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", IO::AIO::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 |
=head1 DESCRIPTION |
32 |
|
33 |
This module implements asynchronous I/O using whatever means your |
34 |
operating system supports. It is implemented as an interface to C<libeio> |
35 |
(L<http://software.schmorp.de/pkg/libeio.html>). |
36 |
|
37 |
Asynchronous means that operations that can normally block your program |
38 |
(e.g. reading from disk) will be done asynchronously: the operation |
39 |
will still block, but you can do something else in the meantime. This |
40 |
is extremely useful for programs that need to stay interactive even |
41 |
when doing heavy I/O (GUI programs, high performance network servers |
42 |
etc.), but can also be used to easily do operations in parallel that are |
43 |
normally done sequentially, e.g. stat'ing many files, which is much faster |
44 |
on a RAID volume or over NFS when you do a number of stat operations |
45 |
concurrently. |
46 |
|
47 |
While most of this works on all types of file descriptors (for |
48 |
example sockets), using these functions on file descriptors that |
49 |
support nonblocking operation (again, sockets, pipes etc.) is |
50 |
very inefficient. Use an event loop for that (such as the L<EV> |
51 |
module): IO::AIO will naturally fit into such an event loop itself. |
52 |
|
53 |
In this version, a number of threads are started that execute your |
54 |
requests and signal their completion. You don't need thread support |
55 |
in perl, and the threads created by this module will not be visible |
56 |
to perl. In the future, this module might make use of the native aio |
57 |
functions available on many operating systems. However, they are often |
58 |
not well-supported or restricted (GNU/Linux doesn't allow them on normal |
59 |
files currently, for example), and they would only support aio_read and |
60 |
aio_write, so the remaining functionality would have to be implemented |
61 |
using threads anyway. |
62 |
|
63 |
Although the module will work in the presence of other (Perl-) threads, |
64 |
it is currently not reentrant in any way, so use appropriate locking |
65 |
yourself, always call C<poll_cb> from within the same thread, or never |
66 |
call C<poll_cb> (or other C<aio_> functions) recursively. |
67 |
|
68 |
=head2 EXAMPLE |
69 |
|
70 |
This is a simple example that uses the EV module and loads |
71 |
F</etc/passwd> asynchronously: |
72 |
|
73 |
use Fcntl; |
74 |
use EV; |
75 |
use IO::AIO; |
76 |
|
77 |
# register the IO::AIO callback with EV |
78 |
my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; |
79 |
|
80 |
# queue the request to open /etc/passwd |
81 |
aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub { |
82 |
my $fh = shift |
83 |
or die "error while opening: $!"; |
84 |
|
85 |
# stat'ing filehandles is generally non-blocking |
86 |
my $size = -s $fh; |
87 |
|
88 |
# queue a request to read the file |
89 |
my $contents; |
90 |
aio_read $fh, 0, $size, $contents, 0, sub { |
91 |
$_[0] == $size |
92 |
or die "short read: $!"; |
93 |
|
94 |
close $fh; |
95 |
|
96 |
# file contents now in $contents |
97 |
print $contents; |
98 |
|
99 |
# exit event loop and program |
100 |
EV::unloop; |
101 |
}; |
102 |
}; |
103 |
|
104 |
# possibly queue up other requests, or open GUI windows, |
105 |
# check for sockets etc. etc. |
106 |
|
107 |
# process events as long as there are some: |
108 |
EV::loop; |
109 |
|
110 |
=head1 REQUEST ANATOMY AND LIFETIME |
111 |
|
112 |
Every C<aio_*> function creates a request. which is a C data structure not |
113 |
directly visible to Perl. |
114 |
|
115 |
If called in non-void context, every request function returns a Perl |
116 |
object representing the request. In void context, nothing is returned, |
117 |
which saves a bit of memory. |
118 |
|
119 |
The perl object is a fairly standard ref-to-hash object. The hash contents |
120 |
are not used by IO::AIO so you are free to store anything you like in it. |
121 |
|
122 |
During their existance, aio requests travel through the following states, |
123 |
in order: |
124 |
|
125 |
=over 4 |
126 |
|
127 |
=item ready |
128 |
|
129 |
Immediately after a request is created it is put into the ready state, |
130 |
waiting for a thread to execute it. |
131 |
|
132 |
=item execute |
133 |
|
134 |
A thread has accepted the request for processing and is currently |
135 |
executing it (e.g. blocking in read). |
136 |
|
137 |
=item pending |
138 |
|
139 |
The request has been executed and is waiting for result processing. |
140 |
|
141 |
While request submission and execution is fully asynchronous, result |
142 |
processing is not and relies on the perl interpreter calling C<poll_cb> |
143 |
(or another function with the same effect). |
144 |
|
145 |
=item result |
146 |
|
147 |
The request results are processed synchronously by C<poll_cb>. |
148 |
|
149 |
The C<poll_cb> function will process all outstanding aio requests by |
150 |
calling their callbacks, freeing memory associated with them and managing |
151 |
any groups they are contained in. |
152 |
|
153 |
=item done |
154 |
|
155 |
Request has reached the end of its lifetime and holds no resources anymore |
156 |
(except possibly for the Perl object, but its connection to the actual |
157 |
aio request is severed and calling its methods will either do nothing or |
158 |
result in a runtime error). |
159 |
|
160 |
=back |
161 |
|
162 |
=cut |
163 |
|
164 |
package IO::AIO; |
165 |
|
166 |
use Carp (); |
167 |
|
168 |
use common::sense; |
169 |
|
170 |
use base 'Exporter'; |
171 |
|
172 |
BEGIN { |
173 |
our $VERSION = '4.0'; |
174 |
|
175 |
our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close |
176 |
aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx |
177 |
aio_scandir aio_symlink aio_readlink aio_realpath aio_sync |
178 |
aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range aio_fallocate |
179 |
aio_pathsync aio_readahead |
180 |
aio_rename aio_link aio_move aio_copy aio_group |
181 |
aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown |
182 |
aio_chmod aio_utime aio_truncate |
183 |
aio_msync aio_mtouch aio_mlock aio_mlockall |
184 |
aio_statvfs |
185 |
aio_wd); |
186 |
|
187 |
our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice)); |
188 |
our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush |
189 |
min_parallel max_parallel max_idle idle_timeout |
190 |
nreqs nready npending nthreads |
191 |
max_poll_time max_poll_reqs |
192 |
sendfile fadvise madvise |
193 |
mmap munmap munlock munlockall); |
194 |
|
195 |
push @AIO_REQ, qw(aio_busy); # not exported |
196 |
|
197 |
@IO::AIO::GRP::ISA = 'IO::AIO::REQ'; |
198 |
|
199 |
require XSLoader; |
200 |
XSLoader::load ("IO::AIO", $VERSION); |
201 |
} |
202 |
|
203 |
=head1 FUNCTIONS |
204 |
|
205 |
=head2 QUICK OVERVIEW |
206 |
|
207 |
This section simply lists the prototypes of the most important functions |
208 |
for quick reference. See the following sections for function-by-function |
209 |
documentation. |
210 |
|
211 |
aio_wd $pathname, $callback->($wd) |
212 |
aio_open $pathname, $flags, $mode, $callback->($fh) |
213 |
aio_close $fh, $callback->($status) |
214 |
aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
215 |
aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
216 |
aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
217 |
aio_readahead $fh,$offset,$length, $callback->($retval) |
218 |
aio_stat $fh_or_path, $callback->($status) |
219 |
aio_lstat $fh, $callback->($status) |
220 |
aio_statvfs $fh_or_path, $callback->($statvfs) |
221 |
aio_utime $fh_or_path, $atime, $mtime, $callback->($status) |
222 |
aio_chown $fh_or_path, $uid, $gid, $callback->($status) |
223 |
aio_truncate $fh_or_path, $offset, $callback->($status) |
224 |
aio_chmod $fh_or_path, $mode, $callback->($status) |
225 |
aio_unlink $pathname, $callback->($status) |
226 |
aio_mknod $path, $mode, $dev, $callback->($status) |
227 |
aio_link $srcpath, $dstpath, $callback->($status) |
228 |
aio_symlink $srcpath, $dstpath, $callback->($status) |
229 |
aio_readlink $path, $callback->($link) |
230 |
aio_realpath $path, $callback->($link) |
231 |
aio_rename $srcpath, $dstpath, $callback->($status) |
232 |
aio_mkdir $pathname, $mode, $callback->($status) |
233 |
aio_rmdir $pathname, $callback->($status) |
234 |
aio_readdir $pathname, $callback->($entries) |
235 |
aio_readdirx $pathname, $flags, $callback->($entries, $flags) |
236 |
IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST |
237 |
IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN |
238 |
aio_load $path, $data, $callback->($status) |
239 |
aio_copy $srcpath, $dstpath, $callback->($status) |
240 |
aio_move $srcpath, $dstpath, $callback->($status) |
241 |
aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
242 |
aio_rmtree $path, $callback->($status) |
243 |
aio_sync $callback->($status) |
244 |
aio_syncfs $fh, $callback->($status) |
245 |
aio_fsync $fh, $callback->($status) |
246 |
aio_fdatasync $fh, $callback->($status) |
247 |
aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) |
248 |
aio_pathsync $path, $callback->($status) |
249 |
aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) |
250 |
aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) |
251 |
aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) |
252 |
aio_mlockall $flags, $callback->($status) |
253 |
aio_group $callback->(...) |
254 |
aio_nop $callback->() |
255 |
|
256 |
$prev_pri = aioreq_pri [$pri] |
257 |
aioreq_nice $pri_adjust |
258 |
|
259 |
IO::AIO::poll_wait |
260 |
IO::AIO::poll_cb |
261 |
IO::AIO::poll |
262 |
IO::AIO::flush |
263 |
IO::AIO::max_poll_reqs $nreqs |
264 |
IO::AIO::max_poll_time $seconds |
265 |
IO::AIO::min_parallel $nthreads |
266 |
IO::AIO::max_parallel $nthreads |
267 |
IO::AIO::max_idle $nthreads |
268 |
IO::AIO::idle_timeout $seconds |
269 |
IO::AIO::max_outstanding $maxreqs |
270 |
IO::AIO::nreqs |
271 |
IO::AIO::nready |
272 |
IO::AIO::npending |
273 |
|
274 |
IO::AIO::sendfile $ofh, $ifh, $offset, $count |
275 |
IO::AIO::fadvise $fh, $offset, $len, $advice |
276 |
IO::AIO::madvise $scalar, $offset, $length, $advice |
277 |
IO::AIO::mprotect $scalar, $offset, $length, $protect |
278 |
IO::AIO::munlock $scalar, $offset = 0, $length = undef |
279 |
IO::AIO::munlockall |
280 |
|
281 |
=head2 AIO REQUEST FUNCTIONS |
282 |
|
283 |
All the C<aio_*> calls are more or less thin wrappers around the syscall |
284 |
with the same name (sans C<aio_>). The arguments are similar or identical, |
285 |
and they all accept an additional (and optional) C<$callback> argument |
286 |
which must be a code reference. This code reference will get called with |
287 |
the syscall return code (e.g. most syscalls return C<-1> on error, unlike |
288 |
perl, which usually delivers "false") as its sole argument after the given |
289 |
syscall has been executed asynchronously. |
290 |
|
291 |
All functions expecting a filehandle keep a copy of the filehandle |
292 |
internally until the request has finished. |
293 |
|
294 |
All functions return request objects of type L<IO::AIO::REQ> that allow |
295 |
further manipulation of those requests while they are in-flight. |
296 |
|
297 |
The pathnames you pass to these routines I<must> be absolute and |
298 |
encoded as octets. The reason for the former is that at the time the |
299 |
request is being executed, the current working directory could have |
300 |
changed. Alternatively, you can make sure that you never change the |
301 |
current working directory anywhere in the program and then use relative |
302 |
paths. |
303 |
|
304 |
To encode pathnames as octets, either make sure you either: a) always pass |
305 |
in filenames you got from outside (command line, readdir etc.) without |
306 |
tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode |
307 |
your pathnames to the locale (or other) encoding in effect in the user |
308 |
environment, d) use Glib::filename_from_unicode on unicode filenames or e) |
309 |
use something else to ensure your scalar has the correct contents. |
310 |
|
311 |
This works, btw. independent of the internal UTF-8 bit, which IO::AIO |
312 |
handles correctly whether it is set or not. |
313 |
|
314 |
=over 4 |
315 |
|
316 |
=item $prev_pri = aioreq_pri [$pri] |
317 |
|
318 |
Returns the priority value that would be used for the next request and, if |
319 |
C<$pri> is given, sets the priority for the next aio request. |
320 |
|
321 |
The default priority is C<0>, the minimum and maximum priorities are C<-4> |
322 |
and C<4>, respectively. Requests with higher priority will be serviced |
323 |
first. |
324 |
|
325 |
The priority will be reset to C<0> after each call to one of the C<aio_*> |
326 |
functions. |
327 |
|
328 |
Example: open a file with low priority, then read something from it with |
329 |
higher priority so the read request is serviced before other low priority |
330 |
open requests (potentially spamming the cache): |
331 |
|
332 |
aioreq_pri -3; |
333 |
aio_open ..., sub { |
334 |
return unless $_[0]; |
335 |
|
336 |
aioreq_pri -2; |
337 |
aio_read $_[0], ..., sub { |
338 |
... |
339 |
}; |
340 |
}; |
341 |
|
342 |
|
343 |
=item aioreq_nice $pri_adjust |
344 |
|
345 |
Similar to C<aioreq_pri>, but subtracts the given value from the current |
346 |
priority, so the effect is cumulative. |
347 |
|
348 |
|
349 |
=item aio_open $pathname, $flags, $mode, $callback->($fh) |
350 |
|
351 |
Asynchronously open or create a file and call the callback with a newly |
352 |
created filehandle for the file. |
353 |
|
354 |
The pathname passed to C<aio_open> must be absolute. See API NOTES, above, |
355 |
for an explanation. |
356 |
|
357 |
The C<$flags> argument is a bitmask. See the C<Fcntl> module for a |
358 |
list. They are the same as used by C<sysopen>. |
359 |
|
360 |
Likewise, C<$mode> specifies the mode of the newly created file, if it |
361 |
didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>, |
362 |
except that it is mandatory (i.e. use C<0> if you don't create new files, |
363 |
and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified |
364 |
by the umask in effect then the request is being executed, so better never |
365 |
change the umask. |
366 |
|
367 |
Example: |
368 |
|
369 |
aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub { |
370 |
if ($_[0]) { |
371 |
print "open successful, fh is $_[0]\n"; |
372 |
... |
373 |
} else { |
374 |
die "open failed: $!\n"; |
375 |
} |
376 |
}; |
377 |
|
378 |
In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>, |
379 |
C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the |
380 |
following POSIX and non-POSIX constants are available (missing ones on |
381 |
your system are, as usual, C<0>): |
382 |
|
383 |
C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>, |
384 |
C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>, |
385 |
C<O_RSYNC>, C<O_SYNC> and C<O_TTY_INIT>. |
386 |
|
387 |
|
388 |
=item aio_close $fh, $callback->($status) |
389 |
|
390 |
Asynchronously close a file and call the callback with the result |
391 |
code. |
392 |
|
393 |
Unfortunately, you can't do this to perl. Perl I<insists> very strongly on |
394 |
closing the file descriptor associated with the filehandle itself. |
395 |
|
396 |
Therefore, C<aio_close> will not close the filehandle - instead it will |
397 |
use dup2 to overwrite the file descriptor with the write-end of a pipe |
398 |
(the pipe fd will be created on demand and will be cached). |
399 |
|
400 |
Or in other words: the file descriptor will be closed, but it will not be |
401 |
free for reuse until the perl filehandle is closed. |
402 |
|
403 |
=cut |
404 |
|
405 |
=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
406 |
|
407 |
=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
408 |
|
409 |
Reads or writes C<$length> bytes from or to the specified C<$fh> and |
410 |
C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> |
411 |
and calls the callback without the actual number of bytes read (or -1 on |
412 |
error, just like the syscall). |
413 |
|
414 |
C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to |
415 |
offset plus the actual number of bytes read. |
416 |
|
417 |
If C<$offset> is undefined, then the current file descriptor offset will |
418 |
be used (and updated), otherwise the file descriptor offset will not be |
419 |
changed by these calls. |
420 |
|
421 |
If C<$length> is undefined in C<aio_write>, use the remaining length of |
422 |
C<$data>. |
423 |
|
424 |
If C<$dataoffset> is less than zero, it will be counted from the end of |
425 |
C<$data>. |
426 |
|
427 |
The C<$data> scalar I<MUST NOT> be modified in any way while the request |
428 |
is outstanding. Modifying it can result in segfaults or World War III (if |
429 |
the necessary/optional hardware is installed). |
430 |
|
431 |
Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at |
432 |
offset C<0> within the scalar: |
433 |
|
434 |
aio_read $fh, 7, 15, $buffer, 0, sub { |
435 |
$_[0] > 0 or die "read error: $!"; |
436 |
print "read $_[0] bytes: <$buffer>\n"; |
437 |
}; |
438 |
|
439 |
|
440 |
=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
441 |
|
442 |
Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts |
443 |
reading at byte offset C<$in_offset>, and starts writing at the current |
444 |
file offset of C<$out_fh>. Because of that, it is not safe to issue more |
445 |
than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each |
446 |
other. The same C<$in_fh> works fine though, as this function does not |
447 |
move or use the file offset of C<$in_fh>. |
448 |
|
449 |
Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than |
450 |
are written, and there is no way to find out how many more bytes have been |
451 |
read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the |
452 |
number of bytes written to C<$out_fh>. Only if the result value equals |
453 |
C<$length> one can assume that C<$length> bytes have been read. |
454 |
|
455 |
Unlike with other C<aio_> functions, it makes a lot of sense to use |
456 |
C<aio_sendfile> on non-blocking sockets, as long as one end (typically |
457 |
the C<$in_fh>) is a file - the file I/O will then be asynchronous, while |
458 |
the socket I/O will be non-blocking. Note, however, that you can run |
459 |
into a trap where C<aio_sendfile> reads some data with readahead, then |
460 |
fails to write all data, and when the socket is ready the next time, the |
461 |
data in the cache is already lost, forcing C<aio_sendfile> to again hit |
462 |
the disk. Explicit C<aio_read> + C<aio_write> let's you better control |
463 |
resource usage. |
464 |
|
465 |
This call tries to make use of a native C<sendfile>-like syscall to |
466 |
provide zero-copy operation. For this to work, C<$out_fh> should refer to |
467 |
a socket, and C<$in_fh> should refer to an mmap'able file. |
468 |
|
469 |
If a native sendfile cannot be found or it fails with C<ENOSYS>, |
470 |
C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or |
471 |
C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any |
472 |
type of filehandle regardless of the limitations of the operating system. |
473 |
|
474 |
As native sendfile syscalls (as practically any non-POSIX interface hacked |
475 |
together in a hurry to improve benchmark numbers) tend to be rather buggy |
476 |
on many systems, this implementation tries to work around some known bugs |
477 |
in Linux and FreeBSD kernels (probably others, too), but that might fail, |
478 |
so you really really should check the return value of C<aio_sendfile> - |
479 |
fewre bytes than expected might have been transferred. |
480 |
|
481 |
|
482 |
=item aio_readahead $fh,$offset,$length, $callback->($retval) |
483 |
|
484 |
C<aio_readahead> populates the page cache with data from a file so that |
485 |
subsequent reads from that file will not block on disk I/O. The C<$offset> |
486 |
argument specifies the starting point from which data is to be read and |
487 |
C<$length> specifies the number of bytes to be read. I/O is performed in |
488 |
whole pages, so that offset is effectively rounded down to a page boundary |
489 |
and bytes are read up to the next page boundary greater than or equal to |
490 |
(off-set+length). C<aio_readahead> does not read beyond the end of the |
491 |
file. The current file offset of the file is left unchanged. |
492 |
|
493 |
If that syscall doesn't exist (likely if your OS isn't Linux) it will be |
494 |
emulated by simply reading the data, which would have a similar effect. |
495 |
|
496 |
|
497 |
=item aio_stat $fh_or_path, $callback->($status) |
498 |
|
499 |
=item aio_lstat $fh, $callback->($status) |
500 |
|
501 |
Works like perl's C<stat> or C<lstat> in void context. The callback will |
502 |
be called after the stat and the results will be available using C<stat _> |
503 |
or C<-s _> etc... |
504 |
|
505 |
The pathname passed to C<aio_stat> must be absolute. See API NOTES, above, |
506 |
for an explanation. |
507 |
|
508 |
Currently, the stats are always 64-bit-stats, i.e. instead of returning an |
509 |
error when stat'ing a large file, the results will be silently truncated |
510 |
unless perl itself is compiled with large file support. |
511 |
|
512 |
To help interpret the mode and dev/rdev stat values, IO::AIO offers the |
513 |
following constants and functions (if not implemented, the constants will |
514 |
be C<0> and the functions will either C<croak> or fall back on traditional |
515 |
behaviour). |
516 |
|
517 |
C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>, |
518 |
C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>, |
519 |
C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>. |
520 |
|
521 |
Example: Print the length of F</etc/passwd>: |
522 |
|
523 |
aio_stat "/etc/passwd", sub { |
524 |
$_[0] and die "stat failed: $!"; |
525 |
print "size is ", -s _, "\n"; |
526 |
}; |
527 |
|
528 |
|
529 |
=item aio_statvfs $fh_or_path, $callback->($statvfs) |
530 |
|
531 |
Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on |
532 |
whether a file handle or path was passed. |
533 |
|
534 |
On success, the callback is passed a hash reference with the following |
535 |
members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>, |
536 |
C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef> |
537 |
is passed. |
538 |
|
539 |
The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and |
540 |
C<ST_NOSUID>. |
541 |
|
542 |
The following non-POSIX IO::AIO::ST_* flag masks are defined to |
543 |
their correct value when available, or to C<0> on systems that do |
544 |
not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>, |
545 |
C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>, |
546 |
C<ST_NODIRATIME> and C<ST_RELATIME>. |
547 |
|
548 |
Example: stat C</wd> and dump out the data if successful. |
549 |
|
550 |
aio_statvfs "/wd", sub { |
551 |
my $f = $_[0] |
552 |
or die "statvfs: $!"; |
553 |
|
554 |
use Data::Dumper; |
555 |
say Dumper $f; |
556 |
}; |
557 |
|
558 |
# result: |
559 |
{ |
560 |
bsize => 1024, |
561 |
bfree => 4333064312, |
562 |
blocks => 10253828096, |
563 |
files => 2050765568, |
564 |
flag => 4096, |
565 |
favail => 2042092649, |
566 |
bavail => 4333064312, |
567 |
ffree => 2042092649, |
568 |
namemax => 255, |
569 |
frsize => 1024, |
570 |
fsid => 1810 |
571 |
} |
572 |
|
573 |
|
574 |
=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status) |
575 |
|
576 |
Works like perl's C<utime> function (including the special case of $atime |
577 |
and $mtime being undef). Fractional times are supported if the underlying |
578 |
syscalls support them. |
579 |
|
580 |
When called with a pathname, uses utimes(2) if available, otherwise |
581 |
utime(2). If called on a file descriptor, uses futimes(2) if available, |
582 |
otherwise returns ENOSYS, so this is not portable. |
583 |
|
584 |
Examples: |
585 |
|
586 |
# set atime and mtime to current time (basically touch(1)): |
587 |
aio_utime "path", undef, undef; |
588 |
# set atime to current time and mtime to beginning of the epoch: |
589 |
aio_utime "path", time, undef; # undef==0 |
590 |
|
591 |
|
592 |
=item aio_chown $fh_or_path, $uid, $gid, $callback->($status) |
593 |
|
594 |
Works like perl's C<chown> function, except that C<undef> for either $uid |
595 |
or $gid is being interpreted as "do not change" (but -1 can also be used). |
596 |
|
597 |
Examples: |
598 |
|
599 |
# same as "chown root path" in the shell: |
600 |
aio_chown "path", 0, -1; |
601 |
# same as above: |
602 |
aio_chown "path", 0, undef; |
603 |
|
604 |
|
605 |
=item aio_truncate $fh_or_path, $offset, $callback->($status) |
606 |
|
607 |
Works like truncate(2) or ftruncate(2). |
608 |
|
609 |
|
610 |
=item aio_chmod $fh_or_path, $mode, $callback->($status) |
611 |
|
612 |
Works like perl's C<chmod> function. |
613 |
|
614 |
|
615 |
=item aio_unlink $pathname, $callback->($status) |
616 |
|
617 |
Asynchronously unlink (delete) a file and call the callback with the |
618 |
result code. |
619 |
|
620 |
|
621 |
=item aio_mknod $path, $mode, $dev, $callback->($status) |
622 |
|
623 |
[EXPERIMENTAL] |
624 |
|
625 |
Asynchronously create a device node (or fifo). See mknod(2). |
626 |
|
627 |
The only (POSIX-) portable way of calling this function is: |
628 |
|
629 |
aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ... |
630 |
|
631 |
See C<aio_stat> for info about some potentially helpful extra constants |
632 |
and functions. |
633 |
|
634 |
=item aio_link $srcpath, $dstpath, $callback->($status) |
635 |
|
636 |
Asynchronously create a new link to the existing object at C<$srcpath> at |
637 |
the path C<$dstpath> and call the callback with the result code. |
638 |
|
639 |
|
640 |
=item aio_symlink $srcpath, $dstpath, $callback->($status) |
641 |
|
642 |
Asynchronously create a new symbolic link to the existing object at C<$srcpath> at |
643 |
the path C<$dstpath> and call the callback with the result code. |
644 |
|
645 |
|
646 |
=item aio_readlink $path, $callback->($link) |
647 |
|
648 |
Asynchronously read the symlink specified by C<$path> and pass it to |
649 |
the callback. If an error occurs, nothing or undef gets passed to the |
650 |
callback. |
651 |
|
652 |
|
653 |
=item aio_realpath $path, $callback->($path) |
654 |
|
655 |
Asynchronously make the path absolute and resolve any symlinks in |
656 |
C<$path>. The resulting path only consists of directories (Same as |
657 |
L<Cwd::realpath>). |
658 |
|
659 |
This request can be used to get the absolute path of the current working |
660 |
directory by passing it a path of F<.> (a single dot). |
661 |
|
662 |
|
663 |
=item aio_rename $srcpath, $dstpath, $callback->($status) |
664 |
|
665 |
Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as |
666 |
rename(2) and call the callback with the result code. |
667 |
|
668 |
|
669 |
=item aio_mkdir $pathname, $mode, $callback->($status) |
670 |
|
671 |
Asynchronously mkdir (create) a directory and call the callback with |
672 |
the result code. C<$mode> will be modified by the umask at the time the |
673 |
request is executed, so do not change your umask. |
674 |
|
675 |
|
676 |
=item aio_rmdir $pathname, $callback->($status) |
677 |
|
678 |
Asynchronously rmdir (delete) a directory and call the callback with the |
679 |
result code. |
680 |
|
681 |
|
682 |
=item aio_readdir $pathname, $callback->($entries) |
683 |
|
684 |
Unlike the POSIX call of the same name, C<aio_readdir> reads an entire |
685 |
directory (i.e. opendir + readdir + closedir). The entries will not be |
686 |
sorted, and will B<NOT> include the C<.> and C<..> entries. |
687 |
|
688 |
The callback is passed a single argument which is either C<undef> or an |
689 |
array-ref with the filenames. |
690 |
|
691 |
|
692 |
=item aio_readdirx $pathname, $flags, $callback->($entries, $flags) |
693 |
|
694 |
Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to |
695 |
tune behaviour and output format. In case of an error, C<$entries> will be |
696 |
C<undef>. |
697 |
|
698 |
The flags are a combination of the following constants, ORed together (the |
699 |
flags will also be passed to the callback, possibly modified): |
700 |
|
701 |
=over 4 |
702 |
|
703 |
=item IO::AIO::READDIR_DENTS |
704 |
|
705 |
When this flag is off, then the callback gets an arrayref consisting of |
706 |
names only (as with C<aio_readdir>), otherwise it gets an arrayref with |
707 |
C<[$name, $type, $inode]> arrayrefs, each describing a single directory |
708 |
entry in more detail. |
709 |
|
710 |
C<$name> is the name of the entry. |
711 |
|
712 |
C<$type> is one of the C<IO::AIO::DT_xxx> constants: |
713 |
|
714 |
C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>, |
715 |
C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>, |
716 |
C<IO::AIO::DT_WHT>. |
717 |
|
718 |
C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to |
719 |
know, you have to run stat yourself. Also, for speed reasons, the C<$type> |
720 |
scalars are read-only: you can not modify them. |
721 |
|
722 |
C<$inode> is the inode number (which might not be exact on systems with 64 |
723 |
bit inode numbers and 32 bit perls). This field has unspecified content on |
724 |
systems that do not deliver the inode information. |
725 |
|
726 |
=item IO::AIO::READDIR_DIRS_FIRST |
727 |
|
728 |
When this flag is set, then the names will be returned in an order where |
729 |
likely directories come first, in optimal stat order. This is useful when |
730 |
you need to quickly find directories, or you want to find all directories |
731 |
while avoiding to stat() each entry. |
732 |
|
733 |
If the system returns type information in readdir, then this is used |
734 |
to find directories directly. Otherwise, likely directories are names |
735 |
beginning with ".", or otherwise names with no dots, of which names with |
736 |
short names are tried first. |
737 |
|
738 |
=item IO::AIO::READDIR_STAT_ORDER |
739 |
|
740 |
When this flag is set, then the names will be returned in an order |
741 |
suitable for stat()'ing each one. That is, when you plan to stat() |
742 |
all files in the given directory, then the returned order will likely |
743 |
be fastest. |
744 |
|
745 |
If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then |
746 |
the likely dirs come first, resulting in a less optimal stat order. |
747 |
|
748 |
=item IO::AIO::READDIR_FOUND_UNKNOWN |
749 |
|
750 |
This flag should not be set when calling C<aio_readdirx>. Instead, it |
751 |
is being set by C<aio_readdirx>, when any of the C<$type>'s found were |
752 |
C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all |
753 |
C<$type>'s are known, which can be used to speed up some algorithms. |
754 |
|
755 |
=back |
756 |
|
757 |
|
758 |
=item aio_load $path, $data, $callback->($status) |
759 |
|
760 |
This is a composite request that tries to fully load the given file into |
761 |
memory. Status is the same as with aio_read. |
762 |
|
763 |
=cut |
764 |
|
765 |
sub aio_load($$;$) { |
766 |
my ($path, undef, $cb) = @_; |
767 |
my $data = \$_[1]; |
768 |
|
769 |
my $pri = aioreq_pri; |
770 |
my $grp = aio_group $cb; |
771 |
|
772 |
aioreq_pri $pri; |
773 |
add $grp aio_open $path, O_RDONLY, 0, sub { |
774 |
my $fh = shift |
775 |
or return $grp->result (-1); |
776 |
|
777 |
aioreq_pri $pri; |
778 |
add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub { |
779 |
$grp->result ($_[0]); |
780 |
}; |
781 |
}; |
782 |
|
783 |
$grp |
784 |
} |
785 |
|
786 |
=item aio_copy $srcpath, $dstpath, $callback->($status) |
787 |
|
788 |
Try to copy the I<file> (directories not supported as either source or |
789 |
destination) from C<$srcpath> to C<$dstpath> and call the callback with |
790 |
a status of C<0> (ok) or C<-1> (error, see C<$!>). |
791 |
|
792 |
This is a composite request that creates the destination file with |
793 |
mode 0200 and copies the contents of the source file into it using |
794 |
C<aio_sendfile>, followed by restoring atime, mtime, access mode and |
795 |
uid/gid, in that order. |
796 |
|
797 |
If an error occurs, the partial destination file will be unlinked, if |
798 |
possible, except when setting atime, mtime, access mode and uid/gid, where |
799 |
errors are being ignored. |
800 |
|
801 |
=cut |
802 |
|
803 |
sub aio_copy($$;$) { |
804 |
my ($src, $dst, $cb) = @_; |
805 |
|
806 |
my $pri = aioreq_pri; |
807 |
my $grp = aio_group $cb; |
808 |
|
809 |
aioreq_pri $pri; |
810 |
add $grp aio_open $src, O_RDONLY, 0, sub { |
811 |
if (my $src_fh = $_[0]) { |
812 |
my @stat = stat $src_fh; # hmm, might block over nfs? |
813 |
|
814 |
aioreq_pri $pri; |
815 |
add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub { |
816 |
if (my $dst_fh = $_[0]) { |
817 |
aioreq_pri $pri; |
818 |
add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub { |
819 |
if ($_[0] == $stat[7]) { |
820 |
$grp->result (0); |
821 |
close $src_fh; |
822 |
|
823 |
my $ch = sub { |
824 |
aioreq_pri $pri; |
825 |
add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub { |
826 |
aioreq_pri $pri; |
827 |
add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub { |
828 |
aioreq_pri $pri; |
829 |
add $grp aio_close $dst_fh; |
830 |
} |
831 |
}; |
832 |
}; |
833 |
|
834 |
aioreq_pri $pri; |
835 |
add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub { |
836 |
if ($_[0] < 0 && $! == ENOSYS) { |
837 |
aioreq_pri $pri; |
838 |
add $grp aio_utime $dst, $stat[8], $stat[9], $ch; |
839 |
} else { |
840 |
$ch->(); |
841 |
} |
842 |
}; |
843 |
} else { |
844 |
$grp->result (-1); |
845 |
close $src_fh; |
846 |
close $dst_fh; |
847 |
|
848 |
aioreq $pri; |
849 |
add $grp aio_unlink $dst; |
850 |
} |
851 |
}; |
852 |
} else { |
853 |
$grp->result (-1); |
854 |
} |
855 |
}, |
856 |
|
857 |
} else { |
858 |
$grp->result (-1); |
859 |
} |
860 |
}; |
861 |
|
862 |
$grp |
863 |
} |
864 |
|
865 |
=item aio_move $srcpath, $dstpath, $callback->($status) |
866 |
|
867 |
Try to move the I<file> (directories not supported as either source or |
868 |
destination) from C<$srcpath> to C<$dstpath> and call the callback with |
869 |
a status of C<0> (ok) or C<-1> (error, see C<$!>). |
870 |
|
871 |
This is a composite request that tries to rename(2) the file first; if |
872 |
rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if |
873 |
that is successful, unlinks the C<$srcpath>. |
874 |
|
875 |
=cut |
876 |
|
877 |
sub aio_move($$;$) { |
878 |
my ($src, $dst, $cb) = @_; |
879 |
|
880 |
my $pri = aioreq_pri; |
881 |
my $grp = aio_group $cb; |
882 |
|
883 |
aioreq_pri $pri; |
884 |
add $grp aio_rename $src, $dst, sub { |
885 |
if ($_[0] && $! == EXDEV) { |
886 |
aioreq_pri $pri; |
887 |
add $grp aio_copy $src, $dst, sub { |
888 |
$grp->result ($_[0]); |
889 |
|
890 |
unless ($_[0]) { |
891 |
aioreq_pri $pri; |
892 |
add $grp aio_unlink $src; |
893 |
} |
894 |
}; |
895 |
} else { |
896 |
$grp->result ($_[0]); |
897 |
} |
898 |
}; |
899 |
|
900 |
$grp |
901 |
} |
902 |
|
903 |
=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
904 |
|
905 |
Scans a directory (similar to C<aio_readdir>) but additionally tries to |
906 |
efficiently separate the entries of directory C<$path> into two sets of |
907 |
names, directories you can recurse into (directories), and ones you cannot |
908 |
recurse into (everything else, including symlinks to directories). |
909 |
|
910 |
C<aio_scandir> is a composite request that creates of many sub requests_ |
911 |
C<$maxreq> specifies the maximum number of outstanding aio requests that |
912 |
this function generates. If it is C<< <= 0 >>, then a suitable default |
913 |
will be chosen (currently 4). |
914 |
|
915 |
On error, the callback is called without arguments, otherwise it receives |
916 |
two array-refs with path-relative entry names. |
917 |
|
918 |
Example: |
919 |
|
920 |
aio_scandir $dir, 0, sub { |
921 |
my ($dirs, $nondirs) = @_; |
922 |
print "real directories: @$dirs\n"; |
923 |
print "everything else: @$nondirs\n"; |
924 |
}; |
925 |
|
926 |
Implementation notes. |
927 |
|
928 |
The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. |
929 |
|
930 |
If readdir returns file type information, then this is used directly to |
931 |
find directories. |
932 |
|
933 |
Otherwise, after reading the directory, the modification time, size etc. |
934 |
of the directory before and after the readdir is checked, and if they |
935 |
match (and isn't the current time), the link count will be used to decide |
936 |
how many entries are directories (if >= 2). Otherwise, no knowledge of the |
937 |
number of subdirectories will be assumed. |
938 |
|
939 |
Then entries will be sorted into likely directories a non-initial dot |
940 |
currently) and likely non-directories (see C<aio_readdirx>). Then every |
941 |
entry plus an appended C</.> will be C<stat>'ed, likely directories first, |
942 |
in order of their inode numbers. If that succeeds, it assumes that the |
943 |
entry is a directory or a symlink to directory (which will be checked |
944 |
separately). This is often faster than stat'ing the entry itself because |
945 |
filesystems might detect the type of the entry without reading the inode |
946 |
data (e.g. ext2fs filetype feature), even on systems that cannot return |
947 |
the filetype information on readdir. |
948 |
|
949 |
If the known number of directories (link count - 2) has been reached, the |
950 |
rest of the entries is assumed to be non-directories. |
951 |
|
952 |
This only works with certainty on POSIX (= UNIX) filesystems, which |
953 |
fortunately are the vast majority of filesystems around. |
954 |
|
955 |
It will also likely work on non-POSIX filesystems with reduced efficiency |
956 |
as those tend to return 0 or 1 as link counts, which disables the |
957 |
directory counting heuristic. |
958 |
|
959 |
=cut |
960 |
|
961 |
sub aio_scandir($$;$) { |
962 |
my ($path, $maxreq, $cb) = @_; |
963 |
|
964 |
my $pri = aioreq_pri; |
965 |
|
966 |
my $grp = aio_group $cb; |
967 |
|
968 |
$maxreq = 4 if $maxreq <= 0; |
969 |
|
970 |
# stat once |
971 |
aioreq_pri $pri; |
972 |
add $grp aio_stat $path, sub { |
973 |
return $grp->result () if $_[0]; |
974 |
my $now = time; |
975 |
my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
976 |
|
977 |
# read the directory entries |
978 |
aioreq_pri $pri; |
979 |
add $grp aio_readdirx $path, READDIR_DIRS_FIRST, sub { |
980 |
my $entries = shift |
981 |
or return $grp->result (); |
982 |
|
983 |
# stat the dir another time |
984 |
aioreq_pri $pri; |
985 |
add $grp aio_stat $path, sub { |
986 |
my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
987 |
|
988 |
my $ndirs; |
989 |
|
990 |
# take the slow route if anything looks fishy |
991 |
if ($hash1 ne $hash2 or (stat _)[9] == $now) { |
992 |
$ndirs = -1; |
993 |
} else { |
994 |
# if nlink == 2, we are finished |
995 |
# for non-posix-fs's, we rely on nlink < 2 |
996 |
$ndirs = (stat _)[3] - 2 |
997 |
or return $grp->result ([], $entries); |
998 |
} |
999 |
|
1000 |
my (@dirs, @nondirs); |
1001 |
|
1002 |
my $statgrp = add $grp aio_group sub { |
1003 |
$grp->result (\@dirs, \@nondirs); |
1004 |
}; |
1005 |
|
1006 |
limit $statgrp $maxreq; |
1007 |
feed $statgrp sub { |
1008 |
return unless @$entries; |
1009 |
my $entry = shift @$entries; |
1010 |
|
1011 |
aioreq_pri $pri; |
1012 |
add $statgrp aio_stat "$path/$entry/.", sub { |
1013 |
if ($_[0] < 0) { |
1014 |
push @nondirs, $entry; |
1015 |
} else { |
1016 |
# need to check for real directory |
1017 |
aioreq_pri $pri; |
1018 |
add $statgrp aio_lstat "$path/$entry", sub { |
1019 |
if (-d _) { |
1020 |
push @dirs, $entry; |
1021 |
|
1022 |
unless (--$ndirs) { |
1023 |
push @nondirs, @$entries; |
1024 |
feed $statgrp; |
1025 |
} |
1026 |
} else { |
1027 |
push @nondirs, $entry; |
1028 |
} |
1029 |
} |
1030 |
} |
1031 |
}; |
1032 |
}; |
1033 |
}; |
1034 |
}; |
1035 |
}; |
1036 |
|
1037 |
$grp |
1038 |
} |
1039 |
|
1040 |
=item aio_rmtree $path, $callback->($status) |
1041 |
|
1042 |
Delete a directory tree starting (and including) C<$path>, return the |
1043 |
status of the final C<rmdir> only. This is a composite request that |
1044 |
uses C<aio_scandir> to recurse into and rmdir directories, and unlink |
1045 |
everything else. |
1046 |
|
1047 |
=cut |
1048 |
|
1049 |
sub aio_rmtree; |
1050 |
sub aio_rmtree($;$) { |
1051 |
my ($path, $cb) = @_; |
1052 |
|
1053 |
my $pri = aioreq_pri; |
1054 |
my $grp = aio_group $cb; |
1055 |
|
1056 |
aioreq_pri $pri; |
1057 |
add $grp aio_scandir $path, 0, sub { |
1058 |
my ($dirs, $nondirs) = @_; |
1059 |
|
1060 |
my $dirgrp = aio_group sub { |
1061 |
add $grp aio_rmdir $path, sub { |
1062 |
$grp->result ($_[0]); |
1063 |
}; |
1064 |
}; |
1065 |
|
1066 |
(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs; |
1067 |
(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs; |
1068 |
|
1069 |
add $grp $dirgrp; |
1070 |
}; |
1071 |
|
1072 |
$grp |
1073 |
} |
1074 |
|
1075 |
=item aio_sync $callback->($status) |
1076 |
|
1077 |
Asynchronously call sync and call the callback when finished. |
1078 |
|
1079 |
=item aio_fsync $fh, $callback->($status) |
1080 |
|
1081 |
Asynchronously call fsync on the given filehandle and call the callback |
1082 |
with the fsync result code. |
1083 |
|
1084 |
=item aio_fdatasync $fh, $callback->($status) |
1085 |
|
1086 |
Asynchronously call fdatasync on the given filehandle and call the |
1087 |
callback with the fdatasync result code. |
1088 |
|
1089 |
If this call isn't available because your OS lacks it or it couldn't be |
1090 |
detected, it will be emulated by calling C<fsync> instead. |
1091 |
|
1092 |
=item aio_syncfs $fh, $callback->($status) |
1093 |
|
1094 |
Asynchronously call the syncfs syscall to sync the filesystem associated |
1095 |
to the given filehandle and call the callback with the syncfs result |
1096 |
code. If syncfs is not available, calls sync(), but returns C<-1> and sets |
1097 |
errno to C<ENOSYS> nevertheless. |
1098 |
|
1099 |
=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) |
1100 |
|
1101 |
Sync the data portion of the file specified by C<$offset> and C<$length> |
1102 |
to disk (but NOT the metadata), by calling the Linux-specific |
1103 |
sync_file_range call. If sync_file_range is not available or it returns |
1104 |
ENOSYS, then fdatasync or fsync is being substituted. |
1105 |
|
1106 |
C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>, |
1107 |
C<IO::AIO::SYNC_FILE_RANGE_WRITE> and |
1108 |
C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range |
1109 |
manpage for details. |
1110 |
|
1111 |
=item aio_pathsync $path, $callback->($status) |
1112 |
|
1113 |
This request tries to open, fsync and close the given path. This is a |
1114 |
composite request intended to sync directories after directory operations |
1115 |
(E.g. rename). This might not work on all operating systems or have any |
1116 |
specific effect, but usually it makes sure that directory changes get |
1117 |
written to disc. It works for anything that can be opened for read-only, |
1118 |
not just directories. |
1119 |
|
1120 |
Future versions of this function might fall back to other methods when |
1121 |
C<fsync> on the directory fails (such as calling C<sync>). |
1122 |
|
1123 |
Passes C<0> when everything went ok, and C<-1> on error. |
1124 |
|
1125 |
=cut |
1126 |
|
1127 |
sub aio_pathsync($;$) { |
1128 |
my ($path, $cb) = @_; |
1129 |
|
1130 |
my $pri = aioreq_pri; |
1131 |
my $grp = aio_group $cb; |
1132 |
|
1133 |
aioreq_pri $pri; |
1134 |
add $grp aio_open $path, O_RDONLY, 0, sub { |
1135 |
my ($fh) = @_; |
1136 |
if ($fh) { |
1137 |
aioreq_pri $pri; |
1138 |
add $grp aio_fsync $fh, sub { |
1139 |
$grp->result ($_[0]); |
1140 |
|
1141 |
aioreq_pri $pri; |
1142 |
add $grp aio_close $fh; |
1143 |
}; |
1144 |
} else { |
1145 |
$grp->result (-1); |
1146 |
} |
1147 |
}; |
1148 |
|
1149 |
$grp |
1150 |
} |
1151 |
|
1152 |
=item aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) |
1153 |
|
1154 |
This is a rather advanced IO::AIO call, which only works on mmap(2)ed |
1155 |
scalars (see the C<IO::AIO::mmap> function, although it also works on data |
1156 |
scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the |
1157 |
scalar must only be modified in-place while an aio operation is pending on |
1158 |
it). |
1159 |
|
1160 |
It calls the C<msync> function of your OS, if available, with the memory |
1161 |
area starting at C<$offset> in the string and ending C<$length> bytes |
1162 |
later. If C<$length> is negative, counts from the end, and if C<$length> |
1163 |
is C<undef>, then it goes till the end of the string. The flags can be |
1164 |
a combination of C<IO::AIO::MS_ASYNC>, C<IO::AIO::MS_INVALIDATE> and |
1165 |
C<IO::AIO::MS_SYNC>. |
1166 |
|
1167 |
=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) |
1168 |
|
1169 |
This is a rather advanced IO::AIO call, which works best on mmap(2)ed |
1170 |
scalars. |
1171 |
|
1172 |
It touches (reads or writes) all memory pages in the specified |
1173 |
range inside the scalar. All caveats and parameters are the same |
1174 |
as for C<aio_msync>, above, except for flags, which must be either |
1175 |
C<0> (which reads all pages and ensures they are instantiated) or |
1176 |
C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and |
1177 |
writing an octet from it, which dirties the page). |
1178 |
|
1179 |
=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) |
1180 |
|
1181 |
This is a rather advanced IO::AIO call, which works best on mmap(2)ed |
1182 |
scalars. |
1183 |
|
1184 |
It reads in all the pages of the underlying storage into memory (if any) |
1185 |
and locks them, so they are not getting swapped/paged out or removed. |
1186 |
|
1187 |
If C<$length> is undefined, then the scalar will be locked till the end. |
1188 |
|
1189 |
On systems that do not implement C<mlock>, this function returns C<-1> |
1190 |
and sets errno to C<ENOSYS>. |
1191 |
|
1192 |
Note that the corresponding C<munlock> is synchronous and is |
1193 |
documented under L<MISCELLANEOUS FUNCTIONS>. |
1194 |
|
1195 |
Example: open a file, mmap and mlock it - both will be undone when |
1196 |
C<$data> gets destroyed. |
1197 |
|
1198 |
open my $fh, "<", $path or die "$path: $!"; |
1199 |
my $data; |
1200 |
IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh; |
1201 |
aio_mlock $data; # mlock in background |
1202 |
|
1203 |
=item aio_mlockall $flags, $callback->($status) |
1204 |
|
1205 |
Calls the C<mlockall> function with the given C<$flags> (a combination of |
1206 |
C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>). |
1207 |
|
1208 |
On systems that do not implement C<mlockall>, this function returns C<-1> |
1209 |
and sets errno to C<ENOSYS>. |
1210 |
|
1211 |
Note that the corresponding C<munlockall> is synchronous and is |
1212 |
documented under L<MISCELLANEOUS FUNCTIONS>. |
1213 |
|
1214 |
Example: asynchronously lock all current and future pages into memory. |
1215 |
|
1216 |
aio_mlockall IO::AIO::MCL_FUTURE; |
1217 |
|
1218 |
=item aio_group $callback->(...) |
1219 |
|
1220 |
This is a very special aio request: Instead of doing something, it is a |
1221 |
container for other aio requests, which is useful if you want to bundle |
1222 |
many requests into a single, composite, request with a definite callback |
1223 |
and the ability to cancel the whole request with its subrequests. |
1224 |
|
1225 |
Returns an object of class L<IO::AIO::GRP>. See its documentation below |
1226 |
for more info. |
1227 |
|
1228 |
Example: |
1229 |
|
1230 |
my $grp = aio_group sub { |
1231 |
print "all stats done\n"; |
1232 |
}; |
1233 |
|
1234 |
add $grp |
1235 |
(aio_stat ...), |
1236 |
(aio_stat ...), |
1237 |
...; |
1238 |
|
1239 |
=item aio_nop $callback->() |
1240 |
|
1241 |
This is a special request - it does nothing in itself and is only used for |
1242 |
side effects, such as when you want to add a dummy request to a group so |
1243 |
that finishing the requests in the group depends on executing the given |
1244 |
code. |
1245 |
|
1246 |
While this request does nothing, it still goes through the execution |
1247 |
phase and still requires a worker thread. Thus, the callback will not |
1248 |
be executed immediately but only after other requests in the queue have |
1249 |
entered their execution phase. This can be used to measure request |
1250 |
latency. |
1251 |
|
1252 |
=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED* |
1253 |
|
1254 |
Mainly used for debugging and benchmarking, this aio request puts one of |
1255 |
the request workers to sleep for the given time. |
1256 |
|
1257 |
While it is theoretically handy to have simple I/O scheduling requests |
1258 |
like sleep and file handle readable/writable, the overhead this creates is |
1259 |
immense (it blocks a thread for a long time) so do not use this function |
1260 |
except to put your application under artificial I/O pressure. |
1261 |
|
1262 |
=back |
1263 |
|
1264 |
=head2 IO::AIO::REQ CLASS |
1265 |
|
1266 |
All non-aggregate C<aio_*> functions return an object of this class when |
1267 |
called in non-void context. |
1268 |
|
1269 |
=over 4 |
1270 |
|
1271 |
=item cancel $req |
1272 |
|
1273 |
Cancels the request, if possible. Has the effect of skipping execution |
1274 |
when entering the B<execute> state and skipping calling the callback when |
1275 |
entering the the B<result> state, but will leave the request otherwise |
1276 |
untouched (with the exception of readdir). That means that requests that |
1277 |
currently execute will not be stopped and resources held by the request |
1278 |
will not be freed prematurely. |
1279 |
|
1280 |
=item cb $req $callback->(...) |
1281 |
|
1282 |
Replace (or simply set) the callback registered to the request. |
1283 |
|
1284 |
=back |
1285 |
|
1286 |
=head2 IO::AIO::GRP CLASS |
1287 |
|
1288 |
This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to |
1289 |
objects of this class, too. |
1290 |
|
1291 |
A IO::AIO::GRP object is a special request that can contain multiple other |
1292 |
aio requests. |
1293 |
|
1294 |
You create one by calling the C<aio_group> constructing function with a |
1295 |
callback that will be called when all contained requests have entered the |
1296 |
C<done> state: |
1297 |
|
1298 |
my $grp = aio_group sub { |
1299 |
print "all requests are done\n"; |
1300 |
}; |
1301 |
|
1302 |
You add requests by calling the C<add> method with one or more |
1303 |
C<IO::AIO::REQ> objects: |
1304 |
|
1305 |
$grp->add (aio_unlink "..."); |
1306 |
|
1307 |
add $grp aio_stat "...", sub { |
1308 |
$_[0] or return $grp->result ("error"); |
1309 |
|
1310 |
# add another request dynamically, if first succeeded |
1311 |
add $grp aio_open "...", sub { |
1312 |
$grp->result ("ok"); |
1313 |
}; |
1314 |
}; |
1315 |
|
1316 |
This makes it very easy to create composite requests (see the source of |
1317 |
C<aio_move> for an application) that work and feel like simple requests. |
1318 |
|
1319 |
=over 4 |
1320 |
|
1321 |
=item * The IO::AIO::GRP objects will be cleaned up during calls to |
1322 |
C<IO::AIO::poll_cb>, just like any other request. |
1323 |
|
1324 |
=item * They can be canceled like any other request. Canceling will cancel not |
1325 |
only the request itself, but also all requests it contains. |
1326 |
|
1327 |
=item * They can also can also be added to other IO::AIO::GRP objects. |
1328 |
|
1329 |
=item * You must not add requests to a group from within the group callback (or |
1330 |
any later time). |
1331 |
|
1332 |
=back |
1333 |
|
1334 |
Their lifetime, simplified, looks like this: when they are empty, they |
1335 |
will finish very quickly. If they contain only requests that are in the |
1336 |
C<done> state, they will also finish. Otherwise they will continue to |
1337 |
exist. |
1338 |
|
1339 |
That means after creating a group you have some time to add requests |
1340 |
(precisely before the callback has been invoked, which is only done within |
1341 |
the C<poll_cb>). And in the callbacks of those requests, you can add |
1342 |
further requests to the group. And only when all those requests have |
1343 |
finished will the the group itself finish. |
1344 |
|
1345 |
=over 4 |
1346 |
|
1347 |
=item add $grp ... |
1348 |
|
1349 |
=item $grp->add (...) |
1350 |
|
1351 |
Add one or more requests to the group. Any type of L<IO::AIO::REQ> can |
1352 |
be added, including other groups, as long as you do not create circular |
1353 |
dependencies. |
1354 |
|
1355 |
Returns all its arguments. |
1356 |
|
1357 |
=item $grp->cancel_subs |
1358 |
|
1359 |
Cancel all subrequests and clears any feeder, but not the group request |
1360 |
itself. Useful when you queued a lot of events but got a result early. |
1361 |
|
1362 |
The group request will finish normally (you cannot add requests to the |
1363 |
group). |
1364 |
|
1365 |
=item $grp->result (...) |
1366 |
|
1367 |
Set the result value(s) that will be passed to the group callback when all |
1368 |
subrequests have finished and set the groups errno to the current value |
1369 |
of errno (just like calling C<errno> without an error number). By default, |
1370 |
no argument will be passed and errno is zero. |
1371 |
|
1372 |
=item $grp->errno ([$errno]) |
1373 |
|
1374 |
Sets the group errno value to C<$errno>, or the current value of errno |
1375 |
when the argument is missing. |
1376 |
|
1377 |
Every aio request has an associated errno value that is restored when |
1378 |
the callback is invoked. This method lets you change this value from its |
1379 |
default (0). |
1380 |
|
1381 |
Calling C<result> will also set errno, so make sure you either set C<$!> |
1382 |
before the call to C<result>, or call c<errno> after it. |
1383 |
|
1384 |
=item feed $grp $callback->($grp) |
1385 |
|
1386 |
Sets a feeder/generator on this group: every group can have an attached |
1387 |
generator that generates requests if idle. The idea behind this is that, |
1388 |
although you could just queue as many requests as you want in a group, |
1389 |
this might starve other requests for a potentially long time. For example, |
1390 |
C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests, |
1391 |
delaying any later requests for a long time. |
1392 |
|
1393 |
To avoid this, and allow incremental generation of requests, you can |
1394 |
instead a group and set a feeder on it that generates those requests. The |
1395 |
feed callback will be called whenever there are few enough (see C<limit>, |
1396 |
below) requests active in the group itself and is expected to queue more |
1397 |
requests. |
1398 |
|
1399 |
The feed callback can queue as many requests as it likes (i.e. C<add> does |
1400 |
not impose any limits). |
1401 |
|
1402 |
If the feed does not queue more requests when called, it will be |
1403 |
automatically removed from the group. |
1404 |
|
1405 |
If the feed limit is C<0> when this method is called, it will be set to |
1406 |
C<2> automatically. |
1407 |
|
1408 |
Example: |
1409 |
|
1410 |
# stat all files in @files, but only ever use four aio requests concurrently: |
1411 |
|
1412 |
my $grp = aio_group sub { print "finished\n" }; |
1413 |
limit $grp 4; |
1414 |
feed $grp sub { |
1415 |
my $file = pop @files |
1416 |
or return; |
1417 |
|
1418 |
add $grp aio_stat $file, sub { ... }; |
1419 |
}; |
1420 |
|
1421 |
=item limit $grp $num |
1422 |
|
1423 |
Sets the feeder limit for the group: The feeder will be called whenever |
1424 |
the group contains less than this many requests. |
1425 |
|
1426 |
Setting the limit to C<0> will pause the feeding process. |
1427 |
|
1428 |
The default value for the limit is C<0>, but note that setting a feeder |
1429 |
automatically bumps it up to C<2>. |
1430 |
|
1431 |
=back |
1432 |
|
1433 |
=head2 SUPPORT FUNCTIONS |
1434 |
|
1435 |
=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION |
1436 |
|
1437 |
=over 4 |
1438 |
|
1439 |
=item $fileno = IO::AIO::poll_fileno |
1440 |
|
1441 |
Return the I<request result pipe file descriptor>. This filehandle must be |
1442 |
polled for reading by some mechanism outside this module (e.g. EV, Glib, |
1443 |
select and so on, see below or the SYNOPSIS). If the pipe becomes readable |
1444 |
you have to call C<poll_cb> to check the results. |
1445 |
|
1446 |
See C<poll_cb> for an example. |
1447 |
|
1448 |
=item IO::AIO::poll_cb |
1449 |
|
1450 |
Process some outstanding events on the result pipe. You have to call |
1451 |
this regularly. Returns C<0> if all events could be processed (or there |
1452 |
were no events to process), or C<-1> if it returned earlier for whatever |
1453 |
reason. Returns immediately when no events are outstanding. The amount of |
1454 |
events processed depends on the settings of C<IO::AIO::max_poll_req> and |
1455 |
C<IO::AIO::max_poll_time>. |
1456 |
|
1457 |
If not all requests were processed for whatever reason, the filehandle |
1458 |
will still be ready when C<poll_cb> returns, so normally you don't have to |
1459 |
do anything special to have it called later. |
1460 |
|
1461 |
Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes |
1462 |
ready, it can be beneficial to call this function from loops which submit |
1463 |
a lot of requests, to make sure the results get processed when they become |
1464 |
available and not just when the loop is finished and the event loop takes |
1465 |
over again. This function returns very fast when there are no outstanding |
1466 |
requests. |
1467 |
|
1468 |
Example: Install an Event watcher that automatically calls |
1469 |
IO::AIO::poll_cb with high priority (more examples can be found in the |
1470 |
SYNOPSIS section, at the top of this document): |
1471 |
|
1472 |
Event->io (fd => IO::AIO::poll_fileno, |
1473 |
poll => 'r', async => 1, |
1474 |
cb => \&IO::AIO::poll_cb); |
1475 |
|
1476 |
=item IO::AIO::poll_wait |
1477 |
|
1478 |
If there are any outstanding requests and none of them in the result |
1479 |
phase, wait till the result filehandle becomes ready for reading (simply |
1480 |
does a C<select> on the filehandle. This is useful if you want to |
1481 |
synchronously wait for some requests to finish). |
1482 |
|
1483 |
See C<nreqs> for an example. |
1484 |
|
1485 |
=item IO::AIO::poll |
1486 |
|
1487 |
Waits until some requests have been handled. |
1488 |
|
1489 |
Returns the number of requests processed, but is otherwise strictly |
1490 |
equivalent to: |
1491 |
|
1492 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1493 |
|
1494 |
=item IO::AIO::flush |
1495 |
|
1496 |
Wait till all outstanding AIO requests have been handled. |
1497 |
|
1498 |
Strictly equivalent to: |
1499 |
|
1500 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1501 |
while IO::AIO::nreqs; |
1502 |
|
1503 |
=item IO::AIO::max_poll_reqs $nreqs |
1504 |
|
1505 |
=item IO::AIO::max_poll_time $seconds |
1506 |
|
1507 |
These set the maximum number of requests (default C<0>, meaning infinity) |
1508 |
that are being processed by C<IO::AIO::poll_cb> in one call, respectively |
1509 |
the maximum amount of time (default C<0>, meaning infinity) spent in |
1510 |
C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount |
1511 |
of time C<poll_cb> is allowed to use). |
1512 |
|
1513 |
Setting C<max_poll_time> to a non-zero value creates an overhead of one |
1514 |
syscall per request processed, which is not normally a problem unless your |
1515 |
callbacks are really really fast or your OS is really really slow (I am |
1516 |
not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead. |
1517 |
|
1518 |
Setting these is useful if you want to ensure some level of |
1519 |
interactiveness when perl is not fast enough to process all requests in |
1520 |
time. |
1521 |
|
1522 |
For interactive programs, values such as C<0.01> to C<0.1> should be fine. |
1523 |
|
1524 |
Example: Install an Event watcher that automatically calls |
1525 |
IO::AIO::poll_cb with low priority, to ensure that other parts of the |
1526 |
program get the CPU sometimes even under high AIO load. |
1527 |
|
1528 |
# try not to spend much more than 0.1s in poll_cb |
1529 |
IO::AIO::max_poll_time 0.1; |
1530 |
|
1531 |
# use a low priority so other tasks have priority |
1532 |
Event->io (fd => IO::AIO::poll_fileno, |
1533 |
poll => 'r', nice => 1, |
1534 |
cb => &IO::AIO::poll_cb); |
1535 |
|
1536 |
=back |
1537 |
|
1538 |
=head3 CONTROLLING THE NUMBER OF THREADS |
1539 |
|
1540 |
=over |
1541 |
|
1542 |
=item IO::AIO::min_parallel $nthreads |
1543 |
|
1544 |
Set the minimum number of AIO threads to C<$nthreads>. The current |
1545 |
default is C<8>, which means eight asynchronous operations can execute |
1546 |
concurrently at any one time (the number of outstanding requests, |
1547 |
however, is unlimited). |
1548 |
|
1549 |
IO::AIO starts threads only on demand, when an AIO request is queued and |
1550 |
no free thread exists. Please note that queueing up a hundred requests can |
1551 |
create demand for a hundred threads, even if it turns out that everything |
1552 |
is in the cache and could have been processed faster by a single thread. |
1553 |
|
1554 |
It is recommended to keep the number of threads relatively low, as some |
1555 |
Linux kernel versions will scale negatively with the number of threads |
1556 |
(higher parallelity => MUCH higher latency). With current Linux 2.6 |
1557 |
versions, 4-32 threads should be fine. |
1558 |
|
1559 |
Under most circumstances you don't need to call this function, as the |
1560 |
module selects a default that is suitable for low to moderate load. |
1561 |
|
1562 |
=item IO::AIO::max_parallel $nthreads |
1563 |
|
1564 |
Sets the maximum number of AIO threads to C<$nthreads>. If more than the |
1565 |
specified number of threads are currently running, this function kills |
1566 |
them. This function blocks until the limit is reached. |
1567 |
|
1568 |
While C<$nthreads> are zero, aio requests get queued but not executed |
1569 |
until the number of threads has been increased again. |
1570 |
|
1571 |
This module automatically runs C<max_parallel 0> at program end, to ensure |
1572 |
that all threads are killed and that there are no outstanding requests. |
1573 |
|
1574 |
Under normal circumstances you don't need to call this function. |
1575 |
|
1576 |
=item IO::AIO::max_idle $nthreads |
1577 |
|
1578 |
Limit the number of threads (default: 4) that are allowed to idle |
1579 |
(i.e., threads that did not get a request to process within the idle |
1580 |
timeout (default: 10 seconds). That means if a thread becomes idle while |
1581 |
C<$nthreads> other threads are also idle, it will free its resources and |
1582 |
exit. |
1583 |
|
1584 |
This is useful when you allow a large number of threads (e.g. 100 or 1000) |
1585 |
to allow for extremely high load situations, but want to free resources |
1586 |
under normal circumstances (1000 threads can easily consume 30MB of RAM). |
1587 |
|
1588 |
The default is probably ok in most situations, especially if thread |
1589 |
creation is fast. If thread creation is very slow on your system you might |
1590 |
want to use larger values. |
1591 |
|
1592 |
=item IO::AIO::idle_timeout $seconds |
1593 |
|
1594 |
Sets the minimum idle timeout (default 10) after which worker threads are |
1595 |
allowed to exit. SEe C<IO::AIO::max_idle>. |
1596 |
|
1597 |
=item IO::AIO::max_outstanding $maxreqs |
1598 |
|
1599 |
Sets the maximum number of outstanding requests to C<$nreqs>. If |
1600 |
you do queue up more than this number of requests, the next call to |
1601 |
C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as |
1602 |
C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no |
1603 |
longer exceeded. |
1604 |
|
1605 |
In other words, this setting does not enforce a queue limit, but can be |
1606 |
used to make poll functions block if the limit is exceeded. |
1607 |
|
1608 |
This is a very bad function to use in interactive programs because it |
1609 |
blocks, and a bad way to reduce concurrency because it is inexact: Better |
1610 |
use an C<aio_group> together with a feed callback. |
1611 |
|
1612 |
It's main use is in scripts without an event loop - when you want to stat |
1613 |
a lot of files, you can write somehting like this: |
1614 |
|
1615 |
IO::AIO::max_outstanding 32; |
1616 |
|
1617 |
for my $path (...) { |
1618 |
aio_stat $path , ...; |
1619 |
IO::AIO::poll_cb; |
1620 |
} |
1621 |
|
1622 |
IO::AIO::flush; |
1623 |
|
1624 |
The call to C<poll_cb> inside the loop will normally return instantly, but |
1625 |
as soon as more thna C<32> reqeusts are in-flight, it will block until |
1626 |
some requests have been handled. This keeps the loop from pushing a large |
1627 |
number of C<aio_stat> requests onto the queue. |
1628 |
|
1629 |
The default value for C<max_outstanding> is very large, so there is no |
1630 |
practical limit on the number of outstanding requests. |
1631 |
|
1632 |
=back |
1633 |
|
1634 |
=head3 STATISTICAL INFORMATION |
1635 |
|
1636 |
=over |
1637 |
|
1638 |
=item IO::AIO::nreqs |
1639 |
|
1640 |
Returns the number of requests currently in the ready, execute or pending |
1641 |
states (i.e. for which their callback has not been invoked yet). |
1642 |
|
1643 |
Example: wait till there are no outstanding requests anymore: |
1644 |
|
1645 |
IO::AIO::poll_wait, IO::AIO::poll_cb |
1646 |
while IO::AIO::nreqs; |
1647 |
|
1648 |
=item IO::AIO::nready |
1649 |
|
1650 |
Returns the number of requests currently in the ready state (not yet |
1651 |
executed). |
1652 |
|
1653 |
=item IO::AIO::npending |
1654 |
|
1655 |
Returns the number of requests currently in the pending state (executed, |
1656 |
but not yet processed by poll_cb). |
1657 |
|
1658 |
=back |
1659 |
|
1660 |
=head3 MISCELLANEOUS FUNCTIONS |
1661 |
|
1662 |
IO::AIO implements some functions that might be useful, but are not |
1663 |
asynchronous. |
1664 |
|
1665 |
=over 4 |
1666 |
|
1667 |
=item IO::AIO::sendfile $ofh, $ifh, $offset, $count |
1668 |
|
1669 |
Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>, |
1670 |
but is blocking (this makes most sense if you know the input data is |
1671 |
likely cached already and the output filehandle is set to non-blocking |
1672 |
operations). |
1673 |
|
1674 |
Returns the number of bytes copied, or C<-1> on error. |
1675 |
|
1676 |
=item IO::AIO::fadvise $fh, $offset, $len, $advice |
1677 |
|
1678 |
Simply calls the C<posix_fadvise> function (see its |
1679 |
manpage for details). The following advice constants are |
1680 |
available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>, |
1681 |
C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>, |
1682 |
C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>. |
1683 |
|
1684 |
On systems that do not implement C<posix_fadvise>, this function returns |
1685 |
ENOSYS, otherwise the return value of C<posix_fadvise>. |
1686 |
|
1687 |
=item IO::AIO::madvise $scalar, $offset, $len, $advice |
1688 |
|
1689 |
Simply calls the C<posix_madvise> function (see its |
1690 |
manpage for details). The following advice constants are |
1691 |
available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>, |
1692 |
C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>. |
1693 |
|
1694 |
On systems that do not implement C<posix_madvise>, this function returns |
1695 |
ENOSYS, otherwise the return value of C<posix_madvise>. |
1696 |
|
1697 |
=item IO::AIO::mprotect $scalar, $offset, $len, $protect |
1698 |
|
1699 |
Simply calls the C<mprotect> function on the preferably AIO::mmap'ed |
1700 |
$scalar (see its manpage for details). The following protect |
1701 |
constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>, |
1702 |
C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>. |
1703 |
|
1704 |
On systems that do not implement C<mprotect>, this function returns |
1705 |
ENOSYS, otherwise the return value of C<mprotect>. |
1706 |
|
1707 |
=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset] |
1708 |
|
1709 |
Memory-maps a file (or anonymous memory range) and attaches it to the |
1710 |
given C<$scalar>, which will act like a string scalar. |
1711 |
|
1712 |
The only operations allowed on the scalar are C<substr>/C<vec> that don't |
1713 |
change the string length, and most read-only operations such as copying it |
1714 |
or searching it with regexes and so on. |
1715 |
|
1716 |
Anything else is unsafe and will, at best, result in memory leaks. |
1717 |
|
1718 |
The memory map associated with the C<$scalar> is automatically removed |
1719 |
when the C<$scalar> is destroyed, or when the C<IO::AIO::mmap> or |
1720 |
C<IO::AIO::munmap> functions are called. |
1721 |
|
1722 |
This calls the C<mmap>(2) function internally. See your system's manual |
1723 |
page for details on the C<$length>, C<$prot> and C<$flags> parameters. |
1724 |
|
1725 |
The C<$length> must be larger than zero and smaller than the actual |
1726 |
filesize. |
1727 |
|
1728 |
C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>, |
1729 |
C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>, |
1730 |
|
1731 |
C<$flags> can be a combination of C<IO::AIO::MAP_SHARED> or |
1732 |
C<IO::AIO::MAP_PRIVATE>, or a number of system-specific flags (when |
1733 |
not available, the are defined as 0): C<IO::AIO::MAP_ANONYMOUS> |
1734 |
(which is set to C<MAP_ANON> if your system only provides this |
1735 |
constant), C<IO::AIO::MAP_HUGETLB>, C<IO::AIO::MAP_LOCKED>, |
1736 |
C<IO::AIO::MAP_NORESERVE>, C<IO::AIO::MAP_POPULATE> or |
1737 |
C<IO::AIO::MAP_NONBLOCK> |
1738 |
|
1739 |
If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed. |
1740 |
|
1741 |
C<$offset> is the offset from the start of the file - it generally must be |
1742 |
a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>. |
1743 |
|
1744 |
Example: |
1745 |
|
1746 |
use Digest::MD5; |
1747 |
use IO::AIO; |
1748 |
|
1749 |
open my $fh, "<verybigfile" |
1750 |
or die "$!"; |
1751 |
|
1752 |
IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh |
1753 |
or die "verybigfile: $!"; |
1754 |
|
1755 |
my $fast_md5 = md5 $data; |
1756 |
|
1757 |
=item IO::AIO::munmap $scalar |
1758 |
|
1759 |
Removes a previous mmap and undefines the C<$scalar>. |
1760 |
|
1761 |
=item IO::AIO::munlock $scalar, $offset = 0, $length = undef |
1762 |
|
1763 |
Calls the C<munlock> function, undoing the effects of a previous |
1764 |
C<aio_mlock> call (see its description for details). |
1765 |
|
1766 |
=item IO::AIO::munlockall |
1767 |
|
1768 |
Calls the C<munlockall> function. |
1769 |
|
1770 |
On systems that do not implement C<munlockall>, this function returns |
1771 |
ENOSYS, otherwise the return value of C<munlockall>. |
1772 |
|
1773 |
=back |
1774 |
|
1775 |
=cut |
1776 |
|
1777 |
min_parallel 8; |
1778 |
|
1779 |
END { flush } |
1780 |
|
1781 |
1; |
1782 |
|
1783 |
=head1 EVENT LOOP INTEGRATION |
1784 |
|
1785 |
It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO |
1786 |
automatically into many event loops: |
1787 |
|
1788 |
# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...) |
1789 |
use AnyEvent::AIO; |
1790 |
|
1791 |
You can also integrate IO::AIO manually into many event loops, here are |
1792 |
some examples of how to do this: |
1793 |
|
1794 |
# EV integration |
1795 |
my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; |
1796 |
|
1797 |
# Event integration |
1798 |
Event->io (fd => IO::AIO::poll_fileno, |
1799 |
poll => 'r', |
1800 |
cb => \&IO::AIO::poll_cb); |
1801 |
|
1802 |
# Glib/Gtk2 integration |
1803 |
add_watch Glib::IO IO::AIO::poll_fileno, |
1804 |
in => sub { IO::AIO::poll_cb; 1 }; |
1805 |
|
1806 |
# Tk integration |
1807 |
Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", |
1808 |
readable => \&IO::AIO::poll_cb); |
1809 |
|
1810 |
# Danga::Socket integration |
1811 |
Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
1812 |
\&IO::AIO::poll_cb); |
1813 |
|
1814 |
=head2 FORK BEHAVIOUR |
1815 |
|
1816 |
Usage of pthreads in a program changes the semantics of fork |
1817 |
considerably. Specifically, only async-safe functions can be called after |
1818 |
fork. Perl doesn't know about this, so in general, you cannot call fork |
1819 |
with defined behaviour in perl if pthreads are involved. IO::AIO uses |
1820 |
pthreads, so this applies, but many other extensions and (for inexplicable |
1821 |
reasons) perl itself often is linked against pthreads, so this limitation |
1822 |
applies to quite a lot of perls. |
1823 |
|
1824 |
This module no longer tries to fight your OS, or POSIX. That means IO::AIO |
1825 |
only works in the process that loaded it. Forking is fully supported, but |
1826 |
using IO::AIO in the child is not. |
1827 |
|
1828 |
You might get around by not I<using> IO::AIO before (or after) |
1829 |
forking. You could also try to call the L<IO::AIO::reinit> function in the |
1830 |
child: |
1831 |
|
1832 |
=over 4 |
1833 |
|
1834 |
=item IO::AIO::reinit |
1835 |
|
1836 |
Abandons all current requests and I/O threads and simply reinitialises all |
1837 |
data structures. This is not an operation supported by any standards, but |
1838 |
happens to work on GNU/Linux and some newer BSD systems. |
1839 |
|
1840 |
The only reasonable use for this function is to call it after forking, if |
1841 |
C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in |
1842 |
the process will result in undefined behaviour. Calling it at any time |
1843 |
will also result in any undefined (by POSIX) behaviour. |
1844 |
|
1845 |
=back |
1846 |
|
1847 |
=head2 MEMORY USAGE |
1848 |
|
1849 |
Per-request usage: |
1850 |
|
1851 |
Each aio request uses - depending on your architecture - around 100-200 |
1852 |
bytes of memory. In addition, stat requests need a stat buffer (possibly |
1853 |
a few hundred bytes), readdir requires a result buffer and so on. Perl |
1854 |
scalars and other data passed into aio requests will also be locked and |
1855 |
will consume memory till the request has entered the done state. |
1856 |
|
1857 |
This is not awfully much, so queuing lots of requests is not usually a |
1858 |
problem. |
1859 |
|
1860 |
Per-thread usage: |
1861 |
|
1862 |
In the execution phase, some aio requests require more memory for |
1863 |
temporary buffers, and each thread requires a stack and other data |
1864 |
structures (usually around 16k-128k, depending on the OS). |
1865 |
|
1866 |
=head1 KNOWN BUGS |
1867 |
|
1868 |
Known bugs will be fixed in the next release. |
1869 |
|
1870 |
=head1 SEE ALSO |
1871 |
|
1872 |
L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a |
1873 |
more natural syntax. |
1874 |
|
1875 |
=head1 AUTHOR |
1876 |
|
1877 |
Marc Lehmann <schmorp@schmorp.de> |
1878 |
http://home.schmorp.de/ |
1879 |
|
1880 |
=cut |
1881 |
|