPLE
92
93This is a simple example that uses the Event module and loads
94F</etc/passwd> asynchronously:
95
96 use Fcntl;
97 use Event;
98 use IO::AIO;
99
100 # register the IO::AIO callback with Event
101 Event->io (fd => IO::AIO::poll_fileno,
102 poll => 'r',
103 cb => \&IO::AIO::poll_cb);
104
105 # queue the request to open /etc/passwd
106 aio_open "/etc/passwd", O_RDONLY, 0, sub {
107 my $fh = shift
108 or die "error while opening: $!";
109
110 # stat'ing filehandles is generally non-blocking
111 my $size = -s $fh;
112
113 # queue a request to read the file
114 my $contents;
115 aio_read $fh, 0, $size, $contents, 0, sub {
116 $_[0] == $size
117 or die "short read: $!";
118
119 close $fh;
120
121 # file contents now in $contents
122 print $contents;
123
124 # exit event loop and program
125 Event::unloop;
126 };
127 };
128
129 # possibly queue up other requests, or open GUI windows,
130 # check for sockets etc. etc.
131
132 # process events as long as there are some:
133 Event::loop;
70 134
71=head1 REQUEST ANATOMY AND LIFETIME 135=head1 REQUEST ANATOMY AND LIFETIME
72 136
73Every C<aio_*> function creates a request. which is a C data structure not 137Every C<aio_*> function creates a request. which is a C data structure not
74directly visible to Perl. 138directly visible to Perl.
… …
116Request has reached the end of its lifetime and holds no resources anymore 180Request has reached the end of its lifetime and holds no resources anymore
117(except possibly for the Perl object, but its connection to the actual 181(except possibly for the Perl object, but its connection to the actual
118aio request is severed and calling its methods will either do nothing or 182aio request is severed and calling its methods will either do nothing or
119result in a runtime error). 183result in a runtime error).
120 184
185=back
186
121=cut 187=cut
122 188
123package IO::AIO; 189package IO::AIO;
190
191use Carp ();
124 192
125no warnings; 193no warnings;
126use strict 'vars'; 194use strict 'vars';
127 195
128use base 'Exporter'; 196use base 'Exporter';
129 197
130BEGIN { 198BEGIN {
131 our $VERSION = '2.0'; 199 our $VERSION = '2.6';
132 200
133 our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat 201 our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close
134 aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink 202 aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir
135 aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move 203 aio_scandir aio_symlink aio_readlink aio_sync aio_fsync
136 aio_group aio_nop); 204 aio_fdatasync aio_pathsync aio_readahead
205 aio_rename aio_link aio_move aio_copy aio_group
206 aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
207 aio_chmod aio_utime aio_truncate);
208
137 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice)); 209 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice aio_block));
138 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush 210 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
139 min_parallel max_parallel max_outstanding nreqs); 211 min_parallel max_parallel max_idle
212 nreqs nready npending nthreads
213 max_poll_time max_poll_reqs);
140 214
141 @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; 215 @IO::AIO::GRP::ISA = 'IO::AIO::REQ';
142 216
143 require XSLoader; 217 require XSLoader;
144 XSLoader::load ("IO::AIO", $VERSION); 218 XSLoader::load ("IO::AIO", $VERSION);
145} 219}
146 220
147=head1 FUNCTIONS 221=head1 FUNCTIONS
148 222
149=head2 AIO FUNCTIONS 223=head2 AIO REQUEST FUNCTIONS
150 224
151All the C<aio_*> calls are more or less thin wrappers around the syscall 225All the C<aio_*> calls are more or less thin wrappers around the syscall
152with the same name (sans C<aio_>). The arguments are similar or identical, 226with the same name (sans C<aio_>). The arguments are similar or identical,
153and they all accept an additional (and optional) C<$callback> argument 227and they all accept an additional (and optional) C<$callback> argument
154which must be a code reference. This code reference will get called with 228which must be a code reference. This code reference will get called with
… …
157syscall has been executed asynchronously. 231syscall has been executed asynchronously.
158 232
159All functions expecting a filehandle keep a copy of the filehandle 233All functions expecting a filehandle keep a copy of the filehandle
160internally until the request has finished. 234internally until the request has finished.
161 235
162All requests return objects of type L<IO::AIO::REQ> that allow further 236All functions return request objects of type L<IO::AIO::REQ> that allow
163manipulation of those requests while they are in-flight. 237further manipulation of those requests while they are in-flight.
164 238
165The pathnames you pass to these routines I<must> be absolute and 239The pathnames you pass to these routines I<must> be absolute and
166encoded in byte form. The reason for the former is that at the time the 240encoded as octets. The reason for the former is that at the time the
167request is being executed, the current working directory could have 241request is being executed, the current working directory could have
168changed. Alternatively, you can make sure that you never change the 242changed. Alternatively, you can make sure that you never change the
169current working directory. 243current working directory anywhere in the program and then use relative
244paths.
170 245
171To encode pathnames to byte form, either make sure you either: a) 246To encode pathnames as octets, either make sure you either: a) always pass
172always pass in filenames you got from outside (command line, readdir 247in filenames you got from outside (command line, readdir etc.) without
173etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode 248tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
174your pathnames to the locale (or other) encoding in effect in the user 249your pathnames to the locale (or other) encoding in effect in the user
175environment, d) use Glib::filename_from_unicode on unicode filenames or e) 250environment, d) use Glib::filename_from_unicode on unicode filenames or e)
176use something else. 251use something else to ensure your scalar has the correct contents.
252
253This works, btw. independent of the internal UTF-8 bit, which IO::AIO
254handles correctly wether it is set or not.
177 255
178=over 4 256=over 4
179 257
180=item aioreq_pri $pri 258=item $prev_pri = aioreq_pri [$pri]
181 259
182Sets the priority for the next aio request. The default priority 260Returns the priority value that would be used for the next request and, if
261C<$pri> is given, sets the priority for the next aio request.
262
183is C<0>, the minimum and maximum priorities are C<-4> and C<4>, 263The default priority is C<0>, the minimum and maximum priorities are C<-4>
184respectively. Requests with higher priority will be serviced first. 264and C<4>, respectively. Requests with higher priority will be serviced
265first.
185 266
186The priority will be reset to C<0> after each call to one of the C<aio_> 267The priority will be reset to C<0> after each call to one of the C<aio_*>
187functions. 268functions.
188 269
189Example: open a file with low priority, then read something from it with 270Example: open a file with low priority, then read something from it with
190higher priority so the read request is serviced before other low priority 271higher priority so the read request is serviced before other low priority
191open requests (potentially spamming the cache): 272open requests (potentially spamming the cache):
… …
198 aio_read $_[0], ..., sub { 279 aio_read $_[0], ..., sub {
199 ... 280 ...
200 }; 281 };
201 }; 282 };
202 283
284
203=item aioreq_nice $pri_adjust 285=item aioreq_nice $pri_adjust
204 286
205Similar to C<aioreq_pri>, but subtracts the given value from the current 287Similar to C<aioreq_pri>, but subtracts the given value from the current
206priority, so effects are cumulative. 288priority, so the effect is cumulative.
289
207 290
208=item aio_open $pathname, $flags, $mode, $callback->($fh) 291=item aio_open $pathname, $flags, $mode, $callback->($fh)
209 292
210Asynchronously open or create a file and call the callback with a newly 293Asynchronously open or create a file and call the callback with a newly
211created filehandle for the file. 294created filehandle for the file.
… …
217list. They are the same as used by C<sysopen>. 300list. They are the same as used by C<sysopen>.
218 301
219Likewise, C<$mode> specifies the mode of the newly created file, if it 302Likewise, C<$mode> specifies the mode of the newly created file, if it
220didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>, 303didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
221except that it is mandatory (i.e. use C<0> if you don't create new files, 304except that it is mandatory (i.e. use C<0> if you don't create new files,
222and C<0666> or C<0777> if you do). 305and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
306by the umask in effect then the request is being executed, so better never
307change the umask.
223 308
224Example: 309Example:
225 310
226 aio_open "/etc/passwd", O_RDONLY, 0, sub { 311 aio_open "/etc/passwd", O_RDONLY, 0, sub {
227 if ($_[0]) { 312 if ($_[0]) {
… …
230 } else { 315 } else {
231 die "open failed: $!\n"; 316 die "open failed: $!\n";
232 } 317 }
233 }; 318 };
234 319
320
235=item aio_close $fh, $callback->($status) 321=item aio_close $fh, $callback->($status)
236 322
237Asynchronously close a file and call the callback with the result 323Asynchronously close a file and call the callback with the result
238code. I<WARNING:> although accepted, you should not pass in a perl 324code.
239filehandle here, as perl will likely close the file descriptor another
240time when the filehandle is destroyed. Normally, you can safely call perls
241C<close> or just let filehandles go out of scope.
242 325
243This is supposed to be a bug in the API, so that might change. It's 326Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
244therefore best to avoid this function. 327closing the file descriptor associated with the filehandle itself. Here is
328what aio_close will try:
329
330 1. dup()licate the fd
331 2. asynchronously close() the duplicated fd
332 3. dup()licate the fd once more
333 4. let perl close() the filehandle
334 5. asynchronously close the duplicated fd
335
336The idea is that the first close() flushes stuff to disk that closing an
337fd will flush, so when perl closes the fd, nothing much will need to be
338flushed. The second async. close() will then flush stuff to disk that
339closing the last fd to the file will flush.
340
341Just FYI, SuSv3 has this to say on close:
342
343 All outstanding record locks owned by the process on the file
344 associated with the file descriptor shall be removed.
345
346 If fildes refers to a socket, close() shall cause the socket to be
347 destroyed. ... close() shall block for up to the current linger
348 interval until all data is transmitted.
349 [this actually sounds like a specification bug, but who knows]
350
351And at least Linux additionally actually flushes stuff on every close,
352even when the file itself is still open.
353
354Sounds enourmously inefficient and complicated? Yes... please show me how
355to nuke perl's fd out of existence...
356
357=cut
358
359sub aio_close($;$) {
360 aio_block {
361 my ($fh, $cb) = @_;
362
363 my $pri = aioreq_pri;
364 my $grp = aio_group $cb;
365
366 my $fd = fileno $fh;
367
368 defined $fd or Carp::croak "aio_close called with fd-less filehandle";
369
370 # if the dups fail we will simply get EBADF
371 my $fd2 = _dup $fd;
372 aioreq_pri $pri;
373 add $grp _aio_close $fd2, sub {
374 my $fd2 = _dup $fd;
375 close $fh;
376 aioreq_pri $pri;
377 add $grp _aio_close $fd2, sub {
378 $grp->result ($_[0]);
379 };
380 };
381
382 $grp
383 }
384}
385
245 386
246=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 387=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
247 388
248=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 389=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
249 390
250Reads or writes C<length> bytes from the specified C<fh> and C<offset> 391Reads or writes C<$length> bytes from the specified C<$fh> and C<$offset>
251into the scalar given by C<data> and offset C<dataoffset> and calls the 392into the scalar given by C<$data> and offset C<$dataoffset> and calls the
252callback without the actual number of bytes read (or -1 on error, just 393callback without the actual number of bytes read (or -1 on error, just
253like the syscall). 394like the syscall).
254 395
396If C<$offset> is undefined, then the current file descriptor offset will
397be used (and updated), otherwise the file descriptor offset will not be
398changed by these calls.
399
400If C<$length> is undefined in C<aio_write>, use the remaining length of C<$data>.
401
402If C<$dataoffset> is less than zero, it will be counted from the end of
403C<$data>.
404
255The C<$data> scalar I<MUST NOT> be modified in any way while the request 405The C<$data> scalar I<MUST NOT> be modified in any way while the request
256is outstanding. Modifying it can result in segfaults or WW3 (if the 406is outstanding. Modifying it can result in segfaults or World War III (if
257necessary/optional hardware is installed). 407the necessary/optional hardware is installed).
258 408
259Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at 409Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
260offset C<0> within the scalar: 410offset C<0> within the scalar:
261 411
262 aio_read $fh, 7, 15, $buffer, 0, sub { 412 aio_read $fh, 7, 15, $buffer, 0, sub {
263 $_[0] > 0 or die "read error: $!"; 413 $_[0] > 0 or die "read error: $!";
264 print "read $_[0] bytes: <$buffer>\n"; 414 print "read $_[0] bytes: <$buffer>\n";
265 }; 415 };
266 416
267=item aio_move $srcpath, $dstpath, $callback->($status)
268
269Try to move the I<file> (directories not supported as either source or
270destination) from C<$srcpath> to C<$dstpath> and call the callback with
271the C<0> (error) or C<-1> ok.
272
273This is a composite request that tries to rename(2) the file first. If
274rename files with C<EXDEV>, it creates the destination file with mode 0200
275and copies the contents of the source file into it using C<aio_sendfile>,
276followed by restoring atime, mtime, access mode and uid/gid, in that
277order, and unlinking the C<$srcpath>.
278
279If an error occurs, the partial destination file will be unlinked, if
280possible, except when setting atime, mtime, access mode and uid/gid, where
281errors are being ignored.
282
283=cut
284
285sub aio_move($$$) {
286 my ($src, $dst, $cb) = @_;
287
288 my $grp = aio_group $cb;
289
290 add $grp aio_rename $src, $dst, sub {
291 if ($_[0] && $! == EXDEV) {
292 add $grp aio_open $src, O_RDONLY, 0, sub {
293 if (my $src_fh = $_[0]) {
294 my @stat = stat $src_fh;
295
296 add $grp aio_open $dst, O_WRONLY, 0200, sub {
297 if (my $dst_fh = $_[0]) {
298 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
299 close $src_fh;
300
301 if ($_[0] == $stat[7]) {
302 utime $stat[8], $stat[9], $dst;
303 chmod $stat[2] & 07777, $dst_fh;
304 chown $stat[4], $stat[5], $dst_fh;
305 close $dst_fh;
306
307 add $grp aio_unlink $src, sub {
308 $grp->result ($_[0]);
309 };
310 } else {
311 my $errno = $!;
312 add $grp aio_unlink $dst, sub {
313 $! = $errno;
314 $grp->result (-1);
315 };
316 }
317 };
318 } else {
319 $grp->result (-1);
320 }
321 },
322
323 } else {
324 $grp->result (-1);
325 }
326 };
327 } else {
328 $grp->result ($_[0]);
329 }
330 };
331
332 $grp
333}
334 417
335=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) 418=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
336 419
337Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts 420Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
338reading at byte offset C<$in_offset>, and starts writing at the current 421reading at byte offset C<$in_offset>, and starts writing at the current
… …
352C<$in_fh> than are written, and there is no way to find out how many 435C<$in_fh> than are written, and there is no way to find out how many
353bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only 436bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
354provides the number of bytes written to C<$out_fh>. Only if the result 437provides the number of bytes written to C<$out_fh>. Only if the result
355value equals C<$length> one can assume that C<$length> bytes have been 438value equals C<$length> one can assume that C<$length> bytes have been
356read. 439read.
440
357 441
358=item aio_readahead $fh,$offset,$length, $callback->($retval) 442=item aio_readahead $fh,$offset,$length, $callback->($retval)
359 443
360C<aio_readahead> populates the page cache with data from a file so that 444C<aio_readahead> populates the page cache with data from a file so that
361subsequent reads from that file will not block on disk I/O. The C<$offset> 445subsequent reads from that file will not block on disk I/O. The C<$offset>
… …
367file. The current file offset of the file is left unchanged. 451file. The current file offset of the file is left unchanged.
368 452
369If that syscall doesn't exist (likely if your OS isn't Linux) it will be 453If that syscall doesn't exist (likely if your OS isn't Linux) it will be
370emulated by simply reading the data, which would have a similar effect. 454emulated by simply reading the data, which would have a similar effect.
371 455
456
372=item aio_stat $fh_or_path, $callback->($status) 457=item aio_stat $fh_or_path, $callback->($status)
373 458
374=item aio_lstat $fh, $callback->($status) 459=item aio_lstat $fh, $callback->($status)
375 460
376Works like perl's C<stat> or C<lstat> in void context. The callback will 461Works like perl's C<stat> or C<lstat> in void context. The callback will
… …
389 aio_stat "/etc/passwd", sub { 474 aio_stat "/etc/passwd", sub {
390 $_[0] and die "stat failed: $!"; 475 $_[0] and die "stat failed: $!";
391 print "size is ", -s _, "\n"; 476 print "size is ", -s _, "\n";
392 }; 477 };
393 478
479
480=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
481
482Works like perl's C<utime> function (including the special case of $atime
483and $mtime being undef). Fractional times are supported if the underlying
484syscalls support them.
485
486When called with a pathname, uses utimes(2) if available, otherwise
487utime(2). If called on a file descriptor, uses futimes(2) if available,
488otherwise returns ENOSYS, so this is not portable.
489
490Examples:
491
492 # set atime and mtime to current time (basically touch(1)):
493 aio_utime "path", undef, undef;
494 # set atime to current time and mtime to beginning of the epoch:
495 aio_utime "path", time, undef; # undef==0
496
497
498=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
499
500Works like perl's C<chown> function, except that C<undef> for either $uid
501or $gid is being interpreted as "do not change" (but -1 can also be used).
502
503Examples:
504
505 # same as "chown root path" in the shell:
506 aio_chown "path", 0, -1;
507 # same as above:
508 aio_chown "path", 0, undef;
509
510
511=item aio_truncate $fh_or_path, $offset, $callback->($status)
512
513Works like truncate(2) or ftruncate(2).
514
515
516=item aio_chmod $fh_or_path, $mode, $callback->($status)
517
518Works like perl's C<chmod> function.
519
520
394=item aio_unlink $pathname, $callback->($status) 521=item aio_unlink $pathname, $callback->($status)
395 522
396Asynchronously unlink (delete) a file and call the callback with the 523Asynchronously unlink (delete) a file and call the callback with the
397result code. 524result code.
398 525
526
527=item aio_mknod $path, $mode, $dev, $callback->($status)
528
529[EXPERIMENTAL]
530
531Asynchronously create a device node (or fifo). See mknod(2).
532
533The only (POSIX-) portable way of calling this function is:
534
535 aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ...
536
537
399=item aio_link $srcpath, $dstpath, $callback->($status) 538=item aio_link $srcpath, $dstpath, $callback->($status)
400 539
401Asynchronously create a new link to the existing object at C<$srcpath> at 540Asynchronously create a new link to the existing object at C<$srcpath> at
402the path C<$dstpath> and call the callback with the result code. 541the path C<$dstpath> and call the callback with the result code.
403 542
543
404=item aio_symlink $srcpath, $dstpath, $callback->($status) 544=item aio_symlink $srcpath, $dstpath, $callback->($status)
405 545
406Asynchronously create a new symbolic link to the existing object at C<$srcpath> at 546Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
407the path C<$dstpath> and call the callback with the result code. 547the path C<$dstpath> and call the callback with the result code.
408 548
549
550=item aio_readlink $path, $callback->($link)
551
552Asynchronously read the symlink specified by C<$path> and pass it to
553the callback. If an error occurs, nothing or undef gets passed to the
554callback.
555
556
409=item aio_rename $srcpath, $dstpath, $callback->($status) 557=item aio_rename $srcpath, $dstpath, $callback->($status)
410 558
411Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as 559Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
412rename(2) and call the callback with the result code. 560rename(2) and call the callback with the result code.
413 561
562
563=item aio_mkdir $pathname, $mode, $callback->($status)
564
565Asynchronously mkdir (create) a directory and call the callback with
566the result code. C<$mode> will be modified by the umask at the time the
567request is executed, so do not change your umask.
568
569
414=item aio_rmdir $pathname, $callback->($status) 570=item aio_rmdir $pathname, $callback->($status)
415 571
416Asynchronously rmdir (delete) a directory and call the callback with the 572Asynchronously rmdir (delete) a directory and call the callback with the
417result code. 573result code.
574
418 575
419=item aio_readdir $pathname, $callback->($entries) 576=item aio_readdir $pathname, $callback->($entries)
420 577
421Unlike the POSIX call of the same name, C<aio_readdir> reads an entire 578Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
422directory (i.e. opendir + readdir + closedir). The entries will not be 579directory (i.e. opendir + readdir + closedir). The entries will not be
423sorted, and will B<NOT> include the C<.> and C<..> entries. 580sorted, and will B<NOT> include the C<.> and C<..> entries.
424 581
425The callback a single argument which is either C<undef> or an array-ref 582The callback a single argument which is either C<undef> or an array-ref
426with the filenames. 583with the filenames.
584
585
586=item aio_load $path, $data, $callback->($status)
587
588This is a composite request that tries to fully load the given file into
589memory. Status is the same as with aio_read.
590
591=cut
592
593sub aio_load($$;$) {
594 aio_block {
595 my ($path, undef, $cb) = @_;
596 my $data = \$_[1];
597
598 my $pri = aioreq_pri;
599 my $grp = aio_group $cb;
600
601 aioreq_pri $pri;
602 add $grp aio_open $path, O_RDONLY, 0, sub {
603 my $fh = shift
604 or return $grp->result (-1);
605
606 aioreq_pri $pri;
607 add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
608 $grp->result ($_[0]);
609 };
610 };
611
612 $grp
613 }
614}
615
616=item aio_copy $srcpath, $dstpath, $callback->($status)
617
618Try to copy the I<file> (directories not supported as either source or
619destination) from C<$srcpath> to C<$dstpath> and call the callback with
620the C<0> (error) or C<-1> ok.
621
622This is a composite request that it creates the destination file with
623mode 0200 and copies the contents of the source file into it using
624C<aio_sendfile>, followed by restoring atime, mtime, access mode and
625uid/gid, in that order.
626
627If an error occurs, the partial destination file will be unlinked, if
628possible, except when setting atime, mtime, access mode and uid/gid, where
629errors are being ignored.
630
631=cut
632
633sub aio_copy($$;$) {
634 aio_block {
635 my ($src, $dst, $cb) = @_;
636
637 my $pri = aioreq_pri;
638 my $grp = aio_group $cb;
639
640 aioreq_pri $pri;
641 add $grp aio_open $src, O_RDONLY, 0, sub {
642 if (my $src_fh = $_[0]) {
643 my @stat = stat $src_fh;
644
645 aioreq_pri $pri;
646 add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
647 if (my $dst_fh = $_[0]) {
648 aioreq_pri $pri;
649 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
650 if ($_[0] == $stat[7]) {
651 $grp->result (0);
652 close $src_fh;
653
654 # those should not normally block. should. should.
655 utime $stat[8], $stat[9], $dst;
656 chmod $stat[2] & 07777, $dst_fh;
657 chown $stat[4], $stat[5], $dst_fh;
658
659 aioreq_pri $pri;
660 add $grp aio_close $dst_fh;
661 } else {
662 $grp->result (-1);
663 close $src_fh;
664 close $dst_fh;
665
666 aioreq $pri;
667 add $grp aio_unlink $dst;
668 }
669 };
670 } else {
671 $grp->result (-1);
672 }
673 },
674
675 } else {
676 $grp->result (-1);
677 }
678 };
679
680 $grp
681 }
682}
683
684=item aio_move $srcpath, $dstpath, $callback->($status)
685
686Try to move the I<file> (directories not supported as either source or
687destination) from C<$srcpath> to C<$dstpath> and call the callback with
688the C<0> (error) or C<-1> ok.
689
690This is a composite request that tries to rename(2) the file first. If
691rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if
692that is successful, unlinking the C<$srcpath>.
693
694=cut
695
696sub aio_move($$;$) {
697 aio_block {
698 my ($src, $dst, $cb) = @_;
699
700 my $pri = aioreq_pri;
701 my $grp = aio_group $cb;
702
703 aioreq_pri $pri;
704 add $grp aio_rename $src, $dst, sub {
705 if ($_[0] && $! == EXDEV) {
706 aioreq_pri $pri;
707 add $grp aio_copy $src, $dst, sub {
708 $grp->result ($_[0]);
709
710 if (!$_[0]) {
711 aioreq_pri $pri;
712 add $grp aio_unlink $src;
713 }
714 };
715 } else {
716 $grp->result ($_[0]);
717 }
718 };
719
720 $grp
721 }
722}
427 723
428=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 724=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
429 725
430Scans a directory (similar to C<aio_readdir>) but additionally tries to 726Scans a directory (similar to C<aio_readdir>) but additionally tries to
431efficiently separate the entries of directory C<$path> into two sets of 727efficiently separate the entries of directory C<$path> into two sets of
… …
433recurse into (everything else, including symlinks to directories). 729recurse into (everything else, including symlinks to directories).
434 730
435C<aio_scandir> is a composite request that creates of many sub requests_ 731C<aio_scandir> is a composite request that creates of many sub requests_
436C<$maxreq> specifies the maximum number of outstanding aio requests that 732C<$maxreq> specifies the maximum number of outstanding aio requests that
437this function generates. If it is C<< <= 0 >>, then a suitable default 733this function generates. If it is C<< <= 0 >>, then a suitable default
438will be chosen (currently 6). 734will be chosen (currently 4).
439 735
440On error, the callback is called without arguments, otherwise it receives 736On error, the callback is called without arguments, otherwise it receives
441two array-refs with path-relative entry names. 737two array-refs with path-relative entry names.
442 738
443Example: 739Example:
… …
477as those tend to return 0 or 1 as link counts, which disables the 773as those tend to return 0 or 1 as link counts, which disables the
478directory counting heuristic. 774directory counting heuristic.
479 775
480=cut 776=cut
481 777
482sub aio_scandir($$$) { 778sub aio_scandir($$;$) {
779 aio_block {
483 my ($path, $maxreq, $cb) = @_; 780 my ($path, $maxreq, $cb) = @_;
484 781
782 my $pri = aioreq_pri;
783
485 my $grp = aio_group $cb; 784 my $grp = aio_group $cb;
486 785
487 $maxreq = 6 if $maxreq <= 0; 786 $maxreq = 4 if $maxreq <= 0;
488 787
489 # stat once 788 # stat once
789 aioreq_pri $pri;
490 add $grp aio_stat $path, sub { 790 add $grp aio_stat $path, sub {
491 return $grp->result () if $_[0]; 791 return $grp->result () if $_[0];
492 my $now = time; 792 my $now = time;
493 my $hash1 = join ":", (stat _)[0,1,3,7,9]; 793 my $hash1 = join ":", (stat _)[0,1,3,7,9];
494 794
495 # read the directory entries 795 # read the directory entries
796 aioreq_pri $pri;
496 add $grp aio_readdir $path, sub { 797 add $grp aio_readdir $path, sub {
497 my $entries = shift 798 my $entries = shift
498 or return $grp->result (); 799 or return $grp->result ();
499 800
500 # stat the dir another time 801 # stat the dir another time
802 aioreq_pri $pri;
501 add $grp aio_stat $path, sub { 803 add $grp aio_stat $path, sub {
502 my $hash2 = join ":", (stat _)[0,1,3,7,9]; 804 my $hash2 = join ":", (stat _)[0,1,3,7,9];
503 805
504 my $ndirs; 806 my $ndirs;
505 807
506 # take the slow route if anything looks fishy 808 # take the slow route if anything looks fishy
507 if ($hash1 ne $hash2 or (stat _)[9] == $now) { 809 if ($hash1 ne $hash2 or (stat _)[9] == $now) {
508 $ndirs = -1; 810 $ndirs = -1;
509 } else { 811 } else {
510 # if nlink == 2, we are finished 812 # if nlink == 2, we are finished
511 # on non-posix-fs's, we rely on nlink < 2 813 # on non-posix-fs's, we rely on nlink < 2
512 $ndirs = (stat _)[3] - 2 814 $ndirs = (stat _)[3] - 2
513 or return $grp->result ([], $entries); 815 or return $grp->result ([], $entries);
514 } 816 }
515 817
516 # sort into likely dirs and likely nondirs 818 # sort into likely dirs and likely nondirs
517 # dirs == files without ".", short entries first 819 # dirs == files without ".", short entries first
518 $entries = [map $_->[0], 820 $entries = [map $_->[0],
519 sort { $b->[1] cmp $a->[1] } 821 sort { $b->[1] cmp $a->[1] }
520 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length], 822 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
521 @$entries]; 823 @$entries];
522 824
523 my (@dirs, @nondirs); 825 my (@dirs, @nondirs);
524 826
525 my $statgrp = add $grp aio_group sub { 827 my $statgrp = add $grp aio_group sub {
526 $grp->result (\@dirs, \@nondirs); 828 $grp->result (\@dirs, \@nondirs);
527 }; 829 };
528 830
529 limit $statgrp $maxreq; 831 limit $statgrp $maxreq;
530 feed $statgrp sub { 832 feed $statgrp sub {
531 return unless @$entries; 833 return unless @$entries;
532 my $entry = pop @$entries; 834 my $entry = pop @$entries;
533 835
836 aioreq_pri $pri;
534 add $statgrp aio_stat "$path/$entry/.", sub { 837 add $statgrp aio_stat "$path/$entry/.", sub {
535 if ($_[0] < 0) { 838 if ($_[0] < 0) {
536 push @nondirs, $entry; 839 push @nondirs, $entry;
537 } else { 840 } else {
538 # need to check for real directory 841 # need to check for real directory
842 aioreq_pri $pri;
539 add $statgrp aio_lstat "$path/$entry", sub { 843 add $statgrp aio_lstat "$path/$entry", sub {
540 if (-d _) { 844 if (-d _) {
541 push @dirs, $entry; 845 push @dirs, $entry;
542 846
543 unless (--$ndirs) { 847 unless (--$ndirs) {
544 push @nondirs, @$entries; 848 push @nondirs, @$entries;
545 feed $statgrp; 849 feed $statgrp;
850 }
851 } else {
852 push @nondirs, $entry;
546 } 853 }
547 } else {
548 push @nondirs, $entry;
549 } 854 }
550 } 855 }
551 } 856 };
552 }; 857 };
553 }; 858 };
554 }; 859 };
555 }; 860 };
861
862 $grp
556 }; 863 }
557
558 $grp
559} 864}
865
866=item aio_rmtree $path, $callback->($status)
867
868Delete a directory tree starting (and including) C<$path>, return the
869status of the final C<rmdir> only. This is a composite request that
870uses C<aio_scandir> to recurse into and rmdir directories, and unlink
871everything else.
872
873=cut
874
875sub aio_rmtree;
876sub aio_rmtree($;$) {
877 aio_block {
878 my ($path, $cb) = @_;
879
880 my $pri = aioreq_pri;
881 my $grp = aio_group $cb;
882
883 aioreq_pri $pri;
884 add $grp aio_scandir $path, 0, sub {
885 my ($dirs, $nondirs) = @_;
886
887 my $dirgrp = aio_group sub {
888 add $grp aio_rmdir $path, sub {
889 $grp->result ($_[0]);
890 };
891 };
892
893 (aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
894 (aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
895
896 add $grp $dirgrp;
897 };
898
899 $grp
900 }
901}
902
903=item aio_sync $callback->($status)
904
905Asynchronously call sync and call the callback when finished.
560 906
561=item aio_fsync $fh, $callback->($status) 907=item aio_fsync $fh, $callback->($status)
562 908
563Asynchronously call fsync on the given filehandle and call the callback 909Asynchronously call fsync on the given filehandle and call the callback
564with the fsync result code. 910with the fsync result code.
… …
568Asynchronously call fdatasync on the given filehandle and call the 914Asynchronously call fdatasync on the given filehandle and call the
569callback with the fdatasync result code. 915callback with the fdatasync result code.
570 916
571If this call isn't available because your OS lacks it or it couldn't be 917If this call isn't available because your OS lacks it or it couldn't be
572detected, it will be emulated by calling C<fsync> instead. 918detected, it will be emulated by calling C<fsync> instead.
919
920=item aio_pathsync $path, $callback->($status)
921
922This request tries to open, fsync and close the given path. This is a
923composite request intended tosync directories after directory operations
924(E.g. rename). This might not work on all operating systems or have any
925specific effect, but usually it makes sure that directory changes get
926written to disc. It works for anything that can be opened for read-only,
927not just directories.
928
929Passes C<0> when everything went ok, and C<-1> on error.
930
931=cut
932
933sub aio_pathsync($;$) {
934 aio_block {
935 my ($path, $cb) = @_;
936
937 my $pri = aioreq_pri;
938 my $grp = aio_group $cb;
939
940 aioreq_pri $pri;
941 add $grp aio_open $path, O_RDONLY, 0, sub {
942 my ($fh) = @_;
943 if ($fh) {
944 aioreq_pri $pri;
945 add $grp aio_fsync $fh, sub {
946 $grp->result ($_[0]);
947
948 aioreq_pri $pri;
949 add $grp aio_close $fh;
950 };
951 } else {
952 $grp->result (-1);
953 }
954 };
955
956 $grp
957 }
958}
573 959
574=item aio_group $callback->(...) 960=item aio_group $callback->(...)
575 961
576This is a very special aio request: Instead of doing something, it is a 962This is a very special aio request: Instead of doing something, it is a
577container for other aio requests, which is useful if you want to bundle 963container for other aio requests, which is useful if you want to bundle
… …
682=item * They can also can also be added to other IO::AIO::GRP objects. 1068=item * They can also can also be added to other IO::AIO::GRP objects.
683 1069
684=item * You must not add requests to a group from within the group callback (or 1070=item * You must not add requests to a group from within the group callback (or
685any later time). 1071any later time).
686 1072
687=item * This does not harmonise well with C<max_outstanding>, so best do
688not combine C<aio_group> with it. Groups and feeders are recommended for
689this kind of concurrency-limiting.
690
691=back 1073=back
692 1074
693Their lifetime, simplified, looks like this: when they are empty, they 1075Their lifetime, simplified, looks like this: when they are empty, they
694will finish very quickly. If they contain only requests that are in the 1076will finish very quickly. If they contain only requests that are in the
695C<done> state, they will also finish. Otherwise they will continue to 1077C<done> state, they will also finish. Otherwise they will continue to
… …
718itself. Useful when you queued a lot of events but got a result early. 1100itself. Useful when you queued a lot of events but got a result early.
719 1101
720=item $grp->result (...) 1102=item $grp->result (...)
721 1103
722Set the result value(s) that will be passed to the group callback when all 1104Set the result value(s) that will be passed to the group callback when all
723subrequests have finished. By default, no argument will be passed. 1105subrequests have finished and set the groups errno to the current value
1106of errno (just like calling C<errno> without an error number). By default,
1107no argument will be passed and errno is zero.
1108
1109=item $grp->errno ([$errno])
1110
1111Sets the group errno value to C<$errno>, or the current value of errno
1112when the argument is missing.
1113
1114Every aio request has an associated errno value that is restored when
1115the callback is invoked. This method lets you change this value from its
1116default (0).
1117
1118Calling C<result> will also set errno, so make sure you either set C<$!>
1119before the call to C<result>, or call c<errno> after it.
724 1120
725=item feed $grp $callback->($grp) 1121=item feed $grp $callback->($grp)
726 1122
727Sets a feeder/generator on this group: every group can have an attached 1123Sets a feeder/generator on this group: every group can have an attached
728generator that generates requests if idle. The idea behind this is that, 1124generator that generates requests if idle. The idea behind this is that,
… …
767 1163
768=back 1164=back
769 1165
770=head2 SUPPORT FUNCTIONS 1166=head2 SUPPORT FUNCTIONS
771 1167
1168=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
1169
772=over 4 1170=over 4
773 1171
774=item $fileno = IO::AIO::poll_fileno 1172=item $fileno = IO::AIO::poll_fileno
775 1173
776Return the I<request result pipe file descriptor>. This filehandle must be 1174Return the I<request result pipe file descriptor>. This filehandle must be
… …
780 1178
781See C<poll_cb> for an example. 1179See C<poll_cb> for an example.
782 1180
783=item IO::AIO::poll_cb 1181=item IO::AIO::poll_cb
784 1182
785Process all outstanding events on the result pipe. You have to call this 1183Process some outstanding events on the result pipe. You have to call this
786regularly. Returns the number of events processed. Returns immediately 1184regularly. Returns the number of events processed. Returns immediately
787when no events are outstanding. 1185when no events are outstanding. The amount of events processed depends on
1186the settings of C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
1187
1188If not all requests were processed for whatever reason, the filehandle
1189will still be ready when C<poll_cb> returns.
788 1190
789Example: Install an Event watcher that automatically calls 1191Example: Install an Event watcher that automatically calls
790IO::AIO::poll_cb with high priority: 1192IO::AIO::poll_cb with high priority:
791 1193
792 Event->io (fd => IO::AIO::poll_fileno, 1194 Event->io (fd => IO::AIO::poll_fileno,
793 poll => 'r', async => 1, 1195 poll => 'r', async => 1,
794 cb => \&IO::AIO::poll_cb); 1196 cb => \&IO::AIO::poll_cb);
795 1197
1198=item IO::AIO::max_poll_reqs $nreqs
1199
1200=item IO::AIO::max_poll_time $seconds
1201
1202These set the maximum number of requests (default C<0>, meaning infinity)
1203that are being processed by C<IO::AIO::poll_cb> in one call, respectively
1204the maximum amount of time (default C<0>, meaning infinity) spent in
1205C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
1206of time C<poll_cb> is allowed to use).
1207
1208Setting C<max_poll_time> to a non-zero value creates an overhead of one
1209syscall per request processed, which is not normally a problem unless your
1210callbacks are really really fast or your OS is really really slow (I am
1211not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
1212
1213Setting these is useful if you want to ensure some level of
1214interactiveness when perl is not fast enough to process all requests in
1215time.
1216
1217For interactive programs, values such as C<0.01> to C<0.1> should be fine.
1218
1219Example: Install an Event watcher that automatically calls
1220IO::AIO::poll_cb with low priority, to ensure that other parts of the
1221program get the CPU sometimes even under high AIO load.
1222
1223 # try not to spend much more than 0.1s in poll_cb
1224 IO::AIO::max_poll_time 0.1;
1225
1226 # use a low priority so other tasks have priority
1227 Event->io (fd => IO::AIO::poll_fileno,
1228 poll => 'r', nice => 1,
1229 cb => &IO::AIO::poll_cb);
1230
796=item IO::AIO::poll_wait 1231=item IO::AIO::poll_wait
797 1232
1233If there are any outstanding requests and none of them in the result
798Wait till the result filehandle becomes ready for reading (simply does a 1234phase, wait till the result filehandle becomes ready for reading (simply
799C<select> on the filehandle. This is useful if you want to synchronously wait 1235does a C<select> on the filehandle. This is useful if you want to
800for some requests to finish). 1236synchronously wait for some requests to finish).
801 1237
802See C<nreqs> for an example. 1238See C<nreqs> for an example.
803 1239
1240=item IO::AIO::poll
1241
1242Waits until some requests have been handled.
1243
1244Returns the number of requests processed, but is otherwise strictly
1245equivalent to:
1246
1247 IO::AIO::poll_wait, IO::AIO::poll_cb
1248
804=item IO::AIO::nreqs 1249=item IO::AIO::flush
805 1250
806Returns the number of requests currently outstanding (i.e. for which their 1251Wait till all outstanding AIO requests have been handled.
807callback has not been invoked yet).
808 1252
809Example: wait till there are no outstanding requests anymore: 1253Strictly equivalent to:
810 1254
811 IO::AIO::poll_wait, IO::AIO::poll_cb 1255 IO::AIO::poll_wait, IO::AIO::poll_cb
812 while IO::AIO::nreqs; 1256 while IO::AIO::nreqs;
813 1257
814=item IO::AIO::flush 1258=back
815 1259
816Wait till all outstanding AIO requests have been handled. 1260=head3 CONTROLLING THE NUMBER OF THREADS
817 1261
818Strictly equivalent to: 1262=over
819
820 IO::AIO::poll_wait, IO::AIO::poll_cb
821 while IO::AIO::nreqs;
822
823=item IO::AIO::poll
824
825Waits until some requests have been handled.
826
827Strictly equivalent to:
828
829 IO::AIO::poll_wait, IO::AIO::poll_cb
830 if IO::AIO::nreqs;
831 1263
832=item IO::AIO::min_parallel $nthreads 1264=item IO::AIO::min_parallel $nthreads
833 1265
834Set the minimum number of AIO threads to C<$nthreads>. The current 1266Set the minimum number of AIO threads to C<$nthreads>. The current
835default is C<8>, which means eight asynchronous operations can execute 1267default is C<8>, which means eight asynchronous operations can execute
836concurrently at any one time (the number of outstanding requests, 1268concurrently at any one time (the number of outstanding requests,
837however, is unlimited). 1269however, is unlimited).
838 1270
839IO::AIO starts threads only on demand, when an AIO request is queued and 1271IO::AIO starts threads only on demand, when an AIO request is queued and
840no free thread exists. 1272no free thread exists. Please note that queueing up a hundred requests can
1273create demand for a hundred threads, even if it turns out that everything
1274is in the cache and could have been processed faster by a single thread.
841 1275
842It is recommended to keep the number of threads relatively low, as some 1276It is recommended to keep the number of threads relatively low, as some
843Linux kernel versions will scale negatively with the number of threads 1277Linux kernel versions will scale negatively with the number of threads
844(higher parallelity => MUCH higher latency). With current Linux 2.6 1278(higher parallelity => MUCH higher latency). With current Linux 2.6
845versions, 4-32 threads should be fine. 1279versions, 4-32 threads should be fine.
… …
859This module automatically runs C<max_parallel 0> at program end, to ensure 1293This module automatically runs C<max_parallel 0> at program end, to ensure
860that all threads are killed and that there are no outstanding requests. 1294that all threads are killed and that there are no outstanding requests.
861 1295
862Under normal circumstances you don't need to call this function. 1296Under normal circumstances you don't need to call this function.
863 1297
1298=item IO::AIO::max_idle $nthreads
1299
1300Limit the number of threads (default: 4) that are allowed to idle (i.e.,
1301threads that did not get a request to process within 10 seconds). That
1302means if a thread becomes idle while C<$nthreads> other threads are also
1303idle, it will free its resources and exit.
1304
1305This is useful when you allow a large number of threads (e.g. 100 or 1000)
1306to allow for extremely high load situations, but want to free resources
1307under normal circumstances (1000 threads can easily consume 30MB of RAM).
1308
1309The default is probably ok in most situations, especially if thread
1310creation is fast. If thread creation is very slow on your system you might
1311want to use larger values.
1312
864=item $oldnreqs = IO::AIO::max_outstanding $nreqs 1313=item $oldmaxreqs = IO::AIO::max_outstanding $maxreqs
865 1314
866[DEPRECATED] 1315This is a very bad function to use in interactive programs because it
1316blocks, and a bad way to reduce concurrency because it is inexact: Better
1317use an C<aio_group> together with a feed callback.
867 1318
868Sets the maximum number of outstanding requests to C<$nreqs>. If you 1319Sets the maximum number of outstanding requests to C<$nreqs>. If you
869try to queue up more than this number of requests, the caller will block until 1320do queue up more than this number of requests, the next call to the
870some requests have been handled. 1321C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
1322function will block until the limit is no longer exceeded.
871 1323
872The default is very large, so normally there is no practical limit. If you 1324The default value is very large, so there is no practical limit on the
873queue up many requests in a loop it often improves speed if you set 1325number of outstanding requests.
874this to a relatively low number, such as C<100>.
875 1326
876This function does not work well together with C<aio_group>'s, and their 1327You can still queue as many requests as you want. Therefore,
877feeder interface is better suited to limiting concurrency, so do not use 1328C<max_oustsanding> is mainly useful in simple scripts (with low values) or
878this function. 1329as a stop gap to shield against fatal memory overflow (with large values).
879
880Under normal circumstances you don't need to call this function.
881 1330
882=back 1331=back
883 1332
1333=head3 STATISTICAL INFORMATION
1334
1335=over
1336
1337=item IO::AIO::nreqs
1338
1339Returns the number of requests currently in the ready, execute or pending
1340states (i.e. for which their callback has not been invoked yet).
1341
1342Example: wait till there are no outstanding requests anymore:
1343
1344 IO::AIO::poll_wait, IO::AIO::poll_cb
1345 while IO::AIO::nreqs;
1346
1347=item IO::AIO::nready
1348
1349Returns the number of requests currently in the ready state (not yet
1350executed).
1351
1352=item IO::AIO::npending
1353
1354Returns the number of requests currently in the pending state (executed,
1355but not yet processed by poll_cb).
1356
1357=back
1358
884=cut 1359=cut
885 1360
886# support function to convert a fd into a perl filehandle
887sub _fd2fh {
888 return undef if $_[0] < 0;
889
890 # try to generate nice filehandles
891 my $sym = "IO::AIO::fd#$_[0]";
892 local *$sym;
893
894 open *$sym, "+<&=$_[0]" # usually works under any unix
895 or open *$sym, "<&=$_[0]" # cygwin needs this
896 or open *$sym, ">&=$_[0]" # or this
897 or return undef;
898
899 *$sym
900}
901
902min_parallel 8; 1361min_parallel 8;
903 1362
904END { 1363END { flush }
905 max_parallel 0;
906}
907 1364
9081; 13651;
909 1366
910=head2 FORK BEHAVIOUR 1367=head2 FORK BEHAVIOUR
911 1368
… …
931bytes of memory. In addition, stat requests need a stat buffer (possibly 1388bytes of memory. In addition, stat requests need a stat buffer (possibly
932a few hundred bytes), readdir requires a result buffer and so on. Perl 1389a few hundred bytes), readdir requires a result buffer and so on. Perl
933scalars and other data passed into aio requests will also be locked and 1390scalars and other data passed into aio requests will also be locked and
934will consume memory till the request has entered the done state. 1391will consume memory till the request has entered the done state.
935 1392
936This is now awfully much, so queuing lots of requests is not usually a 1393This is not awfully much, so queuing lots of requests is not usually a
937problem. 1394problem.
938 1395
939Per-thread usage: 1396Per-thread usage:
940 1397
941In the execution phase, some aio requests require more memory for 1398In the execution phase, some aio requests require more memory for

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines