ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/IO-AIO/README
(Generate patch)

Comparing IO-AIO/README (file contents):
Revision 1.8 by root, Sun Jul 31 17:24:05 2005 UTC vs.
Revision 1.53 by root, Thu Oct 11 03:20:52 2012 UTC

… …
2 IO::AIO - Asynchronous Input/Output 2 IO::AIO - Asynchronous Input/Output
3 3
4SYNOPSIS 4SYNOPSIS
5 use IO::AIO; 5 use IO::AIO;
6 6
7 aio_open "/etc/passwd", O_RDONLY, 0, sub { 7 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
8 my ($fh) = @_; 8 my $fh = shift
9 or die "/etc/passwd: $!";
9 ... 10 ...
10 }; 11 };
11 12
12 aio_unlink "/tmp/file", sub { }; 13 aio_unlink "/tmp/file", sub { };
13 14
14 aio_read $fh, 30000, 1024, $buffer, 0, sub { 15 aio_read $fh, 30000, 1024, $buffer, 0, sub {
15 $_[0] > 0 or die "read error: $!"; 16 $_[0] > 0 or die "read error: $!";
16 }; 17 };
17 18
18 # Event 19 # version 2+ has request and group objects
19 Event->io (fd => IO::AIO::poll_fileno, 20 use IO::AIO 2;
20 poll => 'r',
21 cb => \&IO::AIO::poll_cb);
22 21
23 # Glib/Gtk2 22 aioreq_pri 4; # give next request a very high priority
24 add_watch Glib::IO IO::AIO::poll_fileno, 23 my $req = aio_unlink "/tmp/file", sub { };
25 in => sub { IO::AIO::poll_cb; 1 }; 24 $req->cancel; # cancel request if still in queue
26 25
27 # Tk 26 my $grp = aio_group sub { print "all stats done\n" };
28 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", 27 add $grp aio_stat "..." for ...;
29 readable => \&IO::AIO::poll_cb);
30
31 # Danga::Socket
32 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
33 \&IO::AIO::poll_cb);
34 28
35DESCRIPTION 29DESCRIPTION
36 This module implements asynchronous I/O using whatever means your 30 This module implements asynchronous I/O using whatever means your
37 operating system supports. 31 operating system supports. It is implemented as an interface to "libeio"
32 (<http://software.schmorp.de/pkg/libeio.html>).
38 33
34 Asynchronous means that operations that can normally block your program
35 (e.g. reading from disk) will be done asynchronously: the operation will
36 still block, but you can do something else in the meantime. This is
37 extremely useful for programs that need to stay interactive even when
38 doing heavy I/O (GUI programs, high performance network servers etc.),
39 but can also be used to easily do operations in parallel that are
40 normally done sequentially, e.g. stat'ing many files, which is much
41 faster on a RAID volume or over NFS when you do a number of stat
42 operations concurrently.
43
44 While most of this works on all types of file descriptors (for example
45 sockets), using these functions on file descriptors that support
46 nonblocking operation (again, sockets, pipes etc.) is very inefficient.
47 Use an event loop for that (such as the EV module): IO::AIO will
48 naturally fit into such an event loop itself.
49
39 Currently, a number of threads are started that execute your read/writes 50 In this version, a number of threads are started that execute your
40 and signal their completion. You don't need thread support in your libc 51 requests and signal their completion. You don't need thread support in
41 or perl, and the threads created by this module will not be visible to 52 perl, and the threads created by this module will not be visible to
42 the pthreads library. In the future, this module might make use of the 53 perl. In the future, this module might make use of the native aio
43 native aio functions available on many operating systems. However, they 54 functions available on many operating systems. However, they are often
44 are often not well-supported (Linux doesn't allow them on normal files 55 not well-supported or restricted (GNU/Linux doesn't allow them on normal
45 currently, for example), and they would only support aio_read and 56 files currently, for example), and they would only support aio_read and
46 aio_write, so the remaining functionality would have to be implemented 57 aio_write, so the remaining functionality would have to be implemented
47 using threads anyway. 58 using threads anyway.
48 59
49 Although the module will work with in the presence of other threads, it 60 Although the module will work in the presence of other (Perl-) threads,
50 is currently not reentrant, so use appropriate locking yourself, always 61 it is currently not reentrant in any way, so use appropriate locking
51 call "poll_cb" from within the same thread, or never call "poll_cb" (or 62 yourself, always call "poll_cb" from within the same thread, or never
52 other "aio_" functions) recursively. 63 call "poll_cb" (or other "aio_" functions) recursively.
64
65 EXAMPLE
66 This is a simple example that uses the EV module and loads /etc/passwd
67 asynchronously:
68
69 use Fcntl;
70 use EV;
71 use IO::AIO;
72
73 # register the IO::AIO callback with EV
74 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
75
76 # queue the request to open /etc/passwd
77 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
78 my $fh = shift
79 or die "error while opening: $!";
80
81 # stat'ing filehandles is generally non-blocking
82 my $size = -s $fh;
83
84 # queue a request to read the file
85 my $contents;
86 aio_read $fh, 0, $size, $contents, 0, sub {
87 $_[0] == $size
88 or die "short read: $!";
89
90 close $fh;
91
92 # file contents now in $contents
93 print $contents;
94
95 # exit event loop and program
96 EV::unloop;
97 };
98 };
99
100 # possibly queue up other requests, or open GUI windows,
101 # check for sockets etc. etc.
102
103 # process events as long as there are some:
104 EV::loop;
105
106REQUEST ANATOMY AND LIFETIME
107 Every "aio_*" function creates a request. which is a C data structure
108 not directly visible to Perl.
109
110 If called in non-void context, every request function returns a Perl
111 object representing the request. In void context, nothing is returned,
112 which saves a bit of memory.
113
114 The perl object is a fairly standard ref-to-hash object. The hash
115 contents are not used by IO::AIO so you are free to store anything you
116 like in it.
117
118 During their existance, aio requests travel through the following
119 states, in order:
120
121 ready
122 Immediately after a request is created it is put into the ready
123 state, waiting for a thread to execute it.
124
125 execute
126 A thread has accepted the request for processing and is currently
127 executing it (e.g. blocking in read).
128
129 pending
130 The request has been executed and is waiting for result processing.
131
132 While request submission and execution is fully asynchronous, result
133 processing is not and relies on the perl interpreter calling
134 "poll_cb" (or another function with the same effect).
135
136 result
137 The request results are processed synchronously by "poll_cb".
138
139 The "poll_cb" function will process all outstanding aio requests by
140 calling their callbacks, freeing memory associated with them and
141 managing any groups they are contained in.
142
143 done
144 Request has reached the end of its lifetime and holds no resources
145 anymore (except possibly for the Perl object, but its connection to
146 the actual aio request is severed and calling its methods will
147 either do nothing or result in a runtime error).
53 148
54FUNCTIONS 149FUNCTIONS
55 AIO FUNCTIONS 150 QUICK OVERVIEW
151 This section simply lists the prototypes most of the functions for quick
152 reference. See the following sections for function-by-function
153 documentation.
154
155 aio_wd $pathname, $callback->($wd)
156 aio_open $pathname, $flags, $mode, $callback->($fh)
157 aio_close $fh, $callback->($status)
158 aio_seek $fh,$offset,$whence, $callback->($offs)
159 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
160 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
161 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
162 aio_readahead $fh,$offset,$length, $callback->($retval)
163 aio_stat $fh_or_path, $callback->($status)
164 aio_lstat $fh, $callback->($status)
165 aio_statvfs $fh_or_path, $callback->($statvfs)
166 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
167 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
168 aio_chmod $fh_or_path, $mode, $callback->($status)
169 aio_truncate $fh_or_path, $offset, $callback->($status)
170 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
171 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
172 aio_unlink $pathname, $callback->($status)
173 aio_mknod $pathname, $mode, $dev, $callback->($status)
174 aio_link $srcpath, $dstpath, $callback->($status)
175 aio_symlink $srcpath, $dstpath, $callback->($status)
176 aio_readlink $pathname, $callback->($link)
177 aio_realpath $pathname, $callback->($link)
178 aio_rename $srcpath, $dstpath, $callback->($status)
179 aio_mkdir $pathname, $mode, $callback->($status)
180 aio_rmdir $pathname, $callback->($status)
181 aio_readdir $pathname, $callback->($entries)
182 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
183 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
184 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
185 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
186 aio_load $pathname, $data, $callback->($status)
187 aio_copy $srcpath, $dstpath, $callback->($status)
188 aio_move $srcpath, $dstpath, $callback->($status)
189 aio_rmtree $pathname, $callback->($status)
190 aio_sync $callback->($status)
191 aio_syncfs $fh, $callback->($status)
192 aio_fsync $fh, $callback->($status)
193 aio_fdatasync $fh, $callback->($status)
194 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
195 aio_pathsync $pathname, $callback->($status)
196 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
197 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
198 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
199 aio_mlockall $flags, $callback->($status)
200 aio_group $callback->(...)
201 aio_nop $callback->()
202
203 $prev_pri = aioreq_pri [$pri]
204 aioreq_nice $pri_adjust
205
206 IO::AIO::poll_wait
207 IO::AIO::poll_cb
208 IO::AIO::poll
209 IO::AIO::flush
210 IO::AIO::max_poll_reqs $nreqs
211 IO::AIO::max_poll_time $seconds
212 IO::AIO::min_parallel $nthreads
213 IO::AIO::max_parallel $nthreads
214 IO::AIO::max_idle $nthreads
215 IO::AIO::idle_timeout $seconds
216 IO::AIO::max_outstanding $maxreqs
217 IO::AIO::nreqs
218 IO::AIO::nready
219 IO::AIO::npending
220
221 IO::AIO::sendfile $ofh, $ifh, $offset, $count
222 IO::AIO::fadvise $fh, $offset, $len, $advice
223 IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
224 IO::AIO::munmap $scalar
225 IO::AIO::madvise $scalar, $offset, $length, $advice
226 IO::AIO::mprotect $scalar, $offset, $length, $protect
227 IO::AIO::munlock $scalar, $offset = 0, $length = undef
228 IO::AIO::munlockall
229
230 API NOTES
56 All the "aio_*" calls are more or less thin wrappers around the syscall 231 All the "aio_*" calls are more or less thin wrappers around the syscall
57 with the same name (sans "aio_"). The arguments are similar or 232 with the same name (sans "aio_"). The arguments are similar or
58 identical, and they all accept an additional (and optional) $callback 233 identical, and they all accept an additional (and optional) $callback
59 argument which must be a code reference. This code reference will get 234 argument which must be a code reference. This code reference will be
60 called with the syscall return code (e.g. most syscalls return -1 on 235 called after the syscall has been executed in an asynchronous fashion.
61 error, unlike perl, which usually delivers "false") as it's sole 236 The results of the request will be passed as arguments to the callback
62 argument when the given syscall has been executed asynchronously. 237 (and, if an error occured, in $!) - for most requests the syscall return
238 code (e.g. most syscalls return -1 on error, unlike perl, which usually
239 delivers "false").
240
241 Some requests (such as "aio_readdir") pass the actual results and
242 communicate failures by passing "undef".
63 243
64 All functions expecting a filehandle keep a copy of the filehandle 244 All functions expecting a filehandle keep a copy of the filehandle
65 internally until the request has finished. 245 internally until the request has finished.
66 246
247 All functions return request objects of type IO::AIO::REQ that allow
248 further manipulation of those requests while they are in-flight.
249
67 The filenames you pass to these routines *must* be absolute. The reason 250 The pathnames you pass to these routines *should* be absolute. The
68 for this is that at the time the request is being executed, the current 251 reason for this is that at the time the request is being executed, the
69 working directory could have changed. Alternatively, you can make sure 252 current working directory could have changed. Alternatively, you can
70 that you never change the current working directory. 253 make sure that you never change the current working directory anywhere
254 in the program and then use relative paths. You can also take advantage
255 of IO::AIOs working directory abstraction, that lets you specify paths
256 relative to some previously-opened "working directory object" - see the
257 description of the "IO::AIO::WD" class later in this document.
71 258
259 To encode pathnames as octets, either make sure you either: a) always
260 pass in filenames you got from outside (command line, readdir etc.)
261 without tinkering, b) are in your native filesystem encoding, c) use the
262 Encode module and encode your pathnames to the locale (or other)
263 encoding in effect in the user environment, d) use
264 Glib::filename_from_unicode on unicode filenames or e) use something
265 else to ensure your scalar has the correct contents.
266
267 This works, btw. independent of the internal UTF-8 bit, which IO::AIO
268 handles correctly whether it is set or not.
269
270 AIO REQUEST FUNCTIONS
271 $prev_pri = aioreq_pri [$pri]
272 Returns the priority value that would be used for the next request
273 and, if $pri is given, sets the priority for the next aio request.
274
275 The default priority is 0, the minimum and maximum priorities are -4
276 and 4, respectively. Requests with higher priority will be serviced
277 first.
278
279 The priority will be reset to 0 after each call to one of the
280 "aio_*" functions.
281
282 Example: open a file with low priority, then read something from it
283 with higher priority so the read request is serviced before other
284 low priority open requests (potentially spamming the cache):
285
286 aioreq_pri -3;
287 aio_open ..., sub {
288 return unless $_[0];
289
290 aioreq_pri -2;
291 aio_read $_[0], ..., sub {
292 ...
293 };
294 };
295
296 aioreq_nice $pri_adjust
297 Similar to "aioreq_pri", but subtracts the given value from the
298 current priority, so the effect is cumulative.
299
72 aio_open $pathname, $flags, $mode, $callback 300 aio_open $pathname, $flags, $mode, $callback->($fh)
73 Asynchronously open or create a file and call the callback with a 301 Asynchronously open or create a file and call the callback with a
74 newly created filehandle for the file. 302 newly created filehandle for the file (or "undef" in case of an
303 error).
75 304
76 The pathname passed to "aio_open" must be absolute. See API NOTES, 305 The pathname passed to "aio_open" must be absolute. See API NOTES,
77 above, for an explanation. 306 above, for an explanation.
78 307
79 The $flags argument is a bitmask. See the "Fcntl" module for a list. 308 The $flags argument is a bitmask. See the "Fcntl" module for a list.
80 They are the same as used by "sysopen". 309 They are the same as used by "sysopen".
81 310
82 Likewise, $mode specifies the mode of the newly created file, if it 311 Likewise, $mode specifies the mode of the newly created file, if it
83 didn't exist and "O_CREAT" has been given, just like perl's 312 didn't exist and "O_CREAT" has been given, just like perl's
84 "sysopen", except that it is mandatory (i.e. use 0 if you don't 313 "sysopen", except that it is mandatory (i.e. use 0 if you don't
85 create new files, and 0666 or 0777 if you do). 314 create new files, and 0666 or 0777 if you do). Note that the $mode
315 will be modified by the umask in effect then the request is being
316 executed, so better never change the umask.
86 317
87 Example: 318 Example:
88 319
89 aio_open "/etc/passwd", O_RDONLY, 0, sub { 320 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
90 if ($_[0]) { 321 if ($_[0]) {
91 print "open successful, fh is $_[0]\n"; 322 print "open successful, fh is $_[0]\n";
92 ... 323 ...
93 } else { 324 } else {
94 die "open failed: $!\n"; 325 die "open failed: $!\n";
95 } 326 }
96 }; 327 };
97 328
329 In addition to all the common open modes/flags ("O_RDONLY",
330 "O_WRONLY", "O_RDWR", "O_CREAT", "O_TRUNC", "O_EXCL" and
331 "O_APPEND"), the following POSIX and non-POSIX constants are
332 available (missing ones on your system are, as usual, 0):
333
334 "O_ASYNC", "O_DIRECT", "O_NOATIME", "O_CLOEXEC", "O_NOCTTY",
335 "O_NOFOLLOW", "O_NONBLOCK", "O_EXEC", "O_SEARCH", "O_DIRECTORY",
336 "O_DSYNC", "O_RSYNC", "O_SYNC" and "O_TTY_INIT".
337
98 aio_close $fh, $callback 338 aio_close $fh, $callback->($status)
99 Asynchronously close a file and call the callback with the result 339 Asynchronously close a file and call the callback with the result
100 code. *WARNING:* although accepted, you should not pass in a perl 340 code.
101 filehandle here, as perl will likely close the file descriptor
102 another time when the filehandle is destroyed. Normally, you can
103 safely call perls "close" or just let filehandles go out of scope.
104 341
105 This is supposed to be a bug in the API, so that might change. It's 342 Unfortunately, you can't do this to perl. Perl *insists* very
106 therefore best to avoid this function. 343 strongly on closing the file descriptor associated with the
344 filehandle itself.
107 345
346 Therefore, "aio_close" will not close the filehandle - instead it
347 will use dup2 to overwrite the file descriptor with the write-end of
348 a pipe (the pipe fd will be created on demand and will be cached).
349
350 Or in other words: the file descriptor will be closed, but it will
351 not be free for reuse until the perl filehandle is closed.
352
353 aio_seek $fh, $offset, $whence, $callback->($offs)
354 Seeks the filehandle to the new $offset, similarly to perl's
355 "sysseek". The $whence can use the traditional values (0 for
356 "IO::AIO::SEEK_SET", 1 for "IO::AIO::SEEK_CUR" or 2 for
357 "IO::AIO::SEEK_END").
358
359 The resulting absolute offset will be passed to the callback, or -1
360 in case of an error.
361
362 In theory, the $whence constants could be different than the
363 corresponding values from Fcntl, but perl guarantees they are the
364 same, so don't panic.
365
366 As a GNU/Linux (and maybe Solaris) extension, also the constants
367 "IO::AIO::SEEK_DATA" and "IO::AIO::SEEK_HOLE" are available, if they
368 could be found. No guarantees about suitability for use in
369 "aio_seek" or Perl's "sysseek" can be made though, although I would
370 naively assume they "just work".
371
108 aio_read $fh,$offset,$length, $data,$dataoffset,$callback 372 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
109 aio_write $fh,$offset,$length, $data,$dataoffset,$callback 373 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
110 Reads or writes "length" bytes from the specified "fh" and "offset" 374 Reads or writes $length bytes from or to the specified $fh and
111 into the scalar given by "data" and offset "dataoffset" and calls 375 $offset into the scalar given by $data and offset $dataoffset and
112 the callback without the actual number of bytes read (or -1 on 376 calls the callback without the actual number of bytes read (or -1 on
113 error, just like the syscall). 377 error, just like the syscall).
378
379 "aio_read" will, like "sysread", shrink or grow the $data scalar to
380 offset plus the actual number of bytes read.
381
382 If $offset is undefined, then the current file descriptor offset
383 will be used (and updated), otherwise the file descriptor offset
384 will not be changed by these calls.
385
386 If $length is undefined in "aio_write", use the remaining length of
387 $data.
388
389 If $dataoffset is less than zero, it will be counted from the end of
390 $data.
391
392 The $data scalar *MUST NOT* be modified in any way while the request
393 is outstanding. Modifying it can result in segfaults or World War
394 III (if the necessary/optional hardware is installed).
114 395
115 Example: Read 15 bytes at offset 7 into scalar $buffer, starting at 396 Example: Read 15 bytes at offset 7 into scalar $buffer, starting at
116 offset 0 within the scalar: 397 offset 0 within the scalar:
117 398
118 aio_read $fh, 7, 15, $buffer, 0, sub { 399 aio_read $fh, 7, 15, $buffer, 0, sub {
119 $_[0] > 0 or die "read error: $!"; 400 $_[0] > 0 or die "read error: $!";
120 print "read $_[0] bytes: <$buffer>\n"; 401 print "read $_[0] bytes: <$buffer>\n";
121 }; 402 };
122 403
404 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
405 Tries to copy $length bytes from $in_fh to $out_fh. It starts
406 reading at byte offset $in_offset, and starts writing at the current
407 file offset of $out_fh. Because of that, it is not safe to issue
408 more than one "aio_sendfile" per $out_fh, as they will interfere
409 with each other. The same $in_fh works fine though, as this function
410 does not move or use the file offset of $in_fh.
411
412 Please note that "aio_sendfile" can read more bytes from $in_fh than
413 are written, and there is no way to find out how many more bytes
414 have been read from "aio_sendfile" alone, as "aio_sendfile" only
415 provides the number of bytes written to $out_fh. Only if the result
416 value equals $length one can assume that $length bytes have been
417 read.
418
419 Unlike with other "aio_" functions, it makes a lot of sense to use
420 "aio_sendfile" on non-blocking sockets, as long as one end
421 (typically the $in_fh) is a file - the file I/O will then be
422 asynchronous, while the socket I/O will be non-blocking. Note,
423 however, that you can run into a trap where "aio_sendfile" reads
424 some data with readahead, then fails to write all data, and when the
425 socket is ready the next time, the data in the cache is already
426 lost, forcing "aio_sendfile" to again hit the disk. Explicit
427 "aio_read" + "aio_write" let's you better control resource usage.
428
429 This call tries to make use of a native "sendfile"-like syscall to
430 provide zero-copy operation. For this to work, $out_fh should refer
431 to a socket, and $in_fh should refer to an mmap'able file.
432
433 If a native sendfile cannot be found or it fails with "ENOSYS",
434 "EINVAL", "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or
435 "ENOTSOCK", it will be emulated, so you can call "aio_sendfile" on
436 any type of filehandle regardless of the limitations of the
437 operating system.
438
439 As native sendfile syscalls (as practically any non-POSIX interface
440 hacked together in a hurry to improve benchmark numbers) tend to be
441 rather buggy on many systems, this implementation tries to work
442 around some known bugs in Linux and FreeBSD kernels (probably
443 others, too), but that might fail, so you really really should check
444 the return value of "aio_sendfile" - fewre bytes than expected might
445 have been transferred.
446
123 aio_readahead $fh,$offset,$length, $callback 447 aio_readahead $fh,$offset,$length, $callback->($retval)
124 Asynchronously reads the specified byte range into the page cache,
125 using the "readahead" syscall. If that syscall doesn't exist (likely
126 if your OS isn't Linux) the status will be -1 and $! is set to
127 "ENOSYS".
128
129 "aio_readahead" populates the page cache with data from a file so 448 "aio_readahead" populates the page cache with data from a file so
130 that subsequent reads from that file will not block on disk I/O. The 449 that subsequent reads from that file will not block on disk I/O. The
131 $offset argument specifies the starting point from which data is to 450 $offset argument specifies the starting point from which data is to
132 be read and $length specifies the number of bytes to be read. I/O is 451 be read and $length specifies the number of bytes to be read. I/O is
133 performed in whole pages, so that offset is effectively rounded down 452 performed in whole pages, so that offset is effectively rounded down
134 to a page boundary and bytes are read up to the next page boundary 453 to a page boundary and bytes are read up to the next page boundary
135 greater than or equal to (off-set+length). "aio_readahead" does not 454 greater than or equal to (off-set+length). "aio_readahead" does not
136 read beyond the end of the file. The current file offset of the file 455 read beyond the end of the file. The current file offset of the file
137 is left unchanged. 456 is left unchanged.
138 457
458 If that syscall doesn't exist (likely if your OS isn't Linux) it
459 will be emulated by simply reading the data, which would have a
460 similar effect.
461
139 aio_stat $fh_or_path, $callback 462 aio_stat $fh_or_path, $callback->($status)
140 aio_lstat $fh, $callback 463 aio_lstat $fh, $callback->($status)
141 Works like perl's "stat" or "lstat" in void context. The callback 464 Works like perl's "stat" or "lstat" in void context. The callback
142 will be called after the stat and the results will be available 465 will be called after the stat and the results will be available
143 using "stat _" or "-s _" etc... 466 using "stat _" or "-s _" etc...
144 467
145 The pathname passed to "aio_stat" must be absolute. See API NOTES, 468 The pathname passed to "aio_stat" must be absolute. See API NOTES,
… …
148 Currently, the stats are always 64-bit-stats, i.e. instead of 471 Currently, the stats are always 64-bit-stats, i.e. instead of
149 returning an error when stat'ing a large file, the results will be 472 returning an error when stat'ing a large file, the results will be
150 silently truncated unless perl itself is compiled with large file 473 silently truncated unless perl itself is compiled with large file
151 support. 474 support.
152 475
476 To help interpret the mode and dev/rdev stat values, IO::AIO offers
477 the following constants and functions (if not implemented, the
478 constants will be 0 and the functions will either "croak" or fall
479 back on traditional behaviour).
480
481 "S_IFMT", "S_IFIFO", "S_IFCHR", "S_IFBLK", "S_IFLNK", "S_IFREG",
482 "S_IFDIR", "S_IFWHT", "S_IFSOCK", "IO::AIO::major $dev_t",
483 "IO::AIO::minor $dev_t", "IO::AIO::makedev $major, $minor".
484
153 Example: Print the length of /etc/passwd: 485 Example: Print the length of /etc/passwd:
154 486
155 aio_stat "/etc/passwd", sub { 487 aio_stat "/etc/passwd", sub {
156 $_[0] and die "stat failed: $!"; 488 $_[0] and die "stat failed: $!";
157 print "size is ", -s _, "\n"; 489 print "size is ", -s _, "\n";
158 }; 490 };
159 491
492 aio_statvfs $fh_or_path, $callback->($statvfs)
493 Works like the POSIX "statvfs" or "fstatvfs" syscalls, depending on
494 whether a file handle or path was passed.
495
496 On success, the callback is passed a hash reference with the
497 following members: "bsize", "frsize", "blocks", "bfree", "bavail",
498 "files", "ffree", "favail", "fsid", "flag" and "namemax". On
499 failure, "undef" is passed.
500
501 The following POSIX IO::AIO::ST_* constants are defined: "ST_RDONLY"
502 and "ST_NOSUID".
503
504 The following non-POSIX IO::AIO::ST_* flag masks are defined to
505 their correct value when available, or to 0 on systems that do not
506 support them: "ST_NODEV", "ST_NOEXEC", "ST_SYNCHRONOUS",
507 "ST_MANDLOCK", "ST_WRITE", "ST_APPEND", "ST_IMMUTABLE",
508 "ST_NOATIME", "ST_NODIRATIME" and "ST_RELATIME".
509
510 Example: stat "/wd" and dump out the data if successful.
511
512 aio_statvfs "/wd", sub {
513 my $f = $_[0]
514 or die "statvfs: $!";
515
516 use Data::Dumper;
517 say Dumper $f;
518 };
519
520 # result:
521 {
522 bsize => 1024,
523 bfree => 4333064312,
524 blocks => 10253828096,
525 files => 2050765568,
526 flag => 4096,
527 favail => 2042092649,
528 bavail => 4333064312,
529 ffree => 2042092649,
530 namemax => 255,
531 frsize => 1024,
532 fsid => 1810
533 }
534
535 Here is a (likely partial) list of fsid values used by Linux - it is
536 safe to hardcode these when the $^O is "linux":
537
538 0x0000adf5 adfs
539 0x0000adff affs
540 0x5346414f afs
541 0x09041934 anon-inode filesystem
542 0x00000187 autofs
543 0x42465331 befs
544 0x1badface bfs
545 0x42494e4d binfmt_misc
546 0x9123683e btrfs
547 0x0027e0eb cgroupfs
548 0xff534d42 cifs
549 0x73757245 coda
550 0x012ff7b7 coh
551 0x28cd3d45 cramfs
552 0x453dcd28 cramfs-wend (wrong endianness)
553 0x64626720 debugfs
554 0x00001373 devfs
555 0x00001cd1 devpts
556 0x0000f15f ecryptfs
557 0x00414a53 efs
558 0x0000137d ext
559 0x0000ef53 ext2/ext3
560 0x0000ef51 ext2
561 0x00004006 fat
562 0x65735546 fuseblk
563 0x65735543 fusectl
564 0x0bad1dea futexfs
565 0x01161970 gfs2
566 0x47504653 gpfs
567 0x00004244 hfs
568 0xf995e849 hpfs
569 0x958458f6 hugetlbfs
570 0x2bad1dea inotifyfs
571 0x00009660 isofs
572 0x000072b6 jffs2
573 0x3153464a jfs
574 0x6b414653 k-afs
575 0x0bd00bd0 lustre
576 0x0000137f minix
577 0x0000138f minix 30 char names
578 0x00002468 minix v2
579 0x00002478 minix v2 30 char names
580 0x00004d5a minix v3
581 0x19800202 mqueue
582 0x00004d44 msdos
583 0x0000564c novell
584 0x00006969 nfs
585 0x6e667364 nfsd
586 0x00003434 nilfs
587 0x5346544e ntfs
588 0x00009fa1 openprom
589 0x7461636F ocfs2
590 0x00009fa0 proc
591 0x6165676c pstorefs
592 0x0000002f qnx4
593 0x858458f6 ramfs
594 0x52654973 reiserfs
595 0x00007275 romfs
596 0x67596969 rpc_pipefs
597 0x73636673 securityfs
598 0xf97cff8c selinux
599 0x0000517b smb
600 0x534f434b sockfs
601 0x73717368 squashfs
602 0x62656572 sysfs
603 0x012ff7b6 sysv2
604 0x012ff7b5 sysv4
605 0x01021994 tmpfs
606 0x15013346 udf
607 0x00011954 ufs
608 0x54190100 ufs byteswapped
609 0x00009fa2 usbdevfs
610 0x01021997 v9fs
611 0xa501fcf5 vxfs
612 0xabba1974 xenfs
613 0x012ff7b4 xenix
614 0x58465342 xfs
615 0x012fd16d xia
616
617 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
618 Works like perl's "utime" function (including the special case of
619 $atime and $mtime being undef). Fractional times are supported if
620 the underlying syscalls support them.
621
622 When called with a pathname, uses utimes(2) if available, otherwise
623 utime(2). If called on a file descriptor, uses futimes(2) if
624 available, otherwise returns ENOSYS, so this is not portable.
625
626 Examples:
627
628 # set atime and mtime to current time (basically touch(1)):
629 aio_utime "path", undef, undef;
630 # set atime to current time and mtime to beginning of the epoch:
631 aio_utime "path", time, undef; # undef==0
632
633 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
634 Works like perl's "chown" function, except that "undef" for either
635 $uid or $gid is being interpreted as "do not change" (but -1 can
636 also be used).
637
638 Examples:
639
640 # same as "chown root path" in the shell:
641 aio_chown "path", 0, -1;
642 # same as above:
643 aio_chown "path", 0, undef;
644
645 aio_truncate $fh_or_path, $offset, $callback->($status)
646 Works like truncate(2) or ftruncate(2).
647
648 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
649 Allocates or freed disk space according to the $mode argument. See
650 the linux "fallocate" docuemntation for details.
651
652 $mode can currently be 0 or "IO::AIO::FALLOC_FL_KEEP_SIZE" to
653 allocate space, or "IO::AIO::FALLOC_FL_PUNCH_HOLE |
654 IO::AIO::FALLOC_FL_KEEP_SIZE", to deallocate a file range.
655
656 The file system block size used by "fallocate" is presumably the
657 "f_bsize" returned by "statvfs".
658
659 If "fallocate" isn't available or cannot be emulated (currently no
660 emulation will be attempted), passes -1 and sets $! to "ENOSYS".
661
662 aio_chmod $fh_or_path, $mode, $callback->($status)
663 Works like perl's "chmod" function.
664
160 aio_unlink $pathname, $callback 665 aio_unlink $pathname, $callback->($status)
161 Asynchronously unlink (delete) a file and call the callback with the 666 Asynchronously unlink (delete) a file and call the callback with the
162 result code. 667 result code.
163 668
669 aio_mknod $pathname, $mode, $dev, $callback->($status)
670 [EXPERIMENTAL]
671
672 Asynchronously create a device node (or fifo). See mknod(2).
673
674 The only (POSIX-) portable way of calling this function is:
675
676 aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
677
678 See "aio_stat" for info about some potentially helpful extra
679 constants and functions.
680
681 aio_link $srcpath, $dstpath, $callback->($status)
682 Asynchronously create a new link to the existing object at $srcpath
683 at the path $dstpath and call the callback with the result code.
684
685 aio_symlink $srcpath, $dstpath, $callback->($status)
686 Asynchronously create a new symbolic link to the existing object at
687 $srcpath at the path $dstpath and call the callback with the result
688 code.
689
690 aio_readlink $pathname, $callback->($link)
691 Asynchronously read the symlink specified by $path and pass it to
692 the callback. If an error occurs, nothing or undef gets passed to
693 the callback.
694
695 aio_realpath $pathname, $callback->($path)
696 Asynchronously make the path absolute and resolve any symlinks in
697 $path. The resulting path only consists of directories (Same as
698 Cwd::realpath).
699
700 This request can be used to get the absolute path of the current
701 working directory by passing it a path of . (a single dot).
702
703 aio_rename $srcpath, $dstpath, $callback->($status)
704 Asynchronously rename the object at $srcpath to $dstpath, just as
705 rename(2) and call the callback with the result code.
706
707 aio_mkdir $pathname, $mode, $callback->($status)
708 Asynchronously mkdir (create) a directory and call the callback with
709 the result code. $mode will be modified by the umask at the time the
710 request is executed, so do not change your umask.
711
712 aio_rmdir $pathname, $callback->($status)
713 Asynchronously rmdir (delete) a directory and call the callback with
714 the result code.
715
716 aio_readdir $pathname, $callback->($entries)
717 Unlike the POSIX call of the same name, "aio_readdir" reads an
718 entire directory (i.e. opendir + readdir + closedir). The entries
719 will not be sorted, and will NOT include the "." and ".." entries.
720
721 The callback is passed a single argument which is either "undef" or
722 an array-ref with the filenames.
723
724 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
725 Quite similar to "aio_readdir", but the $flags argument allows one
726 to tune behaviour and output format. In case of an error, $entries
727 will be "undef".
728
729 The flags are a combination of the following constants, ORed
730 together (the flags will also be passed to the callback, possibly
731 modified):
732
733 IO::AIO::READDIR_DENTS
734 When this flag is off, then the callback gets an arrayref
735 consisting of names only (as with "aio_readdir"), otherwise it
736 gets an arrayref with "[$name, $type, $inode]" arrayrefs, each
737 describing a single directory entry in more detail.
738
739 $name is the name of the entry.
740
741 $type is one of the "IO::AIO::DT_xxx" constants:
742
743 "IO::AIO::DT_UNKNOWN", "IO::AIO::DT_FIFO", "IO::AIO::DT_CHR",
744 "IO::AIO::DT_DIR", "IO::AIO::DT_BLK", "IO::AIO::DT_REG",
745 "IO::AIO::DT_LNK", "IO::AIO::DT_SOCK", "IO::AIO::DT_WHT".
746
747 "IO::AIO::DT_UNKNOWN" means just that: readdir does not know. If
748 you need to know, you have to run stat yourself. Also, for speed
749 reasons, the $type scalars are read-only: you can not modify
750 them.
751
752 $inode is the inode number (which might not be exact on systems
753 with 64 bit inode numbers and 32 bit perls). This field has
754 unspecified content on systems that do not deliver the inode
755 information.
756
757 IO::AIO::READDIR_DIRS_FIRST
758 When this flag is set, then the names will be returned in an
759 order where likely directories come first, in optimal stat
760 order. This is useful when you need to quickly find directories,
761 or you want to find all directories while avoiding to stat()
762 each entry.
763
764 If the system returns type information in readdir, then this is
765 used to find directories directly. Otherwise, likely directories
766 are names beginning with ".", or otherwise names with no dots,
767 of which names with short names are tried first.
768
769 IO::AIO::READDIR_STAT_ORDER
770 When this flag is set, then the names will be returned in an
771 order suitable for stat()'ing each one. That is, when you plan
772 to stat() all files in the given directory, then the returned
773 order will likely be fastest.
774
775 If both this flag and "IO::AIO::READDIR_DIRS_FIRST" are
776 specified, then the likely dirs come first, resulting in a less
777 optimal stat order.
778
779 IO::AIO::READDIR_FOUND_UNKNOWN
780 This flag should not be set when calling "aio_readdirx".
781 Instead, it is being set by "aio_readdirx", when any of the
782 $type's found were "IO::AIO::DT_UNKNOWN". The absence of this
783 flag therefore indicates that all $type's are known, which can
784 be used to speed up some algorithms.
785
786 aio_load $pathname, $data, $callback->($status)
787 This is a composite request that tries to fully load the given file
788 into memory. Status is the same as with aio_read.
789
790 aio_copy $srcpath, $dstpath, $callback->($status)
791 Try to copy the *file* (directories not supported as either source
792 or destination) from $srcpath to $dstpath and call the callback with
793 a status of 0 (ok) or -1 (error, see $!).
794
795 This is a composite request that creates the destination file with
796 mode 0200 and copies the contents of the source file into it using
797 "aio_sendfile", followed by restoring atime, mtime, access mode and
798 uid/gid, in that order.
799
800 If an error occurs, the partial destination file will be unlinked,
801 if possible, except when setting atime, mtime, access mode and
802 uid/gid, where errors are being ignored.
803
804 aio_move $srcpath, $dstpath, $callback->($status)
805 Try to move the *file* (directories not supported as either source
806 or destination) from $srcpath to $dstpath and call the callback with
807 a status of 0 (ok) or -1 (error, see $!).
808
809 This is a composite request that tries to rename(2) the file first;
810 if rename fails with "EXDEV", it copies the file with "aio_copy"
811 and, if that is successful, unlinks the $srcpath.
812
813 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
814 Scans a directory (similar to "aio_readdir") but additionally tries
815 to efficiently separate the entries of directory $path into two sets
816 of names, directories you can recurse into (directories), and ones
817 you cannot recurse into (everything else, including symlinks to
818 directories).
819
820 "aio_scandir" is a composite request that creates of many sub
821 requests_ $maxreq specifies the maximum number of outstanding aio
822 requests that this function generates. If it is "<= 0", then a
823 suitable default will be chosen (currently 4).
824
825 On error, the callback is called without arguments, otherwise it
826 receives two array-refs with path-relative entry names.
827
828 Example:
829
830 aio_scandir $dir, 0, sub {
831 my ($dirs, $nondirs) = @_;
832 print "real directories: @$dirs\n";
833 print "everything else: @$nondirs\n";
834 };
835
836 Implementation notes.
837
838 The "aio_readdir" cannot be avoided, but "stat()"'ing every entry
839 can.
840
841 If readdir returns file type information, then this is used directly
842 to find directories.
843
844 Otherwise, after reading the directory, the modification time, size
845 etc. of the directory before and after the readdir is checked, and
846 if they match (and isn't the current time), the link count will be
847 used to decide how many entries are directories (if >= 2).
848 Otherwise, no knowledge of the number of subdirectories will be
849 assumed.
850
851 Then entries will be sorted into likely directories a non-initial
852 dot currently) and likely non-directories (see "aio_readdirx"). Then
853 every entry plus an appended "/." will be "stat"'ed, likely
854 directories first, in order of their inode numbers. If that
855 succeeds, it assumes that the entry is a directory or a symlink to
856 directory (which will be checked separately). This is often faster
857 than stat'ing the entry itself because filesystems might detect the
858 type of the entry without reading the inode data (e.g. ext2fs
859 filetype feature), even on systems that cannot return the filetype
860 information on readdir.
861
862 If the known number of directories (link count - 2) has been
863 reached, the rest of the entries is assumed to be non-directories.
864
865 This only works with certainty on POSIX (= UNIX) filesystems, which
866 fortunately are the vast majority of filesystems around.
867
868 It will also likely work on non-POSIX filesystems with reduced
869 efficiency as those tend to return 0 or 1 as link counts, which
870 disables the directory counting heuristic.
871
872 aio_rmtree $pathname, $callback->($status)
873 Delete a directory tree starting (and including) $path, return the
874 status of the final "rmdir" only. This is a composite request that
875 uses "aio_scandir" to recurse into and rmdir directories, and unlink
876 everything else.
877
878 aio_sync $callback->($status)
879 Asynchronously call sync and call the callback when finished.
880
164 aio_fsync $fh, $callback 881 aio_fsync $fh, $callback->($status)
165 Asynchronously call fsync on the given filehandle and call the 882 Asynchronously call fsync on the given filehandle and call the
166 callback with the fsync result code. 883 callback with the fsync result code.
167 884
168 aio_fdatasync $fh, $callback 885 aio_fdatasync $fh, $callback->($status)
169 Asynchronously call fdatasync on the given filehandle and call the 886 Asynchronously call fdatasync on the given filehandle and call the
170 callback with the fdatasync result code. Might set $! to "ENOSYS" if 887 callback with the fdatasync result code.
171 "fdatasync" is not available. 888
889 If this call isn't available because your OS lacks it or it couldn't
890 be detected, it will be emulated by calling "fsync" instead.
891
892 aio_syncfs $fh, $callback->($status)
893 Asynchronously call the syncfs syscall to sync the filesystem
894 associated to the given filehandle and call the callback with the
895 syncfs result code. If syncfs is not available, calls sync(), but
896 returns -1 and sets errno to "ENOSYS" nevertheless.
897
898 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
899 Sync the data portion of the file specified by $offset and $length
900 to disk (but NOT the metadata), by calling the Linux-specific
901 sync_file_range call. If sync_file_range is not available or it
902 returns ENOSYS, then fdatasync or fsync is being substituted.
903
904 $flags can be a combination of
905 "IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE",
906 "IO::AIO::SYNC_FILE_RANGE_WRITE" and
907 "IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER": refer to the sync_file_range
908 manpage for details.
909
910 aio_pathsync $pathname, $callback->($status)
911 This request tries to open, fsync and close the given path. This is
912 a composite request intended to sync directories after directory
913 operations (E.g. rename). This might not work on all operating
914 systems or have any specific effect, but usually it makes sure that
915 directory changes get written to disc. It works for anything that
916 can be opened for read-only, not just directories.
917
918 Future versions of this function might fall back to other methods
919 when "fsync" on the directory fails (such as calling "sync").
920
921 Passes 0 when everything went ok, and -1 on error.
922
923 aio_msync $scalar, $offset = 0, $length = undef, flags = 0,
924 $callback->($status)
925 This is a rather advanced IO::AIO call, which only works on
926 mmap(2)ed scalars (see the "IO::AIO::mmap" function, although it
927 also works on data scalars managed by the Sys::Mmap or Mmap modules,
928 note that the scalar must only be modified in-place while an aio
929 operation is pending on it).
930
931 It calls the "msync" function of your OS, if available, with the
932 memory area starting at $offset in the string and ending $length
933 bytes later. If $length is negative, counts from the end, and if
934 $length is "undef", then it goes till the end of the string. The
935 flags can be a combination of "IO::AIO::MS_ASYNC",
936 "IO::AIO::MS_INVALIDATE" and "IO::AIO::MS_SYNC".
937
938 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0,
939 $callback->($status)
940 This is a rather advanced IO::AIO call, which works best on
941 mmap(2)ed scalars.
942
943 It touches (reads or writes) all memory pages in the specified range
944 inside the scalar. All caveats and parameters are the same as for
945 "aio_msync", above, except for flags, which must be either 0 (which
946 reads all pages and ensures they are instantiated) or
947 "IO::AIO::MT_MODIFY", which modifies the memory page s(by reading
948 and writing an octet from it, which dirties the page).
949
950 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
951 This is a rather advanced IO::AIO call, which works best on
952 mmap(2)ed scalars.
953
954 It reads in all the pages of the underlying storage into memory (if
955 any) and locks them, so they are not getting swapped/paged out or
956 removed.
957
958 If $length is undefined, then the scalar will be locked till the
959 end.
960
961 On systems that do not implement "mlock", this function returns -1
962 and sets errno to "ENOSYS".
963
964 Note that the corresponding "munlock" is synchronous and is
965 documented under "MISCELLANEOUS FUNCTIONS".
966
967 Example: open a file, mmap and mlock it - both will be undone when
968 $data gets destroyed.
969
970 open my $fh, "<", $path or die "$path: $!";
971 my $data;
972 IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
973 aio_mlock $data; # mlock in background
974
975 aio_mlockall $flags, $callback->($status)
976 Calls the "mlockall" function with the given $flags (a combination
977 of "IO::AIO::MCL_CURRENT" and "IO::AIO::MCL_FUTURE").
978
979 On systems that do not implement "mlockall", this function returns
980 -1 and sets errno to "ENOSYS".
981
982 Note that the corresponding "munlockall" is synchronous and is
983 documented under "MISCELLANEOUS FUNCTIONS".
984
985 Example: asynchronously lock all current and future pages into
986 memory.
987
988 aio_mlockall IO::AIO::MCL_FUTURE;
989
990 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
991 Queries the extents of the given file (by calling the Linux "FIEMAP"
992 ioctl, see <http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for
993 details). If the ioctl is not available on your OS, then this
994 request will fail with "ENOSYS".
995
996 $start is the starting offset to query extents for, $length is the
997 size of the range to query - if it is "undef", then the whole file
998 will be queried.
999
1000 $flags is a combination of flags ("IO::AIO::FIEMAP_FLAG_SYNC" or
1001 "IO::AIO::FIEMAP_FLAG_XATTR" - "IO::AIO::FIEMAP_FLAGS_COMPAT" is
1002 also exported), and is normally 0 or "IO::AIO::FIEMAP_FLAG_SYNC" to
1003 query the data portion.
1004
1005 $count is the maximum number of extent records to return. If it is
1006 "undef", then IO::AIO queries all extents of the range. As a very
1007 special case, if it is 0, then the callback receives the number of
1008 extents instead of the extents themselves (which is unreliable, see
1009 below).
1010
1011 If an error occurs, the callback receives no arguments. The special
1012 "errno" value "IO::AIO::EBADR" is available to test for flag errors.
1013
1014 Otherwise, the callback receives an array reference with extent
1015 structures. Each extent structure is an array reference itself, with
1016 the following members:
1017
1018 [$logical, $physical, $length, $flags]
1019
1020 Flags is any combination of the following flag values (typically
1021 either 0 or "IO::AIO::FIEMAP_EXTENT_LAST" (1)):
1022
1023 "IO::AIO::FIEMAP_EXTENT_LAST", "IO::AIO::FIEMAP_EXTENT_UNKNOWN",
1024 "IO::AIO::FIEMAP_EXTENT_DELALLOC", "IO::AIO::FIEMAP_EXTENT_ENCODED",
1025 "IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED",
1026 "IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED",
1027 "IO::AIO::FIEMAP_EXTENT_DATA_INLINE",
1028 "IO::AIO::FIEMAP_EXTENT_DATA_TAIL",
1029 "IO::AIO::FIEMAP_EXTENT_UNWRITTEN", "IO::AIO::FIEMAP_EXTENT_MERGED"
1030 or "IO::AIO::FIEMAP_EXTENT_SHARED".
1031
1032 At the time of this writing (Linux 3.2), this requets is unreliable
1033 unless $count is "undef", as the kernel has all sorts of bugs
1034 preventing it to return all extents of a range for files with large
1035 number of extents. The code works around all these issues if $count
1036 is undef.
1037
1038 aio_group $callback->(...)
1039 This is a very special aio request: Instead of doing something, it
1040 is a container for other aio requests, which is useful if you want
1041 to bundle many requests into a single, composite, request with a
1042 definite callback and the ability to cancel the whole request with
1043 its subrequests.
1044
1045 Returns an object of class IO::AIO::GRP. See its documentation below
1046 for more info.
1047
1048 Example:
1049
1050 my $grp = aio_group sub {
1051 print "all stats done\n";
1052 };
1053
1054 add $grp
1055 (aio_stat ...),
1056 (aio_stat ...),
1057 ...;
1058
1059 aio_nop $callback->()
1060 This is a special request - it does nothing in itself and is only
1061 used for side effects, such as when you want to add a dummy request
1062 to a group so that finishing the requests in the group depends on
1063 executing the given code.
1064
1065 While this request does nothing, it still goes through the execution
1066 phase and still requires a worker thread. Thus, the callback will
1067 not be executed immediately but only after other requests in the
1068 queue have entered their execution phase. This can be used to
1069 measure request latency.
1070
1071 IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
1072 Mainly used for debugging and benchmarking, this aio request puts
1073 one of the request workers to sleep for the given time.
1074
1075 While it is theoretically handy to have simple I/O scheduling
1076 requests like sleep and file handle readable/writable, the overhead
1077 this creates is immense (it blocks a thread for a long time) so do
1078 not use this function except to put your application under
1079 artificial I/O pressure.
1080
1081 IO::AIO::WD - multiple working directories
1082 Your process only has one current working directory, which is used by
1083 all threads. This makes it hard to use relative paths (some other
1084 component could call "chdir" at any time, and it is hard to control when
1085 the path will be used by IO::AIO).
1086
1087 One solution for this is to always use absolute paths. This usually
1088 works, but can be quite slow (the kernel has to walk the whole path on
1089 every access), and can also be a hassle to implement.
1090
1091 Newer POSIX systems have a number of functions (openat, fdopendir,
1092 futimensat and so on) that make it possible to specify working
1093 directories per operation.
1094
1095 For portability, and because the clowns who "designed", or shall I
1096 write, perpetrated this new interface were obviously half-drunk, this
1097 abstraction cannot be perfect, though.
1098
1099 IO::AIO allows you to convert directory paths into a so-called
1100 IO::AIO::WD object. This object stores the canonicalised, absolute
1101 version of the path, and on systems that allow it, also a directory file
1102 descriptor.
1103
1104 Everywhere where a pathname is accepted by IO::AIO (e.g. in "aio_stat"
1105 or "aio_unlink"), one can specify an array reference with an IO::AIO::WD
1106 object and a pathname instead (or the IO::AIO::WD object alone, which
1107 gets interpreted as "[$wd, "."]"). If the pathname is absolute, the
1108 IO::AIO::WD object is ignored, otherwise the pathname is resolved
1109 relative to that IO::AIO::WD object.
1110
1111 For example, to get a wd object for /etc and then stat passwd inside,
1112 you would write:
1113
1114 aio_wd "/etc", sub {
1115 my $etcdir = shift;
1116
1117 # although $etcdir can be undef on error, there is generally no reason
1118 # to check for errors here, as aio_stat will fail with ENOENT
1119 # when $etcdir is undef.
1120
1121 aio_stat [$etcdir, "passwd"], sub {
1122 # yay
1123 };
1124 };
1125
1126 That "aio_wd" is a request and not a normal function shows that creating
1127 an IO::AIO::WD object is itself a potentially blocking operation, which
1128 is why it is done asynchronously.
1129
1130 To stat the directory obtained with "aio_wd" above, one could write
1131 either of the following three request calls:
1132
1133 aio_lstat "/etc" , sub { ... # pathname as normal string
1134 aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
1135 aio_lstat $wd , sub { ... # shorthand for the previous
1136
1137 As with normal pathnames, IO::AIO keeps a copy of the working directory
1138 object and the pathname string, so you could write the following without
1139 causing any issues due to $path getting reused:
1140
1141 my $path = [$wd, undef];
1142
1143 for my $name (qw(abc def ghi)) {
1144 $path->[1] = $name;
1145 aio_stat $path, sub {
1146 # ...
1147 };
1148 }
1149
1150 There are some caveats: when directories get renamed (or deleted), the
1151 pathname string doesn't change, so will point to the new directory (or
1152 nowhere at all), while the directory fd, if available on the system,
1153 will still point to the original directory. Most functions accepting a
1154 pathname will use the directory fd on newer systems, and the string on
1155 older systems. Some functions (such as realpath) will always rely on the
1156 string form of the pathname.
1157
1158 So this fucntionality is mainly useful to get some protection against
1159 "chdir", to easily get an absolute path out of a relative path for
1160 future reference, and to speed up doing many operations in the same
1161 directory (e.g. when stat'ing all files in a directory).
1162
1163 The following functions implement this working directory abstraction:
1164
1165 aio_wd $pathname, $callback->($wd)
1166 Asynchonously canonicalise the given pathname and convert it to an
1167 IO::AIO::WD object representing it. If possible and supported on the
1168 system, also open a directory fd to speed up pathname resolution
1169 relative to this working directory.
1170
1171 If something goes wrong, then "undef" is passwd to the callback
1172 instead of a working directory object and $! is set appropriately.
1173 Since passing "undef" as working directory component of a pathname
1174 fails the request with "ENOENT", there is often no need for error
1175 checking in the "aio_wd" callback, as future requests using the
1176 value will fail in the expected way.
1177
1178 If this call isn't available because your OS lacks it or it couldn't
1179 be detected, it will be emulated by calling "fsync" instead.
1180
1181 IO::AIO::CWD
1182 This is a compiletime constant (object) that represents the process
1183 current working directory.
1184
1185 Specifying this object as working directory object for a pathname is
1186 as if the pathname would be specified directly, without a directory
1187 object, e.g., these calls are functionally identical:
1188
1189 aio_stat "somefile", sub { ... };
1190 aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
1191
1192 IO::AIO::REQ CLASS
1193 All non-aggregate "aio_*" functions return an object of this class when
1194 called in non-void context.
1195
1196 cancel $req
1197 Cancels the request, if possible. Has the effect of skipping
1198 execution when entering the execute state and skipping calling the
1199 callback when entering the the result state, but will leave the
1200 request otherwise untouched (with the exception of readdir). That
1201 means that requests that currently execute will not be stopped and
1202 resources held by the request will not be freed prematurely.
1203
1204 cb $req $callback->(...)
1205 Replace (or simply set) the callback registered to the request.
1206
1207 IO::AIO::GRP CLASS
1208 This class is a subclass of IO::AIO::REQ, so all its methods apply to
1209 objects of this class, too.
1210
1211 A IO::AIO::GRP object is a special request that can contain multiple
1212 other aio requests.
1213
1214 You create one by calling the "aio_group" constructing function with a
1215 callback that will be called when all contained requests have entered
1216 the "done" state:
1217
1218 my $grp = aio_group sub {
1219 print "all requests are done\n";
1220 };
1221
1222 You add requests by calling the "add" method with one or more
1223 "IO::AIO::REQ" objects:
1224
1225 $grp->add (aio_unlink "...");
1226
1227 add $grp aio_stat "...", sub {
1228 $_[0] or return $grp->result ("error");
1229
1230 # add another request dynamically, if first succeeded
1231 add $grp aio_open "...", sub {
1232 $grp->result ("ok");
1233 };
1234 };
1235
1236 This makes it very easy to create composite requests (see the source of
1237 "aio_move" for an application) that work and feel like simple requests.
1238
1239 * The IO::AIO::GRP objects will be cleaned up during calls to
1240 "IO::AIO::poll_cb", just like any other request.
1241
1242 * They can be canceled like any other request. Canceling will cancel
1243 not only the request itself, but also all requests it contains.
1244
1245 * They can also can also be added to other IO::AIO::GRP objects.
1246
1247 * You must not add requests to a group from within the group callback
1248 (or any later time).
1249
1250 Their lifetime, simplified, looks like this: when they are empty, they
1251 will finish very quickly. If they contain only requests that are in the
1252 "done" state, they will also finish. Otherwise they will continue to
1253 exist.
1254
1255 That means after creating a group you have some time to add requests
1256 (precisely before the callback has been invoked, which is only done
1257 within the "poll_cb"). And in the callbacks of those requests, you can
1258 add further requests to the group. And only when all those requests have
1259 finished will the the group itself finish.
1260
1261 add $grp ...
1262 $grp->add (...)
1263 Add one or more requests to the group. Any type of IO::AIO::REQ can
1264 be added, including other groups, as long as you do not create
1265 circular dependencies.
1266
1267 Returns all its arguments.
1268
1269 $grp->cancel_subs
1270 Cancel all subrequests and clears any feeder, but not the group
1271 request itself. Useful when you queued a lot of events but got a
1272 result early.
1273
1274 The group request will finish normally (you cannot add requests to
1275 the group).
1276
1277 $grp->result (...)
1278 Set the result value(s) that will be passed to the group callback
1279 when all subrequests have finished and set the groups errno to the
1280 current value of errno (just like calling "errno" without an error
1281 number). By default, no argument will be passed and errno is zero.
1282
1283 $grp->errno ([$errno])
1284 Sets the group errno value to $errno, or the current value of errno
1285 when the argument is missing.
1286
1287 Every aio request has an associated errno value that is restored
1288 when the callback is invoked. This method lets you change this value
1289 from its default (0).
1290
1291 Calling "result" will also set errno, so make sure you either set $!
1292 before the call to "result", or call c<errno> after it.
1293
1294 feed $grp $callback->($grp)
1295 Sets a feeder/generator on this group: every group can have an
1296 attached generator that generates requests if idle. The idea behind
1297 this is that, although you could just queue as many requests as you
1298 want in a group, this might starve other requests for a potentially
1299 long time. For example, "aio_scandir" might generate hundreds of
1300 thousands of "aio_stat" requests, delaying any later requests for a
1301 long time.
1302
1303 To avoid this, and allow incremental generation of requests, you can
1304 instead a group and set a feeder on it that generates those
1305 requests. The feed callback will be called whenever there are few
1306 enough (see "limit", below) requests active in the group itself and
1307 is expected to queue more requests.
1308
1309 The feed callback can queue as many requests as it likes (i.e. "add"
1310 does not impose any limits).
1311
1312 If the feed does not queue more requests when called, it will be
1313 automatically removed from the group.
1314
1315 If the feed limit is 0 when this method is called, it will be set to
1316 2 automatically.
1317
1318 Example:
1319
1320 # stat all files in @files, but only ever use four aio requests concurrently:
1321
1322 my $grp = aio_group sub { print "finished\n" };
1323 limit $grp 4;
1324 feed $grp sub {
1325 my $file = pop @files
1326 or return;
1327
1328 add $grp aio_stat $file, sub { ... };
1329 };
1330
1331 limit $grp $num
1332 Sets the feeder limit for the group: The feeder will be called
1333 whenever the group contains less than this many requests.
1334
1335 Setting the limit to 0 will pause the feeding process.
1336
1337 The default value for the limit is 0, but note that setting a feeder
1338 automatically bumps it up to 2.
172 1339
173 SUPPORT FUNCTIONS 1340 SUPPORT FUNCTIONS
1341 EVENT PROCESSING AND EVENT LOOP INTEGRATION
174 $fileno = IO::AIO::poll_fileno 1342 $fileno = IO::AIO::poll_fileno
175 Return the *request result pipe file descriptor*. This filehandle 1343 Return the *request result pipe file descriptor*. This filehandle
176 must be polled for reading by some mechanism outside this module 1344 must be polled for reading by some mechanism outside this module
177 (e.g. Event or select, see below or the SYNOPSIS). If the pipe 1345 (e.g. EV, Glib, select and so on, see below or the SYNOPSIS). If the
178 becomes readable you have to call "poll_cb" to check the results. 1346 pipe becomes readable you have to call "poll_cb" to check the
1347 results.
179 1348
180 See "poll_cb" for an example. 1349 See "poll_cb" for an example.
181 1350
182 IO::AIO::poll_cb 1351 IO::AIO::poll_cb
183 Process all outstanding events on the result pipe. You have to call 1352 Process some outstanding events on the result pipe. You have to call
184 this regularly. Returns the number of events processed. Returns 1353 this regularly. Returns 0 if all events could be processed (or there
1354 were no events to process), or -1 if it returned earlier for
185 immediately when no events are outstanding. 1355 whatever reason. Returns immediately when no events are outstanding.
1356 The amount of events processed depends on the settings of
1357 "IO::AIO::max_poll_req" and "IO::AIO::max_poll_time".
1358
1359 If not all requests were processed for whatever reason, the
1360 filehandle will still be ready when "poll_cb" returns, so normally
1361 you don't have to do anything special to have it called later.
1362
1363 Apart from calling "IO::AIO::poll_cb" when the event filehandle
1364 becomes ready, it can be beneficial to call this function from loops
1365 which submit a lot of requests, to make sure the results get
1366 processed when they become available and not just when the loop is
1367 finished and the event loop takes over again. This function returns
1368 very fast when there are no outstanding requests.
186 1369
187 Example: Install an Event watcher that automatically calls 1370 Example: Install an Event watcher that automatically calls
188 IO::AIO::poll_cb with high priority: 1371 IO::AIO::poll_cb with high priority (more examples can be found in
1372 the SYNOPSIS section, at the top of this document):
189 1373
190 Event->io (fd => IO::AIO::poll_fileno, 1374 Event->io (fd => IO::AIO::poll_fileno,
191 poll => 'r', async => 1, 1375 poll => 'r', async => 1,
192 cb => \&IO::AIO::poll_cb); 1376 cb => \&IO::AIO::poll_cb);
193 1377
194 IO::AIO::poll_wait 1378 IO::AIO::poll_wait
1379 If there are any outstanding requests and none of them in the result
195 Wait till the result filehandle becomes ready for reading (simply 1380 phase, wait till the result filehandle becomes ready for reading
196 does a "select" on the filehandle. This is useful if you want to 1381 (simply does a "select" on the filehandle. This is useful if you
197 synchronously wait for some requests to finish). 1382 want to synchronously wait for some requests to finish).
198 1383
199 See "nreqs" for an example. 1384 See "nreqs" for an example.
200 1385
1386 IO::AIO::poll
1387 Waits until some requests have been handled.
1388
1389 Returns the number of requests processed, but is otherwise strictly
1390 equivalent to:
1391
1392 IO::AIO::poll_wait, IO::AIO::poll_cb
1393
201 IO::AIO::nreqs 1394 IO::AIO::flush
202 Returns the number of requests currently outstanding (i.e. for which 1395 Wait till all outstanding AIO requests have been handled.
203 their callback has not been invoked yet).
204 1396
205 Example: wait till there are no outstanding requests anymore: 1397 Strictly equivalent to:
206 1398
207 IO::AIO::poll_wait, IO::AIO::poll_cb 1399 IO::AIO::poll_wait, IO::AIO::poll_cb
208 while IO::AIO::nreqs; 1400 while IO::AIO::nreqs;
209 1401
210 IO::AIO::flush 1402 IO::AIO::max_poll_reqs $nreqs
211 Wait till all outstanding AIO requests have been handled. 1403 IO::AIO::max_poll_time $seconds
1404 These set the maximum number of requests (default 0, meaning
1405 infinity) that are being processed by "IO::AIO::poll_cb" in one
1406 call, respectively the maximum amount of time (default 0, meaning
1407 infinity) spent in "IO::AIO::poll_cb" to process requests (more
1408 correctly the mininum amount of time "poll_cb" is allowed to use).
212 1409
213 Strictly equivalent to: 1410 Setting "max_poll_time" to a non-zero value creates an overhead of
1411 one syscall per request processed, which is not normally a problem
1412 unless your callbacks are really really fast or your OS is really
1413 really slow (I am not mentioning Solaris here). Using
1414 "max_poll_reqs" incurs no overhead.
214 1415
215 IO::AIO::poll_wait, IO::AIO::poll_cb 1416 Setting these is useful if you want to ensure some level of
216 while IO::AIO::nreqs; 1417 interactiveness when perl is not fast enough to process all requests
1418 in time.
217 1419
218 IO::AIO::poll 1420 For interactive programs, values such as 0.01 to 0.1 should be fine.
219 Waits until some requests have been handled.
220 1421
221 Strictly equivalent to: 1422 Example: Install an Event watcher that automatically calls
1423 IO::AIO::poll_cb with low priority, to ensure that other parts of
1424 the program get the CPU sometimes even under high AIO load.
222 1425
223 IO::AIO::poll_wait, IO::AIO::poll_cb 1426 # try not to spend much more than 0.1s in poll_cb
224 if IO::AIO::nreqs; 1427 IO::AIO::max_poll_time 0.1;
225 1428
1429 # use a low priority so other tasks have priority
1430 Event->io (fd => IO::AIO::poll_fileno,
1431 poll => 'r', nice => 1,
1432 cb => &IO::AIO::poll_cb);
1433
1434 CONTROLLING THE NUMBER OF THREADS
226 IO::AIO::min_parallel $nthreads 1435 IO::AIO::min_parallel $nthreads
227 Set the minimum number of AIO threads to $nthreads. The default is 1436 Set the minimum number of AIO threads to $nthreads. The current
228 1, which means a single asynchronous operation can be done at one 1437 default is 8, which means eight asynchronous operations can execute
229 time (the number of outstanding operations, however, is unlimited). 1438 concurrently at any one time (the number of outstanding requests,
1439 however, is unlimited).
230 1440
1441 IO::AIO starts threads only on demand, when an AIO request is queued
1442 and no free thread exists. Please note that queueing up a hundred
1443 requests can create demand for a hundred threads, even if it turns
1444 out that everything is in the cache and could have been processed
1445 faster by a single thread.
1446
231 It is recommended to keep the number of threads low, as some Linux 1447 It is recommended to keep the number of threads relatively low, as
232 kernel versions will scale negatively with the number of threads 1448 some Linux kernel versions will scale negatively with the number of
233 (higher parallelity => MUCH higher latency). With current Linux 2.6 1449 threads (higher parallelity => MUCH higher latency). With current
234 versions, 4-32 threads should be fine. 1450 Linux 2.6 versions, 4-32 threads should be fine.
235 1451
236 Under normal circumstances you don't need to call this function, as 1452 Under most circumstances you don't need to call this function, as
237 this module automatically starts some threads (the exact number 1453 the module selects a default that is suitable for low to moderate
238 might change, and is currently 4). 1454 load.
239 1455
240 IO::AIO::max_parallel $nthreads 1456 IO::AIO::max_parallel $nthreads
241 Sets the maximum number of AIO threads to $nthreads. If more than 1457 Sets the maximum number of AIO threads to $nthreads. If more than
242 the specified number of threads are currently running, kill them. 1458 the specified number of threads are currently running, this function
243 This function blocks until the limit is reached. 1459 kills them. This function blocks until the limit is reached.
1460
1461 While $nthreads are zero, aio requests get queued but not executed
1462 until the number of threads has been increased again.
244 1463
245 This module automatically runs "max_parallel 0" at program end, to 1464 This module automatically runs "max_parallel 0" at program end, to
246 ensure that all threads are killed and that there are no outstanding 1465 ensure that all threads are killed and that there are no outstanding
247 requests. 1466 requests.
248 1467
249 Under normal circumstances you don't need to call this function. 1468 Under normal circumstances you don't need to call this function.
250 1469
1470 IO::AIO::max_idle $nthreads
1471 Limit the number of threads (default: 4) that are allowed to idle
1472 (i.e., threads that did not get a request to process within the idle
1473 timeout (default: 10 seconds). That means if a thread becomes idle
1474 while $nthreads other threads are also idle, it will free its
1475 resources and exit.
1476
1477 This is useful when you allow a large number of threads (e.g. 100 or
1478 1000) to allow for extremely high load situations, but want to free
1479 resources under normal circumstances (1000 threads can easily
1480 consume 30MB of RAM).
1481
1482 The default is probably ok in most situations, especially if thread
1483 creation is fast. If thread creation is very slow on your system you
1484 might want to use larger values.
1485
1486 IO::AIO::idle_timeout $seconds
1487 Sets the minimum idle timeout (default 10) after which worker
1488 threads are allowed to exit. SEe "IO::AIO::max_idle".
1489
251 $oldnreqs = IO::AIO::max_outstanding $nreqs 1490 IO::AIO::max_outstanding $maxreqs
252 Sets the maximum number of outstanding requests to $nreqs. If you 1491 Sets the maximum number of outstanding requests to $nreqs. If you do
253 try to queue up more than this number of requests, the caller will 1492 queue up more than this number of requests, the next call to
1493 "IO::AIO::poll_cb" (and other functions calling "poll_cb", such as
1494 "IO::AIO::flush" or "IO::AIO::poll") will block until the limit is
1495 no longer exceeded.
1496
1497 In other words, this setting does not enforce a queue limit, but can
1498 be used to make poll functions block if the limit is exceeded.
1499
1500 This is a very bad function to use in interactive programs because
1501 it blocks, and a bad way to reduce concurrency because it is
1502 inexact: Better use an "aio_group" together with a feed callback.
1503
1504 It's main use is in scripts without an event loop - when you want to
1505 stat a lot of files, you can write somehting like this:
1506
1507 IO::AIO::max_outstanding 32;
1508
1509 for my $path (...) {
1510 aio_stat $path , ...;
1511 IO::AIO::poll_cb;
1512 }
1513
1514 IO::AIO::flush;
1515
1516 The call to "poll_cb" inside the loop will normally return
1517 instantly, but as soon as more thna 32 reqeusts are in-flight, it
254 block until some requests have been handled. 1518 will block until some requests have been handled. This keeps the
1519 loop from pushing a large number of "aio_stat" requests onto the
1520 queue.
255 1521
256 The default is very large, so normally there is no practical limit. 1522 The default value for "max_outstanding" is very large, so there is
257 If you queue up many requests in a loop it it often improves speed 1523 no practical limit on the number of outstanding requests.
258 if you set this to a relatively low number, such as 100.
259 1524
260 Under normal circumstances you don't need to call this function. 1525 STATISTICAL INFORMATION
1526 IO::AIO::nreqs
1527 Returns the number of requests currently in the ready, execute or
1528 pending states (i.e. for which their callback has not been invoked
1529 yet).
1530
1531 Example: wait till there are no outstanding requests anymore:
1532
1533 IO::AIO::poll_wait, IO::AIO::poll_cb
1534 while IO::AIO::nreqs;
1535
1536 IO::AIO::nready
1537 Returns the number of requests currently in the ready state (not yet
1538 executed).
1539
1540 IO::AIO::npending
1541 Returns the number of requests currently in the pending state
1542 (executed, but not yet processed by poll_cb).
1543
1544 MISCELLANEOUS FUNCTIONS
1545 IO::AIO implements some functions that might be useful, but are not
1546 asynchronous.
1547
1548 IO::AIO::sendfile $ofh, $ifh, $offset, $count
1549 Calls the "eio_sendfile_sync" function, which is like
1550 "aio_sendfile", but is blocking (this makes most sense if you know
1551 the input data is likely cached already and the output filehandle is
1552 set to non-blocking operations).
1553
1554 Returns the number of bytes copied, or -1 on error.
1555
1556 IO::AIO::fadvise $fh, $offset, $len, $advice
1557 Simply calls the "posix_fadvise" function (see its manpage for
1558 details). The following advice constants are available:
1559 "IO::AIO::FADV_NORMAL", "IO::AIO::FADV_SEQUENTIAL",
1560 "IO::AIO::FADV_RANDOM", "IO::AIO::FADV_NOREUSE",
1561 "IO::AIO::FADV_WILLNEED", "IO::AIO::FADV_DONTNEED".
1562
1563 On systems that do not implement "posix_fadvise", this function
1564 returns ENOSYS, otherwise the return value of "posix_fadvise".
1565
1566 IO::AIO::madvise $scalar, $offset, $len, $advice
1567 Simply calls the "posix_madvise" function (see its manpage for
1568 details). The following advice constants are available:
1569 "IO::AIO::MADV_NORMAL", "IO::AIO::MADV_SEQUENTIAL",
1570 "IO::AIO::MADV_RANDOM", "IO::AIO::MADV_WILLNEED",
1571 "IO::AIO::MADV_DONTNEED".
1572
1573 On systems that do not implement "posix_madvise", this function
1574 returns ENOSYS, otherwise the return value of "posix_madvise".
1575
1576 IO::AIO::mprotect $scalar, $offset, $len, $protect
1577 Simply calls the "mprotect" function on the preferably AIO::mmap'ed
1578 $scalar (see its manpage for details). The following protect
1579 constants are available: "IO::AIO::PROT_NONE", "IO::AIO::PROT_READ",
1580 "IO::AIO::PROT_WRITE", "IO::AIO::PROT_EXEC".
1581
1582 On systems that do not implement "mprotect", this function returns
1583 ENOSYS, otherwise the return value of "mprotect".
1584
1585 IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
1586 Memory-maps a file (or anonymous memory range) and attaches it to
1587 the given $scalar, which will act like a string scalar. Returns true
1588 on success, and false otherwise.
1589
1590 The only operations allowed on the scalar are "substr"/"vec" that
1591 don't change the string length, and most read-only operations such
1592 as copying it or searching it with regexes and so on.
1593
1594 Anything else is unsafe and will, at best, result in memory leaks.
1595
1596 The memory map associated with the $scalar is automatically removed
1597 when the $scalar is destroyed, or when the "IO::AIO::mmap" or
1598 "IO::AIO::munmap" functions are called.
1599
1600 This calls the "mmap"(2) function internally. See your system's
1601 manual page for details on the $length, $prot and $flags parameters.
1602
1603 The $length must be larger than zero and smaller than the actual
1604 filesize.
1605
1606 $prot is a combination of "IO::AIO::PROT_NONE",
1607 "IO::AIO::PROT_EXEC", "IO::AIO::PROT_READ" and/or
1608 "IO::AIO::PROT_WRITE",
1609
1610 $flags can be a combination of "IO::AIO::MAP_SHARED" or
1611 "IO::AIO::MAP_PRIVATE", or a number of system-specific flags (when
1612 not available, the are defined as 0): "IO::AIO::MAP_ANONYMOUS"
1613 (which is set to "MAP_ANON" if your system only provides this
1614 constant), "IO::AIO::MAP_HUGETLB", "IO::AIO::MAP_LOCKED",
1615 "IO::AIO::MAP_NORESERVE", "IO::AIO::MAP_POPULATE" or
1616 "IO::AIO::MAP_NONBLOCK"
1617
1618 If $fh is "undef", then a file descriptor of -1 is passed.
1619
1620 $offset is the offset from the start of the file - it generally must
1621 be a multiple of "IO::AIO::PAGESIZE" and defaults to 0.
1622
1623 Example:
1624
1625 use Digest::MD5;
1626 use IO::AIO;
1627
1628 open my $fh, "<verybigfile"
1629 or die "$!";
1630
1631 IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
1632 or die "verybigfile: $!";
1633
1634 my $fast_md5 = md5 $data;
1635
1636 IO::AIO::munmap $scalar
1637 Removes a previous mmap and undefines the $scalar.
1638
1639 IO::AIO::munlock $scalar, $offset = 0, $length = undef
1640 Calls the "munlock" function, undoing the effects of a previous
1641 "aio_mlock" call (see its description for details).
1642
1643 IO::AIO::munlockall
1644 Calls the "munlockall" function.
1645
1646 On systems that do not implement "munlockall", this function returns
1647 ENOSYS, otherwise the return value of "munlockall".
1648
1649 IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
1650 Calls the GNU/Linux splice(2) syscall, if available. If $r_off or
1651 $w_off are "undef", then "NULL" is passed for these, otherwise they
1652 should be the file offset.
1653
1654 $r_fh and $w_fh should not refer to the same file, as splice might
1655 silently corrupt the data in this case.
1656
1657 The following symbol flag values are available:
1658 "IO::AIO::SPLICE_F_MOVE", "IO::AIO::SPLICE_F_NONBLOCK",
1659 "IO::AIO::SPLICE_F_MORE" and "IO::AIO::SPLICE_F_GIFT".
1660
1661 See the splice(2) manpage for details.
1662
1663 IO::AIO::tee $r_fh, $w_fh, $length, $flags
1664 Calls the GNU/Linux tee(2) syscall, see it's manpage and the
1665 description for "IO::AIO::splice" above for details.
1666
1667EVENT LOOP INTEGRATION
1668 It is recommended to use AnyEvent::AIO to integrate IO::AIO
1669 automatically into many event loops:
1670
1671 # AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
1672 use AnyEvent::AIO;
1673
1674 You can also integrate IO::AIO manually into many event loops, here are
1675 some examples of how to do this:
1676
1677 # EV integration
1678 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
1679
1680 # Event integration
1681 Event->io (fd => IO::AIO::poll_fileno,
1682 poll => 'r',
1683 cb => \&IO::AIO::poll_cb);
1684
1685 # Glib/Gtk2 integration
1686 add_watch Glib::IO IO::AIO::poll_fileno,
1687 in => sub { IO::AIO::poll_cb; 1 };
1688
1689 # Tk integration
1690 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
1691 readable => \&IO::AIO::poll_cb);
1692
1693 # Danga::Socket integration
1694 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
1695 \&IO::AIO::poll_cb);
1696
1697 FORK BEHAVIOUR
1698 Usage of pthreads in a program changes the semantics of fork
1699 considerably. Specifically, only async-safe functions can be called
1700 after fork. Perl doesn't know about this, so in general, you cannot call
1701 fork with defined behaviour in perl if pthreads are involved. IO::AIO
1702 uses pthreads, so this applies, but many other extensions and (for
1703 inexplicable reasons) perl itself often is linked against pthreads, so
1704 this limitation applies to quite a lot of perls.
1705
1706 This module no longer tries to fight your OS, or POSIX. That means
1707 IO::AIO only works in the process that loaded it. Forking is fully
1708 supported, but using IO::AIO in the child is not.
1709
1710 You might get around by not *using* IO::AIO before (or after) forking.
1711 You could also try to call the IO::AIO::reinit function in the child:
1712
1713 IO::AIO::reinit
1714 Abandons all current requests and I/O threads and simply
1715 reinitialises all data structures. This is not an operation
1716 supported by any standards, but happens to work on GNU/Linux and
1717 some newer BSD systems.
1718
1719 The only reasonable use for this function is to call it after
1720 forking, if "IO::AIO" was used in the parent. Calling it while
1721 IO::AIO is active in the process will result in undefined behaviour.
1722 Calling it at any time will also result in any undefined (by POSIX)
1723 behaviour.
1724
1725 MEMORY USAGE
1726 Per-request usage:
1727
1728 Each aio request uses - depending on your architecture - around 100-200
1729 bytes of memory. In addition, stat requests need a stat buffer (possibly
1730 a few hundred bytes), readdir requires a result buffer and so on. Perl
1731 scalars and other data passed into aio requests will also be locked and
1732 will consume memory till the request has entered the done state.
1733
1734 This is not awfully much, so queuing lots of requests is not usually a
1735 problem.
1736
1737 Per-thread usage:
1738
1739 In the execution phase, some aio requests require more memory for
1740 temporary buffers, and each thread requires a stack and other data
1741 structures (usually around 16k-128k, depending on the OS).
1742
1743KNOWN BUGS
1744 Known bugs will be fixed in the next release.
261 1745
262SEE ALSO 1746SEE ALSO
263 Coro, Linux::AIO. 1747 AnyEvent::AIO for easy integration into event loops, Coro::AIO for a
1748 more natural syntax.
264 1749
265AUTHOR 1750AUTHOR
266 Marc Lehmann <schmorp@schmorp.de> 1751 Marc Lehmann <schmorp@schmorp.de>
267 http://home.schmorp.de/ 1752 http://home.schmorp.de/
268 1753

Diff Legend

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