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

Comparing IO-AIO/README (file contents):
Revision 1.18 by root, Thu Oct 26 16:28:33 2006 UTC vs.
Revision 1.51 by root, Sat Apr 7 00:50:33 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
… …
23 $req->cancel; # cancel request if still in queue 24 $req->cancel; # cancel request if still in queue
24 25
25 my $grp = aio_group sub { print "all stats done\n" }; 26 my $grp = aio_group sub { print "all stats done\n" };
26 add $grp aio_stat "..." for ...; 27 add $grp aio_stat "..." for ...;
27 28
28 # AnyEvent integration
29 open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
30 my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
31
32 # Event integration
33 Event->io (fd => IO::AIO::poll_fileno,
34 poll => 'r',
35 cb => \&IO::AIO::poll_cb);
36
37 # Glib/Gtk2 integration
38 add_watch Glib::IO IO::AIO::poll_fileno,
39 in => sub { IO::AIO::poll_cb; 1 };
40
41 # Tk integration
42 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
43 readable => \&IO::AIO::poll_cb);
44
45 # Danga::Socket integration
46 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
47 \&IO::AIO::poll_cb);
48
49DESCRIPTION 29DESCRIPTION
50 This module implements asynchronous I/O using whatever means your 30 This module implements asynchronous I/O using whatever means your
51 operating system supports. 31 operating system supports. It is implemented as an interface to "libeio"
32 (<http://software.schmorp.de/pkg/libeio.html>).
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.
52 49
53 In this version, a number of threads are started that execute your 50 In this version, a number of threads are started that execute your
54 requests and signal their completion. You don't need thread support in 51 requests and signal their completion. You don't need thread support in
55 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
56 perl. In the future, this module might make use of the native aio 53 perl. In the future, this module might make use of the native aio
57 functions available on many operating systems. However, they are often 54 functions available on many operating systems. However, they are often
58 not well-supported or restricted (Linux doesn't allow them on normal 55 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 56 files currently, for example), and they would only support aio_read and
60 aio_write, so the remaining functionality would have to be implemented 57 aio_write, so the remaining functionality would have to be implemented
61 using threads anyway. 58 using threads anyway.
62 59
63 Although the module will work with in the presence of other (Perl-) 60 Although the module will work in the presence of other (Perl-) threads,
64 threads, it is currently not reentrant in any way, so use appropriate 61 it is currently not reentrant in any way, so use appropriate locking
65 locking yourself, always call "poll_cb" from within the same thread, or 62 yourself, always call "poll_cb" from within the same thread, or never
66 never call "poll_cb" (or 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;
67 105
68REQUEST ANATOMY AND LIFETIME 106REQUEST ANATOMY AND LIFETIME
69 Every "aio_*" function creates a request. which is a C data structure 107 Every "aio_*" function creates a request. which is a C data structure
70 not directly visible to Perl. 108 not directly visible to Perl.
71 109
… …
107 anymore (except possibly for the Perl object, but its connection to 145 anymore (except possibly for the Perl object, but its connection to
108 the actual aio request is severed and calling its methods will 146 the actual aio request is severed and calling its methods will
109 either do nothing or result in a runtime error). 147 either do nothing or result in a runtime error).
110 148
111FUNCTIONS 149FUNCTIONS
112 AIO FUNCTIONS 150 QUICK OVERVIEW
113 All the "aio_*" calls are more or less thin wrappers around the 151 This section simply lists the prototypes of the most important functions
114 syscall with the same name (sans "aio_"). The arguments are similar 152 for quick reference. See the following sections for function-by-function
115 or identical, and they all accept an additional (and optional) 153 documentation.
116 $callback argument which must be a code reference. This code
117 reference will get called with the syscall return code (e.g. most
118 syscalls return -1 on error, unlike perl, which usually delivers
119 "false") as it's sole argument when the given syscall has been
120 executed asynchronously.
121 154
122 All functions expecting a filehandle keep a copy of the filehandle 155 aio_wd $pathname, $callback->($wd)
123 internally until the request has finished.
124
125 All requests return objects of type IO::AIO::REQ that allow further
126 manipulation of those requests while they are in-flight.
127
128 The pathnames you pass to these routines *must* be absolute and
129 encoded in byte form. The reason for the former is that at the time
130 the request is being executed, the current working directory could
131 have changed. Alternatively, you can make sure that you never change
132 the current working directory.
133
134 To encode pathnames to byte form, either make sure you either: a)
135 always pass in filenames you got from outside (command line, readdir
136 etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and
137 encode your pathnames to the locale (or other) encoding in effect in
138 the user environment, d) use Glib::filename_from_unicode on unicode
139 filenames or e) use something else.
140
141 $prev_pri = aioreq_pri [$pri]
142 Returns the priority value that would be used for the next
143 request and, if $pri is given, sets the priority for the next
144 aio request.
145
146 The default priority is 0, the minimum and maximum priorities
147 are -4 and 4, respectively. Requests with higher priority will
148 be serviced first.
149
150 The priority will be reset to 0 after each call to one of the
151 "aio_*" functions.
152
153 Example: open a file with low priority, then read something from
154 it with higher priority so the read request is serviced before
155 other low priority open requests (potentially spamming the
156 cache):
157
158 aioreq_pri -3;
159 aio_open ..., sub {
160 return unless $_[0];
161
162 aioreq_pri -2;
163 aio_read $_[0], ..., sub {
164 ...
165 };
166 };
167
168 aioreq_nice $pri_adjust
169 Similar to "aioreq_pri", but subtracts the given value from the
170 current priority, so effects are cumulative.
171
172 aio_open $pathname, $flags, $mode, $callback->($fh) 156 aio_open $pathname, $flags, $mode, $callback->($fh)
173 Asynchronously open or create a file and call the callback with
174 a newly created filehandle for the file.
175
176 The pathname passed to "aio_open" must be absolute. See API
177 NOTES, above, for an explanation.
178
179 The $flags argument is a bitmask. See the "Fcntl" module for a
180 list. They are the same as used by "sysopen".
181
182 Likewise, $mode specifies the mode of the newly created file, if
183 it didn't exist and "O_CREAT" has been given, just like perl's
184 "sysopen", except that it is mandatory (i.e. use 0 if you don't
185 create new files, and 0666 or 0777 if you do).
186
187 Example:
188
189 aio_open "/etc/passwd", O_RDONLY, 0, sub {
190 if ($_[0]) {
191 print "open successful, fh is $_[0]\n";
192 ...
193 } else {
194 die "open failed: $!\n";
195 }
196 };
197
198 aio_close $fh, $callback->($status) 157 aio_close $fh, $callback->($status)
199 Asynchronously close a file and call the callback with the 158 aio_seek $fh,$offset,$whence, $callback->($offs)
200 result code. *WARNING:* although accepted, you should not pass
201 in a perl filehandle here, as perl will likely close the file
202 descriptor another time when the filehandle is destroyed.
203 Normally, you can safely call perls "close" or just let
204 filehandles go out of scope.
205
206 This is supposed to be a bug in the API, so that might change.
207 It's therefore best to avoid this function.
208
209 aio_read $fh,$offset,$length, $data,$dataoffset, 159 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
210 $callback->($retval)
211 aio_write $fh,$offset,$length, $data,$dataoffset, 160 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
212 $callback->($retval)
213 Reads or writes "length" bytes from the specified "fh" and
214 "offset" into the scalar given by "data" and offset "dataoffset"
215 and calls the callback without the actual number of bytes read
216 (or -1 on error, just like the syscall).
217
218 The $data scalar *MUST NOT* be modified in any way while the
219 request is outstanding. Modifying it can result in segfaults or
220 WW3 (if the necessary/optional hardware is installed).
221
222 Example: Read 15 bytes at offset 7 into scalar $buffer, starting
223 at offset 0 within the scalar:
224
225 aio_read $fh, 7, 15, $buffer, 0, sub {
226 $_[0] > 0 or die "read error: $!";
227 print "read $_[0] bytes: <$buffer>\n";
228 };
229
230 aio_move $srcpath, $dstpath, $callback->($status)
231 Try to move the *file* (directories not supported as either
232 source or destination) from $srcpath to $dstpath and call the
233 callback with the 0 (error) or -1 ok.
234
235 This is a composite request that tries to rename(2) the file
236 first. If rename files with "EXDEV", it creates the destination
237 file with mode 0200 and copies the contents of the source file
238 into it using "aio_sendfile", followed by restoring atime,
239 mtime, access mode and uid/gid, in that order, and unlinking the
240 $srcpath.
241
242 If an error occurs, the partial destination file will be
243 unlinked, if possible, except when setting atime, mtime, access
244 mode and uid/gid, where errors are being ignored.
245
246 aio_sendfile $out_fh, $in_fh, $in_offset, $length, 161 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
247 $callback->($retval)
248 Tries to copy $length bytes from $in_fh to $out_fh. It starts
249 reading at byte offset $in_offset, and starts writing at the
250 current file offset of $out_fh. Because of that, it is not safe
251 to issue more than one "aio_sendfile" per $out_fh, as they will
252 interfere with each other.
253
254 This call tries to make use of a native "sendfile" syscall to
255 provide zero-copy operation. For this to work, $out_fh should
256 refer to a socket, and $in_fh should refer to mmap'able file.
257
258 If the native sendfile call fails or is not implemented, it will
259 be emulated, so you can call "aio_sendfile" on any type of
260 filehandle regardless of the limitations of the operating
261 system.
262
263 Please note, however, that "aio_sendfile" can read more bytes
264 from $in_fh than are written, and there is no way to find out
265 how many bytes have been read from "aio_sendfile" alone, as
266 "aio_sendfile" only provides the number of bytes written to
267 $out_fh. Only if the result value equals $length one can assume
268 that $length bytes have been read.
269
270 aio_readahead $fh,$offset,$length, $callback->($retval) 162 aio_readahead $fh,$offset,$length, $callback->($retval)
271 "aio_readahead" populates the page cache with data from a file
272 so that subsequent reads from that file will not block on disk
273 I/O. The $offset argument specifies the starting point from
274 which data is to be read and $length specifies the number of
275 bytes to be read. I/O is performed in whole pages, so that
276 offset is effectively rounded down to a page boundary and bytes
277 are read up to the next page boundary greater than or equal to
278 (off-set+length). "aio_readahead" does not read beyond the end
279 of the file. The current file offset of the file is left
280 unchanged.
281
282 If that syscall doesn't exist (likely if your OS isn't Linux) it
283 will be emulated by simply reading the data, which would have a
284 similar effect.
285
286 aio_stat $fh_or_path, $callback->($status) 163 aio_stat $fh_or_path, $callback->($status)
287 aio_lstat $fh, $callback->($status) 164 aio_lstat $fh, $callback->($status)
288 Works like perl's "stat" or "lstat" in void context. The 165 aio_statvfs $fh_or_path, $callback->($statvfs)
289 callback will be called after the stat and the results will be 166 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
290 available using "stat _" or "-s _" etc... 167 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
291 168 aio_chmod $fh_or_path, $mode, $callback->($status)
292 The pathname passed to "aio_stat" must be absolute. See API 169 aio_truncate $fh_or_path, $offset, $callback->($status)
293 NOTES, above, for an explanation.
294
295 Currently, the stats are always 64-bit-stats, i.e. instead of
296 returning an error when stat'ing a large file, the results will
297 be silently truncated unless perl itself is compiled with large
298 file support.
299
300 Example: Print the length of /etc/passwd:
301
302 aio_stat "/etc/passwd", sub {
303 $_[0] and die "stat failed: $!";
304 print "size is ", -s _, "\n";
305 };
306
307 aio_unlink $pathname, $callback->($status) 170 aio_unlink $pathname, $callback->($status)
308 Asynchronously unlink (delete) a file and call the callback with 171 aio_mknod $pathname, $mode, $dev, $callback->($status)
309 the result code.
310
311 aio_link $srcpath, $dstpath, $callback->($status) 172 aio_link $srcpath, $dstpath, $callback->($status)
312 Asynchronously create a new link to the existing object at
313 $srcpath at the path $dstpath and call the callback with the
314 result code.
315
316 aio_symlink $srcpath, $dstpath, $callback->($status) 173 aio_symlink $srcpath, $dstpath, $callback->($status)
317 Asynchronously create a new symbolic link to the existing object 174 aio_readlink $pathname, $callback->($link)
318 at $srcpath at the path $dstpath and call the callback with the 175 aio_realpath $pathname, $callback->($link)
319 result code.
320
321 aio_rename $srcpath, $dstpath, $callback->($status) 176 aio_rename $srcpath, $dstpath, $callback->($status)
322 Asynchronously rename the object at $srcpath to $dstpath, just 177 aio_mkdir $pathname, $mode, $callback->($status)
323 as rename(2) and call the callback with the result code.
324
325 aio_rmdir $pathname, $callback->($status) 178 aio_rmdir $pathname, $callback->($status)
326 Asynchronously rmdir (delete) a directory and call the callback
327 with the result code.
328
329 aio_readdir $pathname, $callback->($entries) 179 aio_readdir $pathname, $callback->($entries)
330 Unlike the POSIX call of the same name, "aio_readdir" reads an 180 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
331 entire directory (i.e. opendir + readdir + closedir). The 181 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
332 entries will not be sorted, and will NOT include the "." and 182 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
333 ".." entries.
334
335 The callback a single argument which is either "undef" or an
336 array-ref with the filenames.
337
338 aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 183 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
339 Scans a directory (similar to "aio_readdir") but additionally 184 aio_load $pathname, $data, $callback->($status)
340 tries to efficiently separate the entries of directory $path 185 aio_copy $srcpath, $dstpath, $callback->($status)
341 into two sets of names, directories you can recurse into 186 aio_move $srcpath, $dstpath, $callback->($status)
342 (directories), and ones you cannot recurse into (everything 187 aio_rmtree $pathname, $callback->($status)
343 else, including symlinks to directories). 188 aio_sync $callback->($status)
344 189 aio_syncfs $fh, $callback->($status)
345 "aio_scandir" is a composite request that creates of many sub
346 requests_ $maxreq specifies the maximum number of outstanding
347 aio requests that this function generates. If it is "<= 0", then
348 a suitable default will be chosen (currently 6).
349
350 On error, the callback is called without arguments, otherwise it
351 receives two array-refs with path-relative entry names.
352
353 Example:
354
355 aio_scandir $dir, 0, sub {
356 my ($dirs, $nondirs) = @_;
357 print "real directories: @$dirs\n";
358 print "everything else: @$nondirs\n";
359 };
360
361 Implementation notes.
362
363 The "aio_readdir" cannot be avoided, but "stat()"'ing every
364 entry can.
365
366 After reading the directory, the modification time, size etc. of
367 the directory before and after the readdir is checked, and if
368 they match (and isn't the current time), the link count will be
369 used to decide how many entries are directories (if >= 2).
370 Otherwise, no knowledge of the number of subdirectories will be
371 assumed.
372
373 Then entries will be sorted into likely directories (everything
374 without a non-initial dot currently) and likely non-directories
375 (everything else). Then every entry plus an appended "/." will
376 be "stat"'ed, likely directories first. If that succeeds, it
377 assumes that the entry is a directory or a symlink to directory
378 (which will be checked seperately). This is often faster than
379 stat'ing the entry itself because filesystems might detect the
380 type of the entry without reading the inode data (e.g. ext2fs
381 filetype feature).
382
383 If the known number of directories (link count - 2) has been
384 reached, the rest of the entries is assumed to be
385 non-directories.
386
387 This only works with certainty on POSIX (= UNIX) filesystems,
388 which fortunately are the vast majority of filesystems around.
389
390 It will also likely work on non-POSIX filesystems with reduced
391 efficiency as those tend to return 0 or 1 as link counts, which
392 disables the directory counting heuristic.
393
394 aio_fsync $fh, $callback->($status) 190 aio_fsync $fh, $callback->($status)
395 Asynchronously call fsync on the given filehandle and call the
396 callback with the fsync result code.
397
398 aio_fdatasync $fh, $callback->($status) 191 aio_fdatasync $fh, $callback->($status)
399 Asynchronously call fdatasync on the given filehandle and call 192 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
400 the callback with the fdatasync result code. 193 aio_pathsync $pathname, $callback->($status)
401 194 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
402 If this call isn't available because your OS lacks it or it 195 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
403 couldn't be detected, it will be emulated by calling "fsync" 196 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
404 instead. 197 aio_mlockall $flags, $callback->($status)
405
406 aio_group $callback->(...) 198 aio_group $callback->(...)
407 This is a very special aio request: Instead of doing something,
408 it is a container for other aio requests, which is useful if you
409 want to bundle many requests into a single, composite, request
410 with a definite callback and the ability to cancel the whole
411 request with its subrequests.
412
413 Returns an object of class IO::AIO::GRP. See its documentation
414 below for more info.
415
416 Example:
417
418 my $grp = aio_group sub {
419 print "all stats done\n";
420 };
421
422 add $grp
423 (aio_stat ...),
424 (aio_stat ...),
425 ...;
426
427 aio_nop $callback->() 199 aio_nop $callback->()
428 This is a special request - it does nothing in itself and is
429 only used for side effects, such as when you want to add a dummy
430 request to a group so that finishing the requests in the group
431 depends on executing the given code.
432 200
433 While this request does nothing, it still goes through the 201 $prev_pri = aioreq_pri [$pri]
434 execution phase and still requires a worker thread. Thus, the 202 aioreq_nice $pri_adjust
435 callback will not be executed immediately but only after other
436 requests in the queue have entered their execution phase. This
437 can be used to measure request latency.
438 203
439 IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED* 204 IO::AIO::poll_wait
440 Mainly used for debugging and benchmarking, this aio request 205 IO::AIO::poll_cb
441 puts one of the request workers to sleep for the given time. 206 IO::AIO::poll
207 IO::AIO::flush
208 IO::AIO::max_poll_reqs $nreqs
209 IO::AIO::max_poll_time $seconds
210 IO::AIO::min_parallel $nthreads
211 IO::AIO::max_parallel $nthreads
212 IO::AIO::max_idle $nthreads
213 IO::AIO::idle_timeout $seconds
214 IO::AIO::max_outstanding $maxreqs
215 IO::AIO::nreqs
216 IO::AIO::nready
217 IO::AIO::npending
442 218
443 While it is theoretically handy to have simple I/O scheduling 219 IO::AIO::sendfile $ofh, $ifh, $offset, $count
444 requests like sleep and file handle readable/writable, the 220 IO::AIO::fadvise $fh, $offset, $len, $advice
445 overhead this creates is immense (it blocks a thread for a long 221 IO::AIO::madvise $scalar, $offset, $length, $advice
446 time) so do not use this function except to put your application 222 IO::AIO::mprotect $scalar, $offset, $length, $protect
447 under artificial I/O pressure. 223 IO::AIO::munlock $scalar, $offset = 0, $length = undef
224 IO::AIO::munlockall
448 225
449 IO::AIO::REQ CLASS 226 API NOTES
450 All non-aggregate "aio_*" functions return an object of this class 227 All the "aio_*" calls are more or less thin wrappers around the syscall
451 when called in non-void context. 228 with the same name (sans "aio_"). The arguments are similar or
229 identical, and they all accept an additional (and optional) $callback
230 argument which must be a code reference. This code reference will be
231 called after the syscall has been executed in an asynchronous fashion.
232 The results of the request will be passed as arguments to the callback
233 (and, if an error occured, in $!) - for most requests the syscall return
234 code (e.g. most syscalls return -1 on error, unlike perl, which usually
235 delivers "false").
452 236
453 cancel $req 237 Some requests (such as "aio_readdir") pass the actual results and
454 Cancels the request, if possible. Has the effect of skipping 238 communicate failures by passing "undef".
455 execution when entering the execute state and skipping calling
456 the callback when entering the the result state, but will leave
457 the request otherwise untouched. That means that requests that
458 currently execute will not be stopped and resources held by the
459 request will not be freed prematurely.
460 239
461 cb $req $callback->(...) 240 All functions expecting a filehandle keep a copy of the filehandle
462 Replace (or simply set) the callback registered to the request. 241 internally until the request has finished.
463 242
464 IO::AIO::GRP CLASS 243 All functions return request objects of type IO::AIO::REQ that allow
465 This class is a subclass of IO::AIO::REQ, so all its methods apply 244 further manipulation of those requests while they are in-flight.
466 to objects of this class, too.
467 245
468 A IO::AIO::GRP object is a special request that can contain multiple 246 The pathnames you pass to these routines *should* be absolute. The
469 other aio requests. 247 reason for this is that at the time the request is being executed, the
248 current working directory could have changed. Alternatively, you can
249 make sure that you never change the current working directory anywhere
250 in the program and then use relative paths. You can also take advantage
251 of IO::AIOs working directory abstraction, that lets you specify paths
252 relative to some previously-opened "working directory object" - see the
253 description of the "IO::AIO::WD" class later in this document.
470 254
471 You create one by calling the "aio_group" constructing function with 255 To encode pathnames as octets, either make sure you either: a) always
472 a callback that will be called when all contained requests have 256 pass in filenames you got from outside (command line, readdir etc.)
473 entered the "done" state: 257 without tinkering, b) are in your native filesystem encoding, c) use the
258 Encode module and encode your pathnames to the locale (or other)
259 encoding in effect in the user environment, d) use
260 Glib::filename_from_unicode on unicode filenames or e) use something
261 else to ensure your scalar has the correct contents.
474 262
475 my $grp = aio_group sub { 263 This works, btw. independent of the internal UTF-8 bit, which IO::AIO
476 print "all requests are done\n"; 264 handles correctly whether it is set or not.
477 };
478 265
479 You add requests by calling the "add" method with one or more 266 AIO REQUEST FUNCTIONS
480 "IO::AIO::REQ" objects: 267 $prev_pri = aioreq_pri [$pri]
268 Returns the priority value that would be used for the next request
269 and, if $pri is given, sets the priority for the next aio request.
481 270
482 $grp->add (aio_unlink "..."); 271 The default priority is 0, the minimum and maximum priorities are -4
272 and 4, respectively. Requests with higher priority will be serviced
273 first.
483 274
484 add $grp aio_stat "...", sub { 275 The priority will be reset to 0 after each call to one of the
485 $_[0] or return $grp->result ("error"); 276 "aio_*" functions.
486 277
487 # add another request dynamically, if first succeeded 278 Example: open a file with low priority, then read something from it
279 with higher priority so the read request is serviced before other
280 low priority open requests (potentially spamming the cache):
281
282 aioreq_pri -3;
488 add $grp aio_open "...", sub { 283 aio_open ..., sub {
489 $grp->result ("ok"); 284 return unless $_[0];
285
286 aioreq_pri -2;
287 aio_read $_[0], ..., sub {
288 ...
490 }; 289 };
491 }; 290 };
492 291
292 aioreq_nice $pri_adjust
293 Similar to "aioreq_pri", but subtracts the given value from the
294 current priority, so the effect is cumulative.
295
296 aio_open $pathname, $flags, $mode, $callback->($fh)
297 Asynchronously open or create a file and call the callback with a
298 newly created filehandle for the file.
299
300 The pathname passed to "aio_open" must be absolute. See API NOTES,
301 above, for an explanation.
302
303 The $flags argument is a bitmask. See the "Fcntl" module for a list.
304 They are the same as used by "sysopen".
305
306 Likewise, $mode specifies the mode of the newly created file, if it
307 didn't exist and "O_CREAT" has been given, just like perl's
308 "sysopen", except that it is mandatory (i.e. use 0 if you don't
309 create new files, and 0666 or 0777 if you do). Note that the $mode
310 will be modified by the umask in effect then the request is being
311 executed, so better never change the umask.
312
313 Example:
314
315 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
316 if ($_[0]) {
317 print "open successful, fh is $_[0]\n";
318 ...
319 } else {
320 die "open failed: $!\n";
321 }
322 };
323
324 In addition to all the common open modes/flags ("O_RDONLY",
325 "O_WRONLY", "O_RDWR", "O_CREAT", "O_TRUNC", "O_EXCL" and
326 "O_APPEND"), the following POSIX and non-POSIX constants are
327 available (missing ones on your system are, as usual, 0):
328
329 "O_ASYNC", "O_DIRECT", "O_NOATIME", "O_CLOEXEC", "O_NOCTTY",
330 "O_NOFOLLOW", "O_NONBLOCK", "O_EXEC", "O_SEARCH", "O_DIRECTORY",
331 "O_DSYNC", "O_RSYNC", "O_SYNC" and "O_TTY_INIT".
332
333 aio_close $fh, $callback->($status)
334 Asynchronously close a file and call the callback with the result
335 code.
336
337 Unfortunately, you can't do this to perl. Perl *insists* very
338 strongly on closing the file descriptor associated with the
339 filehandle itself.
340
341 Therefore, "aio_close" will not close the filehandle - instead it
342 will use dup2 to overwrite the file descriptor with the write-end of
343 a pipe (the pipe fd will be created on demand and will be cached).
344
345 Or in other words: the file descriptor will be closed, but it will
346 not be free for reuse until the perl filehandle is closed.
347
348 aio_seek $fh, $offset, $whence, $callback->($offs)
349 Seeks the filehandle to the new $offset, similarly to perl's
350 "sysseek". The $whence can use the traditional values (0 for
351 "IO::AIO::SEEK_SET", 1 for "IO::AIO::SEEK_CUR" or 2 for
352 "IO::AIO::SEEK_END").
353
354 The resulting absolute offset will be passed to the callback, or -1
355 in case of an error.
356
357 In theory, the $whence constants could be different than the
358 corresponding values from Fcntl, but perl guarantees they are the
359 same, so don't panic.
360
361 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
362 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
363 Reads or writes $length bytes from or to the specified $fh and
364 $offset into the scalar given by $data and offset $dataoffset and
365 calls the callback without the actual number of bytes read (or -1 on
366 error, just like the syscall).
367
368 "aio_read" will, like "sysread", shrink or grow the $data scalar to
369 offset plus the actual number of bytes read.
370
371 If $offset is undefined, then the current file descriptor offset
372 will be used (and updated), otherwise the file descriptor offset
373 will not be changed by these calls.
374
375 If $length is undefined in "aio_write", use the remaining length of
376 $data.
377
378 If $dataoffset is less than zero, it will be counted from the end of
379 $data.
380
381 The $data scalar *MUST NOT* be modified in any way while the request
382 is outstanding. Modifying it can result in segfaults or World War
383 III (if the necessary/optional hardware is installed).
384
385 Example: Read 15 bytes at offset 7 into scalar $buffer, starting at
386 offset 0 within the scalar:
387
388 aio_read $fh, 7, 15, $buffer, 0, sub {
389 $_[0] > 0 or die "read error: $!";
390 print "read $_[0] bytes: <$buffer>\n";
391 };
392
393 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
394 Tries to copy $length bytes from $in_fh to $out_fh. It starts
395 reading at byte offset $in_offset, and starts writing at the current
396 file offset of $out_fh. Because of that, it is not safe to issue
397 more than one "aio_sendfile" per $out_fh, as they will interfere
398 with each other. The same $in_fh works fine though, as this function
399 does not move or use the file offset of $in_fh.
400
401 Please note that "aio_sendfile" can read more bytes from $in_fh than
402 are written, and there is no way to find out how many more bytes
403 have been read from "aio_sendfile" alone, as "aio_sendfile" only
404 provides the number of bytes written to $out_fh. Only if the result
405 value equals $length one can assume that $length bytes have been
406 read.
407
408 Unlike with other "aio_" functions, it makes a lot of sense to use
409 "aio_sendfile" on non-blocking sockets, as long as one end
410 (typically the $in_fh) is a file - the file I/O will then be
411 asynchronous, while the socket I/O will be non-blocking. Note,
412 however, that you can run into a trap where "aio_sendfile" reads
413 some data with readahead, then fails to write all data, and when the
414 socket is ready the next time, the data in the cache is already
415 lost, forcing "aio_sendfile" to again hit the disk. Explicit
416 "aio_read" + "aio_write" let's you better control resource usage.
417
418 This call tries to make use of a native "sendfile"-like syscall to
419 provide zero-copy operation. For this to work, $out_fh should refer
420 to a socket, and $in_fh should refer to an mmap'able file.
421
422 If a native sendfile cannot be found or it fails with "ENOSYS",
423 "EINVAL", "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or
424 "ENOTSOCK", it will be emulated, so you can call "aio_sendfile" on
425 any type of filehandle regardless of the limitations of the
426 operating system.
427
428 As native sendfile syscalls (as practically any non-POSIX interface
429 hacked together in a hurry to improve benchmark numbers) tend to be
430 rather buggy on many systems, this implementation tries to work
431 around some known bugs in Linux and FreeBSD kernels (probably
432 others, too), but that might fail, so you really really should check
433 the return value of "aio_sendfile" - fewre bytes than expected might
434 have been transferred.
435
436 aio_readahead $fh,$offset,$length, $callback->($retval)
437 "aio_readahead" populates the page cache with data from a file so
438 that subsequent reads from that file will not block on disk I/O. The
439 $offset argument specifies the starting point from which data is to
440 be read and $length specifies the number of bytes to be read. I/O is
441 performed in whole pages, so that offset is effectively rounded down
442 to a page boundary and bytes are read up to the next page boundary
443 greater than or equal to (off-set+length). "aio_readahead" does not
444 read beyond the end of the file. The current file offset of the file
445 is left unchanged.
446
447 If that syscall doesn't exist (likely if your OS isn't Linux) it
448 will be emulated by simply reading the data, which would have a
449 similar effect.
450
451 aio_stat $fh_or_path, $callback->($status)
452 aio_lstat $fh, $callback->($status)
453 Works like perl's "stat" or "lstat" in void context. The callback
454 will be called after the stat and the results will be available
455 using "stat _" or "-s _" etc...
456
457 The pathname passed to "aio_stat" must be absolute. See API NOTES,
458 above, for an explanation.
459
460 Currently, the stats are always 64-bit-stats, i.e. instead of
461 returning an error when stat'ing a large file, the results will be
462 silently truncated unless perl itself is compiled with large file
463 support.
464
465 To help interpret the mode and dev/rdev stat values, IO::AIO offers
466 the following constants and functions (if not implemented, the
467 constants will be 0 and the functions will either "croak" or fall
468 back on traditional behaviour).
469
470 "S_IFMT", "S_IFIFO", "S_IFCHR", "S_IFBLK", "S_IFLNK", "S_IFREG",
471 "S_IFDIR", "S_IFWHT", "S_IFSOCK", "IO::AIO::major $dev_t",
472 "IO::AIO::minor $dev_t", "IO::AIO::makedev $major, $minor".
473
474 Example: Print the length of /etc/passwd:
475
476 aio_stat "/etc/passwd", sub {
477 $_[0] and die "stat failed: $!";
478 print "size is ", -s _, "\n";
479 };
480
481 aio_statvfs $fh_or_path, $callback->($statvfs)
482 Works like the POSIX "statvfs" or "fstatvfs" syscalls, depending on
483 whether a file handle or path was passed.
484
485 On success, the callback is passed a hash reference with the
486 following members: "bsize", "frsize", "blocks", "bfree", "bavail",
487 "files", "ffree", "favail", "fsid", "flag" and "namemax". On
488 failure, "undef" is passed.
489
490 The following POSIX IO::AIO::ST_* constants are defined: "ST_RDONLY"
491 and "ST_NOSUID".
492
493 The following non-POSIX IO::AIO::ST_* flag masks are defined to
494 their correct value when available, or to 0 on systems that do not
495 support them: "ST_NODEV", "ST_NOEXEC", "ST_SYNCHRONOUS",
496 "ST_MANDLOCK", "ST_WRITE", "ST_APPEND", "ST_IMMUTABLE",
497 "ST_NOATIME", "ST_NODIRATIME" and "ST_RELATIME".
498
499 Example: stat "/wd" and dump out the data if successful.
500
501 aio_statvfs "/wd", sub {
502 my $f = $_[0]
503 or die "statvfs: $!";
504
505 use Data::Dumper;
506 say Dumper $f;
507 };
508
509 # result:
510 {
511 bsize => 1024,
512 bfree => 4333064312,
513 blocks => 10253828096,
514 files => 2050765568,
515 flag => 4096,
516 favail => 2042092649,
517 bavail => 4333064312,
518 ffree => 2042092649,
519 namemax => 255,
520 frsize => 1024,
521 fsid => 1810
522 }
523
524 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
525 Works like perl's "utime" function (including the special case of
526 $atime and $mtime being undef). Fractional times are supported if
527 the underlying syscalls support them.
528
529 When called with a pathname, uses utimes(2) if available, otherwise
530 utime(2). If called on a file descriptor, uses futimes(2) if
531 available, otherwise returns ENOSYS, so this is not portable.
532
533 Examples:
534
535 # set atime and mtime to current time (basically touch(1)):
536 aio_utime "path", undef, undef;
537 # set atime to current time and mtime to beginning of the epoch:
538 aio_utime "path", time, undef; # undef==0
539
540 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
541 Works like perl's "chown" function, except that "undef" for either
542 $uid or $gid is being interpreted as "do not change" (but -1 can
543 also be used).
544
545 Examples:
546
547 # same as "chown root path" in the shell:
548 aio_chown "path", 0, -1;
549 # same as above:
550 aio_chown "path", 0, undef;
551
552 aio_truncate $fh_or_path, $offset, $callback->($status)
553 Works like truncate(2) or ftruncate(2).
554
555 aio_chmod $fh_or_path, $mode, $callback->($status)
556 Works like perl's "chmod" function.
557
558 aio_unlink $pathname, $callback->($status)
559 Asynchronously unlink (delete) a file and call the callback with the
560 result code.
561
562 aio_mknod $pathname, $mode, $dev, $callback->($status)
563 [EXPERIMENTAL]
564
565 Asynchronously create a device node (or fifo). See mknod(2).
566
567 The only (POSIX-) portable way of calling this function is:
568
569 aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
570
571 See "aio_stat" for info about some potentially helpful extra
572 constants and functions.
573
574 aio_link $srcpath, $dstpath, $callback->($status)
575 Asynchronously create a new link to the existing object at $srcpath
576 at the path $dstpath and call the callback with the result code.
577
578 aio_symlink $srcpath, $dstpath, $callback->($status)
579 Asynchronously create a new symbolic link to the existing object at
580 $srcpath at the path $dstpath and call the callback with the result
581 code.
582
583 aio_readlink $pathname, $callback->($link)
584 Asynchronously read the symlink specified by $path and pass it to
585 the callback. If an error occurs, nothing or undef gets passed to
586 the callback.
587
588 aio_realpath $pathname, $callback->($path)
589 Asynchronously make the path absolute and resolve any symlinks in
590 $path. The resulting path only consists of directories (Same as
591 Cwd::realpath).
592
593 This request can be used to get the absolute path of the current
594 working directory by passing it a path of . (a single dot).
595
596 aio_rename $srcpath, $dstpath, $callback->($status)
597 Asynchronously rename the object at $srcpath to $dstpath, just as
598 rename(2) and call the callback with the result code.
599
600 aio_mkdir $pathname, $mode, $callback->($status)
601 Asynchronously mkdir (create) a directory and call the callback with
602 the result code. $mode will be modified by the umask at the time the
603 request is executed, so do not change your umask.
604
605 aio_rmdir $pathname, $callback->($status)
606 Asynchronously rmdir (delete) a directory and call the callback with
607 the result code.
608
609 aio_readdir $pathname, $callback->($entries)
610 Unlike the POSIX call of the same name, "aio_readdir" reads an
611 entire directory (i.e. opendir + readdir + closedir). The entries
612 will not be sorted, and will NOT include the "." and ".." entries.
613
614 The callback is passed a single argument which is either "undef" or
615 an array-ref with the filenames.
616
617 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
618 Quite similar to "aio_readdir", but the $flags argument allows one
619 to tune behaviour and output format. In case of an error, $entries
620 will be "undef".
621
622 The flags are a combination of the following constants, ORed
623 together (the flags will also be passed to the callback, possibly
624 modified):
625
626 IO::AIO::READDIR_DENTS
627 When this flag is off, then the callback gets an arrayref
628 consisting of names only (as with "aio_readdir"), otherwise it
629 gets an arrayref with "[$name, $type, $inode]" arrayrefs, each
630 describing a single directory entry in more detail.
631
632 $name is the name of the entry.
633
634 $type is one of the "IO::AIO::DT_xxx" constants:
635
636 "IO::AIO::DT_UNKNOWN", "IO::AIO::DT_FIFO", "IO::AIO::DT_CHR",
637 "IO::AIO::DT_DIR", "IO::AIO::DT_BLK", "IO::AIO::DT_REG",
638 "IO::AIO::DT_LNK", "IO::AIO::DT_SOCK", "IO::AIO::DT_WHT".
639
640 "IO::AIO::DT_UNKNOWN" means just that: readdir does not know. If
641 you need to know, you have to run stat yourself. Also, for speed
642 reasons, the $type scalars are read-only: you can not modify
643 them.
644
645 $inode is the inode number (which might not be exact on systems
646 with 64 bit inode numbers and 32 bit perls). This field has
647 unspecified content on systems that do not deliver the inode
648 information.
649
650 IO::AIO::READDIR_DIRS_FIRST
651 When this flag is set, then the names will be returned in an
652 order where likely directories come first, in optimal stat
653 order. This is useful when you need to quickly find directories,
654 or you want to find all directories while avoiding to stat()
655 each entry.
656
657 If the system returns type information in readdir, then this is
658 used to find directories directly. Otherwise, likely directories
659 are names beginning with ".", or otherwise names with no dots,
660 of which names with short names are tried first.
661
662 IO::AIO::READDIR_STAT_ORDER
663 When this flag is set, then the names will be returned in an
664 order suitable for stat()'ing each one. That is, when you plan
665 to stat() all files in the given directory, then the returned
666 order will likely be fastest.
667
668 If both this flag and "IO::AIO::READDIR_DIRS_FIRST" are
669 specified, then the likely dirs come first, resulting in a less
670 optimal stat order.
671
672 IO::AIO::READDIR_FOUND_UNKNOWN
673 This flag should not be set when calling "aio_readdirx".
674 Instead, it is being set by "aio_readdirx", when any of the
675 $type's found were "IO::AIO::DT_UNKNOWN". The absence of this
676 flag therefore indicates that all $type's are known, which can
677 be used to speed up some algorithms.
678
679 aio_load $pathname, $data, $callback->($status)
680 This is a composite request that tries to fully load the given file
681 into memory. Status is the same as with aio_read.
682
683 aio_copy $srcpath, $dstpath, $callback->($status)
684 Try to copy the *file* (directories not supported as either source
685 or destination) from $srcpath to $dstpath and call the callback with
686 a status of 0 (ok) or -1 (error, see $!).
687
688 This is a composite request that creates the destination file with
689 mode 0200 and copies the contents of the source file into it using
690 "aio_sendfile", followed by restoring atime, mtime, access mode and
691 uid/gid, in that order.
692
693 If an error occurs, the partial destination file will be unlinked,
694 if possible, except when setting atime, mtime, access mode and
695 uid/gid, where errors are being ignored.
696
697 aio_move $srcpath, $dstpath, $callback->($status)
698 Try to move the *file* (directories not supported as either source
699 or destination) from $srcpath to $dstpath and call the callback with
700 a status of 0 (ok) or -1 (error, see $!).
701
702 This is a composite request that tries to rename(2) the file first;
703 if rename fails with "EXDEV", it copies the file with "aio_copy"
704 and, if that is successful, unlinks the $srcpath.
705
706 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
707 Scans a directory (similar to "aio_readdir") but additionally tries
708 to efficiently separate the entries of directory $path into two sets
709 of names, directories you can recurse into (directories), and ones
710 you cannot recurse into (everything else, including symlinks to
711 directories).
712
713 "aio_scandir" is a composite request that creates of many sub
714 requests_ $maxreq specifies the maximum number of outstanding aio
715 requests that this function generates. If it is "<= 0", then a
716 suitable default will be chosen (currently 4).
717
718 On error, the callback is called without arguments, otherwise it
719 receives two array-refs with path-relative entry names.
720
721 Example:
722
723 aio_scandir $dir, 0, sub {
724 my ($dirs, $nondirs) = @_;
725 print "real directories: @$dirs\n";
726 print "everything else: @$nondirs\n";
727 };
728
729 Implementation notes.
730
731 The "aio_readdir" cannot be avoided, but "stat()"'ing every entry
732 can.
733
734 If readdir returns file type information, then this is used directly
735 to find directories.
736
737 Otherwise, after reading the directory, the modification time, size
738 etc. of the directory before and after the readdir is checked, and
739 if they match (and isn't the current time), the link count will be
740 used to decide how many entries are directories (if >= 2).
741 Otherwise, no knowledge of the number of subdirectories will be
742 assumed.
743
744 Then entries will be sorted into likely directories a non-initial
745 dot currently) and likely non-directories (see "aio_readdirx"). Then
746 every entry plus an appended "/." will be "stat"'ed, likely
747 directories first, in order of their inode numbers. If that
748 succeeds, it assumes that the entry is a directory or a symlink to
749 directory (which will be checked separately). This is often faster
750 than stat'ing the entry itself because filesystems might detect the
751 type of the entry without reading the inode data (e.g. ext2fs
752 filetype feature), even on systems that cannot return the filetype
753 information on readdir.
754
755 If the known number of directories (link count - 2) has been
756 reached, the rest of the entries is assumed to be non-directories.
757
758 This only works with certainty on POSIX (= UNIX) filesystems, which
759 fortunately are the vast majority of filesystems around.
760
761 It will also likely work on non-POSIX filesystems with reduced
762 efficiency as those tend to return 0 or 1 as link counts, which
763 disables the directory counting heuristic.
764
765 aio_rmtree $pathname, $callback->($status)
766 Delete a directory tree starting (and including) $path, return the
767 status of the final "rmdir" only. This is a composite request that
768 uses "aio_scandir" to recurse into and rmdir directories, and unlink
769 everything else.
770
771 aio_sync $callback->($status)
772 Asynchronously call sync and call the callback when finished.
773
774 aio_fsync $fh, $callback->($status)
775 Asynchronously call fsync on the given filehandle and call the
776 callback with the fsync result code.
777
778 aio_fdatasync $fh, $callback->($status)
779 Asynchronously call fdatasync on the given filehandle and call the
780 callback with the fdatasync result code.
781
782 If this call isn't available because your OS lacks it or it couldn't
783 be detected, it will be emulated by calling "fsync" instead.
784
785 aio_syncfs $fh, $callback->($status)
786 Asynchronously call the syncfs syscall to sync the filesystem
787 associated to the given filehandle and call the callback with the
788 syncfs result code. If syncfs is not available, calls sync(), but
789 returns -1 and sets errno to "ENOSYS" nevertheless.
790
791 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
792 Sync the data portion of the file specified by $offset and $length
793 to disk (but NOT the metadata), by calling the Linux-specific
794 sync_file_range call. If sync_file_range is not available or it
795 returns ENOSYS, then fdatasync or fsync is being substituted.
796
797 $flags can be a combination of
798 "IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE",
799 "IO::AIO::SYNC_FILE_RANGE_WRITE" and
800 "IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER": refer to the sync_file_range
801 manpage for details.
802
803 aio_pathsync $pathname, $callback->($status)
804 This request tries to open, fsync and close the given path. This is
805 a composite request intended to sync directories after directory
806 operations (E.g. rename). This might not work on all operating
807 systems or have any specific effect, but usually it makes sure that
808 directory changes get written to disc. It works for anything that
809 can be opened for read-only, not just directories.
810
811 Future versions of this function might fall back to other methods
812 when "fsync" on the directory fails (such as calling "sync").
813
814 Passes 0 when everything went ok, and -1 on error.
815
816 aio_msync $scalar, $offset = 0, $length = undef, flags = 0,
817 $callback->($status)
818 This is a rather advanced IO::AIO call, which only works on
819 mmap(2)ed scalars (see the "IO::AIO::mmap" function, although it
820 also works on data scalars managed by the Sys::Mmap or Mmap modules,
821 note that the scalar must only be modified in-place while an aio
822 operation is pending on it).
823
824 It calls the "msync" function of your OS, if available, with the
825 memory area starting at $offset in the string and ending $length
826 bytes later. If $length is negative, counts from the end, and if
827 $length is "undef", then it goes till the end of the string. The
828 flags can be a combination of "IO::AIO::MS_ASYNC",
829 "IO::AIO::MS_INVALIDATE" and "IO::AIO::MS_SYNC".
830
831 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0,
832 $callback->($status)
833 This is a rather advanced IO::AIO call, which works best on
834 mmap(2)ed scalars.
835
836 It touches (reads or writes) all memory pages in the specified range
837 inside the scalar. All caveats and parameters are the same as for
838 "aio_msync", above, except for flags, which must be either 0 (which
839 reads all pages and ensures they are instantiated) or
840 "IO::AIO::MT_MODIFY", which modifies the memory page s(by reading
841 and writing an octet from it, which dirties the page).
842
843 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
844 This is a rather advanced IO::AIO call, which works best on
845 mmap(2)ed scalars.
846
847 It reads in all the pages of the underlying storage into memory (if
848 any) and locks them, so they are not getting swapped/paged out or
849 removed.
850
851 If $length is undefined, then the scalar will be locked till the
852 end.
853
854 On systems that do not implement "mlock", this function returns -1
855 and sets errno to "ENOSYS".
856
857 Note that the corresponding "munlock" is synchronous and is
858 documented under "MISCELLANEOUS FUNCTIONS".
859
860 Example: open a file, mmap and mlock it - both will be undone when
861 $data gets destroyed.
862
863 open my $fh, "<", $path or die "$path: $!";
864 my $data;
865 IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
866 aio_mlock $data; # mlock in background
867
868 aio_mlockall $flags, $callback->($status)
869 Calls the "mlockall" function with the given $flags (a combination
870 of "IO::AIO::MCL_CURRENT" and "IO::AIO::MCL_FUTURE").
871
872 On systems that do not implement "mlockall", this function returns
873 -1 and sets errno to "ENOSYS".
874
875 Note that the corresponding "munlockall" is synchronous and is
876 documented under "MISCELLANEOUS FUNCTIONS".
877
878 Example: asynchronously lock all current and future pages into
879 memory.
880
881 aio_mlockall IO::AIO::MCL_FUTURE;
882
883 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
884 Queries the extents of the given file (by calling the Linux FIEMAP
885 ioctl, see <http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for
886 details). If the "ioctl" is not available on your OS, then this
887 rquiest will fail with "ENOSYS".
888
889 $start is the starting offset to query extents for, $length is the
890 size of the range to query - if it is "undef", then the whole file
891 will be queried.
892
893 $flags is a combination of flags ("IO::AIO::FIEMAP_FLAG_SYNC" or
894 "IO::AIO::FIEMAP_FLAG_XATTR" - "IO::AIO::FIEMAP_FLAGS_COMPAT" is
895 also exported), and is normally 0 or "IO::AIO::FIEMAP_FLAG_SYNC" to
896 query the data portion.
897
898 $count is the maximum number of extent records to return. If it is
899 "undef", then IO::AIO queries all extents of the file. As a very
900 special case, if it is 0, then the callback receives the number of
901 extents instead of the extents themselves.
902
903 If an error occurs, the callback receives no arguments. The special
904 "errno" value "IO::AIO::EBADR" is available to test for flag errors.
905
906 Otherwise, the callback receives an array reference with extent
907 structures. Each extent structure is an array reference itself, with
908 the following members:
909
910 [$logical, $physical, $length, $flags]
911
912 Flags is any combination of the following flag values (typically
913 either 0 or "IO::AIO::FIEMAP_EXTENT_LAST"):
914
915 "IO::AIO::FIEMAP_EXTENT_LAST", "IO::AIO::FIEMAP_EXTENT_UNKNOWN",
916 "IO::AIO::FIEMAP_EXTENT_DELALLOC", "IO::AIO::FIEMAP_EXTENT_ENCODED",
917 "IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED",
918 "IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED",
919 "IO::AIO::FIEMAP_EXTENT_DATA_INLINE",
920 "IO::AIO::FIEMAP_EXTENT_DATA_TAIL",
921 "IO::AIO::FIEMAP_EXTENT_UNWRITTEN", "IO::AIO::FIEMAP_EXTENT_MERGED"
922 or "IO::AIO::FIEMAP_EXTENT_SHARED".
923
924 aio_group $callback->(...)
925 This is a very special aio request: Instead of doing something, it
926 is a container for other aio requests, which is useful if you want
927 to bundle many requests into a single, composite, request with a
928 definite callback and the ability to cancel the whole request with
929 its subrequests.
930
931 Returns an object of class IO::AIO::GRP. See its documentation below
932 for more info.
933
934 Example:
935
936 my $grp = aio_group sub {
937 print "all stats done\n";
938 };
939
940 add $grp
941 (aio_stat ...),
942 (aio_stat ...),
943 ...;
944
945 aio_nop $callback->()
946 This is a special request - it does nothing in itself and is only
947 used for side effects, such as when you want to add a dummy request
948 to a group so that finishing the requests in the group depends on
949 executing the given code.
950
951 While this request does nothing, it still goes through the execution
952 phase and still requires a worker thread. Thus, the callback will
953 not be executed immediately but only after other requests in the
954 queue have entered their execution phase. This can be used to
955 measure request latency.
956
957 IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
958 Mainly used for debugging and benchmarking, this aio request puts
959 one of the request workers to sleep for the given time.
960
961 While it is theoretically handy to have simple I/O scheduling
962 requests like sleep and file handle readable/writable, the overhead
963 this creates is immense (it blocks a thread for a long time) so do
964 not use this function except to put your application under
965 artificial I/O pressure.
966
967 IO::AIO::WD - multiple working directories
968 Your process only has one current working directory, which is used by
969 all threads. This makes it hard to use relative paths (some other
970 component could call "chdir" at any time, and it is hard to control when
971 the path will be used by IO::AIO).
972
973 One solution for this is to always use absolute paths. This usually
974 works, but can be quite slow (the kernel has to walk the whole path on
975 every access), and can also be a hassle to implement.
976
977 Newer POSIX systems have a number of functions (openat, fdopendir,
978 futimensat and so on) that make it possible to specify working
979 directories per operation.
980
981 For portability, and because the clowns who "designed", or shall I
982 write, perpetrated this new interface were obviously half-drunk, this
983 abstraction cannot be perfect, though.
984
985 IO::AIO allows you to convert directory paths into a so-called
986 IO::AIO::WD object. This object stores the canonicalised, absolute
987 version of the path, and on systems that allow it, also a directory file
988 descriptor.
989
990 Everywhere where a pathname is accepted by IO::AIO (e.g. in "aio_stat"
991 or "aio_unlink"), one can specify an array reference with an IO::AIO::WD
992 object and a pathname instead (or the IO::AIO::WD object alone, which
993 gets interpreted as "[$wd, "."]"). If the pathname is absolute, the
994 IO::AIO::WD object is ignored, otherwise the pathname is resolved
995 relative to that IO::AIO::WD object.
996
997 For example, to get a wd object for /etc and then stat passwd inside,
998 you would write:
999
1000 aio_wd "/etc", sub {
1001 my $etcdir = shift;
1002
1003 # although $etcdir can be undef on error, there is generally no reason
1004 # to check for errors here, as aio_stat will fail with ENOENT
1005 # when $etcdir is undef.
1006
1007 aio_stat [$etcdir, "passwd"], sub {
1008 # yay
1009 };
1010 };
1011
1012 That "aio_wd" is a request and not a normal function shows that creating
1013 an IO::AIO::WD object is itself a potentially blocking operation, which
1014 is why it is done asynchronously.
1015
1016 To stat the directory obtained with "aio_wd" above, one could write
1017 either of the following three request calls:
1018
1019 aio_lstat "/etc" , sub { ... # pathname as normal string
1020 aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
1021 aio_lstat $wd , sub { ... # shorthand for the previous
1022
1023 As with normal pathnames, IO::AIO keeps a copy of the working directory
1024 object and the pathname string, so you could write the following without
1025 causing any issues due to $path getting reused:
1026
1027 my $path = [$wd, undef];
1028
1029 for my $name (qw(abc def ghi)) {
1030 $path->[1] = $name;
1031 aio_stat $path, sub {
1032 # ...
1033 };
1034 }
1035
1036 There are some caveats: when directories get renamed (or deleted), the
1037 pathname string doesn't change, so will point to the new directory (or
1038 nowhere at all), while the directory fd, if available on the system,
1039 will still point to the original directory. Most functions accepting a
1040 pathname will use the directory fd on newer systems, and the string on
1041 older systems. Some functions (such as realpath) will always rely on the
1042 string form of the pathname.
1043
1044 So this fucntionality is mainly useful to get some protection against
1045 "chdir", to easily get an absolute path out of a relative path for
1046 future reference, and to speed up doing many operations in the same
1047 directory (e.g. when stat'ing all files in a directory).
1048
1049 The following functions implement this working directory abstraction:
1050
1051 aio_wd $pathname, $callback->($wd)
1052 Asynchonously canonicalise the given pathname and convert it to an
1053 IO::AIO::WD object representing it. If possible and supported on the
1054 system, also open a directory fd to speed up pathname resolution
1055 relative to this working directory.
1056
1057 If something goes wrong, then "undef" is passwd to the callback
1058 instead of a working directory object and $! is set appropriately.
1059 Since passing "undef" as working directory component of a pathname
1060 fails the request with "ENOENT", there is often no need for error
1061 checking in the "aio_wd" callback, as future requests using the
1062 value will fail in the expected way.
1063
1064 If this call isn't available because your OS lacks it or it couldn't
1065 be detected, it will be emulated by calling "fsync" instead.
1066
1067 IO::AIO::CWD
1068 This is a compiletime constant (object) that represents the process
1069 current working directory.
1070
1071 Specifying this object as working directory object for a pathname is
1072 as if the pathname would be specified directly, without a directory
1073 object, e.g., these calls are functionally identical:
1074
1075 aio_stat "somefile", sub { ... };
1076 aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
1077
1078 IO::AIO::REQ CLASS
1079 All non-aggregate "aio_*" functions return an object of this class when
1080 called in non-void context.
1081
1082 cancel $req
1083 Cancels the request, if possible. Has the effect of skipping
1084 execution when entering the execute state and skipping calling the
1085 callback when entering the the result state, but will leave the
1086 request otherwise untouched (with the exception of readdir). That
1087 means that requests that currently execute will not be stopped and
1088 resources held by the request will not be freed prematurely.
1089
1090 cb $req $callback->(...)
1091 Replace (or simply set) the callback registered to the request.
1092
1093 IO::AIO::GRP CLASS
1094 This class is a subclass of IO::AIO::REQ, so all its methods apply to
1095 objects of this class, too.
1096
1097 A IO::AIO::GRP object is a special request that can contain multiple
1098 other aio requests.
1099
1100 You create one by calling the "aio_group" constructing function with a
1101 callback that will be called when all contained requests have entered
1102 the "done" state:
1103
1104 my $grp = aio_group sub {
1105 print "all requests are done\n";
1106 };
1107
1108 You add requests by calling the "add" method with one or more
1109 "IO::AIO::REQ" objects:
1110
1111 $grp->add (aio_unlink "...");
1112
1113 add $grp aio_stat "...", sub {
1114 $_[0] or return $grp->result ("error");
1115
1116 # add another request dynamically, if first succeeded
1117 add $grp aio_open "...", sub {
1118 $grp->result ("ok");
1119 };
1120 };
1121
493 This makes it very easy to create composite requests (see the source 1122 This makes it very easy to create composite requests (see the source of
494 of "aio_move" for an application) that work and feel like simple 1123 "aio_move" for an application) that work and feel like simple requests.
1124
1125 * The IO::AIO::GRP objects will be cleaned up during calls to
1126 "IO::AIO::poll_cb", just like any other request.
1127
1128 * They can be canceled like any other request. Canceling will cancel
1129 not only the request itself, but also all requests it contains.
1130
1131 * They can also can also be added to other IO::AIO::GRP objects.
1132
1133 * You must not add requests to a group from within the group callback
1134 (or any later time).
1135
1136 Their lifetime, simplified, looks like this: when they are empty, they
1137 will finish very quickly. If they contain only requests that are in the
1138 "done" state, they will also finish. Otherwise they will continue to
1139 exist.
1140
1141 That means after creating a group you have some time to add requests
1142 (precisely before the callback has been invoked, which is only done
1143 within the "poll_cb"). And in the callbacks of those requests, you can
1144 add further requests to the group. And only when all those requests have
1145 finished will the the group itself finish.
1146
1147 add $grp ...
1148 $grp->add (...)
1149 Add one or more requests to the group. Any type of IO::AIO::REQ can
1150 be added, including other groups, as long as you do not create
1151 circular dependencies.
1152
1153 Returns all its arguments.
1154
1155 $grp->cancel_subs
1156 Cancel all subrequests and clears any feeder, but not the group
1157 request itself. Useful when you queued a lot of events but got a
1158 result early.
1159
1160 The group request will finish normally (you cannot add requests to
1161 the group).
1162
1163 $grp->result (...)
1164 Set the result value(s) that will be passed to the group callback
1165 when all subrequests have finished and set the groups errno to the
1166 current value of errno (just like calling "errno" without an error
1167 number). By default, no argument will be passed and errno is zero.
1168
1169 $grp->errno ([$errno])
1170 Sets the group errno value to $errno, or the current value of errno
1171 when the argument is missing.
1172
1173 Every aio request has an associated errno value that is restored
1174 when the callback is invoked. This method lets you change this value
1175 from its default (0).
1176
1177 Calling "result" will also set errno, so make sure you either set $!
1178 before the call to "result", or call c<errno> after it.
1179
1180 feed $grp $callback->($grp)
1181 Sets a feeder/generator on this group: every group can have an
1182 attached generator that generates requests if idle. The idea behind
1183 this is that, although you could just queue as many requests as you
1184 want in a group, this might starve other requests for a potentially
1185 long time. For example, "aio_scandir" might generate hundreds of
1186 thousands of "aio_stat" requests, delaying any later requests for a
1187 long time.
1188
1189 To avoid this, and allow incremental generation of requests, you can
1190 instead a group and set a feeder on it that generates those
1191 requests. The feed callback will be called whenever there are few
1192 enough (see "limit", below) requests active in the group itself and
1193 is expected to queue more requests.
1194
1195 The feed callback can queue as many requests as it likes (i.e. "add"
1196 does not impose any limits).
1197
1198 If the feed does not queue more requests when called, it will be
1199 automatically removed from the group.
1200
1201 If the feed limit is 0 when this method is called, it will be set to
1202 2 automatically.
1203
1204 Example:
1205
1206 # stat all files in @files, but only ever use four aio requests concurrently:
1207
1208 my $grp = aio_group sub { print "finished\n" };
1209 limit $grp 4;
1210 feed $grp sub {
1211 my $file = pop @files
1212 or return;
1213
1214 add $grp aio_stat $file, sub { ... };
1215 };
1216
1217 limit $grp $num
1218 Sets the feeder limit for the group: The feeder will be called
1219 whenever the group contains less than this many requests.
1220
1221 Setting the limit to 0 will pause the feeding process.
1222
1223 The default value for the limit is 0, but note that setting a feeder
1224 automatically bumps it up to 2.
1225
1226 SUPPORT FUNCTIONS
1227 EVENT PROCESSING AND EVENT LOOP INTEGRATION
1228 $fileno = IO::AIO::poll_fileno
1229 Return the *request result pipe file descriptor*. This filehandle
1230 must be polled for reading by some mechanism outside this module
1231 (e.g. EV, Glib, select and so on, see below or the SYNOPSIS). If the
1232 pipe becomes readable you have to call "poll_cb" to check the
1233 results.
1234
1235 See "poll_cb" for an example.
1236
1237 IO::AIO::poll_cb
1238 Process some outstanding events on the result pipe. You have to call
1239 this regularly. Returns 0 if all events could be processed (or there
1240 were no events to process), or -1 if it returned earlier for
1241 whatever reason. Returns immediately when no events are outstanding.
1242 The amount of events processed depends on the settings of
1243 "IO::AIO::max_poll_req" and "IO::AIO::max_poll_time".
1244
1245 If not all requests were processed for whatever reason, the
1246 filehandle will still be ready when "poll_cb" returns, so normally
1247 you don't have to do anything special to have it called later.
1248
1249 Apart from calling "IO::AIO::poll_cb" when the event filehandle
1250 becomes ready, it can be beneficial to call this function from loops
1251 which submit a lot of requests, to make sure the results get
1252 processed when they become available and not just when the loop is
1253 finished and the event loop takes over again. This function returns
1254 very fast when there are no outstanding requests.
1255
1256 Example: Install an Event watcher that automatically calls
1257 IO::AIO::poll_cb with high priority (more examples can be found in
1258 the SYNOPSIS section, at the top of this document):
1259
1260 Event->io (fd => IO::AIO::poll_fileno,
1261 poll => 'r', async => 1,
1262 cb => \&IO::AIO::poll_cb);
1263
1264 IO::AIO::poll_wait
1265 If there are any outstanding requests and none of them in the result
1266 phase, wait till the result filehandle becomes ready for reading
1267 (simply does a "select" on the filehandle. This is useful if you
1268 want to synchronously wait for some requests to finish).
1269
1270 See "nreqs" for an example.
1271
1272 IO::AIO::poll
1273 Waits until some requests have been handled.
1274
1275 Returns the number of requests processed, but is otherwise strictly
1276 equivalent to:
1277
1278 IO::AIO::poll_wait, IO::AIO::poll_cb
1279
1280 IO::AIO::flush
1281 Wait till all outstanding AIO requests have been handled.
1282
1283 Strictly equivalent to:
1284
1285 IO::AIO::poll_wait, IO::AIO::poll_cb
1286 while IO::AIO::nreqs;
1287
1288 IO::AIO::max_poll_reqs $nreqs
1289 IO::AIO::max_poll_time $seconds
1290 These set the maximum number of requests (default 0, meaning
1291 infinity) that are being processed by "IO::AIO::poll_cb" in one
1292 call, respectively the maximum amount of time (default 0, meaning
1293 infinity) spent in "IO::AIO::poll_cb" to process requests (more
1294 correctly the mininum amount of time "poll_cb" is allowed to use).
1295
1296 Setting "max_poll_time" to a non-zero value creates an overhead of
1297 one syscall per request processed, which is not normally a problem
1298 unless your callbacks are really really fast or your OS is really
1299 really slow (I am not mentioning Solaris here). Using
1300 "max_poll_reqs" incurs no overhead.
1301
1302 Setting these is useful if you want to ensure some level of
1303 interactiveness when perl is not fast enough to process all requests
1304 in time.
1305
1306 For interactive programs, values such as 0.01 to 0.1 should be fine.
1307
1308 Example: Install an Event watcher that automatically calls
1309 IO::AIO::poll_cb with low priority, to ensure that other parts of
1310 the program get the CPU sometimes even under high AIO load.
1311
1312 # try not to spend much more than 0.1s in poll_cb
1313 IO::AIO::max_poll_time 0.1;
1314
1315 # use a low priority so other tasks have priority
1316 Event->io (fd => IO::AIO::poll_fileno,
1317 poll => 'r', nice => 1,
1318 cb => &IO::AIO::poll_cb);
1319
1320 CONTROLLING THE NUMBER OF THREADS
1321 IO::AIO::min_parallel $nthreads
1322 Set the minimum number of AIO threads to $nthreads. The current
1323 default is 8, which means eight asynchronous operations can execute
1324 concurrently at any one time (the number of outstanding requests,
1325 however, is unlimited).
1326
1327 IO::AIO starts threads only on demand, when an AIO request is queued
1328 and no free thread exists. Please note that queueing up a hundred
1329 requests can create demand for a hundred threads, even if it turns
1330 out that everything is in the cache and could have been processed
1331 faster by a single thread.
1332
1333 It is recommended to keep the number of threads relatively low, as
1334 some Linux kernel versions will scale negatively with the number of
1335 threads (higher parallelity => MUCH higher latency). With current
1336 Linux 2.6 versions, 4-32 threads should be fine.
1337
1338 Under most circumstances you don't need to call this function, as
1339 the module selects a default that is suitable for low to moderate
1340 load.
1341
1342 IO::AIO::max_parallel $nthreads
1343 Sets the maximum number of AIO threads to $nthreads. If more than
1344 the specified number of threads are currently running, this function
1345 kills them. This function blocks until the limit is reached.
1346
1347 While $nthreads are zero, aio requests get queued but not executed
1348 until the number of threads has been increased again.
1349
1350 This module automatically runs "max_parallel 0" at program end, to
1351 ensure that all threads are killed and that there are no outstanding
495 requests. 1352 requests.
496 1353
497 * The IO::AIO::GRP objects will be cleaned up during calls to
498 "IO::AIO::poll_cb", just like any other request.
499 * They can be canceled like any other request. Canceling will cancel
500 not only the request itself, but also all requests it contains.
501 * They can also can also be added to other IO::AIO::GRP objects.
502 * You must not add requests to a group from within the group
503 callback (or any later time).
504
505 Their lifetime, simplified, looks like this: when they are empty,
506 they will finish very quickly. If they contain only requests that
507 are in the "done" state, they will also finish. Otherwise they will
508 continue to exist.
509
510 That means after creating a group you have some time to add
511 requests. And in the callbacks of those requests, you can add
512 further requests to the group. And only when all those requests have
513 finished will the the group itself finish.
514
515 add $grp ...
516 $grp->add (...)
517 Add one or more requests to the group. Any type of IO::AIO::REQ
518 can be added, including other groups, as long as you do not
519 create circular dependencies.
520
521 Returns all its arguments.
522
523 $grp->cancel_subs
524 Cancel all subrequests and clears any feeder, but not the group
525 request itself. Useful when you queued a lot of events but got a
526 result early.
527
528 $grp->result (...)
529 Set the result value(s) that will be passed to the group
530 callback when all subrequests have finished and set thre groups
531 errno to the current value of errno (just like calling "errno"
532 without an error number). By default, no argument will be passed
533 and errno is zero.
534
535 $grp->errno ([$errno])
536 Sets the group errno value to $errno, or the current value of
537 errno when the argument is missing.
538
539 Every aio request has an associated errno value that is restored
540 when the callback is invoked. This method lets you change this
541 value from its default (0).
542
543 Calling "result" will also set errno, so make sure you either
544 set $! before the call to "result", or call c<errno> after it.
545
546 feed $grp $callback->($grp)
547 Sets a feeder/generator on this group: every group can have an
548 attached generator that generates requests if idle. The idea
549 behind this is that, although you could just queue as many
550 requests as you want in a group, this might starve other
551 requests for a potentially long time. For example, "aio_scandir"
552 might generate hundreds of thousands "aio_stat" requests,
553 delaying any later requests for a long time.
554
555 To avoid this, and allow incremental generation of requests, you
556 can instead a group and set a feeder on it that generates those
557 requests. The feed callback will be called whenever there are
558 few enough (see "limit", below) requests active in the group
559 itself and is expected to queue more requests.
560
561 The feed callback can queue as many requests as it likes (i.e.
562 "add" does not impose any limits).
563
564 If the feed does not queue more requests when called, it will be
565 automatically removed from the group.
566
567 If the feed limit is 0, it will be set to 2 automatically.
568
569 Example:
570
571 # stat all files in @files, but only ever use four aio requests concurrently:
572
573 my $grp = aio_group sub { print "finished\n" };
574 limit $grp 4;
575 feed $grp sub {
576 my $file = pop @files
577 or return;
578
579 add $grp aio_stat $file, sub { ... };
580 };
581
582 limit $grp $num
583 Sets the feeder limit for the group: The feeder will be called
584 whenever the group contains less than this many requests.
585
586 Setting the limit to 0 will pause the feeding process.
587
588 SUPPORT FUNCTIONS
589 $fileno = IO::AIO::poll_fileno
590 Return the *request result pipe file descriptor*. This
591 filehandle must be polled for reading by some mechanism outside
592 this module (e.g. Event or select, see below or the SYNOPSIS).
593 If the pipe becomes readable you have to call "poll_cb" to check
594 the results.
595
596 See "poll_cb" for an example.
597
598 IO::AIO::poll_cb
599 Process all outstanding events on the result pipe. You have to
600 call this regularly. Returns the number of events processed.
601 Returns immediately when no events are outstanding.
602
603 If not all requests were processed for whatever reason, the
604 filehandle will still be ready when "poll_cb" returns.
605
606 Example: Install an Event watcher that automatically calls
607 IO::AIO::poll_cb with high priority:
608
609 Event->io (fd => IO::AIO::poll_fileno,
610 poll => 'r', async => 1,
611 cb => \&IO::AIO::poll_cb);
612
613 IO::AIO::poll_some $max_requests
614 Similar to "poll_cb", but only processes up to $max_requests
615 requests at a time.
616
617 Useful if you want to ensure some level of interactiveness when
618 perl is not fast enough to process all requests in time.
619
620 Example: Install an Event watcher that automatically calls
621 IO::AIO::poll_some with low priority, to ensure that other parts
622 of the program get the CPU sometimes even under high AIO load.
623
624 Event->io (fd => IO::AIO::poll_fileno,
625 poll => 'r', nice => 1,
626 cb => sub { IO::AIO::poll_some 256 });
627
628 IO::AIO::poll_wait
629 Wait till the result filehandle becomes ready for reading
630 (simply does a "select" on the filehandle. This is useful if you
631 want to synchronously wait for some requests to finish).
632
633 See "nreqs" for an example.
634
635 IO::AIO::nreqs
636 Returns the number of requests currently in the ready, execute
637 or pending states (i.e. for which their callback has not been
638 invoked yet).
639
640 Example: wait till there are no outstanding requests anymore:
641
642 IO::AIO::poll_wait, IO::AIO::poll_cb
643 while IO::AIO::nreqs;
644
645 IO::AIO::nready
646 Returns the number of requests currently in the ready state (not
647 yet executed).
648
649 IO::AIO::npending
650 Returns the number of requests currently in the pending state
651 (executed, but not yet processed by poll_cb).
652
653 IO::AIO::flush
654 Wait till all outstanding AIO requests have been handled.
655
656 Strictly equivalent to:
657
658 IO::AIO::poll_wait, IO::AIO::poll_cb
659 while IO::AIO::nreqs;
660
661 IO::AIO::poll
662 Waits until some requests have been handled.
663
664 Strictly equivalent to:
665
666 IO::AIO::poll_wait, IO::AIO::poll_cb
667 if IO::AIO::nreqs;
668
669 IO::AIO::min_parallel $nthreads
670 Set the minimum number of AIO threads to $nthreads. The current
671 default is 8, which means eight asynchronous operations can
672 execute concurrently at any one time (the number of outstanding
673 requests, however, is unlimited).
674
675 IO::AIO starts threads only on demand, when an AIO request is
676 queued and no free thread exists.
677
678 It is recommended to keep the number of threads relatively low,
679 as some Linux kernel versions will scale negatively with the
680 number of threads (higher parallelity => MUCH higher latency).
681 With current Linux 2.6 versions, 4-32 threads should be fine.
682
683 Under most circumstances you don't need to call this function,
684 as the module selects a default that is suitable for low to
685 moderate load.
686
687 IO::AIO::max_parallel $nthreads
688 Sets the maximum number of AIO threads to $nthreads. If more
689 than the specified number of threads are currently running, this
690 function kills them. This function blocks until the limit is
691 reached.
692
693 While $nthreads are zero, aio requests get queued but not
694 executed until the number of threads has been increased again.
695
696 This module automatically runs "max_parallel 0" at program end,
697 to ensure that all threads are killed and that there are no
698 outstanding requests.
699
700 Under normal circumstances you don't need to call this function. 1354 Under normal circumstances you don't need to call this function.
701 1355
1356 IO::AIO::max_idle $nthreads
1357 Limit the number of threads (default: 4) that are allowed to idle
1358 (i.e., threads that did not get a request to process within the idle
1359 timeout (default: 10 seconds). That means if a thread becomes idle
1360 while $nthreads other threads are also idle, it will free its
1361 resources and exit.
1362
1363 This is useful when you allow a large number of threads (e.g. 100 or
1364 1000) to allow for extremely high load situations, but want to free
1365 resources under normal circumstances (1000 threads can easily
1366 consume 30MB of RAM).
1367
1368 The default is probably ok in most situations, especially if thread
1369 creation is fast. If thread creation is very slow on your system you
1370 might want to use larger values.
1371
1372 IO::AIO::idle_timeout $seconds
1373 Sets the minimum idle timeout (default 10) after which worker
1374 threads are allowed to exit. SEe "IO::AIO::max_idle".
1375
702 $oldmaxreqs = IO::AIO::max_outstanding $maxreqs 1376 IO::AIO::max_outstanding $maxreqs
703 This is a very bad function to use in interactive programs
704 because it blocks, and a bad way to reduce concurrency because
705 it is inexact: Better use an "aio_group" together with a feed
706 callback.
707
708 Sets the maximum number of outstanding requests to $nreqs. If 1377 Sets the maximum number of outstanding requests to $nreqs. If you do
709 you to queue up more than this number of requests, the next call 1378 queue up more than this number of requests, the next call to
710 to the "poll_cb" (and "poll_some" and other functions calling 1379 "IO::AIO::poll_cb" (and other functions calling "poll_cb", such as
711 "poll_cb") function will block until the limit is no longer 1380 "IO::AIO::flush" or "IO::AIO::poll") will block until the limit is
712 exceeded. 1381 no longer exceeded.
713 1382
714 The default value is very large, so there is no practical limit 1383 In other words, this setting does not enforce a queue limit, but can
1384 be used to make poll functions block if the limit is exceeded.
1385
1386 This is a very bad function to use in interactive programs because
1387 it blocks, and a bad way to reduce concurrency because it is
1388 inexact: Better use an "aio_group" together with a feed callback.
1389
1390 It's main use is in scripts without an event loop - when you want to
1391 stat a lot of files, you can write somehting like this:
1392
1393 IO::AIO::max_outstanding 32;
1394
1395 for my $path (...) {
1396 aio_stat $path , ...;
1397 IO::AIO::poll_cb;
1398 }
1399
1400 IO::AIO::flush;
1401
1402 The call to "poll_cb" inside the loop will normally return
1403 instantly, but as soon as more thna 32 reqeusts are in-flight, it
1404 will block until some requests have been handled. This keeps the
1405 loop from pushing a large number of "aio_stat" requests onto the
1406 queue.
1407
1408 The default value for "max_outstanding" is very large, so there is
715 on the number of outstanding requests. 1409 no practical limit on the number of outstanding requests.
716 1410
717 You can still queue as many requests as you want. Therefore, 1411 STATISTICAL INFORMATION
718 "max_oustsanding" is mainly useful in simple scripts (with low 1412 IO::AIO::nreqs
719 values) or as a stop gap to shield against fatal memory overflow 1413 Returns the number of requests currently in the ready, execute or
720 (with large values). 1414 pending states (i.e. for which their callback has not been invoked
1415 yet).
1416
1417 Example: wait till there are no outstanding requests anymore:
1418
1419 IO::AIO::poll_wait, IO::AIO::poll_cb
1420 while IO::AIO::nreqs;
1421
1422 IO::AIO::nready
1423 Returns the number of requests currently in the ready state (not yet
1424 executed).
1425
1426 IO::AIO::npending
1427 Returns the number of requests currently in the pending state
1428 (executed, but not yet processed by poll_cb).
1429
1430 MISCELLANEOUS FUNCTIONS
1431 IO::AIO implements some functions that might be useful, but are not
1432 asynchronous.
1433
1434 IO::AIO::sendfile $ofh, $ifh, $offset, $count
1435 Calls the "eio_sendfile_sync" function, which is like
1436 "aio_sendfile", but is blocking (this makes most sense if you know
1437 the input data is likely cached already and the output filehandle is
1438 set to non-blocking operations).
1439
1440 Returns the number of bytes copied, or -1 on error.
1441
1442 IO::AIO::fadvise $fh, $offset, $len, $advice
1443 Simply calls the "posix_fadvise" function (see its manpage for
1444 details). The following advice constants are available:
1445 "IO::AIO::FADV_NORMAL", "IO::AIO::FADV_SEQUENTIAL",
1446 "IO::AIO::FADV_RANDOM", "IO::AIO::FADV_NOREUSE",
1447 "IO::AIO::FADV_WILLNEED", "IO::AIO::FADV_DONTNEED".
1448
1449 On systems that do not implement "posix_fadvise", this function
1450 returns ENOSYS, otherwise the return value of "posix_fadvise".
1451
1452 IO::AIO::madvise $scalar, $offset, $len, $advice
1453 Simply calls the "posix_madvise" function (see its manpage for
1454 details). The following advice constants are available:
1455 "IO::AIO::MADV_NORMAL", "IO::AIO::MADV_SEQUENTIAL",
1456 "IO::AIO::MADV_RANDOM", "IO::AIO::MADV_WILLNEED",
1457 "IO::AIO::MADV_DONTNEED".
1458
1459 On systems that do not implement "posix_madvise", this function
1460 returns ENOSYS, otherwise the return value of "posix_madvise".
1461
1462 IO::AIO::mprotect $scalar, $offset, $len, $protect
1463 Simply calls the "mprotect" function on the preferably AIO::mmap'ed
1464 $scalar (see its manpage for details). The following protect
1465 constants are available: "IO::AIO::PROT_NONE", "IO::AIO::PROT_READ",
1466 "IO::AIO::PROT_WRITE", "IO::AIO::PROT_EXEC".
1467
1468 On systems that do not implement "mprotect", this function returns
1469 ENOSYS, otherwise the return value of "mprotect".
1470
1471 IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
1472 Memory-maps a file (or anonymous memory range) and attaches it to
1473 the given $scalar, which will act like a string scalar.
1474
1475 The only operations allowed on the scalar are "substr"/"vec" that
1476 don't change the string length, and most read-only operations such
1477 as copying it or searching it with regexes and so on.
1478
1479 Anything else is unsafe and will, at best, result in memory leaks.
1480
1481 The memory map associated with the $scalar is automatically removed
1482 when the $scalar is destroyed, or when the "IO::AIO::mmap" or
1483 "IO::AIO::munmap" functions are called.
1484
1485 This calls the "mmap"(2) function internally. See your system's
1486 manual page for details on the $length, $prot and $flags parameters.
1487
1488 The $length must be larger than zero and smaller than the actual
1489 filesize.
1490
1491 $prot is a combination of "IO::AIO::PROT_NONE",
1492 "IO::AIO::PROT_EXEC", "IO::AIO::PROT_READ" and/or
1493 "IO::AIO::PROT_WRITE",
1494
1495 $flags can be a combination of "IO::AIO::MAP_SHARED" or
1496 "IO::AIO::MAP_PRIVATE", or a number of system-specific flags (when
1497 not available, the are defined as 0): "IO::AIO::MAP_ANONYMOUS"
1498 (which is set to "MAP_ANON" if your system only provides this
1499 constant), "IO::AIO::MAP_HUGETLB", "IO::AIO::MAP_LOCKED",
1500 "IO::AIO::MAP_NORESERVE", "IO::AIO::MAP_POPULATE" or
1501 "IO::AIO::MAP_NONBLOCK"
1502
1503 If $fh is "undef", then a file descriptor of -1 is passed.
1504
1505 $offset is the offset from the start of the file - it generally must
1506 be a multiple of "IO::AIO::PAGESIZE" and defaults to 0.
1507
1508 Example:
1509
1510 use Digest::MD5;
1511 use IO::AIO;
1512
1513 open my $fh, "<verybigfile"
1514 or die "$!";
1515
1516 IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
1517 or die "verybigfile: $!";
1518
1519 my $fast_md5 = md5 $data;
1520
1521 IO::AIO::munmap $scalar
1522 Removes a previous mmap and undefines the $scalar.
1523
1524 IO::AIO::munlock $scalar, $offset = 0, $length = undef
1525 Calls the "munlock" function, undoing the effects of a previous
1526 "aio_mlock" call (see its description for details).
1527
1528 IO::AIO::munlockall
1529 Calls the "munlockall" function.
1530
1531 On systems that do not implement "munlockall", this function returns
1532 ENOSYS, otherwise the return value of "munlockall".
1533
1534EVENT LOOP INTEGRATION
1535 It is recommended to use AnyEvent::AIO to integrate IO::AIO
1536 automatically into many event loops:
1537
1538 # AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
1539 use AnyEvent::AIO;
1540
1541 You can also integrate IO::AIO manually into many event loops, here are
1542 some examples of how to do this:
1543
1544 # EV integration
1545 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
1546
1547 # Event integration
1548 Event->io (fd => IO::AIO::poll_fileno,
1549 poll => 'r',
1550 cb => \&IO::AIO::poll_cb);
1551
1552 # Glib/Gtk2 integration
1553 add_watch Glib::IO IO::AIO::poll_fileno,
1554 in => sub { IO::AIO::poll_cb; 1 };
1555
1556 # Tk integration
1557 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
1558 readable => \&IO::AIO::poll_cb);
1559
1560 # Danga::Socket integration
1561 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
1562 \&IO::AIO::poll_cb);
721 1563
722 FORK BEHAVIOUR 1564 FORK BEHAVIOUR
723 This module should do "the right thing" when the process using it 1565 Usage of pthreads in a program changes the semantics of fork
724 forks: 1566 considerably. Specifically, only async-safe functions can be called
1567 after fork. Perl doesn't know about this, so in general, you cannot call
1568 fork with defined behaviour in perl if pthreads are involved. IO::AIO
1569 uses pthreads, so this applies, but many other extensions and (for
1570 inexplicable reasons) perl itself often is linked against pthreads, so
1571 this limitation applies to quite a lot of perls.
725 1572
726 Before the fork, IO::AIO enters a quiescent state where no requests 1573 This module no longer tries to fight your OS, or POSIX. That means
727 can be added in other threads and no results will be processed. 1574 IO::AIO only works in the process that loaded it. Forking is fully
728 After the fork the parent simply leaves the quiescent state and 1575 supported, but using IO::AIO in the child is not.
729 continues request/result processing, while the child frees the
730 request/result queue (so that the requests started before the fork
731 will only be handled in the parent). Threads will be started on
732 demand until the limit set in the parent process has been reached
733 again.
734 1576
735 In short: the parent will, after a short pause, continue as if fork 1577 You might get around by not *using* IO::AIO before (or after) forking.
736 had not been called, while the child will act as if IO::AIO has not 1578 You could also try to call the IO::AIO::reinit function in the child:
737 been used yet. 1579
1580 IO::AIO::reinit
1581 Abandons all current requests and I/O threads and simply
1582 reinitialises all data structures. This is not an operation
1583 supported by any standards, but happens to work on GNU/Linux and
1584 some newer BSD systems.
1585
1586 The only reasonable use for this function is to call it after
1587 forking, if "IO::AIO" was used in the parent. Calling it while
1588 IO::AIO is active in the process will result in undefined behaviour.
1589 Calling it at any time will also result in any undefined (by POSIX)
1590 behaviour.
738 1591
739 MEMORY USAGE 1592 MEMORY USAGE
740 Per-request usage: 1593 Per-request usage:
741 1594
742 Each aio request uses - depending on your architecture - around 1595 Each aio request uses - depending on your architecture - around 100-200
743 100-200 bytes of memory. In addition, stat requests need a stat 1596 bytes of memory. In addition, stat requests need a stat buffer (possibly
744 buffer (possibly a few hundred bytes), readdir requires a result 1597 a few hundred bytes), readdir requires a result buffer and so on. Perl
745 buffer and so on. Perl scalars and other data passed into aio 1598 scalars and other data passed into aio requests will also be locked and
746 requests will also be locked and will consume memory till the 1599 will consume memory till the request has entered the done state.
747 request has entered the done state.
748 1600
749 This is now awfully much, so queuing lots of requests is not usually 1601 This is not awfully much, so queuing lots of requests is not usually a
750 a problem. 1602 problem.
751 1603
752 Per-thread usage: 1604 Per-thread usage:
753 1605
754 In the execution phase, some aio requests require more memory for 1606 In the execution phase, some aio requests require more memory for
755 temporary buffers, and each thread requires a stack and other data 1607 temporary buffers, and each thread requires a stack and other data
756 structures (usually around 16k-128k, depending on the OS). 1608 structures (usually around 16k-128k, depending on the OS).
757 1609
758KNOWN BUGS 1610KNOWN BUGS
759 Known bugs will be fixed in the next release. 1611 Known bugs will be fixed in the next release.
760 1612
761SEE ALSO 1613SEE ALSO
762 Coro::AIO. 1614 AnyEvent::AIO for easy integration into event loops, Coro::AIO for a
1615 more natural syntax.
763 1616
764AUTHOR 1617AUTHOR
765 Marc Lehmann <schmorp@schmorp.de> 1618 Marc Lehmann <schmorp@schmorp.de>
766 http://home.schmorp.de/ 1619 http://home.schmorp.de/
767 1620

Diff Legend

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