1 |
=head1 NAME |
2 |
|
3 |
IO::AIO - Asynchronous Input/Output |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use IO::AIO; |
8 |
|
9 |
aio_open "/etc/passwd", O_RDONLY, 0, sub { |
10 |
my $fh = shift |
11 |
or die "/etc/passwd: $!"; |
12 |
... |
13 |
}; |
14 |
|
15 |
aio_unlink "/tmp/file", sub { }; |
16 |
|
17 |
aio_read $fh, 30000, 1024, $buffer, 0, sub { |
18 |
$_[0] > 0 or die "read error: $!"; |
19 |
}; |
20 |
|
21 |
# version 2+ has request and group objects |
22 |
use IO::AIO 2; |
23 |
|
24 |
aioreq_pri 4; # give next request a very high priority |
25 |
my $req = aio_unlink "/tmp/file", sub { }; |
26 |
$req->cancel; # cancel request if still in queue |
27 |
|
28 |
my $grp = aio_group sub { print "all stats done\n" }; |
29 |
add $grp aio_stat "..." for ...; |
30 |
|
31 |
# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...) |
32 |
use AnyEvent::AIO; |
33 |
|
34 |
# EV integration |
35 |
my $w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; |
36 |
|
37 |
# Event integration |
38 |
Event->io (fd => IO::AIO::poll_fileno, |
39 |
poll => 'r', |
40 |
cb => \&IO::AIO::poll_cb); |
41 |
|
42 |
# Glib/Gtk2 integration |
43 |
add_watch Glib::IO IO::AIO::poll_fileno, |
44 |
in => sub { IO::AIO::poll_cb; 1 }; |
45 |
|
46 |
# Tk integration |
47 |
Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", |
48 |
readable => \&IO::AIO::poll_cb); |
49 |
|
50 |
# Danga::Socket integration |
51 |
Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
52 |
\&IO::AIO::poll_cb); |
53 |
|
54 |
=head1 DESCRIPTION |
55 |
|
56 |
This module implements asynchronous I/O using whatever means your |
57 |
operating system supports. |
58 |
|
59 |
Asynchronous means that operations that can normally block your program |
60 |
(e.g. reading from disk) will be done asynchronously: the operation |
61 |
will still block, but you can do something else in the meantime. This |
62 |
is extremely useful for programs that need to stay interactive even |
63 |
when doing heavy I/O (GUI programs, high performance network servers |
64 |
etc.), but can also be used to easily do operations in parallel that are |
65 |
normally done sequentially, e.g. stat'ing many files, which is much faster |
66 |
on a RAID volume or over NFS when you do a number of stat operations |
67 |
concurrently. |
68 |
|
69 |
While most of this works on all types of file descriptors (for |
70 |
example sockets), using these functions on file descriptors that |
71 |
support nonblocking operation (again, sockets, pipes etc.) is very |
72 |
inefficient. Use an event loop for that (such as the L<Event|Event> |
73 |
module): IO::AIO will naturally fit into such an event loop itself. |
74 |
|
75 |
In this version, a number of threads are started that execute your |
76 |
requests and signal their completion. You don't need thread support |
77 |
in perl, and the threads created by this module will not be visible |
78 |
to perl. In the future, this module might make use of the native aio |
79 |
functions available on many operating systems. However, they are often |
80 |
not well-supported or restricted (GNU/Linux doesn't allow them on normal |
81 |
files currently, for example), and they would only support aio_read and |
82 |
aio_write, so the remaining functionality would have to be implemented |
83 |
using threads anyway. |
84 |
|
85 |
Although the module will work in the presence of other (Perl-) threads, |
86 |
it is currently not reentrant in any way, so use appropriate locking |
87 |
yourself, always call C<poll_cb> from within the same thread, or never |
88 |
call C<poll_cb> (or other C<aio_> functions) recursively. |
89 |
|
90 |
=head2 EXAMPLE |
91 |
|
92 |
This is a simple example that uses the Event module and loads |
93 |
F</etc/passwd> asynchronously: |
94 |
|
95 |
use Fcntl; |
96 |
use Event; |
97 |
use IO::AIO; |
98 |
|
99 |
# register the IO::AIO callback with Event |
100 |
Event->io (fd => IO::AIO::poll_fileno, |
101 |
poll => 'r', |
102 |
cb => \&IO::AIO::poll_cb); |
103 |
|
104 |
# queue the request to open /etc/passwd |
105 |
aio_open "/etc/passwd", O_RDONLY, 0, sub { |
106 |
my $fh = shift |
107 |
or die "error while opening: $!"; |
108 |
|
109 |
# stat'ing filehandles is generally non-blocking |
110 |
my $size = -s $fh; |
111 |
|
112 |
# queue a request to read the file |
113 |
my $contents; |
114 |
aio_read $fh, 0, $size, $contents, 0, sub { |
115 |
$_[0] == $size |
116 |
or die "short read: $!"; |
117 |
|
118 |
close $fh; |
119 |
|
120 |
# file contents now in $contents |
121 |
print $contents; |
122 |
|
123 |
# exit event loop and program |
124 |
Event::unloop; |
125 |
}; |
126 |
}; |
127 |
|
128 |
# possibly queue up other requests, or open GUI windows, |
129 |
# check for sockets etc. etc. |
130 |
|
131 |
# process events as long as there are some: |
132 |
Event::loop; |
133 |
|
134 |
=head1 REQUEST ANATOMY AND LIFETIME |
135 |
|
136 |
Every C<aio_*> function creates a request. which is a C data structure not |
137 |
directly visible to Perl. |
138 |
|
139 |
If called in non-void context, every request function returns a Perl |
140 |
object representing the request. In void context, nothing is returned, |
141 |
which saves a bit of memory. |
142 |
|
143 |
The perl object is a fairly standard ref-to-hash object. The hash contents |
144 |
are not used by IO::AIO so you are free to store anything you like in it. |
145 |
|
146 |
During their existance, aio requests travel through the following states, |
147 |
in order: |
148 |
|
149 |
=over 4 |
150 |
|
151 |
=item ready |
152 |
|
153 |
Immediately after a request is created it is put into the ready state, |
154 |
waiting for a thread to execute it. |
155 |
|
156 |
=item execute |
157 |
|
158 |
A thread has accepted the request for processing and is currently |
159 |
executing it (e.g. blocking in read). |
160 |
|
161 |
=item pending |
162 |
|
163 |
The request has been executed and is waiting for result processing. |
164 |
|
165 |
While request submission and execution is fully asynchronous, result |
166 |
processing is not and relies on the perl interpreter calling C<poll_cb> |
167 |
(or another function with the same effect). |
168 |
|
169 |
=item result |
170 |
|
171 |
The request results are processed synchronously by C<poll_cb>. |
172 |
|
173 |
The C<poll_cb> function will process all outstanding aio requests by |
174 |
calling their callbacks, freeing memory associated with them and managing |
175 |
any groups they are contained in. |
176 |
|
177 |
=item done |
178 |
|
179 |
Request has reached the end of its lifetime and holds no resources anymore |
180 |
(except possibly for the Perl object, but its connection to the actual |
181 |
aio request is severed and calling its methods will either do nothing or |
182 |
result in a runtime error). |
183 |
|
184 |
=back |
185 |
|
186 |
=cut |
187 |
|
188 |
package IO::AIO; |
189 |
|
190 |
use Carp (); |
191 |
|
192 |
no warnings; |
193 |
use strict 'vars'; |
194 |
|
195 |
use base 'Exporter'; |
196 |
|
197 |
BEGIN { |
198 |
our $VERSION = '3.18'; |
199 |
|
200 |
our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close |
201 |
aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir |
202 |
aio_scandir aio_symlink aio_readlink aio_sync aio_fsync |
203 |
aio_fdatasync aio_sync_file_range aio_pathsync aio_readahead |
204 |
aio_rename aio_link aio_move aio_copy aio_group |
205 |
aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown |
206 |
aio_chmod aio_utime aio_truncate); |
207 |
|
208 |
our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice)); |
209 |
our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush |
210 |
min_parallel max_parallel max_idle |
211 |
nreqs nready npending nthreads |
212 |
max_poll_time max_poll_reqs); |
213 |
|
214 |
push @AIO_REQ, qw(aio_busy); # not exported |
215 |
|
216 |
@IO::AIO::GRP::ISA = 'IO::AIO::REQ'; |
217 |
|
218 |
require XSLoader; |
219 |
XSLoader::load ("IO::AIO", $VERSION); |
220 |
} |
221 |
|
222 |
=head1 FUNCTIONS |
223 |
|
224 |
=head2 AIO REQUEST FUNCTIONS |
225 |
|
226 |
All the C<aio_*> calls are more or less thin wrappers around the syscall |
227 |
with the same name (sans C<aio_>). The arguments are similar or identical, |
228 |
and they all accept an additional (and optional) C<$callback> argument |
229 |
which must be a code reference. This code reference will get called with |
230 |
the syscall return code (e.g. most syscalls return C<-1> on error, unlike |
231 |
perl, which usually delivers "false") as its sole argument after the given |
232 |
syscall has been executed asynchronously. |
233 |
|
234 |
All functions expecting a filehandle keep a copy of the filehandle |
235 |
internally until the request has finished. |
236 |
|
237 |
All functions return request objects of type L<IO::AIO::REQ> that allow |
238 |
further manipulation of those requests while they are in-flight. |
239 |
|
240 |
The pathnames you pass to these routines I<must> be absolute and |
241 |
encoded as octets. The reason for the former is that at the time the |
242 |
request is being executed, the current working directory could have |
243 |
changed. Alternatively, you can make sure that you never change the |
244 |
current working directory anywhere in the program and then use relative |
245 |
paths. |
246 |
|
247 |
To encode pathnames as octets, either make sure you either: a) always pass |
248 |
in filenames you got from outside (command line, readdir etc.) without |
249 |
tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode |
250 |
your pathnames to the locale (or other) encoding in effect in the user |
251 |
environment, d) use Glib::filename_from_unicode on unicode filenames or e) |
252 |
use something else to ensure your scalar has the correct contents. |
253 |
|
254 |
This works, btw. independent of the internal UTF-8 bit, which IO::AIO |
255 |
handles correctly whether it is set or not. |
256 |
|
257 |
=over 4 |
258 |
|
259 |
=item $prev_pri = aioreq_pri [$pri] |
260 |
|
261 |
Returns the priority value that would be used for the next request and, if |
262 |
C<$pri> is given, sets the priority for the next aio request. |
263 |
|
264 |
The default priority is C<0>, the minimum and maximum priorities are C<-4> |
265 |
and C<4>, respectively. Requests with higher priority will be serviced |
266 |
first. |
267 |
|
268 |
The priority will be reset to C<0> after each call to one of the C<aio_*> |
269 |
functions. |
270 |
|
271 |
Example: open a file with low priority, then read something from it with |
272 |
higher priority so the read request is serviced before other low priority |
273 |
open requests (potentially spamming the cache): |
274 |
|
275 |
aioreq_pri -3; |
276 |
aio_open ..., sub { |
277 |
return unless $_[0]; |
278 |
|
279 |
aioreq_pri -2; |
280 |
aio_read $_[0], ..., sub { |
281 |
... |
282 |
}; |
283 |
}; |
284 |
|
285 |
|
286 |
=item aioreq_nice $pri_adjust |
287 |
|
288 |
Similar to C<aioreq_pri>, but subtracts the given value from the current |
289 |
priority, so the effect is cumulative. |
290 |
|
291 |
|
292 |
=item aio_open $pathname, $flags, $mode, $callback->($fh) |
293 |
|
294 |
Asynchronously open or create a file and call the callback with a newly |
295 |
created filehandle for the file. |
296 |
|
297 |
The pathname passed to C<aio_open> must be absolute. See API NOTES, above, |
298 |
for an explanation. |
299 |
|
300 |
The C<$flags> argument is a bitmask. See the C<Fcntl> module for a |
301 |
list. They are the same as used by C<sysopen>. |
302 |
|
303 |
Likewise, C<$mode> specifies the mode of the newly created file, if it |
304 |
didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>, |
305 |
except that it is mandatory (i.e. use C<0> if you don't create new files, |
306 |
and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified |
307 |
by the umask in effect then the request is being executed, so better never |
308 |
change the umask. |
309 |
|
310 |
Example: |
311 |
|
312 |
aio_open "/etc/passwd", O_RDONLY, 0, sub { |
313 |
if ($_[0]) { |
314 |
print "open successful, fh is $_[0]\n"; |
315 |
... |
316 |
} else { |
317 |
die "open failed: $!\n"; |
318 |
} |
319 |
}; |
320 |
|
321 |
|
322 |
=item aio_close $fh, $callback->($status) |
323 |
|
324 |
Asynchronously close a file and call the callback with the result |
325 |
code. |
326 |
|
327 |
Unfortunately, you can't do this to perl. Perl I<insists> very strongly on |
328 |
closing the file descriptor associated with the filehandle itself. |
329 |
|
330 |
Therefore, C<aio_close> will not close the filehandle - instead it will |
331 |
use dup2 to overwrite the file descriptor with the write-end of a pipe |
332 |
(the pipe fd will be created on demand and will be cached). |
333 |
|
334 |
Or in other words: the file descriptor will be closed, but it will not be |
335 |
free for reuse until the perl filehandle is closed. |
336 |
|
337 |
=cut |
338 |
|
339 |
=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
340 |
|
341 |
=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
342 |
|
343 |
Reads or writes C<$length> bytes from or to the specified C<$fh> and |
344 |
C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> |
345 |
and calls the callback without the actual number of bytes read (or -1 on |
346 |
error, just like the syscall). |
347 |
|
348 |
If C<$offset> is undefined, then the current file descriptor offset will |
349 |
be used (and updated), otherwise the file descriptor offset will not be |
350 |
changed by these calls. |
351 |
|
352 |
If C<$length> is undefined in C<aio_write>, use the remaining length of |
353 |
C<$data>. |
354 |
|
355 |
If C<$dataoffset> is less than zero, it will be counted from the end of |
356 |
C<$data>. |
357 |
|
358 |
The C<$data> scalar I<MUST NOT> be modified in any way while the request |
359 |
is outstanding. Modifying it can result in segfaults or World War III (if |
360 |
the necessary/optional hardware is installed). |
361 |
|
362 |
Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at |
363 |
offset C<0> within the scalar: |
364 |
|
365 |
aio_read $fh, 7, 15, $buffer, 0, sub { |
366 |
$_[0] > 0 or die "read error: $!"; |
367 |
print "read $_[0] bytes: <$buffer>\n"; |
368 |
}; |
369 |
|
370 |
|
371 |
=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
372 |
|
373 |
Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts |
374 |
reading at byte offset C<$in_offset>, and starts writing at the current |
375 |
file offset of C<$out_fh>. Because of that, it is not safe to issue more |
376 |
than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each |
377 |
other. |
378 |
|
379 |
This call tries to make use of a native C<sendfile> syscall to provide |
380 |
zero-copy operation. For this to work, C<$out_fh> should refer to a |
381 |
socket, and C<$in_fh> should refer to mmap'able file. |
382 |
|
383 |
If the native sendfile call fails or is not implemented, it will be |
384 |
emulated, so you can call C<aio_sendfile> on any type of filehandle |
385 |
regardless of the limitations of the operating system. |
386 |
|
387 |
Please note, however, that C<aio_sendfile> can read more bytes from |
388 |
C<$in_fh> than are written, and there is no way to find out how many |
389 |
bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only |
390 |
provides the number of bytes written to C<$out_fh>. Only if the result |
391 |
value equals C<$length> one can assume that C<$length> bytes have been |
392 |
read. |
393 |
|
394 |
|
395 |
=item aio_readahead $fh,$offset,$length, $callback->($retval) |
396 |
|
397 |
C<aio_readahead> populates the page cache with data from a file so that |
398 |
subsequent reads from that file will not block on disk I/O. The C<$offset> |
399 |
argument specifies the starting point from which data is to be read and |
400 |
C<$length> specifies the number of bytes to be read. I/O is performed in |
401 |
whole pages, so that offset is effectively rounded down to a page boundary |
402 |
and bytes are read up to the next page boundary greater than or equal to |
403 |
(off-set+length). C<aio_readahead> does not read beyond the end of the |
404 |
file. The current file offset of the file is left unchanged. |
405 |
|
406 |
If that syscall doesn't exist (likely if your OS isn't Linux) it will be |
407 |
emulated by simply reading the data, which would have a similar effect. |
408 |
|
409 |
|
410 |
=item aio_stat $fh_or_path, $callback->($status) |
411 |
|
412 |
=item aio_lstat $fh, $callback->($status) |
413 |
|
414 |
Works like perl's C<stat> or C<lstat> in void context. The callback will |
415 |
be called after the stat and the results will be available using C<stat _> |
416 |
or C<-s _> etc... |
417 |
|
418 |
The pathname passed to C<aio_stat> must be absolute. See API NOTES, above, |
419 |
for an explanation. |
420 |
|
421 |
Currently, the stats are always 64-bit-stats, i.e. instead of returning an |
422 |
error when stat'ing a large file, the results will be silently truncated |
423 |
unless perl itself is compiled with large file support. |
424 |
|
425 |
Example: Print the length of F</etc/passwd>: |
426 |
|
427 |
aio_stat "/etc/passwd", sub { |
428 |
$_[0] and die "stat failed: $!"; |
429 |
print "size is ", -s _, "\n"; |
430 |
}; |
431 |
|
432 |
|
433 |
=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status) |
434 |
|
435 |
Works like perl's C<utime> function (including the special case of $atime |
436 |
and $mtime being undef). Fractional times are supported if the underlying |
437 |
syscalls support them. |
438 |
|
439 |
When called with a pathname, uses utimes(2) if available, otherwise |
440 |
utime(2). If called on a file descriptor, uses futimes(2) if available, |
441 |
otherwise returns ENOSYS, so this is not portable. |
442 |
|
443 |
Examples: |
444 |
|
445 |
# set atime and mtime to current time (basically touch(1)): |
446 |
aio_utime "path", undef, undef; |
447 |
# set atime to current time and mtime to beginning of the epoch: |
448 |
aio_utime "path", time, undef; # undef==0 |
449 |
|
450 |
|
451 |
=item aio_chown $fh_or_path, $uid, $gid, $callback->($status) |
452 |
|
453 |
Works like perl's C<chown> function, except that C<undef> for either $uid |
454 |
or $gid is being interpreted as "do not change" (but -1 can also be used). |
455 |
|
456 |
Examples: |
457 |
|
458 |
# same as "chown root path" in the shell: |
459 |
aio_chown "path", 0, -1; |
460 |
# same as above: |
461 |
aio_chown "path", 0, undef; |
462 |
|
463 |
|
464 |
=item aio_truncate $fh_or_path, $offset, $callback->($status) |
465 |
|
466 |
Works like truncate(2) or ftruncate(2). |
467 |
|
468 |
|
469 |
=item aio_chmod $fh_or_path, $mode, $callback->($status) |
470 |
|
471 |
Works like perl's C<chmod> function. |
472 |
|
473 |
|
474 |
=item aio_unlink $pathname, $callback->($status) |
475 |
|
476 |
Asynchronously unlink (delete) a file and call the callback with the |
477 |
result code. |
478 |
|
479 |
|
480 |
=item aio_mknod $path, $mode, $dev, $callback->($status) |
481 |
|
482 |
[EXPERIMENTAL] |
483 |
|
484 |
Asynchronously create a device node (or fifo). See mknod(2). |
485 |
|
486 |
The only (POSIX-) portable way of calling this function is: |
487 |
|
488 |
aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ... |
489 |
|
490 |
|
491 |
=item aio_link $srcpath, $dstpath, $callback->($status) |
492 |
|
493 |
Asynchronously create a new link to the existing object at C<$srcpath> at |
494 |
the path C<$dstpath> and call the callback with the result code. |
495 |
|
496 |
|
497 |
=item aio_symlink $srcpath, $dstpath, $callback->($status) |
498 |
|
499 |
Asynchronously create a new symbolic link to the existing object at C<$srcpath> at |
500 |
the path C<$dstpath> and call the callback with the result code. |
501 |
|
502 |
|
503 |
=item aio_readlink $path, $callback->($link) |
504 |
|
505 |
Asynchronously read the symlink specified by C<$path> and pass it to |
506 |
the callback. If an error occurs, nothing or undef gets passed to the |
507 |
callback. |
508 |
|
509 |
|
510 |
=item aio_rename $srcpath, $dstpath, $callback->($status) |
511 |
|
512 |
Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as |
513 |
rename(2) and call the callback with the result code. |
514 |
|
515 |
|
516 |
=item aio_mkdir $pathname, $mode, $callback->($status) |
517 |
|
518 |
Asynchronously mkdir (create) a directory and call the callback with |
519 |
the result code. C<$mode> will be modified by the umask at the time the |
520 |
request is executed, so do not change your umask. |
521 |
|
522 |
|
523 |
=item aio_rmdir $pathname, $callback->($status) |
524 |
|
525 |
Asynchronously rmdir (delete) a directory and call the callback with the |
526 |
result code. |
527 |
|
528 |
|
529 |
=item aio_readdir $pathname, $callback->($entries) |
530 |
|
531 |
Unlike the POSIX call of the same name, C<aio_readdir> reads an entire |
532 |
directory (i.e. opendir + readdir + closedir). The entries will not be |
533 |
sorted, and will B<NOT> include the C<.> and C<..> entries. |
534 |
|
535 |
The callback a single argument which is either C<undef> or an array-ref |
536 |
with the filenames. |
537 |
|
538 |
|
539 |
=item aio_load $path, $data, $callback->($status) |
540 |
|
541 |
This is a composite request that tries to fully load the given file into |
542 |
memory. Status is the same as with aio_read. |
543 |
|
544 |
=cut |
545 |
|
546 |
sub aio_load($$;$) { |
547 |
my ($path, undef, $cb) = @_; |
548 |
my $data = \$_[1]; |
549 |
|
550 |
my $pri = aioreq_pri; |
551 |
my $grp = aio_group $cb; |
552 |
|
553 |
aioreq_pri $pri; |
554 |
add $grp aio_open $path, O_RDONLY, 0, sub { |
555 |
my $fh = shift |
556 |
or return $grp->result (-1); |
557 |
|
558 |
aioreq_pri $pri; |
559 |
add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub { |
560 |
$grp->result ($_[0]); |
561 |
}; |
562 |
}; |
563 |
|
564 |
$grp |
565 |
} |
566 |
|
567 |
=item aio_copy $srcpath, $dstpath, $callback->($status) |
568 |
|
569 |
Try to copy the I<file> (directories not supported as either source or |
570 |
destination) from C<$srcpath> to C<$dstpath> and call the callback with |
571 |
the C<0> (error) or C<-1> ok. |
572 |
|
573 |
This is a composite request that creates the destination file with |
574 |
mode 0200 and copies the contents of the source file into it using |
575 |
C<aio_sendfile>, followed by restoring atime, mtime, access mode and |
576 |
uid/gid, in that order. |
577 |
|
578 |
If an error occurs, the partial destination file will be unlinked, if |
579 |
possible, except when setting atime, mtime, access mode and uid/gid, where |
580 |
errors are being ignored. |
581 |
|
582 |
=cut |
583 |
|
584 |
sub aio_copy($$;$) { |
585 |
my ($src, $dst, $cb) = @_; |
586 |
|
587 |
my $pri = aioreq_pri; |
588 |
my $grp = aio_group $cb; |
589 |
|
590 |
aioreq_pri $pri; |
591 |
add $grp aio_open $src, O_RDONLY, 0, sub { |
592 |
if (my $src_fh = $_[0]) { |
593 |
my @stat = stat $src_fh; |
594 |
|
595 |
aioreq_pri $pri; |
596 |
add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub { |
597 |
if (my $dst_fh = $_[0]) { |
598 |
aioreq_pri $pri; |
599 |
add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub { |
600 |
if ($_[0] == $stat[7]) { |
601 |
$grp->result (0); |
602 |
close $src_fh; |
603 |
|
604 |
# those should not normally block. should. should. |
605 |
utime $stat[8], $stat[9], $dst; |
606 |
chmod $stat[2] & 07777, $dst_fh; |
607 |
chown $stat[4], $stat[5], $dst_fh; |
608 |
|
609 |
aioreq_pri $pri; |
610 |
add $grp aio_close $dst_fh; |
611 |
} else { |
612 |
$grp->result (-1); |
613 |
close $src_fh; |
614 |
close $dst_fh; |
615 |
|
616 |
aioreq $pri; |
617 |
add $grp aio_unlink $dst; |
618 |
} |
619 |
}; |
620 |
} else { |
621 |
$grp->result (-1); |
622 |
} |
623 |
}, |
624 |
|
625 |
} else { |
626 |
$grp->result (-1); |
627 |
} |
628 |
}; |
629 |
|
630 |
$grp |
631 |
} |
632 |
|
633 |
=item aio_move $srcpath, $dstpath, $callback->($status) |
634 |
|
635 |
Try to move the I<file> (directories not supported as either source or |
636 |
destination) from C<$srcpath> to C<$dstpath> and call the callback with |
637 |
the C<0> (error) or C<-1> ok. |
638 |
|
639 |
This is a composite request that tries to rename(2) the file first; if |
640 |
rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if |
641 |
that is successful, unlinks the C<$srcpath>. |
642 |
|
643 |
=cut |
644 |
|
645 |
sub aio_move($$;$) { |
646 |
my ($src, $dst, $cb) = @_; |
647 |
|
648 |
my $pri = aioreq_pri; |
649 |
my $grp = aio_group $cb; |
650 |
|
651 |
aioreq_pri $pri; |
652 |
add $grp aio_rename $src, $dst, sub { |
653 |
if ($_[0] && $! == EXDEV) { |
654 |
aioreq_pri $pri; |
655 |
add $grp aio_copy $src, $dst, sub { |
656 |
$grp->result ($_[0]); |
657 |
|
658 |
if (!$_[0]) { |
659 |
aioreq_pri $pri; |
660 |
add $grp aio_unlink $src; |
661 |
} |
662 |
}; |
663 |
} else { |
664 |
$grp->result ($_[0]); |
665 |
} |
666 |
}; |
667 |
|
668 |
$grp |
669 |
} |
670 |
|
671 |
=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
672 |
|
673 |
Scans a directory (similar to C<aio_readdir>) but additionally tries to |
674 |
efficiently separate the entries of directory C<$path> into two sets of |
675 |
names, directories you can recurse into (directories), and ones you cannot |
676 |
recurse into (everything else, including symlinks to directories). |
677 |
|
678 |
C<aio_scandir> is a composite request that creates of many sub requests_ |
679 |
C<$maxreq> specifies the maximum number of outstanding aio requests that |
680 |
this function generates. If it is C<< <= 0 >>, then a suitable default |
681 |
will be chosen (currently 4). |
682 |
|
683 |
On error, the callback is called without arguments, otherwise it receives |
684 |
two array-refs with path-relative entry names. |
685 |
|
686 |
Example: |
687 |
|
688 |
aio_scandir $dir, 0, sub { |
689 |
my ($dirs, $nondirs) = @_; |
690 |
print "real directories: @$dirs\n"; |
691 |
print "everything else: @$nondirs\n"; |
692 |
}; |
693 |
|
694 |
Implementation notes. |
695 |
|
696 |
The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. |
697 |
|
698 |
After reading the directory, the modification time, size etc. of the |
699 |
directory before and after the readdir is checked, and if they match (and |
700 |
isn't the current time), the link count will be used to decide how many |
701 |
entries are directories (if >= 2). Otherwise, no knowledge of the number |
702 |
of subdirectories will be assumed. |
703 |
|
704 |
Then entries will be sorted into likely directories (everything without |
705 |
a non-initial dot currently) and likely non-directories (everything |
706 |
else). Then every entry plus an appended C</.> will be C<stat>'ed, |
707 |
likely directories first. If that succeeds, it assumes that the entry |
708 |
is a directory or a symlink to directory (which will be checked |
709 |
seperately). This is often faster than stat'ing the entry itself because |
710 |
filesystems might detect the type of the entry without reading the inode |
711 |
data (e.g. ext2fs filetype feature). |
712 |
|
713 |
If the known number of directories (link count - 2) has been reached, the |
714 |
rest of the entries is assumed to be non-directories. |
715 |
|
716 |
This only works with certainty on POSIX (= UNIX) filesystems, which |
717 |
fortunately are the vast majority of filesystems around. |
718 |
|
719 |
It will also likely work on non-POSIX filesystems with reduced efficiency |
720 |
as those tend to return 0 or 1 as link counts, which disables the |
721 |
directory counting heuristic. |
722 |
|
723 |
=cut |
724 |
|
725 |
sub aio_scandir($$;$) { |
726 |
my ($path, $maxreq, $cb) = @_; |
727 |
|
728 |
my $pri = aioreq_pri; |
729 |
|
730 |
my $grp = aio_group $cb; |
731 |
|
732 |
$maxreq = 4 if $maxreq <= 0; |
733 |
|
734 |
# stat once |
735 |
aioreq_pri $pri; |
736 |
add $grp aio_stat $path, sub { |
737 |
return $grp->result () if $_[0]; |
738 |
my $now = time; |
739 |
my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
740 |
|
741 |
# read the directory entries |
742 |
aioreq_pri $pri; |
743 |
add $grp aio_readdir $path, sub { |
744 |
my $entries = shift |
745 |
or return $grp->result (); |
746 |
|
747 |
# stat the dir another time |
748 |
aioreq_pri $pri; |
749 |
add $grp aio_stat $path, sub { |
750 |
my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
751 |
|
752 |
my $ndirs; |
753 |
|
754 |
# take the slow route if anything looks fishy |
755 |
if ($hash1 ne $hash2 or (stat _)[9] == $now) { |
756 |
$ndirs = -1; |
757 |
} else { |
758 |
# if nlink == 2, we are finished |
759 |
# on non-posix-fs's, we rely on nlink < 2 |
760 |
$ndirs = (stat _)[3] - 2 |
761 |
or return $grp->result ([], $entries); |
762 |
} |
763 |
|
764 |
# sort into likely dirs and likely nondirs |
765 |
# dirs == files without ".", short entries first |
766 |
$entries = [map $_->[0], |
767 |
sort { $b->[1] cmp $a->[1] } |
768 |
map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length], |
769 |
@$entries]; |
770 |
|
771 |
my (@dirs, @nondirs); |
772 |
|
773 |
my $statgrp = add $grp aio_group sub { |
774 |
$grp->result (\@dirs, \@nondirs); |
775 |
}; |
776 |
|
777 |
limit $statgrp $maxreq; |
778 |
feed $statgrp sub { |
779 |
return unless @$entries; |
780 |
my $entry = pop @$entries; |
781 |
|
782 |
aioreq_pri $pri; |
783 |
add $statgrp aio_stat "$path/$entry/.", sub { |
784 |
if ($_[0] < 0) { |
785 |
push @nondirs, $entry; |
786 |
} else { |
787 |
# need to check for real directory |
788 |
aioreq_pri $pri; |
789 |
add $statgrp aio_lstat "$path/$entry", sub { |
790 |
if (-d _) { |
791 |
push @dirs, $entry; |
792 |
|
793 |
unless (--$ndirs) { |
794 |
push @nondirs, @$entries; |
795 |
feed $statgrp; |
796 |
} |
797 |
} else { |
798 |
push @nondirs, $entry; |
799 |
} |
800 |
} |
801 |
} |
802 |
}; |
803 |
}; |
804 |
}; |
805 |
}; |
806 |
}; |
807 |
|
808 |
$grp |
809 |
} |
810 |
|
811 |
=item aio_rmtree $path, $callback->($status) |
812 |
|
813 |
Delete a directory tree starting (and including) C<$path>, return the |
814 |
status of the final C<rmdir> only. This is a composite request that |
815 |
uses C<aio_scandir> to recurse into and rmdir directories, and unlink |
816 |
everything else. |
817 |
|
818 |
=cut |
819 |
|
820 |
sub aio_rmtree; |
821 |
sub aio_rmtree($;$) { |
822 |
my ($path, $cb) = @_; |
823 |
|
824 |
my $pri = aioreq_pri; |
825 |
my $grp = aio_group $cb; |
826 |
|
827 |
aioreq_pri $pri; |
828 |
add $grp aio_scandir $path, 0, sub { |
829 |
my ($dirs, $nondirs) = @_; |
830 |
|
831 |
my $dirgrp = aio_group sub { |
832 |
add $grp aio_rmdir $path, sub { |
833 |
$grp->result ($_[0]); |
834 |
}; |
835 |
}; |
836 |
|
837 |
(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs; |
838 |
(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs; |
839 |
|
840 |
add $grp $dirgrp; |
841 |
}; |
842 |
|
843 |
$grp |
844 |
} |
845 |
|
846 |
=item aio_sync $callback->($status) |
847 |
|
848 |
Asynchronously call sync and call the callback when finished. |
849 |
|
850 |
=item aio_fsync $fh, $callback->($status) |
851 |
|
852 |
Asynchronously call fsync on the given filehandle and call the callback |
853 |
with the fsync result code. |
854 |
|
855 |
=item aio_fdatasync $fh, $callback->($status) |
856 |
|
857 |
Asynchronously call fdatasync on the given filehandle and call the |
858 |
callback with the fdatasync result code. |
859 |
|
860 |
If this call isn't available because your OS lacks it or it couldn't be |
861 |
detected, it will be emulated by calling C<fsync> instead. |
862 |
|
863 |
=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) |
864 |
|
865 |
Sync the data portion of the file specified by C<$offset> and C<$length> |
866 |
to disk (but NOT the metadata), by calling the Linux-specific |
867 |
sync_file_range call. If sync_file_range is not available or it returns |
868 |
ENOSYS, then fdatasync or fsync is being substituted. |
869 |
|
870 |
C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>, |
871 |
C<IO::AIO::SYNC_FILE_RANGE_WRITE> and |
872 |
C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range |
873 |
manpage for details. |
874 |
|
875 |
=item aio_pathsync $path, $callback->($status) |
876 |
|
877 |
This request tries to open, fsync and close the given path. This is a |
878 |
composite request intended to sync directories after directory operations |
879 |
(E.g. rename). This might not work on all operating systems or have any |
880 |
specific effect, but usually it makes sure that directory changes get |
881 |
written to disc. It works for anything that can be opened for read-only, |
882 |
not just directories. |
883 |
|
884 |
Passes C<0> when everything went ok, and C<-1> on error. |
885 |
|
886 |
=cut |
887 |
|
888 |
sub aio_pathsync($;$) { |
889 |
my ($path, $cb) = @_; |
890 |
|
891 |
my $pri = aioreq_pri; |
892 |
my $grp = aio_group $cb; |
893 |
|
894 |
aioreq_pri $pri; |
895 |
add $grp aio_open $path, O_RDONLY, 0, sub { |
896 |
my ($fh) = @_; |
897 |
if ($fh) { |
898 |
aioreq_pri $pri; |
899 |
add $grp aio_fsync $fh, sub { |
900 |
$grp->result ($_[0]); |
901 |
|
902 |
aioreq_pri $pri; |
903 |
add $grp aio_close $fh; |
904 |
}; |
905 |
} else { |
906 |
$grp->result (-1); |
907 |
} |
908 |
}; |
909 |
|
910 |
$grp |
911 |
} |
912 |
|
913 |
=item aio_group $callback->(...) |
914 |
|
915 |
This is a very special aio request: Instead of doing something, it is a |
916 |
container for other aio requests, which is useful if you want to bundle |
917 |
many requests into a single, composite, request with a definite callback |
918 |
and the ability to cancel the whole request with its subrequests. |
919 |
|
920 |
Returns an object of class L<IO::AIO::GRP>. See its documentation below |
921 |
for more info. |
922 |
|
923 |
Example: |
924 |
|
925 |
my $grp = aio_group sub { |
926 |
print "all stats done\n"; |
927 |
}; |
928 |
|
929 |
add $grp |
930 |
(aio_stat ...), |
931 |
(aio_stat ...), |
932 |
...; |
933 |
|
934 |
=item aio_nop $callback->() |
935 |
|
936 |
This is a special request - it does nothing in itself and is only used for |
937 |
side effects, such as when you want to add a dummy request to a group so |
938 |
that finishing the requests in the group depends on executing the given |
939 |
code. |
940 |
|
941 |
While this request does nothing, it still goes through the execution |
942 |
phase and still requires a worker thread. Thus, the callback will not |
943 |
be executed immediately but only after other requests in the queue have |
944 |
entered their execution phase. This can be used to measure request |
945 |
latency. |
946 |
|
947 |
=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED* |
948 |
|
949 |
Mainly used for debugging and benchmarking, this aio request puts one of |
950 |
the request workers to sleep for the given time. |
951 |
|
952 |
While it is theoretically handy to have simple I/O scheduling requests |
953 |
like sleep and file handle readable/writable, the overhead this creates is |
954 |
immense (it blocks a thread for a long time) so do not use this function |
955 |
except to put your application under artificial I/O pressure. |
956 |
|
957 |
=back |
958 |
|
959 |
=head2 IO::AIO::REQ CLASS |
960 |
|
961 |
All non-aggregate C<aio_*> functions return an object of this class when |
962 |
called in non-void context. |
963 |
|
964 |
=over 4 |
965 |
|
966 |
=item cancel $req |
967 |
|
968 |
Cancels the request, if possible. Has the effect of skipping execution |
969 |
when entering the B<execute> state and skipping calling the callback when |
970 |
entering the the B<result> state, but will leave the request otherwise |
971 |
untouched. That means that requests that currently execute will not be |
972 |
stopped and resources held by the request will not be freed prematurely. |
973 |
|
974 |
=item cb $req $callback->(...) |
975 |
|
976 |
Replace (or simply set) the callback registered to the request. |
977 |
|
978 |
=back |
979 |
|
980 |
=head2 IO::AIO::GRP CLASS |
981 |
|
982 |
This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to |
983 |
objects of this class, too. |
984 |
|
985 |
A IO::AIO::GRP object is a special request that can contain multiple other |
986 |
aio requests. |
987 |
|
988 |
You create one by calling the C<aio_group> constructing function with a |
989 |
callback that will be called when all contained requests have entered the |
990 |
C<done> state: |
991 |
|
992 |
my $grp = aio_group sub { |
993 |
print "all requests are done\n"; |
994 |
}; |
995 |
|
996 |
You add requests by calling the C<add> method with one or more |
997 |
C<IO::AIO::REQ> objects: |
998 |
|
999 |
$grp->add (aio_unlink "..."); |
1000 |
|
1001 |
add $grp aio_stat "...", sub { |
1002 |
$_[0] or return $grp->result ("error"); |
1003 |
|
1004 |
# add another request dynamically, if first succeeded |
1005 |
add $grp aio_open "...", sub { |
1006 |
$grp->result ("ok"); |
1007 |
}; |
1008 |
}; |
1009 |
|
1010 |
This makes it very easy to create composite requests (see the source of |
1011 |
C<aio_move> for an application) that work and feel like simple requests. |
1012 |
|
1013 |
=over 4 |
1014 |
|
1015 |
=item * The IO::AIO::GRP objects will be cleaned up during calls to |
1016 |
C<IO::AIO::poll_cb>, just like any other request. |
1017 |
|
1018 |
=item * They can be canceled like any other request. Canceling will cancel not |
1019 |
only the request itself, but also all requests it contains. |
1020 |
|
1021 |
=item * They can also can also be added to other IO::AIO::GRP objects. |
1022 |
|
1023 |
=item * You must not add requests to a group from within the group callback (or |
1024 |
any later time). |
1025 |
|
1026 |
=back |
1027 |
|
1028 |
Their lifetime, simplified, looks like this: when they are empty, they |
1029 |
will finish very quickly. If they contain only requests that are in the |
1030 |
C<done> state, they will also finish. Otherwise they will continue to |
1031 |
exist. |
1032 |
|
1033 |
That means after creating a group you have some time to add requests |
1034 |
(precisely before the callback has been invoked, which is only done within |
1035 |
the C<poll_cb>). And in the callbacks of those requests, you can add |
1036 |
further requests to the group. And only when all those requests have |
1037 |
finished will the the group itself finish. |
1038 |
|
1039 |
=over 4 |
1040 |
|
1041 |
=item add $grp ... |
1042 |
|
1043 |
=item $grp->add (...) |
1044 |
|
1045 |
Add one or more requests to the group. Any type of L<IO::AIO::REQ> can |
1046 |
be added, including other groups, as long as you do not create circular |
1047 |
dependencies. |
1048 |
|
1049 |
Returns all its arguments. |
1050 |
|
1051 |
=item $grp->cancel_subs |
1052 |
|
1053 |
Cancel all subrequests and clears any feeder, but not the group request |
1054 |
itself. Useful when you queued a lot of events but got a result early. |
1055 |
|
1056 |
=item $grp->result (...) |
1057 |
|
1058 |
Set the result value(s) that will be passed to the group callback when all |
1059 |
subrequests have finished and set the groups errno to the current value |
1060 |
of errno (just like calling C<errno> without an error number). By default, |
1061 |
no argument will be passed and errno is zero. |
1062 |
|
1063 |
=item $grp->errno ([$errno]) |
1064 |
|
1065 |
Sets the group errno value to C<$errno>, or the current value of errno |
1066 |
when the argument is missing. |
1067 |
|
1068 |
Every aio request has an associated errno value that is restored when |
1069 |
the callback is invoked. This method lets you change this value from its |
1070 |
default (0). |
1071 |
|
1072 |
Calling C<result> will also set errno, so make sure you either set C<$!> |
1073 |
before the call to C<result>, or call c<errno> after it. |
1074 |
|
1075 |
=item feed $grp $callback->($grp) |
1076 |
|
1077 |
Sets a feeder/generator on this group: every group can have an attached |
1078 |
generator that generates requests if idle. The idea behind this is that, |
1079 |
although you could just queue as many requests as you want in a group, |
1080 |
this might starve other requests for a potentially long time. For example, |
1081 |
C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests, |
1082 |
delaying any later requests for a long time. |
1083 |
|
1084 |
To avoid this, and allow incremental generation of requests, you can |
1085 |
instead a group and set a feeder on it that generates those requests. The |
1086 |
feed callback will be called whenever there are few enough (see C<limit>, |
1087 |
below) requests active in the group itself and is expected to queue more |
1088 |
requests. |
1089 |
|
1090 |
The feed callback can queue as many requests as it likes (i.e. C<add> does |
1091 |
not impose any limits). |
1092 |
|
1093 |
If the feed does not queue more requests when called, it will be |
1094 |
automatically removed from the group. |
1095 |
|
1096 |
If the feed limit is C<0> when this method is called, it will be set to |
1097 |
C<2> automatically. |
1098 |
|
1099 |
Example: |
1100 |
|
1101 |
# stat all files in @files, but only ever use four aio requests concurrently: |
1102 |
|
1103 |
my $grp = aio_group sub { print "finished\n" }; |
1104 |
limit $grp 4; |
1105 |
feed $grp sub { |
1106 |
my $file = pop @files |
1107 |
or return; |
1108 |
|
1109 |
add $grp aio_stat $file, sub { ... }; |
1110 |
}; |
1111 |
|
1112 |
=item limit $grp $num |
1113 |
|
1114 |
Sets the feeder limit for the group: The feeder will be called whenever |
1115 |
the group contains less than this many requests. |
1116 |
|
1117 |
Setting the limit to C<0> will pause the feeding process. |
1118 |
|
1119 |
The default value for the limit is C<0>, but note that setting a feeder |
1120 |
automatically bumps it up to C<2>. |
1121 |
|
1122 |
=back |
1123 |
|
1124 |
=head2 SUPPORT FUNCTIONS |
1125 |
|
1126 |
=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION |
1127 |
|
1128 |
=over 4 |
1129 |
|
1130 |
=item $fileno = IO::AIO::poll_fileno |
1131 |
|
1132 |
Return the I<request result pipe file descriptor>. This filehandle must be |
1133 |
polled for reading by some mechanism outside this module (e.g. Event or |
1134 |
select, see below or the SYNOPSIS). If the pipe becomes readable you have |
1135 |
to call C<poll_cb> to check the results. |
1136 |
|
1137 |
See C<poll_cb> for an example. |
1138 |
|
1139 |
=item IO::AIO::poll_cb |
1140 |
|
1141 |
Process some outstanding events on the result pipe. You have to call this |
1142 |
regularly. Returns C<0> if all events could be processed, or C<-1> if it |
1143 |
returned earlier for whatever reason. Returns immediately when no events |
1144 |
are outstanding. The amount of events processed depends on the settings of |
1145 |
C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>. |
1146 |
|
1147 |
If not all requests were processed for whatever reason, the filehandle |
1148 |
will still be ready when C<poll_cb> returns, so normally you don't have to |
1149 |
do anything special to have it called later. |
1150 |
|
1151 |
Example: Install an Event watcher that automatically calls |
1152 |
IO::AIO::poll_cb with high priority: |
1153 |
|
1154 |
Event->io (fd => IO::AIO::poll_fileno, |
1155 |
poll => 'r', async => 1, |
1156 |
cb => \&IO::AIO::poll_cb); |
1157 |
|
1158 |
=item IO::AIO::max_poll_reqs $nreqs |
1159 |
|
1160 |
=item IO::AIO::max_poll_time $seconds |
1161 |
|
1162 |
These set the maximum number of requests (default C<0>, meaning infinity) |
1163 |
that are being processed by C<IO::AIO::poll_cb> in one call, respectively |
1164 |
the maximum amount of time (default C<0>, meaning infinity) spent in |
1165 |
C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount |
1166 |
of time C<poll_cb> is allowed to use). |
1167 |
|
1168 |
Setting C<max_poll_time> to a non-zero value creates an overhead of one |
1169 |
syscall per request processed, which is not normally a problem unless your |
1170 |
callbacks are really really fast or your OS is really really slow (I am |
1171 |
not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead. |
1172 |
|
1173 |
Setting these is useful if you want to ensure some level of |
1174 |
interactiveness when perl is not fast enough to process all requests in |
1175 |
time. |
1176 |
|
1177 |
For interactive programs, values such as C<0.01> to C<0.1> should be fine. |
1178 |
|
1179 |
Example: Install an Event watcher that automatically calls |
1180 |
IO::AIO::poll_cb with low priority, to ensure that other parts of the |
1181 |
program get the CPU sometimes even under high AIO load. |
1182 |
|
1183 |
# try not to spend much more than 0.1s in poll_cb |
1184 |
IO::AIO::max_poll_time 0.1; |
1185 |
|
1186 |
# use a low priority so other tasks have priority |
1187 |
Event->io (fd => IO::AIO::poll_fileno, |
1188 |
poll => 'r', nice => 1, |
1189 |
cb => &IO::AIO::poll_cb); |
1190 |
|
1191 |
=item IO::AIO::poll_wait |
1192 |
|
1193 |
If there are any outstanding requests and none of them in the result |
1194 |
phase, wait till the result filehandle becomes ready for reading (simply |
1195 |
does a C<select> on the filehandle. This is useful if you want to |
1196 |
synchronously wait for some requests to finish). |
1197 |
|
1198 |
See C<nreqs> for an example. |
1199 |
|
1200 |
=item IO::AIO::poll |
1201 |
|
1202 |
Waits until some requests have been handled. |
1203 |
|
1204 |
Returns the number of requests processed, but is otherwise strictly |
1205 |
equivalent to: |
1206 |
|
1207 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1208 |
|
1209 |
=item IO::AIO::flush |
1210 |
|
1211 |
Wait till all outstanding AIO requests have been handled. |
1212 |
|
1213 |
Strictly equivalent to: |
1214 |
|
1215 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1216 |
while IO::AIO::nreqs; |
1217 |
|
1218 |
=back |
1219 |
|
1220 |
=head3 CONTROLLING THE NUMBER OF THREADS |
1221 |
|
1222 |
=over |
1223 |
|
1224 |
=item IO::AIO::min_parallel $nthreads |
1225 |
|
1226 |
Set the minimum number of AIO threads to C<$nthreads>. The current |
1227 |
default is C<8>, which means eight asynchronous operations can execute |
1228 |
concurrently at any one time (the number of outstanding requests, |
1229 |
however, is unlimited). |
1230 |
|
1231 |
IO::AIO starts threads only on demand, when an AIO request is queued and |
1232 |
no free thread exists. Please note that queueing up a hundred requests can |
1233 |
create demand for a hundred threads, even if it turns out that everything |
1234 |
is in the cache and could have been processed faster by a single thread. |
1235 |
|
1236 |
It is recommended to keep the number of threads relatively low, as some |
1237 |
Linux kernel versions will scale negatively with the number of threads |
1238 |
(higher parallelity => MUCH higher latency). With current Linux 2.6 |
1239 |
versions, 4-32 threads should be fine. |
1240 |
|
1241 |
Under most circumstances you don't need to call this function, as the |
1242 |
module selects a default that is suitable for low to moderate load. |
1243 |
|
1244 |
=item IO::AIO::max_parallel $nthreads |
1245 |
|
1246 |
Sets the maximum number of AIO threads to C<$nthreads>. If more than the |
1247 |
specified number of threads are currently running, this function kills |
1248 |
them. This function blocks until the limit is reached. |
1249 |
|
1250 |
While C<$nthreads> are zero, aio requests get queued but not executed |
1251 |
until the number of threads has been increased again. |
1252 |
|
1253 |
This module automatically runs C<max_parallel 0> at program end, to ensure |
1254 |
that all threads are killed and that there are no outstanding requests. |
1255 |
|
1256 |
Under normal circumstances you don't need to call this function. |
1257 |
|
1258 |
=item IO::AIO::max_idle $nthreads |
1259 |
|
1260 |
Limit the number of threads (default: 4) that are allowed to idle (i.e., |
1261 |
threads that did not get a request to process within 10 seconds). That |
1262 |
means if a thread becomes idle while C<$nthreads> other threads are also |
1263 |
idle, it will free its resources and exit. |
1264 |
|
1265 |
This is useful when you allow a large number of threads (e.g. 100 or 1000) |
1266 |
to allow for extremely high load situations, but want to free resources |
1267 |
under normal circumstances (1000 threads can easily consume 30MB of RAM). |
1268 |
|
1269 |
The default is probably ok in most situations, especially if thread |
1270 |
creation is fast. If thread creation is very slow on your system you might |
1271 |
want to use larger values. |
1272 |
|
1273 |
=item IO::AIO::max_outstanding $maxreqs |
1274 |
|
1275 |
This is a very bad function to use in interactive programs because it |
1276 |
blocks, and a bad way to reduce concurrency because it is inexact: Better |
1277 |
use an C<aio_group> together with a feed callback. |
1278 |
|
1279 |
Sets the maximum number of outstanding requests to C<$nreqs>. If you |
1280 |
do queue up more than this number of requests, the next call to the |
1281 |
C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>) |
1282 |
function will block until the limit is no longer exceeded. |
1283 |
|
1284 |
The default value is very large, so there is no practical limit on the |
1285 |
number of outstanding requests. |
1286 |
|
1287 |
You can still queue as many requests as you want. Therefore, |
1288 |
C<max_outstanding> is mainly useful in simple scripts (with low values) or |
1289 |
as a stop gap to shield against fatal memory overflow (with large values). |
1290 |
|
1291 |
=back |
1292 |
|
1293 |
=head3 STATISTICAL INFORMATION |
1294 |
|
1295 |
=over |
1296 |
|
1297 |
=item IO::AIO::nreqs |
1298 |
|
1299 |
Returns the number of requests currently in the ready, execute or pending |
1300 |
states (i.e. for which their callback has not been invoked yet). |
1301 |
|
1302 |
Example: wait till there are no outstanding requests anymore: |
1303 |
|
1304 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1305 |
while IO::AIO::nreqs; |
1306 |
|
1307 |
=item IO::AIO::nready |
1308 |
|
1309 |
Returns the number of requests currently in the ready state (not yet |
1310 |
executed). |
1311 |
|
1312 |
=item IO::AIO::npending |
1313 |
|
1314 |
Returns the number of requests currently in the pending state (executed, |
1315 |
but not yet processed by poll_cb). |
1316 |
|
1317 |
=back |
1318 |
|
1319 |
=cut |
1320 |
|
1321 |
min_parallel 8; |
1322 |
|
1323 |
END { flush } |
1324 |
|
1325 |
1; |
1326 |
|
1327 |
=head2 FORK BEHAVIOUR |
1328 |
|
1329 |
This module should do "the right thing" when the process using it forks: |
1330 |
|
1331 |
Before the fork, IO::AIO enters a quiescent state where no requests |
1332 |
can be added in other threads and no results will be processed. After |
1333 |
the fork the parent simply leaves the quiescent state and continues |
1334 |
request/result processing, while the child frees the request/result queue |
1335 |
(so that the requests started before the fork will only be handled in the |
1336 |
parent). Threads will be started on demand until the limit set in the |
1337 |
parent process has been reached again. |
1338 |
|
1339 |
In short: the parent will, after a short pause, continue as if fork had |
1340 |
not been called, while the child will act as if IO::AIO has not been used |
1341 |
yet. |
1342 |
|
1343 |
=head2 MEMORY USAGE |
1344 |
|
1345 |
Per-request usage: |
1346 |
|
1347 |
Each aio request uses - depending on your architecture - around 100-200 |
1348 |
bytes of memory. In addition, stat requests need a stat buffer (possibly |
1349 |
a few hundred bytes), readdir requires a result buffer and so on. Perl |
1350 |
scalars and other data passed into aio requests will also be locked and |
1351 |
will consume memory till the request has entered the done state. |
1352 |
|
1353 |
This is not awfully much, so queuing lots of requests is not usually a |
1354 |
problem. |
1355 |
|
1356 |
Per-thread usage: |
1357 |
|
1358 |
In the execution phase, some aio requests require more memory for |
1359 |
temporary buffers, and each thread requires a stack and other data |
1360 |
structures (usually around 16k-128k, depending on the OS). |
1361 |
|
1362 |
=head1 KNOWN BUGS |
1363 |
|
1364 |
Known bugs will be fixed in the next release. |
1365 |
|
1366 |
=head1 SEE ALSO |
1367 |
|
1368 |
L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a |
1369 |
more natural syntax. |
1370 |
|
1371 |
=head1 AUTHOR |
1372 |
|
1373 |
Marc Lehmann <schmorp@schmorp.de> |
1374 |
http://home.schmorp.de/ |
1375 |
|
1376 |
=cut |
1377 |
|