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

Comparing IO-AIO/README (file contents):
Revision 1.47 by root, Fri May 27 00:44:49 2011 UTC vs.
Revision 1.69 by root, Tue Sep 6 10:56:12 2022 UTC

1NAME 1NAME
2 IO::AIO - Asynchronous Input/Output 2 IO::AIO - Asynchronous/Advanced Input/Output
3 3
4SYNOPSIS 4SYNOPSIS
5 use IO::AIO; 5 use IO::AIO;
6 6
7 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub { 7 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
… …
55 not well-supported or restricted (GNU/Linux doesn't allow them on normal 55 not well-supported or restricted (GNU/Linux doesn't allow them on normal
56 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
57 aio_write, so the remaining functionality would have to be implemented 57 aio_write, so the remaining functionality would have to be implemented
58 using threads anyway. 58 using threads anyway.
59 59
60 In addition to asynchronous I/O, this module also exports some rather
61 arcane interfaces, such as "madvise" or linux's "splice" system call,
62 which is why the "A" in "AIO" can also mean *advanced*.
63
60 Although the module will work in the presence of other (Perl-) threads, 64 Although the module will work in the presence of other (Perl-) threads,
61 it is currently not reentrant in any way, so use appropriate locking 65 it is currently not reentrant in any way, so use appropriate locking
62 yourself, always call "poll_cb" from within the same thread, or never 66 yourself, always call "poll_cb" from within the same thread, or never
63 call "poll_cb" (or other "aio_" functions) recursively. 67 call "poll_cb" (or other "aio_" functions) recursively.
64 68
65 EXAMPLE 69 EXAMPLE
66 This is a simple example that uses the EV module and loads /etc/passwd 70 This is a simple example that uses the EV module and loads /etc/passwd
67 asynchronously: 71 asynchronously:
68 72
69 use Fcntl;
70 use EV; 73 use EV;
71 use IO::AIO; 74 use IO::AIO;
72 75
73 # register the IO::AIO callback with EV 76 # register the IO::AIO callback with EV
74 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; 77 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
… …
91 94
92 # file contents now in $contents 95 # file contents now in $contents
93 print $contents; 96 print $contents;
94 97
95 # exit event loop and program 98 # exit event loop and program
96 EV::unloop; 99 EV::break;
97 }; 100 };
98 }; 101 };
99 102
100 # possibly queue up other requests, or open GUI windows, 103 # possibly queue up other requests, or open GUI windows,
101 # check for sockets etc. etc. 104 # check for sockets etc. etc.
102 105
103 # process events as long as there are some: 106 # process events as long as there are some:
104 EV::loop; 107 EV::run;
105 108
106REQUEST ANATOMY AND LIFETIME 109REQUEST ANATOMY AND LIFETIME
107 Every "aio_*" function creates a request. which is a C data structure 110 Every "aio_*" function creates a request. which is a C data structure
108 not directly visible to Perl. 111 not directly visible to Perl.
109 112
… …
146 the actual aio request is severed and calling its methods will 149 the actual aio request is severed and calling its methods will
147 either do nothing or result in a runtime error). 150 either do nothing or result in a runtime error).
148 151
149FUNCTIONS 152FUNCTIONS
150 QUICK OVERVIEW 153 QUICK OVERVIEW
151 This section simply lists the prototypes of the most important functions 154 This section simply lists the prototypes most of the functions for quick
152 for quick reference. See the following sections for function-by-function 155 reference. See the following sections for function-by-function
153 documentation. 156 documentation.
154 157
158 aio_wd $pathname, $callback->($wd)
155 aio_open $pathname, $flags, $mode, $callback->($fh) 159 aio_open $pathname, $flags, $mode, $callback->($fh)
156 aio_close $fh, $callback->($status) 160 aio_close $fh, $callback->($status)
161 aio_seek $fh,$offset,$whence, $callback->($offs)
157 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 162 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
158 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 163 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
159 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) 164 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
160 aio_readahead $fh,$offset,$length, $callback->($retval) 165 aio_readahead $fh,$offset,$length, $callback->($retval)
161 aio_stat $fh_or_path, $callback->($status) 166 aio_stat $fh_or_path, $callback->($status)
162 aio_lstat $fh, $callback->($status) 167 aio_lstat $fh, $callback->($status)
163 aio_statvfs $fh_or_path, $callback->($statvfs) 168 aio_statvfs $fh_or_path, $callback->($statvfs)
164 aio_utime $fh_or_path, $atime, $mtime, $callback->($status) 169 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
165 aio_chown $fh_or_path, $uid, $gid, $callback->($status) 170 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
171 aio_chmod $fh_or_path, $mode, $callback->($status)
166 aio_truncate $fh_or_path, $offset, $callback->($status) 172 aio_truncate $fh_or_path, $offset, $callback->($status)
167 aio_chmod $fh_or_path, $mode, $callback->($status) 173 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
174 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
168 aio_unlink $pathname, $callback->($status) 175 aio_unlink $pathname, $callback->($status)
169 aio_mknod $path, $mode, $dev, $callback->($status) 176 aio_mknod $pathname, $mode, $dev, $callback->($status)
170 aio_link $srcpath, $dstpath, $callback->($status) 177 aio_link $srcpath, $dstpath, $callback->($status)
171 aio_symlink $srcpath, $dstpath, $callback->($status) 178 aio_symlink $srcpath, $dstpath, $callback->($status)
172 aio_readlink $path, $callback->($link) 179 aio_readlink $pathname, $callback->($link)
180 aio_realpath $pathname, $callback->($path)
173 aio_rename $srcpath, $dstpath, $callback->($status) 181 aio_rename $srcpath, $dstpath, $callback->($status)
182 aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
174 aio_mkdir $pathname, $mode, $callback->($status) 183 aio_mkdir $pathname, $mode, $callback->($status)
175 aio_rmdir $pathname, $callback->($status) 184 aio_rmdir $pathname, $callback->($status)
176 aio_readdir $pathname, $callback->($entries) 185 aio_readdir $pathname, $callback->($entries)
177 aio_readdirx $pathname, $flags, $callback->($entries, $flags) 186 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
178 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST 187 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
179 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN 188 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
189 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
180 aio_load $path, $data, $callback->($status) 190 aio_load $pathname, $data, $callback->($status)
181 aio_copy $srcpath, $dstpath, $callback->($status) 191 aio_copy $srcpath, $dstpath, $callback->($status)
182 aio_move $srcpath, $dstpath, $callback->($status) 192 aio_move $srcpath, $dstpath, $callback->($status)
183 aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
184 aio_rmtree $path, $callback->($status) 193 aio_rmtree $pathname, $callback->($status)
194 aio_fcntl $fh, $cmd, $arg, $callback->($status)
195 aio_ioctl $fh, $request, $buf, $callback->($status)
185 aio_sync $callback->($status) 196 aio_sync $callback->($status)
197 aio_syncfs $fh, $callback->($status)
186 aio_fsync $fh, $callback->($status) 198 aio_fsync $fh, $callback->($status)
187 aio_fdatasync $fh, $callback->($status) 199 aio_fdatasync $fh, $callback->($status)
188 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) 200 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
189 aio_pathsync $path, $callback->($status) 201 aio_pathsync $pathname, $callback->($status)
190 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) 202 aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
191 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) 203 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
192 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) 204 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
193 aio_mlockall $flags, $callback->($status) 205 aio_mlockall $flags, $callback->($status)
194 aio_group $callback->(...) 206 aio_group $callback->(...)
195 aio_nop $callback->() 207 aio_nop $callback->()
… …
209 IO::AIO::idle_timeout $seconds 221 IO::AIO::idle_timeout $seconds
210 IO::AIO::max_outstanding $maxreqs 222 IO::AIO::max_outstanding $maxreqs
211 IO::AIO::nreqs 223 IO::AIO::nreqs
212 IO::AIO::nready 224 IO::AIO::nready
213 IO::AIO::npending 225 IO::AIO::npending
226 IO::AIO::reinit
227
228 $nfd = IO::AIO::get_fdlimit
229 IO::AIO::min_fdlimit $nfd
214 230
215 IO::AIO::sendfile $ofh, $ifh, $offset, $count 231 IO::AIO::sendfile $ofh, $ifh, $offset, $count
216 IO::AIO::fadvise $fh, $offset, $len, $advice 232 IO::AIO::fadvise $fh, $offset, $len, $advice
233 IO::AIO::fexecve $fh, $argv, $envp
234
235 IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
236 IO::AIO::munmap $scalar
237 IO::AIO::mremap $scalar, $new_length, $flags[, $new_address]
217 IO::AIO::madvise $scalar, $offset, $length, $advice 238 IO::AIO::madvise $scalar, $offset, $length, $advice
218 IO::AIO::mprotect $scalar, $offset, $length, $protect 239 IO::AIO::mprotect $scalar, $offset, $length, $protect
219 IO::AIO::munlock $scalar, $offset = 0, $length = undef 240 IO::AIO::munlock $scalar, $offset = 0, $length = undef
220 IO::AIO::munlockall 241 IO::AIO::munlockall
221 242
222 AIO REQUEST FUNCTIONS 243 # stat extensions
244 $counter = IO::AIO::st_gen
245 $seconds = IO::AIO::st_atime, IO::AIO::st_mtime, IO::AIO::st_ctime, IO::AIO::st_btime
246 ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtime
247 $nanoseconds = IO::AIO::st_atimensec, IO::AIO::st_mtimensec, IO::AIO::st_ctimensec, IO::AIO::st_btimensec
248 $seconds = IO::AIO::st_btimesec
249 ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtimensec
250
251 # very much unportable syscalls
252 IO::AIO::accept4 $r_fh, $sockaddr, $sockaddr_len, $flags
253 IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
254 IO::AIO::tee $r_fh, $w_fh, $length, $flags
255
256 $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
257 ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
258
259 $fh = IO::AIO::eventfd [$initval, [$flags]]
260 $fh = IO::AIO::memfd_create $pathname[, $flags]
261
262 $fh = IO::AIO::timerfd_create $clockid[, $flags]
263 ($cur_interval, $cur_value) = IO::AIO::timerfd_settime $fh, $flags, $new_interval, $nbw_value
264 ($cur_interval, $cur_value) = IO::AIO::timerfd_gettime $fh
265
266 $fh = IO::AIO::pidfd_open $pid[, $flags]
267 $status = IO::AIO::pidfd_send_signal $pidfh, $signal[, $siginfo[, $flags]]
268 $fh = IO::AIO::pidfd_getfd $pidfh, $targetfd[, $flags]
269
270 $retval = IO::AIO::mount $special, $path, $fstype, $flags = 0, $data = undef
271 $retval = IO::AIO::umount $path, $flags = 0
272
273 API NOTES
223 All the "aio_*" calls are more or less thin wrappers around the syscall 274 All the "aio_*" calls are more or less thin wrappers around the syscall
224 with the same name (sans "aio_"). The arguments are similar or 275 with the same name (sans "aio_"). The arguments are similar or
225 identical, and they all accept an additional (and optional) $callback 276 identical, and they all accept an additional (and optional) $callback
226 argument which must be a code reference. This code reference will get 277 argument which must be a code reference. This code reference will be
227 called with the syscall return code (e.g. most syscalls return -1 on
228 error, unlike perl, which usually delivers "false") as its sole argument
229 after the given syscall has been executed asynchronously. 278 called after the syscall has been executed in an asynchronous fashion.
279 The results of the request will be passed as arguments to the callback
280 (and, if an error occured, in $!) - for most requests the syscall return
281 code (e.g. most syscalls return -1 on error, unlike perl, which usually
282 delivers "false").
283
284 Some requests (such as "aio_readdir") pass the actual results and
285 communicate failures by passing "undef".
230 286
231 All functions expecting a filehandle keep a copy of the filehandle 287 All functions expecting a filehandle keep a copy of the filehandle
232 internally until the request has finished. 288 internally until the request has finished.
233 289
234 All functions return request objects of type IO::AIO::REQ that allow 290 All functions return request objects of type IO::AIO::REQ that allow
235 further manipulation of those requests while they are in-flight. 291 further manipulation of those requests while they are in-flight.
236 292
237 The pathnames you pass to these routines *must* be absolute and encoded 293 The pathnames you pass to these routines *should* be absolute. The
238 as octets. The reason for the former is that at the time the request is 294 reason for this is that at the time the request is being executed, the
239 being executed, the current working directory could have changed. 295 current working directory could have changed. Alternatively, you can
240 Alternatively, you can make sure that you never change the current 296 make sure that you never change the current working directory anywhere
241 working directory anywhere in the program and then use relative paths. 297 in the program and then use relative paths. You can also take advantage
298 of IO::AIOs working directory abstraction, that lets you specify paths
299 relative to some previously-opened "working directory object" - see the
300 description of the "IO::AIO::WD" class later in this document.
242 301
243 To encode pathnames as octets, either make sure you either: a) always 302 To encode pathnames as octets, either make sure you either: a) always
244 pass in filenames you got from outside (command line, readdir etc.) 303 pass in filenames you got from outside (command line, readdir etc.)
245 without tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module 304 without tinkering, b) are in your native filesystem encoding, c) use the
246 and encode your pathnames to the locale (or other) encoding in effect in 305 Encode module and encode your pathnames to the locale (or other)
247 the user environment, d) use Glib::filename_from_unicode on unicode 306 encoding in effect in the user environment, d) use
248 filenames or e) use something else to ensure your scalar has the correct 307 Glib::filename_from_unicode on unicode filenames or e) use something
249 contents. 308 else to ensure your scalar has the correct contents.
250 309
251 This works, btw. independent of the internal UTF-8 bit, which IO::AIO 310 This works, btw. independent of the internal UTF-8 bit, which IO::AIO
252 handles correctly whether it is set or not. 311 handles correctly whether it is set or not.
253 312
313 AIO REQUEST FUNCTIONS
254 $prev_pri = aioreq_pri [$pri] 314 $prev_pri = aioreq_pri [$pri]
255 Returns the priority value that would be used for the next request 315 Returns the priority value that would be used for the next request
256 and, if $pri is given, sets the priority for the next aio request. 316 and, if $pri is given, sets the priority for the next aio request.
257 317
258 The default priority is 0, the minimum and maximum priorities are -4 318 The default priority is 0, the minimum and maximum priorities are -4
… …
280 Similar to "aioreq_pri", but subtracts the given value from the 340 Similar to "aioreq_pri", but subtracts the given value from the
281 current priority, so the effect is cumulative. 341 current priority, so the effect is cumulative.
282 342
283 aio_open $pathname, $flags, $mode, $callback->($fh) 343 aio_open $pathname, $flags, $mode, $callback->($fh)
284 Asynchronously open or create a file and call the callback with a 344 Asynchronously open or create a file and call the callback with a
285 newly created filehandle for the file. 345 newly created filehandle for the file (or "undef" in case of an
346 error).
286 347
287 The pathname passed to "aio_open" must be absolute. See API NOTES, 348 The pathname passed to "aio_open" must be absolute. See API NOTES,
288 above, for an explanation. 349 above, for an explanation.
289 350
290 The $flags argument is a bitmask. See the "Fcntl" module for a list. 351 The $flags argument is a bitmask. See the "Fcntl" module for a list.
… …
313 "O_APPEND"), the following POSIX and non-POSIX constants are 374 "O_APPEND"), the following POSIX and non-POSIX constants are
314 available (missing ones on your system are, as usual, 0): 375 available (missing ones on your system are, as usual, 0):
315 376
316 "O_ASYNC", "O_DIRECT", "O_NOATIME", "O_CLOEXEC", "O_NOCTTY", 377 "O_ASYNC", "O_DIRECT", "O_NOATIME", "O_CLOEXEC", "O_NOCTTY",
317 "O_NOFOLLOW", "O_NONBLOCK", "O_EXEC", "O_SEARCH", "O_DIRECTORY", 378 "O_NOFOLLOW", "O_NONBLOCK", "O_EXEC", "O_SEARCH", "O_DIRECTORY",
318 "O_DSYNC", "O_RSYNC", "O_SYNC" and "O_TTY_INIT". 379 "O_DSYNC", "O_RSYNC", "O_SYNC", "O_PATH", "O_TMPFILE", "O_TTY_INIT"
380 and "O_ACCMODE".
319 381
320 aio_close $fh, $callback->($status) 382 aio_close $fh, $callback->($status)
321 Asynchronously close a file and call the callback with the result 383 Asynchronously close a file and call the callback with the result
322 code. 384 code.
323 385
… …
329 will use dup2 to overwrite the file descriptor with the write-end of 391 will use dup2 to overwrite the file descriptor with the write-end of
330 a pipe (the pipe fd will be created on demand and will be cached). 392 a pipe (the pipe fd will be created on demand and will be cached).
331 393
332 Or in other words: the file descriptor will be closed, but it will 394 Or in other words: the file descriptor will be closed, but it will
333 not be free for reuse until the perl filehandle is closed. 395 not be free for reuse until the perl filehandle is closed.
396
397 aio_seek $fh, $offset, $whence, $callback->($offs)
398 Seeks the filehandle to the new $offset, similarly to perl's
399 "sysseek". The $whence can use the traditional values (0 for
400 "IO::AIO::SEEK_SET", 1 for "IO::AIO::SEEK_CUR" or 2 for
401 "IO::AIO::SEEK_END").
402
403 The resulting absolute offset will be passed to the callback, or -1
404 in case of an error.
405
406 In theory, the $whence constants could be different than the
407 corresponding values from Fcntl, but perl guarantees they are the
408 same, so don't panic.
409
410 As a GNU/Linux (and maybe Solaris) extension, also the constants
411 "IO::AIO::SEEK_DATA" and "IO::AIO::SEEK_HOLE" are available, if they
412 could be found. No guarantees about suitability for use in
413 "aio_seek" or Perl's "sysseek" can be made though, although I would
414 naively assume they "just work".
334 415
335 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 416 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
336 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 417 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
337 Reads or writes $length bytes from or to the specified $fh and 418 Reads or writes $length bytes from or to the specified $fh and
338 $offset into the scalar given by $data and offset $dataoffset and 419 $offset into the scalar given by $data and offset $dataoffset and
339 calls the callback without the actual number of bytes read (or -1 on 420 calls the callback with the actual number of bytes transferred (or
340 error, just like the syscall). 421 -1 on error, just like the syscall).
341 422
342 "aio_read" will, like "sysread", shrink or grow the $data scalar to 423 "aio_read" will, like "sysread", shrink or grow the $data scalar to
343 offset plus the actual number of bytes read. 424 offset plus the actual number of bytes read.
344 425
345 If $offset is undefined, then the current file descriptor offset 426 If $offset is undefined, then the current file descriptor offset
… …
367 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) 448 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
368 Tries to copy $length bytes from $in_fh to $out_fh. It starts 449 Tries to copy $length bytes from $in_fh to $out_fh. It starts
369 reading at byte offset $in_offset, and starts writing at the current 450 reading at byte offset $in_offset, and starts writing at the current
370 file offset of $out_fh. Because of that, it is not safe to issue 451 file offset of $out_fh. Because of that, it is not safe to issue
371 more than one "aio_sendfile" per $out_fh, as they will interfere 452 more than one "aio_sendfile" per $out_fh, as they will interfere
372 with each other. 453 with each other. The same $in_fh works fine though, as this function
454 does not move or use the file offset of $in_fh.
373 455
374 Please note that "aio_sendfile" can read more bytes from $in_fh than 456 Please note that "aio_sendfile" can read more bytes from $in_fh than
375 are written, and there is no way to find out how many bytes have 457 are written, and there is no way to find out how many more bytes
376 been read from "aio_sendfile" alone, as "aio_sendfile" only provides 458 have been read from "aio_sendfile" alone, as "aio_sendfile" only
377 the number of bytes written to $out_fh. Only if the result value 459 provides the number of bytes written to $out_fh. Only if the result
378 equals $length one can assume that $length bytes have been read. 460 value equals $length one can assume that $length bytes have been
461 read.
379 462
380 Unlike with other "aio_" functions, it makes a lot of sense to use 463 Unlike with other "aio_" functions, it makes a lot of sense to use
381 "aio_sendfile" on non-blocking sockets, as long as one end 464 "aio_sendfile" on non-blocking sockets, as long as one end
382 (typically the $in_fh) is a file - the file I/O will then be 465 (typically the $in_fh) is a file - the file I/O will then be
383 asynchronous, while the socket I/O will be non-blocking. Note, 466 asynchronous, while the socket I/O will be non-blocking. Note,
384 however, that you can run into a trap where "aio_sendfile" reads 467 however, that you can run into a trap where "aio_sendfile" reads
385 some data with readahead, then fails to write all data, and when the 468 some data with readahead, then fails to write all data, and when the
386 socket is ready the next time, the data in the cache is already 469 socket is ready the next time, the data in the cache is already
387 lost, forcing "aio_sendfile" to again hit the disk. Explicit 470 lost, forcing "aio_sendfile" to again hit the disk. Explicit
388 "aio_read" + "aio_write" let's you control resource usage much 471 "aio_read" + "aio_write" let's you better control resource usage.
389 better.
390 472
391 This call tries to make use of a native "sendfile" syscall to 473 This call tries to make use of a native "sendfile"-like syscall to
392 provide zero-copy operation. For this to work, $out_fh should refer 474 provide zero-copy operation. For this to work, $out_fh should refer
393 to a socket, and $in_fh should refer to an mmap'able file. 475 to a socket, and $in_fh should refer to an mmap'able file.
394 476
395 If a native sendfile cannot be found or it fails with "ENOSYS", 477 If a native sendfile cannot be found or it fails with "ENOSYS",
396 "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or "ENOTSOCK", 478 "EINVAL", "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or
397 it will be emulated, so you can call "aio_sendfile" on any type of 479 "ENOTSOCK", it will be emulated, so you can call "aio_sendfile" on
398 filehandle regardless of the limitations of the operating system. 480 any type of filehandle regardless of the limitations of the
481 operating system.
482
483 As native sendfile syscalls (as practically any non-POSIX interface
484 hacked together in a hurry to improve benchmark numbers) tend to be
485 rather buggy on many systems, this implementation tries to work
486 around some known bugs in Linux and FreeBSD kernels (probably
487 others, too), but that might fail, so you really really should check
488 the return value of "aio_sendfile" - fewer bytes than expected might
489 have been transferred.
399 490
400 aio_readahead $fh,$offset,$length, $callback->($retval) 491 aio_readahead $fh,$offset,$length, $callback->($retval)
401 "aio_readahead" populates the page cache with data from a file so 492 "aio_readahead" populates the page cache with data from a file so
402 that subsequent reads from that file will not block on disk I/O. The 493 that subsequent reads from that file will not block on disk I/O. The
403 $offset argument specifies the starting point from which data is to 494 $offset argument specifies the starting point from which data is to
… …
406 to a page boundary and bytes are read up to the next page boundary 497 to a page boundary and bytes are read up to the next page boundary
407 greater than or equal to (off-set+length). "aio_readahead" does not 498 greater than or equal to (off-set+length). "aio_readahead" does not
408 read beyond the end of the file. The current file offset of the file 499 read beyond the end of the file. The current file offset of the file
409 is left unchanged. 500 is left unchanged.
410 501
411 If that syscall doesn't exist (likely if your OS isn't Linux) it 502 If that syscall doesn't exist (likely if your kernel isn't Linux) it
412 will be emulated by simply reading the data, which would have a 503 will be emulated by simply reading the data, which would have a
413 similar effect. 504 similar effect.
414 505
415 aio_stat $fh_or_path, $callback->($status) 506 aio_stat $fh_or_path, $callback->($status)
416 aio_lstat $fh, $callback->($status) 507 aio_lstat $fh, $callback->($status)
417 Works like perl's "stat" or "lstat" in void context. The callback 508 Works almost exactly like perl's "stat" or "lstat" in void context.
418 will be called after the stat and the results will be available 509 The callback will be called after the stat and the results will be
419 using "stat _" or "-s _" etc... 510 available using "stat _" or "-s _" and other tests (with the
511 exception of "-B" and "-T").
420 512
421 The pathname passed to "aio_stat" must be absolute. See API NOTES, 513 The pathname passed to "aio_stat" must be absolute. See API NOTES,
422 above, for an explanation. 514 above, for an explanation.
423 515
424 Currently, the stats are always 64-bit-stats, i.e. instead of 516 Currently, the stats are always 64-bit-stats, i.e. instead of
… …
432 back on traditional behaviour). 524 back on traditional behaviour).
433 525
434 "S_IFMT", "S_IFIFO", "S_IFCHR", "S_IFBLK", "S_IFLNK", "S_IFREG", 526 "S_IFMT", "S_IFIFO", "S_IFCHR", "S_IFBLK", "S_IFLNK", "S_IFREG",
435 "S_IFDIR", "S_IFWHT", "S_IFSOCK", "IO::AIO::major $dev_t", 527 "S_IFDIR", "S_IFWHT", "S_IFSOCK", "IO::AIO::major $dev_t",
436 "IO::AIO::minor $dev_t", "IO::AIO::makedev $major, $minor". 528 "IO::AIO::minor $dev_t", "IO::AIO::makedev $major, $minor".
529
530 To access higher resolution stat timestamps, see "SUBSECOND STAT
531 TIME ACCESS".
437 532
438 Example: Print the length of /etc/passwd: 533 Example: Print the length of /etc/passwd:
439 534
440 aio_stat "/etc/passwd", sub { 535 aio_stat "/etc/passwd", sub {
441 $_[0] and die "stat failed: $!"; 536 $_[0] and die "stat failed: $!";
… …
488 aio_utime $fh_or_path, $atime, $mtime, $callback->($status) 583 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
489 Works like perl's "utime" function (including the special case of 584 Works like perl's "utime" function (including the special case of
490 $atime and $mtime being undef). Fractional times are supported if 585 $atime and $mtime being undef). Fractional times are supported if
491 the underlying syscalls support them. 586 the underlying syscalls support them.
492 587
493 When called with a pathname, uses utimes(2) if available, otherwise 588 When called with a pathname, uses utimensat(2) or utimes(2) if
494 utime(2). If called on a file descriptor, uses futimes(2) if 589 available, otherwise utime(2). If called on a file descriptor, uses
495 available, otherwise returns ENOSYS, so this is not portable. 590 futimens(2) or futimes(2) if available, otherwise returns ENOSYS, so
591 this is not portable.
496 592
497 Examples: 593 Examples:
498 594
499 # set atime and mtime to current time (basically touch(1)): 595 # set atime and mtime to current time (basically touch(1)):
500 aio_utime "path", undef, undef; 596 aio_utime "path", undef, undef;
… …
514 aio_chown "path", 0, undef; 610 aio_chown "path", 0, undef;
515 611
516 aio_truncate $fh_or_path, $offset, $callback->($status) 612 aio_truncate $fh_or_path, $offset, $callback->($status)
517 Works like truncate(2) or ftruncate(2). 613 Works like truncate(2) or ftruncate(2).
518 614
615 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
616 Allocates or frees disk space according to the $mode argument. See
617 the linux "fallocate" documentation for details.
618
619 $mode is usually 0 or "IO::AIO::FALLOC_FL_KEEP_SIZE" to allocate
620 space, or "IO::AIO::FALLOC_FL_PUNCH_HOLE |
621 IO::AIO::FALLOC_FL_KEEP_SIZE", to deallocate a file range.
622
623 IO::AIO also supports "FALLOC_FL_COLLAPSE_RANGE", to remove a range
624 (without leaving a hole), "FALLOC_FL_ZERO_RANGE", to zero a range,
625 "FALLOC_FL_INSERT_RANGE" to insert a range and
626 "FALLOC_FL_UNSHARE_RANGE" to unshare shared blocks (see your
627 fallocate(2) manpage).
628
629 The file system block size used by "fallocate" is presumably the
630 "f_bsize" returned by "statvfs", but different filesystems and
631 filetypes can dictate other limitations.
632
633 If "fallocate" isn't available or cannot be emulated (currently no
634 emulation will be attempted), passes -1 and sets $! to "ENOSYS".
635
519 aio_chmod $fh_or_path, $mode, $callback->($status) 636 aio_chmod $fh_or_path, $mode, $callback->($status)
520 Works like perl's "chmod" function. 637 Works like perl's "chmod" function.
521 638
522 aio_unlink $pathname, $callback->($status) 639 aio_unlink $pathname, $callback->($status)
523 Asynchronously unlink (delete) a file and call the callback with the 640 Asynchronously unlink (delete) a file and call the callback with the
524 result code. 641 result code.
525 642
526 aio_mknod $path, $mode, $dev, $callback->($status) 643 aio_mknod $pathname, $mode, $dev, $callback->($status)
527 [EXPERIMENTAL] 644 [EXPERIMENTAL]
528 645
529 Asynchronously create a device node (or fifo). See mknod(2). 646 Asynchronously create a device node (or fifo). See mknod(2).
530 647
531 The only (POSIX-) portable way of calling this function is: 648 The only (POSIX-) portable way of calling this function is:
532 649
533 aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ... 650 aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
534 651
535 See "aio_stat" for info about some potentially helpful extra 652 See "aio_stat" for info about some potentially helpful extra
536 constants and functions. 653 constants and functions.
537 654
538 aio_link $srcpath, $dstpath, $callback->($status) 655 aio_link $srcpath, $dstpath, $callback->($status)
… …
542 aio_symlink $srcpath, $dstpath, $callback->($status) 659 aio_symlink $srcpath, $dstpath, $callback->($status)
543 Asynchronously create a new symbolic link to the existing object at 660 Asynchronously create a new symbolic link to the existing object at
544 $srcpath at the path $dstpath and call the callback with the result 661 $srcpath at the path $dstpath and call the callback with the result
545 code. 662 code.
546 663
547 aio_readlink $path, $callback->($link) 664 aio_readlink $pathname, $callback->($link)
548 Asynchronously read the symlink specified by $path and pass it to 665 Asynchronously read the symlink specified by $path and pass it to
549 the callback. If an error occurs, nothing or undef gets passed to 666 the callback. If an error occurs, nothing or undef gets passed to
550 the callback. 667 the callback.
551 668
669 aio_realpath $pathname, $callback->($path)
670 Asynchronously make the path absolute and resolve any symlinks in
671 $path. The resulting path only consists of directories (same as
672 Cwd::realpath).
673
674 This request can be used to get the absolute path of the current
675 working directory by passing it a path of . (a single dot).
676
552 aio_rename $srcpath, $dstpath, $callback->($status) 677 aio_rename $srcpath, $dstpath, $callback->($status)
553 Asynchronously rename the object at $srcpath to $dstpath, just as 678 Asynchronously rename the object at $srcpath to $dstpath, just as
554 rename(2) and call the callback with the result code. 679 rename(2) and call the callback with the result code.
680
681 On systems that support the AIO::WD working directory abstraction
682 natively, the case "[$wd, "."]" as $srcpath is specialcased -
683 instead of failing, "rename" is called on the absolute path of $wd.
684
685 aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
686 Basically a version of "aio_rename" with an additional $flags
687 argument. Calling this with "$flags=0" is the same as calling
688 "aio_rename".
689
690 Non-zero flags are currently only supported on GNU/Linux systems
691 that support renameat2. Other systems fail with "ENOSYS" in this
692 case.
693
694 The following constants are available (missing ones are, as usual
695 0), see renameat2(2) for details:
696
697 "IO::AIO::RENAME_NOREPLACE", "IO::AIO::RENAME_EXCHANGE" and
698 "IO::AIO::RENAME_WHITEOUT".
555 699
556 aio_mkdir $pathname, $mode, $callback->($status) 700 aio_mkdir $pathname, $mode, $callback->($status)
557 Asynchronously mkdir (create) a directory and call the callback with 701 Asynchronously mkdir (create) a directory and call the callback with
558 the result code. $mode will be modified by the umask at the time the 702 the result code. $mode will be modified by the umask at the time the
559 request is executed, so do not change your umask. 703 request is executed, so do not change your umask.
560 704
561 aio_rmdir $pathname, $callback->($status) 705 aio_rmdir $pathname, $callback->($status)
562 Asynchronously rmdir (delete) a directory and call the callback with 706 Asynchronously rmdir (delete) a directory and call the callback with
563 the result code. 707 the result code.
564 708
709 On systems that support the AIO::WD working directory abstraction
710 natively, the case "[$wd, "."]" is specialcased - instead of
711 failing, "rmdir" is called on the absolute path of $wd.
712
565 aio_readdir $pathname, $callback->($entries) 713 aio_readdir $pathname, $callback->($entries)
566 Unlike the POSIX call of the same name, "aio_readdir" reads an 714 Unlike the POSIX call of the same name, "aio_readdir" reads an
567 entire directory (i.e. opendir + readdir + closedir). The entries 715 entire directory (i.e. opendir + readdir + closedir). The entries
568 will not be sorted, and will NOT include the "." and ".." entries. 716 will not be sorted, and will NOT include the "." and ".." entries.
569 717
570 The callback is passed a single argument which is either "undef" or 718 The callback is passed a single argument which is either "undef" or
571 an array-ref with the filenames. 719 an array-ref with the filenames.
572 720
573 aio_readdirx $pathname, $flags, $callback->($entries, $flags) 721 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
574 Quite similar to "aio_readdir", but the $flags argument allows to 722 Quite similar to "aio_readdir", but the $flags argument allows one
575 tune behaviour and output format. In case of an error, $entries will 723 to tune behaviour and output format. In case of an error, $entries
576 be "undef". 724 will be "undef".
577 725
578 The flags are a combination of the following constants, ORed 726 The flags are a combination of the following constants, ORed
579 together (the flags will also be passed to the callback, possibly 727 together (the flags will also be passed to the callback, possibly
580 modified): 728 modified):
581 729
582 IO::AIO::READDIR_DENTS 730 IO::AIO::READDIR_DENTS
583 When this flag is off, then the callback gets an arrayref 731 Normally the callback gets an arrayref consisting of names only
584 consisting of names only (as with "aio_readdir"), otherwise it 732 (as with "aio_readdir"). If this flag is set, then the callback
585 gets an arrayref with "[$name, $type, $inode]" arrayrefs, each 733 gets an arrayref with "[$name, $type, $inode]" arrayrefs, each
586 describing a single directory entry in more detail. 734 describing a single directory entry in more detail:
587 735
588 $name is the name of the entry. 736 $name is the name of the entry.
589 737
590 $type is one of the "IO::AIO::DT_xxx" constants: 738 $type is one of the "IO::AIO::DT_xxx" constants:
591 739
592 "IO::AIO::DT_UNKNOWN", "IO::AIO::DT_FIFO", "IO::AIO::DT_CHR", 740 "IO::AIO::DT_UNKNOWN", "IO::AIO::DT_FIFO", "IO::AIO::DT_CHR",
593 "IO::AIO::DT_DIR", "IO::AIO::DT_BLK", "IO::AIO::DT_REG", 741 "IO::AIO::DT_DIR", "IO::AIO::DT_BLK", "IO::AIO::DT_REG",
594 "IO::AIO::DT_LNK", "IO::AIO::DT_SOCK", "IO::AIO::DT_WHT". 742 "IO::AIO::DT_LNK", "IO::AIO::DT_SOCK", "IO::AIO::DT_WHT".
595 743
596 "IO::AIO::DT_UNKNOWN" means just that: readdir does not know. If 744 "IO::AIO::DT_UNKNOWN" means just that: readdir does not know. If
597 you need to know, you have to run stat yourself. Also, for speed 745 you need to know, you have to run stat yourself. Also, for
598 reasons, the $type scalars are read-only: you can not modify 746 speed/memory reasons, the $type scalars are read-only: you must
599 them. 747 not modify them.
600 748
601 $inode is the inode number (which might not be exact on systems 749 $inode is the inode number (which might not be exact on systems
602 with 64 bit inode numbers and 32 bit perls). This field has 750 with 64 bit inode numbers and 32 bit perls). This field has
603 unspecified content on systems that do not deliver the inode 751 unspecified content on systems that do not deliver the inode
604 information. 752 information.
… …
616 of which names with short names are tried first. 764 of which names with short names are tried first.
617 765
618 IO::AIO::READDIR_STAT_ORDER 766 IO::AIO::READDIR_STAT_ORDER
619 When this flag is set, then the names will be returned in an 767 When this flag is set, then the names will be returned in an
620 order suitable for stat()'ing each one. That is, when you plan 768 order suitable for stat()'ing each one. That is, when you plan
621 to stat() all files in the given directory, then the returned 769 to stat() most or all files in the given directory, then the
622 order will likely be fastest. 770 returned order will likely be faster.
623 771
624 If both this flag and "IO::AIO::READDIR_DIRS_FIRST" are 772 If both this flag and "IO::AIO::READDIR_DIRS_FIRST" are
625 specified, then the likely dirs come first, resulting in a less 773 specified, then the likely dirs come first, resulting in a less
626 optimal stat order. 774 optimal stat order for stat'ing all entries, but likely a more
775 optimal order for finding subdirectories.
627 776
628 IO::AIO::READDIR_FOUND_UNKNOWN 777 IO::AIO::READDIR_FOUND_UNKNOWN
629 This flag should not be set when calling "aio_readdirx". 778 This flag should not be set when calling "aio_readdirx".
630 Instead, it is being set by "aio_readdirx", when any of the 779 Instead, it is being set by "aio_readdirx", when any of the
631 $type's found were "IO::AIO::DT_UNKNOWN". The absense of this 780 $type's found were "IO::AIO::DT_UNKNOWN". The absence of this
632 flag therefore indicates that all $type's are known, which can 781 flag therefore indicates that all $type's are known, which can
633 be used to speed up some algorithms. 782 be used to speed up some algorithms.
634 783
784 aio_slurp $pathname, $offset, $length, $data, $callback->($status)
785 Opens, reads and closes the given file. The data is put into $data,
786 which is resized as required.
787
788 If $offset is negative, then it is counted from the end of the file.
789
790 If $length is zero, then the remaining length of the file is used.
791 Also, in this case, the same limitations to modifying $data apply as
792 when IO::AIO::mmap is used, i.e. it must only be modified in-place
793 with "substr". If the size of the file is known, specifying a
794 non-zero $length results in a performance advantage.
795
796 This request is similar to the older "aio_load" request, but since
797 it is a single request, it might be more efficient to use.
798
799 Example: load /etc/passwd into $passwd.
800
801 my $passwd;
802 aio_slurp "/etc/passwd", 0, 0, $passwd, sub {
803 $_[0] >= 0
804 or die "/etc/passwd: $!\n";
805
806 printf "/etc/passwd is %d bytes long, and contains:\n", length $passwd;
807 print $passwd;
808 };
809 IO::AIO::flush;
810
635 aio_load $path, $data, $callback->($status) 811 aio_load $pathname, $data, $callback->($status)
636 This is a composite request that tries to fully load the given file 812 This is a composite request that tries to fully load the given file
637 into memory. Status is the same as with aio_read. 813 into memory. Status is the same as with aio_read.
814
815 Using "aio_slurp" might be more efficient, as it is a single
816 request.
638 817
639 aio_copy $srcpath, $dstpath, $callback->($status) 818 aio_copy $srcpath, $dstpath, $callback->($status)
640 Try to copy the *file* (directories not supported as either source 819 Try to copy the *file* (directories not supported as either source
641 or destination) from $srcpath to $dstpath and call the callback with 820 or destination) from $srcpath to $dstpath and call the callback with
642 a status of 0 (ok) or -1 (error, see $!). 821 a status of 0 (ok) or -1 (error, see $!).
643 822
823 Existing destination files will be truncated.
824
644 This is a composite request that creates the destination file with 825 This is a composite request that creates the destination file with
645 mode 0200 and copies the contents of the source file into it using 826 mode 0200 and copies the contents of the source file into it using
646 "aio_sendfile", followed by restoring atime, mtime, access mode and 827 "aio_sendfile", followed by restoring atime, mtime, access mode and
647 uid/gid, in that order. 828 uid/gid, in that order.
648 829
… …
657 838
658 This is a composite request that tries to rename(2) the file first; 839 This is a composite request that tries to rename(2) the file first;
659 if rename fails with "EXDEV", it copies the file with "aio_copy" 840 if rename fails with "EXDEV", it copies the file with "aio_copy"
660 and, if that is successful, unlinks the $srcpath. 841 and, if that is successful, unlinks the $srcpath.
661 842
662 aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 843 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
663 Scans a directory (similar to "aio_readdir") but additionally tries 844 Scans a directory (similar to "aio_readdir") but additionally tries
664 to efficiently separate the entries of directory $path into two sets 845 to efficiently separate the entries of directory $path into two sets
665 of names, directories you can recurse into (directories), and ones 846 of names, directories you can recurse into (directories), and ones
666 you cannot recurse into (everything else, including symlinks to 847 you cannot recurse into (everything else, including symlinks to
667 directories). 848 directories).
668 849
669 "aio_scandir" is a composite request that creates of many sub 850 "aio_scandir" is a composite request that generates many sub
670 requests_ $maxreq specifies the maximum number of outstanding aio 851 requests. $maxreq specifies the maximum number of outstanding aio
671 requests that this function generates. If it is "<= 0", then a 852 requests that this function generates. If it is "<= 0", then a
672 suitable default will be chosen (currently 4). 853 suitable default will be chosen (currently 4).
673 854
674 On error, the callback is called without arguments, otherwise it 855 On error, the callback is called without arguments, otherwise it
675 receives two array-refs with path-relative entry names. 856 receives two array-refs with path-relative entry names.
… …
700 Then entries will be sorted into likely directories a non-initial 881 Then entries will be sorted into likely directories a non-initial
701 dot currently) and likely non-directories (see "aio_readdirx"). Then 882 dot currently) and likely non-directories (see "aio_readdirx"). Then
702 every entry plus an appended "/." will be "stat"'ed, likely 883 every entry plus an appended "/." will be "stat"'ed, likely
703 directories first, in order of their inode numbers. If that 884 directories first, in order of their inode numbers. If that
704 succeeds, it assumes that the entry is a directory or a symlink to 885 succeeds, it assumes that the entry is a directory or a symlink to
705 directory (which will be checked seperately). This is often faster 886 directory (which will be checked separately). This is often faster
706 than stat'ing the entry itself because filesystems might detect the 887 than stat'ing the entry itself because filesystems might detect the
707 type of the entry without reading the inode data (e.g. ext2fs 888 type of the entry without reading the inode data (e.g. ext2fs
708 filetype feature), even on systems that cannot return the filetype 889 filetype feature), even on systems that cannot return the filetype
709 information on readdir. 890 information on readdir.
710 891
… …
716 897
717 It will also likely work on non-POSIX filesystems with reduced 898 It will also likely work on non-POSIX filesystems with reduced
718 efficiency as those tend to return 0 or 1 as link counts, which 899 efficiency as those tend to return 0 or 1 as link counts, which
719 disables the directory counting heuristic. 900 disables the directory counting heuristic.
720 901
721 aio_rmtree $path, $callback->($status) 902 aio_rmtree $pathname, $callback->($status)
722 Delete a directory tree starting (and including) $path, return the 903 Delete a directory tree starting (and including) $path, return the
723 status of the final "rmdir" only. This is a composite request that 904 status of the final "rmdir" only. This is a composite request that
724 uses "aio_scandir" to recurse into and rmdir directories, and unlink 905 uses "aio_scandir" to recurse into and rmdir directories, and unlink
725 everything else. 906 everything else.
726 907
908 aio_fcntl $fh, $cmd, $arg, $callback->($status)
909 aio_ioctl $fh, $request, $buf, $callback->($status)
910 These work just like the "fcntl" and "ioctl" built-in functions,
911 except they execute asynchronously and pass the return value to the
912 callback.
913
914 Both calls can be used for a lot of things, some of which make more
915 sense to run asynchronously in their own thread, while some others
916 make less sense. For example, calls that block waiting for external
917 events, such as locking, will also lock down an I/O thread while it
918 is waiting, which can deadlock the whole I/O system. At the same
919 time, there might be no alternative to using a thread to wait.
920
921 So in general, you should only use these calls for things that do
922 (filesystem) I/O, not for things that wait for other events
923 (network, other processes), although if you are careful and know
924 what you are doing, you still can.
925
926 The following constants are available and can be used for normal
927 "ioctl" and "fcntl" as well (missing ones are, as usual 0):
928
929 "F_DUPFD_CLOEXEC",
930
931 "F_OFD_GETLK", "F_OFD_SETLK", "F_OFD_GETLKW",
932
933 "FIFREEZE", "FITHAW", "FITRIM", "FICLONE", "FICLONERANGE",
934 "FIDEDUPERANGE".
935
936 "F_ADD_SEALS", "F_GET_SEALS", "F_SEAL_SEAL", "F_SEAL_SHRINK",
937 "F_SEAL_GROW" and "F_SEAL_WRITE".
938
939 "FS_IOC_GETFLAGS", "FS_IOC_SETFLAGS", "FS_IOC_GETVERSION",
940 "FS_IOC_SETVERSION", "FS_IOC_FIEMAP".
941
942 "FS_IOC_FSGETXATTR", "FS_IOC_FSSETXATTR",
943 "FS_IOC_SET_ENCRYPTION_POLICY", "FS_IOC_GET_ENCRYPTION_PWSALT",
944 "FS_IOC_GET_ENCRYPTION_POLICY", "FS_KEY_DESCRIPTOR_SIZE".
945
946 "FS_SECRM_FL", "FS_UNRM_FL", "FS_COMPR_FL", "FS_SYNC_FL",
947 "FS_IMMUTABLE_FL", "FS_APPEND_FL", "FS_NODUMP_FL", "FS_NOATIME_FL",
948 "FS_DIRTY_FL", "FS_COMPRBLK_FL", "FS_NOCOMP_FL", "FS_ENCRYPT_FL",
949 "FS_BTREE_FL", "FS_INDEX_FL", "FS_JOURNAL_DATA_FL", "FS_NOTAIL_FL",
950 "FS_DIRSYNC_FL", "FS_TOPDIR_FL", "FS_FL_USER_MODIFIABLE".
951
952 "FS_XFLAG_REALTIME", "FS_XFLAG_PREALLOC", "FS_XFLAG_IMMUTABLE",
953 "FS_XFLAG_APPEND", "FS_XFLAG_SYNC", "FS_XFLAG_NOATIME",
954 "FS_XFLAG_NODUMP", "FS_XFLAG_RTINHERIT", "FS_XFLAG_PROJINHERIT",
955 "FS_XFLAG_NOSYMLINKS", "FS_XFLAG_EXTSIZE", "FS_XFLAG_EXTSZINHERIT",
956 "FS_XFLAG_NODEFRAG", "FS_XFLAG_FILESTREAM", "FS_XFLAG_DAX",
957 "FS_XFLAG_HASATTR",
958
959 "BLKROSET", "BLKROGET", "BLKRRPART", "BLKGETSIZE", "BLKFLSBUF",
960 "BLKRASET", "BLKRAGET", "BLKFRASET", "BLKFRAGET", "BLKSECTSET",
961 "BLKSECTGET", "BLKSSZGET", "BLKBSZGET", "BLKBSZSET", "BLKGETSIZE64",
962
727 aio_sync $callback->($status) 963 aio_sync $callback->($status)
728 Asynchronously call sync and call the callback when finished. 964 Asynchronously call sync and call the callback when finished.
729 965
730 aio_fsync $fh, $callback->($status) 966 aio_fsync $fh, $callback->($status)
731 Asynchronously call fsync on the given filehandle and call the 967 Asynchronously call fsync on the given filehandle and call the
… …
735 Asynchronously call fdatasync on the given filehandle and call the 971 Asynchronously call fdatasync on the given filehandle and call the
736 callback with the fdatasync result code. 972 callback with the fdatasync result code.
737 973
738 If this call isn't available because your OS lacks it or it couldn't 974 If this call isn't available because your OS lacks it or it couldn't
739 be detected, it will be emulated by calling "fsync" instead. 975 be detected, it will be emulated by calling "fsync" instead.
976
977 aio_syncfs $fh, $callback->($status)
978 Asynchronously call the syncfs syscall to sync the filesystem
979 associated to the given filehandle and call the callback with the
980 syncfs result code. If syncfs is not available, calls sync(), but
981 returns -1 and sets errno to "ENOSYS" nevertheless.
740 982
741 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) 983 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
742 Sync the data portion of the file specified by $offset and $length 984 Sync the data portion of the file specified by $offset and $length
743 to disk (but NOT the metadata), by calling the Linux-specific 985 to disk (but NOT the metadata), by calling the Linux-specific
744 sync_file_range call. If sync_file_range is not available or it 986 sync_file_range call. If sync_file_range is not available or it
… …
748 "IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE", 990 "IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE",
749 "IO::AIO::SYNC_FILE_RANGE_WRITE" and 991 "IO::AIO::SYNC_FILE_RANGE_WRITE" and
750 "IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER": refer to the sync_file_range 992 "IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER": refer to the sync_file_range
751 manpage for details. 993 manpage for details.
752 994
753 aio_pathsync $path, $callback->($status) 995 aio_pathsync $pathname, $callback->($status)
754 This request tries to open, fsync and close the given path. This is 996 This request tries to open, fsync and close the given path. This is
755 a composite request intended to sync directories after directory 997 a composite request intended to sync directories after directory
756 operations (E.g. rename). This might not work on all operating 998 operations (E.g. rename). This might not work on all operating
757 systems or have any specific effect, but usually it makes sure that 999 systems or have any specific effect, but usually it makes sure that
758 directory changes get written to disc. It works for anything that 1000 directory changes get written to disc. It works for anything that
… …
761 Future versions of this function might fall back to other methods 1003 Future versions of this function might fall back to other methods
762 when "fsync" on the directory fails (such as calling "sync"). 1004 when "fsync" on the directory fails (such as calling "sync").
763 1005
764 Passes 0 when everything went ok, and -1 on error. 1006 Passes 0 when everything went ok, and -1 on error.
765 1007
766 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, 1008 aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC,
767 $callback->($status) 1009 $callback->($status)
768 This is a rather advanced IO::AIO call, which only works on 1010 This is a rather advanced IO::AIO call, which only works on
769 mmap(2)ed scalars (see the "IO::AIO::mmap" function, although it 1011 mmap(2)ed scalars (see the "IO::AIO::mmap" function, although it
770 also works on data scalars managed by the Sys::Mmap or Mmap modules, 1012 also works on data scalars managed by the Sys::Mmap or Mmap modules,
771 note that the scalar must only be modified in-place while an aio 1013 note that the scalar must only be modified in-place while an aio
… …
773 1015
774 It calls the "msync" function of your OS, if available, with the 1016 It calls the "msync" function of your OS, if available, with the
775 memory area starting at $offset in the string and ending $length 1017 memory area starting at $offset in the string and ending $length
776 bytes later. If $length is negative, counts from the end, and if 1018 bytes later. If $length is negative, counts from the end, and if
777 $length is "undef", then it goes till the end of the string. The 1019 $length is "undef", then it goes till the end of the string. The
778 flags can be a combination of "IO::AIO::MS_ASYNC", 1020 flags can be either "IO::AIO::MS_ASYNC" or "IO::AIO::MS_SYNC", plus
779 "IO::AIO::MS_INVALIDATE" and "IO::AIO::MS_SYNC". 1021 an optional "IO::AIO::MS_INVALIDATE".
780 1022
781 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, 1023 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0,
782 $callback->($status) 1024 $callback->($status)
783 This is a rather advanced IO::AIO call, which works best on 1025 This is a rather advanced IO::AIO call, which works best on
784 mmap(2)ed scalars. 1026 mmap(2)ed scalars.
785 1027
786 It touches (reads or writes) all memory pages in the specified range 1028 It touches (reads or writes) all memory pages in the specified range
787 inside the scalar. All caveats and parameters are the same as for 1029 inside the scalar. All caveats and parameters are the same as for
788 "aio_msync", above, except for flags, which must be either 0 (which 1030 "aio_msync", above, except for flags, which must be either 0 (which
789 reads all pages and ensures they are instantiated) or 1031 reads all pages and ensures they are instantiated) or
790 "IO::AIO::MT_MODIFY", which modifies the memory page s(by reading 1032 "IO::AIO::MT_MODIFY", which modifies the memory pages (by reading
791 and writing an octet from it, which dirties the page). 1033 and writing an octet from it, which dirties the page).
792 1034
793 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) 1035 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
794 This is a rather advanced IO::AIO call, which works best on 1036 This is a rather advanced IO::AIO call, which works best on
795 mmap(2)ed scalars. 1037 mmap(2)ed scalars.
… …
815 IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh; 1057 IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
816 aio_mlock $data; # mlock in background 1058 aio_mlock $data; # mlock in background
817 1059
818 aio_mlockall $flags, $callback->($status) 1060 aio_mlockall $flags, $callback->($status)
819 Calls the "mlockall" function with the given $flags (a combination 1061 Calls the "mlockall" function with the given $flags (a combination
820 of "IO::AIO::MCL_CURRENT" and "IO::AIO::MCL_FUTURE"). 1062 of "IO::AIO::MCL_CURRENT", "IO::AIO::MCL_FUTURE" and
1063 "IO::AIO::MCL_ONFAULT").
821 1064
822 On systems that do not implement "mlockall", this function returns 1065 On systems that do not implement "mlockall", this function returns
823 -1 and sets errno to "ENOSYS". 1066 -1 and sets errno to "ENOSYS". Similarly, flag combinations not
1067 supported by the system result in a return value of -1 with errno
1068 being set to "EINVAL".
824 1069
825 Note that the corresponding "munlockall" is synchronous and is 1070 Note that the corresponding "munlockall" is synchronous and is
826 documented under "MISCELLANEOUS FUNCTIONS". 1071 documented under "MISCELLANEOUS FUNCTIONS".
827 1072
828 Example: asynchronously lock all current and future pages into 1073 Example: asynchronously lock all current and future pages into
829 memory. 1074 memory.
830 1075
831 aio_mlockall IO::AIO::MCL_FUTURE; 1076 aio_mlockall IO::AIO::MCL_FUTURE;
1077
1078 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
1079 Queries the extents of the given file (by calling the Linux "FIEMAP"
1080 ioctl, see <http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for
1081 details). If the ioctl is not available on your OS, then this
1082 request will fail with "ENOSYS".
1083
1084 $start is the starting offset to query extents for, $length is the
1085 size of the range to query - if it is "undef", then the whole file
1086 will be queried.
1087
1088 $flags is a combination of flags ("IO::AIO::FIEMAP_FLAG_SYNC" or
1089 "IO::AIO::FIEMAP_FLAG_XATTR" - "IO::AIO::FIEMAP_FLAGS_COMPAT" is
1090 also exported), and is normally 0 or "IO::AIO::FIEMAP_FLAG_SYNC" to
1091 query the data portion.
1092
1093 $count is the maximum number of extent records to return. If it is
1094 "undef", then IO::AIO queries all extents of the range. As a very
1095 special case, if it is 0, then the callback receives the number of
1096 extents instead of the extents themselves (which is unreliable, see
1097 below).
1098
1099 If an error occurs, the callback receives no arguments. The special
1100 "errno" value "IO::AIO::EBADR" is available to test for flag errors.
1101
1102 Otherwise, the callback receives an array reference with extent
1103 structures. Each extent structure is an array reference itself, with
1104 the following members:
1105
1106 [$logical, $physical, $length, $flags]
1107
1108 Flags is any combination of the following flag values (typically
1109 either 0 or "IO::AIO::FIEMAP_EXTENT_LAST" (1)):
1110
1111 "IO::AIO::FIEMAP_EXTENT_LAST", "IO::AIO::FIEMAP_EXTENT_UNKNOWN",
1112 "IO::AIO::FIEMAP_EXTENT_DELALLOC", "IO::AIO::FIEMAP_EXTENT_ENCODED",
1113 "IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED",
1114 "IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED",
1115 "IO::AIO::FIEMAP_EXTENT_DATA_INLINE",
1116 "IO::AIO::FIEMAP_EXTENT_DATA_TAIL",
1117 "IO::AIO::FIEMAP_EXTENT_UNWRITTEN", "IO::AIO::FIEMAP_EXTENT_MERGED"
1118 or "IO::AIO::FIEMAP_EXTENT_SHARED".
1119
1120 At the time of this writing (Linux 3.2), this request is unreliable
1121 unless $count is "undef", as the kernel has all sorts of bugs
1122 preventing it to return all extents of a range for files with a
1123 large number of extents. The code (only) works around all these
1124 issues if $count is "undef".
832 1125
833 aio_group $callback->(...) 1126 aio_group $callback->(...)
834 This is a very special aio request: Instead of doing something, it 1127 This is a very special aio request: Instead of doing something, it
835 is a container for other aio requests, which is useful if you want 1128 is a container for other aio requests, which is useful if you want
836 to bundle many requests into a single, composite, request with a 1129 to bundle many requests into a single, composite, request with a
… …
870 While it is theoretically handy to have simple I/O scheduling 1163 While it is theoretically handy to have simple I/O scheduling
871 requests like sleep and file handle readable/writable, the overhead 1164 requests like sleep and file handle readable/writable, the overhead
872 this creates is immense (it blocks a thread for a long time) so do 1165 this creates is immense (it blocks a thread for a long time) so do
873 not use this function except to put your application under 1166 not use this function except to put your application under
874 artificial I/O pressure. 1167 artificial I/O pressure.
1168
1169 IO::AIO::WD - multiple working directories
1170 Your process only has one current working directory, which is used by
1171 all threads. This makes it hard to use relative paths (some other
1172 component could call "chdir" at any time, and it is hard to control when
1173 the path will be used by IO::AIO).
1174
1175 One solution for this is to always use absolute paths. This usually
1176 works, but can be quite slow (the kernel has to walk the whole path on
1177 every access), and can also be a hassle to implement.
1178
1179 Newer POSIX systems have a number of functions (openat, fdopendir,
1180 futimensat and so on) that make it possible to specify working
1181 directories per operation.
1182
1183 For portability, and because the clowns who "designed", or shall I
1184 write, perpetrated this new interface were obviously half-drunk, this
1185 abstraction cannot be perfect, though.
1186
1187 IO::AIO allows you to convert directory paths into a so-called
1188 IO::AIO::WD object. This object stores the canonicalised, absolute
1189 version of the path, and on systems that allow it, also a directory file
1190 descriptor.
1191
1192 Everywhere where a pathname is accepted by IO::AIO (e.g. in "aio_stat"
1193 or "aio_unlink"), one can specify an array reference with an IO::AIO::WD
1194 object and a pathname instead (or the IO::AIO::WD object alone, which
1195 gets interpreted as "[$wd, "."]"). If the pathname is absolute, the
1196 IO::AIO::WD object is ignored, otherwise the pathname is resolved
1197 relative to that IO::AIO::WD object.
1198
1199 For example, to get a wd object for /etc and then stat passwd inside,
1200 you would write:
1201
1202 aio_wd "/etc", sub {
1203 my $etcdir = shift;
1204
1205 # although $etcdir can be undef on error, there is generally no reason
1206 # to check for errors here, as aio_stat will fail with ENOENT
1207 # when $etcdir is undef.
1208
1209 aio_stat [$etcdir, "passwd"], sub {
1210 # yay
1211 };
1212 };
1213
1214 The fact that "aio_wd" is a request and not a normal function shows that
1215 creating an IO::AIO::WD object is itself a potentially blocking
1216 operation, which is why it is done asynchronously.
1217
1218 To stat the directory obtained with "aio_wd" above, one could write
1219 either of the following three request calls:
1220
1221 aio_lstat "/etc" , sub { ... # pathname as normal string
1222 aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
1223 aio_lstat $wd , sub { ... # shorthand for the previous
1224
1225 As with normal pathnames, IO::AIO keeps a copy of the working directory
1226 object and the pathname string, so you could write the following without
1227 causing any issues due to $path getting reused:
1228
1229 my $path = [$wd, undef];
1230
1231 for my $name (qw(abc def ghi)) {
1232 $path->[1] = $name;
1233 aio_stat $path, sub {
1234 # ...
1235 };
1236 }
1237
1238 There are some caveats: when directories get renamed (or deleted), the
1239 pathname string doesn't change, so will point to the new directory (or
1240 nowhere at all), while the directory fd, if available on the system,
1241 will still point to the original directory. Most functions accepting a
1242 pathname will use the directory fd on newer systems, and the string on
1243 older systems. Some functions (such as "aio_realpath") will always rely
1244 on the string form of the pathname.
1245
1246 So this functionality is mainly useful to get some protection against
1247 "chdir", to easily get an absolute path out of a relative path for
1248 future reference, and to speed up doing many operations in the same
1249 directory (e.g. when stat'ing all files in a directory).
1250
1251 The following functions implement this working directory abstraction:
1252
1253 aio_wd $pathname, $callback->($wd)
1254 Asynchonously canonicalise the given pathname and convert it to an
1255 IO::AIO::WD object representing it. If possible and supported on the
1256 system, also open a directory fd to speed up pathname resolution
1257 relative to this working directory.
1258
1259 If something goes wrong, then "undef" is passwd to the callback
1260 instead of a working directory object and $! is set appropriately.
1261 Since passing "undef" as working directory component of a pathname
1262 fails the request with "ENOENT", there is often no need for error
1263 checking in the "aio_wd" callback, as future requests using the
1264 value will fail in the expected way.
1265
1266 IO::AIO::CWD
1267 This is a compile time constant (object) that represents the process
1268 current working directory.
1269
1270 Specifying this object as working directory object for a pathname is
1271 as if the pathname would be specified directly, without a directory
1272 object. For example, these calls are functionally identical:
1273
1274 aio_stat "somefile", sub { ... };
1275 aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
1276
1277 To recover the path associated with an IO::AIO::WD object, you can use
1278 "aio_realpath":
1279
1280 aio_realpath $wd, sub {
1281 warn "path is $_[0]\n";
1282 };
1283
1284 Currently, "aio_statvfs" always, and "aio_rename" and "aio_rmdir"
1285 sometimes, fall back to using an absolue path.
875 1286
876 IO::AIO::REQ CLASS 1287 IO::AIO::REQ CLASS
877 All non-aggregate "aio_*" functions return an object of this class when 1288 All non-aggregate "aio_*" functions return an object of this class when
878 called in non-void context. 1289 called in non-void context.
879 1290
… …
979 Sets a feeder/generator on this group: every group can have an 1390 Sets a feeder/generator on this group: every group can have an
980 attached generator that generates requests if idle. The idea behind 1391 attached generator that generates requests if idle. The idea behind
981 this is that, although you could just queue as many requests as you 1392 this is that, although you could just queue as many requests as you
982 want in a group, this might starve other requests for a potentially 1393 want in a group, this might starve other requests for a potentially
983 long time. For example, "aio_scandir" might generate hundreds of 1394 long time. For example, "aio_scandir" might generate hundreds of
984 thousands "aio_stat" requests, delaying any later requests for a 1395 thousands of "aio_stat" requests, delaying any later requests for a
985 long time. 1396 long time.
986 1397
987 To avoid this, and allow incremental generation of requests, you can 1398 To avoid this, and allow incremental generation of requests, you can
988 instead a group and set a feeder on it that generates those 1399 instead a group and set a feeder on it that generates those
989 requests. The feed callback will be called whenever there are few 1400 requests. The feed callback will be called whenever there are few
… …
1031 results. 1442 results.
1032 1443
1033 See "poll_cb" for an example. 1444 See "poll_cb" for an example.
1034 1445
1035 IO::AIO::poll_cb 1446 IO::AIO::poll_cb
1036 Process some outstanding events on the result pipe. You have to call 1447 Process some requests that have reached the result phase (i.e. they
1448 have been executed but the results are not yet reported). You have
1449 to call this "regularly" to finish outstanding requests.
1450
1037 this regularly. Returns 0 if all events could be processed (or there 1451 Returns 0 if all events could be processed (or there were no events
1038 were no events to process), or -1 if it returned earlier for 1452 to process), or -1 if it returned earlier for whatever reason.
1039 whatever reason. Returns immediately when no events are outstanding. 1453 Returns immediately when no events are outstanding. The amount of
1040 The amount of events processed depends on the settings of 1454 events processed depends on the settings of "IO::AIO::max_poll_req",
1041 "IO::AIO::max_poll_req" and "IO::AIO::max_poll_time". 1455 "IO::AIO::max_poll_time" and "IO::AIO::max_outstanding".
1042 1456
1043 If not all requests were processed for whatever reason, the 1457 If not all requests were processed for whatever reason, the poll
1044 filehandle will still be ready when "poll_cb" returns, so normally 1458 file descriptor will still be ready when "poll_cb" returns, so
1045 you don't have to do anything special to have it called later. 1459 normally you don't have to do anything special to have it called
1460 later.
1046 1461
1047 Apart from calling "IO::AIO::poll_cb" when the event filehandle 1462 Apart from calling "IO::AIO::poll_cb" when the event filehandle
1048 becomes ready, it can be beneficial to call this function from loops 1463 becomes ready, it can be beneficial to call this function from loops
1049 which submit a lot of requests, to make sure the results get 1464 which submit a lot of requests, to make sure the results get
1050 processed when they become available and not just when the loop is 1465 processed when they become available and not just when the loop is
… …
1058 Event->io (fd => IO::AIO::poll_fileno, 1473 Event->io (fd => IO::AIO::poll_fileno,
1059 poll => 'r', async => 1, 1474 poll => 'r', async => 1,
1060 cb => \&IO::AIO::poll_cb); 1475 cb => \&IO::AIO::poll_cb);
1061 1476
1062 IO::AIO::poll_wait 1477 IO::AIO::poll_wait
1063 If there are any outstanding requests and none of them in the result 1478 Wait until either at least one request is in the result phase or no
1064 phase, wait till the result filehandle becomes ready for reading 1479 requests are outstanding anymore.
1065 (simply does a "select" on the filehandle. This is useful if you 1480
1066 want to synchronously wait for some requests to finish). 1481 This is useful if you want to synchronously wait for some requests
1482 to become ready, without actually handling them.
1067 1483
1068 See "nreqs" for an example. 1484 See "nreqs" for an example.
1069 1485
1070 IO::AIO::poll 1486 IO::AIO::poll
1071 Waits until some requests have been handled. 1487 Waits until some requests have been handled.
… …
1080 1496
1081 Strictly equivalent to: 1497 Strictly equivalent to:
1082 1498
1083 IO::AIO::poll_wait, IO::AIO::poll_cb 1499 IO::AIO::poll_wait, IO::AIO::poll_cb
1084 while IO::AIO::nreqs; 1500 while IO::AIO::nreqs;
1501
1502 This function can be useful at program aborts, to make sure
1503 outstanding I/O has been done ("IO::AIO" uses an "END" block which
1504 already calls this function on normal exits), or when you are merely
1505 using "IO::AIO" for its more advanced functions, rather than for
1506 async I/O, e.g.:
1507
1508 my ($dirs, $nondirs);
1509 IO::AIO::aio_scandir "/tmp", 0, sub { ($dirs, $nondirs) = @_ };
1510 IO::AIO::flush;
1511 # $dirs, $nondirs are now set
1085 1512
1086 IO::AIO::max_poll_reqs $nreqs 1513 IO::AIO::max_poll_reqs $nreqs
1087 IO::AIO::max_poll_time $seconds 1514 IO::AIO::max_poll_time $seconds
1088 These set the maximum number of requests (default 0, meaning 1515 These set the maximum number of requests (default 0, meaning
1089 infinity) that are being processed by "IO::AIO::poll_cb" in one 1516 infinity) that are being processed by "IO::AIO::poll_cb" in one
… …
1170 IO::AIO::idle_timeout $seconds 1597 IO::AIO::idle_timeout $seconds
1171 Sets the minimum idle timeout (default 10) after which worker 1598 Sets the minimum idle timeout (default 10) after which worker
1172 threads are allowed to exit. SEe "IO::AIO::max_idle". 1599 threads are allowed to exit. SEe "IO::AIO::max_idle".
1173 1600
1174 IO::AIO::max_outstanding $maxreqs 1601 IO::AIO::max_outstanding $maxreqs
1175 This is a very bad function to use in interactive programs because
1176 it blocks, and a bad way to reduce concurrency because it is
1177 inexact: Better use an "aio_group" together with a feed callback.
1178
1179 Sets the maximum number of outstanding requests to $nreqs. If you do 1602 Sets the maximum number of outstanding requests to $nreqs. If you do
1180 queue up more than this number of requests, the next call to the 1603 queue up more than this number of requests, the next call to
1181 "poll_cb" (and "poll_some" and other functions calling "poll_cb") 1604 "IO::AIO::poll_cb" (and other functions calling "poll_cb", such as
1182 function will block until the limit is no longer exceeded. 1605 "IO::AIO::flush" or "IO::AIO::poll") will block until the limit is
1606 no longer exceeded.
1183 1607
1184 The default value is very large, so there is no practical limit on 1608 In other words, this setting does not enforce a queue limit, but can
1609 be used to make poll functions block if the limit is exceeded.
1610
1611 This is a bad function to use in interactive programs because it
1612 blocks, and a bad way to reduce concurrency because it is inexact.
1613 If you need to issue many requests without being able to call a poll
1614 function on demand, it is better to use an "aio_group" together with
1615 a feed callback.
1616
1617 Its main use is in scripts without an event loop - when you want to
1618 stat a lot of files, you can write something like this:
1619
1620 IO::AIO::max_outstanding 32;
1621
1622 for my $path (...) {
1623 aio_stat $path , ...;
1624 IO::AIO::poll_cb;
1625 }
1626
1627 IO::AIO::flush;
1628
1629 The call to "poll_cb" inside the loop will normally return
1630 instantly, allowing the loop to progress, but as soon as more than
1631 32 requests are in-flight, it will block until some requests have
1632 been handled. This keeps the loop from pushing a large number of
1633 "aio_stat" requests onto the queue (which, with many paths to stat,
1634 can use up a lot of memory).
1635
1636 The default value for "max_outstanding" is very large, so there is
1185 the number of outstanding requests. 1637 no practical limit on the number of outstanding requests.
1186
1187 You can still queue as many requests as you want. Therefore,
1188 "max_outstanding" is mainly useful in simple scripts (with low
1189 values) or as a stop gap to shield against fatal memory overflow
1190 (with large values).
1191 1638
1192 STATISTICAL INFORMATION 1639 STATISTICAL INFORMATION
1193 IO::AIO::nreqs 1640 IO::AIO::nreqs
1194 Returns the number of requests currently in the ready, execute or 1641 Returns the number of requests currently in the ready, execute or
1195 pending states (i.e. for which their callback has not been invoked 1642 pending states (i.e. for which their callback has not been invoked
… …
1206 1653
1207 IO::AIO::npending 1654 IO::AIO::npending
1208 Returns the number of requests currently in the pending state 1655 Returns the number of requests currently in the pending state
1209 (executed, but not yet processed by poll_cb). 1656 (executed, but not yet processed by poll_cb).
1210 1657
1658 SUBSECOND STAT TIME ACCESS
1659 Both "aio_stat"/"aio_lstat" and perl's "stat"/"lstat" functions can
1660 generally find access/modification and change times with subsecond time
1661 accuracy of the system supports it, but perl's built-in functions only
1662 return the integer part.
1663
1664 The following functions return the timestamps of the most recent stat
1665 with subsecond precision on most systems and work both after
1666 "aio_stat"/"aio_lstat" and perl's "stat"/"lstat" calls. Their return
1667 value is only meaningful after a successful "stat"/"lstat" call, or
1668 during/after a successful "aio_stat"/"aio_lstat" callback.
1669
1670 This is similar to the Time::HiRes "stat" functions, but can return full
1671 resolution without rounding and work with standard perl "stat",
1672 alleviating the need to call the special "Time::HiRes" functions, which
1673 do not act like their perl counterparts.
1674
1675 On operating systems or file systems where subsecond time resolution is
1676 not supported or could not be detected, a fractional part of 0 is
1677 returned, so it is always safe to call these functions.
1678
1679 $seconds = IO::AIO::st_atime, IO::AIO::st_mtime, IO::AIO::st_ctime,
1680 IO::AIO::st_btime
1681 Return the access, modication, change or birth time, respectively,
1682 including fractional part. Due to the limited precision of floating
1683 point, the accuracy on most platforms is only a bit better than
1684 milliseconds for times around now - see the *nsec* function family,
1685 below, for full accuracy.
1686
1687 File birth time is only available when the OS and perl support it
1688 (on FreeBSD and NetBSD at the time of this writing, although support
1689 is adaptive, so if your OS/perl gains support, IO::AIO can take
1690 advantage of it). On systems where it isn't available, 0 is
1691 currently returned, but this might change to "undef" in a future
1692 version.
1693
1694 ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtime
1695 Returns access, modification, change and birth time all in one go,
1696 and maybe more times in the future version.
1697
1698 $nanoseconds = IO::AIO::st_atimensec, IO::AIO::st_mtimensec,
1699 IO::AIO::st_ctimensec, IO::AIO::st_btimensec
1700 Return the fractional access, modifcation, change or birth time, in
1701 nanoseconds, as an integer in the range 0 to 999999999.
1702
1703 Note that no accessors are provided for access, modification and
1704 change times - you need to get those from "stat _" if required ("int
1705 IO::AIO::st_atime" and so on will *not* generally give you the
1706 correct value).
1707
1708 $seconds = IO::AIO::st_btimesec
1709 The (integral) seconds part of the file birth time, if available.
1710
1711 ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtimensec
1712 Like the functions above, but returns all four times in one go (and
1713 maybe more in future versions).
1714
1715 $counter = IO::AIO::st_gen
1716 Returns the generation counter (in practice this is just a random
1717 number) of the file. This is only available on platforms which have
1718 this member in their "struct stat" (most BSDs at the time of this
1719 writing) and generally only to the root usert. If unsupported, 0 is
1720 returned, but this might change to "undef" in a future version.
1721
1722 Example: print the high resolution modification time of /etc, using
1723 "stat", and "IO::AIO::aio_stat".
1724
1725 if (stat "/etc") {
1726 printf "stat(/etc) mtime: %f\n", IO::AIO::st_mtime;
1727 }
1728
1729 IO::AIO::aio_stat "/etc", sub {
1730 $_[0]
1731 and return;
1732
1733 printf "aio_stat(/etc) mtime: %d.%09d\n", (stat _)[9], IO::AIO::st_mtimensec;
1734 };
1735
1736 IO::AIO::flush;
1737
1738 Output of the awbove on my system, showing reduced and full accuracy:
1739
1740 stat(/etc) mtime: 1534043702.020808
1741 aio_stat(/etc) mtime: 1534043702.020807792
1742
1211 MISCELLANEOUS FUNCTIONS 1743 MISCELLANEOUS FUNCTIONS
1212 IO::AIO implements some functions that might be useful, but are not 1744 IO::AIO implements some functions that are useful when you want to use
1213 asynchronous. 1745 some "Advanced I/O" function not available to in Perl, without going the
1746 "Asynchronous I/O" route. Many of these have an asynchronous "aio_*"
1747 counterpart.
1748
1749 $retval = IO::AIO::fexecve $fh, $argv, $envp
1750 A more-or-less direct equivalent to the POSIX "fexecve" functions,
1751 which allows you to specify the program to be executed via a file
1752 descriptor (or handle). Returns -1 and sets errno to "ENOSYS" if not
1753 available.
1754
1755 $retval = IO::AIO::mount $special, $path, $fstype, $flags = 0, $data =
1756 undef
1757 Calls the GNU/Linux mount syscall with the given arguments. All
1758 except $flags are strings, and if $data is "undef", a "NULL" will be
1759 passed.
1760
1761 The following values for $flags are available:
1762
1763 "IO::AIO::MS_RDONLY", "IO::AIO::MS_NOSUID", "IO::AIO::MS_NODEV",
1764 "IO::AIO::MS_NOEXEC", "IO::AIO::MS_SYNCHRONOUS",
1765 "IO::AIO::MS_REMOUNT", "IO::AIO::MS_MANDLOCK",
1766 "IO::AIO::MS_DIRSYNC", "IO::AIO::MS_NOATIME",
1767 "IO::AIO::MS_NODIRATIME", "IO::AIO::MS_BIND", "IO::AIO::MS_MOVE",
1768 "IO::AIO::MS_REC", "IO::AIO::MS_SILENT", "IO::AIO::MS_POSIXACL",
1769 "IO::AIO::MS_UNBINDABLE", "IO::AIO::MS_PRIVATE",
1770 "IO::AIO::MS_SLAVE", "IO::AIO::MS_SHARED", "IO::AIO::MS_RELATIME",
1771 "IO::AIO::MS_KERNMOUNT", "IO::AIO::MS_I_VERSION",
1772 "IO::AIO::MS_STRICTATIME", "IO::AIO::MS_LAZYTIME",
1773 "IO::AIO::MS_ACTIVE", "IO::AIO::MS_NOUSER", "IO::AIO::MS_RMT_MASK",
1774 "IO::AIO::MS_MGC_VAL" and "IO::AIO::MS_MGC_MSK".
1775
1776 $retval = IO::AIO::umount $path, $flags = 0
1777 Invokes the GNU/Linux "umount" or "umount2" syscalls. Always calls
1778 "umount" if $flags is 0, otherwqise always tries to call "umount2".
1779
1780 The following $flags are available:
1781
1782 "IO::AIO::MNT_FORCE", "IO::AIO::MNT_DETACH", "IO::AIO::MNT_EXPIRE"
1783 and "IO::AIO::UMOUNT_NOFOLLOW".
1784
1785 $numfd = IO::AIO::get_fdlimit
1786 Tries to find the current file descriptor limit and returns it, or
1787 "undef" and sets $! in case of an error. The limit is one larger
1788 than the highest valid file descriptor number.
1789
1790 IO::AIO::min_fdlimit [$numfd]
1791 Try to increase the current file descriptor limit(s) to at least
1792 $numfd by changing the soft or hard file descriptor resource limit.
1793 If $numfd is missing, it will try to set a very high limit, although
1794 this is not recommended when you know the actual minimum that you
1795 require.
1796
1797 If the limit cannot be raised enough, the function makes a
1798 best-effort attempt to increase the limit as much as possible, using
1799 various tricks, while still failing. You can query the resulting
1800 limit using "IO::AIO::get_fdlimit".
1801
1802 If an error occurs, returns "undef" and sets $!, otherwise returns
1803 true.
1214 1804
1215 IO::AIO::sendfile $ofh, $ifh, $offset, $count 1805 IO::AIO::sendfile $ofh, $ifh, $offset, $count
1216 Calls the "eio_sendfile_sync" function, which is like 1806 Calls the "eio_sendfile_sync" function, which is like
1217 "aio_sendfile", but is blocking (this makes most sense if you know 1807 "aio_sendfile", but is blocking (this makes most sense if you know
1218 the input data is likely cached already and the output filehandle is 1808 the input data is likely cached already and the output filehandle is
… …
1220 1810
1221 Returns the number of bytes copied, or -1 on error. 1811 Returns the number of bytes copied, or -1 on error.
1222 1812
1223 IO::AIO::fadvise $fh, $offset, $len, $advice 1813 IO::AIO::fadvise $fh, $offset, $len, $advice
1224 Simply calls the "posix_fadvise" function (see its manpage for 1814 Simply calls the "posix_fadvise" function (see its manpage for
1225 details). The following advice constants are avaiable: 1815 details). The following advice constants are available:
1226 "IO::AIO::FADV_NORMAL", "IO::AIO::FADV_SEQUENTIAL", 1816 "IO::AIO::FADV_NORMAL", "IO::AIO::FADV_SEQUENTIAL",
1227 "IO::AIO::FADV_RANDOM", "IO::AIO::FADV_NOREUSE", 1817 "IO::AIO::FADV_RANDOM", "IO::AIO::FADV_NOREUSE",
1228 "IO::AIO::FADV_WILLNEED", "IO::AIO::FADV_DONTNEED". 1818 "IO::AIO::FADV_WILLNEED", "IO::AIO::FADV_DONTNEED".
1229 1819
1230 On systems that do not implement "posix_fadvise", this function 1820 On systems that do not implement "posix_fadvise", this function
1231 returns ENOSYS, otherwise the return value of "posix_fadvise". 1821 returns ENOSYS, otherwise the return value of "posix_fadvise".
1232 1822
1233 IO::AIO::madvise $scalar, $offset, $len, $advice 1823 IO::AIO::madvise $scalar, $offset, $len, $advice
1234 Simply calls the "posix_madvise" function (see its manpage for 1824 Simply calls the "posix_madvise" function (see its manpage for
1235 details). The following advice constants are avaiable: 1825 details). The following advice constants are available:
1236 "IO::AIO::MADV_NORMAL", "IO::AIO::MADV_SEQUENTIAL", 1826 "IO::AIO::MADV_NORMAL", "IO::AIO::MADV_SEQUENTIAL",
1237 "IO::AIO::MADV_RANDOM", "IO::AIO::MADV_WILLNEED", 1827 "IO::AIO::MADV_RANDOM", "IO::AIO::MADV_WILLNEED",
1238 "IO::AIO::MADV_DONTNEED". 1828 "IO::AIO::MADV_DONTNEED".
1239 1829
1830 If $offset is negative, counts from the end. If $length is negative,
1831 the remaining length of the $scalar is used. If possible, $length
1832 will be reduced to fit into the $scalar.
1833
1240 On systems that do not implement "posix_madvise", this function 1834 On systems that do not implement "posix_madvise", this function
1241 returns ENOSYS, otherwise the return value of "posix_madvise". 1835 returns ENOSYS, otherwise the return value of "posix_madvise".
1242 1836
1243 IO::AIO::mprotect $scalar, $offset, $len, $protect 1837 IO::AIO::mprotect $scalar, $offset, $len, $protect
1244 Simply calls the "mprotect" function on the preferably AIO::mmap'ed 1838 Simply calls the "mprotect" function on the preferably AIO::mmap'ed
1245 $scalar (see its manpage for details). The following protect 1839 $scalar (see its manpage for details). The following protect
1246 constants are avaiable: "IO::AIO::PROT_NONE", "IO::AIO::PROT_READ", 1840 constants are available: "IO::AIO::PROT_NONE", "IO::AIO::PROT_READ",
1247 "IO::AIO::PROT_WRITE", "IO::AIO::PROT_EXEC". 1841 "IO::AIO::PROT_WRITE", "IO::AIO::PROT_EXEC".
1842
1843 If $offset is negative, counts from the end. If $length is negative,
1844 the remaining length of the $scalar is used. If possible, $length
1845 will be reduced to fit into the $scalar.
1248 1846
1249 On systems that do not implement "mprotect", this function returns 1847 On systems that do not implement "mprotect", this function returns
1250 ENOSYS, otherwise the return value of "mprotect". 1848 ENOSYS, otherwise the return value of "mprotect".
1251 1849
1252 IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset] 1850 IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
1253 Memory-maps a file (or anonymous memory range) and attaches it to 1851 Memory-maps a file (or anonymous memory range) and attaches it to
1254 the given $scalar, which will act like a string scalar. 1852 the given $scalar, which will act like a string scalar. Returns true
1853 on success, and false otherwise.
1255 1854
1855 The scalar must exist, but its contents do not matter - this means
1856 you cannot use a nonexistant array or hash element. When in doubt,
1857 "undef" the scalar first.
1858
1256 The only operations allowed on the scalar are "substr"/"vec" that 1859 The only operations allowed on the mmapped scalar are
1257 don't change the string length, and most read-only operations such 1860 "substr"/"vec", which don't change the string length, and most
1258 as copying it or searching it with regexes and so on. 1861 read-only operations such as copying it or searching it with regexes
1862 and so on.
1259 1863
1260 Anything else is unsafe and will, at best, result in memory leaks. 1864 Anything else is unsafe and will, at best, result in memory leaks.
1261 1865
1262 The memory map associated with the $scalar is automatically removed 1866 The memory map associated with the $scalar is automatically removed
1263 when the $scalar is destroyed, or when the "IO::AIO::mmap" or 1867 when the $scalar is undef'd or destroyed, or when the
1264 "IO::AIO::munmap" functions are called. 1868 "IO::AIO::mmap" or "IO::AIO::munmap" functions are called on it.
1265 1869
1266 This calls the "mmap"(2) function internally. See your system's 1870 This calls the "mmap"(2) function internally. See your system's
1267 manual page for details on the $length, $prot and $flags parameters. 1871 manual page for details on the $length, $prot and $flags parameters.
1268 1872
1269 The $length must be larger than zero and smaller than the actual 1873 The $length must be larger than zero and smaller than the actual
… …
1273 "IO::AIO::PROT_EXEC", "IO::AIO::PROT_READ" and/or 1877 "IO::AIO::PROT_EXEC", "IO::AIO::PROT_READ" and/or
1274 "IO::AIO::PROT_WRITE", 1878 "IO::AIO::PROT_WRITE",
1275 1879
1276 $flags can be a combination of "IO::AIO::MAP_SHARED" or 1880 $flags can be a combination of "IO::AIO::MAP_SHARED" or
1277 "IO::AIO::MAP_PRIVATE", or a number of system-specific flags (when 1881 "IO::AIO::MAP_PRIVATE", or a number of system-specific flags (when
1278 not available, the are defined as 0): "IO::AIO::MAP_ANONYMOUS" 1882 not available, the are 0): "IO::AIO::MAP_ANONYMOUS" (which is set to
1279 (which is set to "MAP_ANON" if your system only provides this 1883 "MAP_ANON" if your system only provides this constant),
1884 "IO::AIO::MAP_LOCKED", "IO::AIO::MAP_NORESERVE",
1280 constant), "IO::AIO::MAP_HUGETLB", "IO::AIO::MAP_LOCKED", 1885 "IO::AIO::MAP_POPULATE", "IO::AIO::MAP_NONBLOCK",
1281 "IO::AIO::MAP_NORESERVE", "IO::AIO::MAP_POPULATE" or 1886 "IO::AIO::MAP_FIXED", "IO::AIO::MAP_GROWSDOWN",
1282 "IO::AIO::MAP_NONBLOCK" 1887 "IO::AIO::MAP_32BIT", "IO::AIO::MAP_HUGETLB", "IO::AIO::MAP_STACK",
1888 "IO::AIO::MAP_FIXED_NOREPLACE", "IO::AIO::MAP_SHARED_VALIDATE",
1889 "IO::AIO::MAP_SYNC" or "IO::AIO::MAP_UNINITIALIZED".
1283 1890
1284 If $fh is "undef", then a file descriptor of -1 is passed. 1891 If $fh is "undef", then a file descriptor of -1 is passed.
1285 1892
1286 $offset is the offset from the start of the file - it generally must 1893 $offset is the offset from the start of the file - it generally must
1287 be a multiple of "IO::AIO::PAGESIZE" and defaults to 0. 1894 be a multiple of "IO::AIO::PAGESIZE" and defaults to 0.
… …
1299 1906
1300 my $fast_md5 = md5 $data; 1907 my $fast_md5 = md5 $data;
1301 1908
1302 IO::AIO::munmap $scalar 1909 IO::AIO::munmap $scalar
1303 Removes a previous mmap and undefines the $scalar. 1910 Removes a previous mmap and undefines the $scalar.
1911
1912 IO::AIO::mremap $scalar, $new_length, $flags = MREMAP_MAYMOVE[,
1913 $new_address = 0]
1914 Calls the Linux-specific mremap(2) system call. The $scalar must
1915 have been mapped by "IO::AIO::mmap", and $flags must currently
1916 either be 0 or "IO::AIO::MREMAP_MAYMOVE".
1917
1918 Returns true if successful, and false otherwise. If the underlying
1919 mmapped region has changed address, then the true value has the
1920 numerical value 1, otherwise it has the numerical value 0:
1921
1922 my $success = IO::AIO::mremap $mmapped, 8192, IO::AIO::MREMAP_MAYMOVE
1923 or die "mremap: $!";
1924
1925 if ($success*1) {
1926 warn "scalar has chanegd address in memory\n";
1927 }
1928
1929 "IO::AIO::MREMAP_FIXED" and the $new_address argument are currently
1930 implemented, but not supported and might go away in a future
1931 version.
1932
1933 On systems where this call is not supported or is not emulated, this
1934 call returns falls and sets $! to "ENOSYS".
1935
1936 IO::AIO::mlockall $flags
1937 Calls the "eio_mlockall_sync" function, which is like
1938 "aio_mlockall", but is blocking.
1304 1939
1305 IO::AIO::munlock $scalar, $offset = 0, $length = undef 1940 IO::AIO::munlock $scalar, $offset = 0, $length = undef
1306 Calls the "munlock" function, undoing the effects of a previous 1941 Calls the "munlock" function, undoing the effects of a previous
1307 "aio_mlock" call (see its description for details). 1942 "aio_mlock" call (see its description for details).
1308 1943
1309 IO::AIO::munlockall 1944 IO::AIO::munlockall
1310 Calls the "munlockall" function. 1945 Calls the "munlockall" function.
1311 1946
1312 On systems that do not implement "munlockall", this function returns 1947 On systems that do not implement "munlockall", this function returns
1313 ENOSYS, otherwise the return value of "munlockall". 1948 ENOSYS, otherwise the return value of "munlockall".
1949
1950 $fh = IO::AIO::accept4 $r_fh, $sockaddr, $sockaddr_maxlen, $flags
1951 Uses the GNU/Linux accept4(2) syscall, if available, to accept a
1952 socket and return the new file handle on success, or sets $! and
1953 returns "undef" on error.
1954
1955 The remote name of the new socket will be stored in $sockaddr, which
1956 will be extended to allow for at least $sockaddr_maxlen octets. If
1957 the socket name does not fit into $sockaddr_maxlen octets, this is
1958 signaled by returning a longer string in $sockaddr, which might or
1959 might not be truncated.
1960
1961 To accept name-less sockets, use "undef" for $sockaddr and 0 for
1962 $sockaddr_maxlen.
1963
1964 The main reasons to use this syscall rather than portable accept(2)
1965 are that you can specify "SOCK_NONBLOCK" and/or "SOCK_CLOEXEC" flags
1966 and you can accept name-less sockets by specifying 0 for
1967 $sockaddr_maxlen, which is sadly not possible with perl's interface
1968 to "accept".
1969
1970 IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
1971 Calls the GNU/Linux splice(2) syscall, if available. If $r_off or
1972 $w_off are "undef", then "NULL" is passed for these, otherwise they
1973 should be the file offset.
1974
1975 $r_fh and $w_fh should not refer to the same file, as splice might
1976 silently corrupt the data in this case.
1977
1978 The following symbol flag values are available:
1979 "IO::AIO::SPLICE_F_MOVE", "IO::AIO::SPLICE_F_NONBLOCK",
1980 "IO::AIO::SPLICE_F_MORE" and "IO::AIO::SPLICE_F_GIFT".
1981
1982 See the splice(2) manpage for details.
1983
1984 IO::AIO::tee $r_fh, $w_fh, $length, $flags
1985 Calls the GNU/Linux tee(2) syscall, see its manpage and the
1986 description for "IO::AIO::splice" above for details.
1987
1988 $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
1989 Attempts to query or change the pipe buffer size. Obviously works
1990 only on pipes, and currently works only on GNU/Linux systems, and
1991 fails with -1/"ENOSYS" everywhere else. If anybody knows how to
1992 influence pipe buffer size on other systems, drop me a note.
1993
1994 ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
1995 This is a direct interface to the Linux pipe2(2) system call. If
1996 $flags is missing or 0, then this should be the same as a call to
1997 perl's built-in "pipe" function and create a new pipe, and works on
1998 systems that lack the pipe2 syscall. On win32, this case invokes
1999 "_pipe (..., 4096, O_BINARY)".
2000
2001 If $flags is non-zero, it tries to invoke the pipe2 system call with
2002 the given flags (Linux 2.6.27, glibc 2.9).
2003
2004 On success, the read and write file handles are returned.
2005
2006 On error, nothing will be returned. If the pipe2 syscall is missing
2007 and $flags is non-zero, fails with "ENOSYS".
2008
2009 Please refer to pipe2(2) for more info on the $flags, but at the
2010 time of this writing, "IO::AIO::O_CLOEXEC", "IO::AIO::O_NONBLOCK"
2011 and "IO::AIO::O_DIRECT" (Linux 3.4, for packet-based pipes) were
2012 supported.
2013
2014 Example: create a pipe race-free w.r.t. threads and fork:
2015
2016 my ($rfh, $wfh) = IO::AIO::pipe2 IO::AIO::O_CLOEXEC
2017 or die "pipe2: $!\n";
2018
2019 $fh = IO::AIO::memfd_create $pathname[, $flags]
2020 This is a direct interface to the Linux memfd_create(2) system call.
2021 The (unhelpful) default for $flags is 0, but your default should be
2022 "IO::AIO::MFD_CLOEXEC".
2023
2024 On success, the new memfd filehandle is returned, otherwise returns
2025 "undef". If the memfd_create syscall is missing, fails with
2026 "ENOSYS".
2027
2028 Please refer to memfd_create(2) for more info on this call.
2029
2030 The following $flags values are available: "IO::AIO::MFD_CLOEXEC",
2031 "IO::AIO::MFD_ALLOW_SEALING", "IO::AIO::MFD_HUGETLB",
2032 "IO::AIO::MFD_HUGETLB_2MB" and "IO::AIO::MFD_HUGETLB_1GB".
2033
2034 Example: create a new memfd.
2035
2036 my $fh = IO::AIO::memfd_create "somenameforprocfd", IO::AIO::MFD_CLOEXEC
2037 or die "memfd_create: $!\n";
2038
2039 $fh = IO::AIO::pidfd_open $pid[, $flags]
2040 This is an interface to the Linux pidfd_open(2) system call. The
2041 default for $flags is 0.
2042
2043 On success, a new pidfd filehandle is returned (that is already set
2044 to close-on-exec), otherwise returns "undef". If the syscall is
2045 missing, fails with "ENOSYS".
2046
2047 Example: open pid 6341 as pidfd.
2048
2049 my $fh = IO::AIO::pidfd_open 6341
2050 or die "pidfd_open: $!\n";
2051
2052 $status = IO::AIO::pidfd_send_signal $pidfh, $signal[, $siginfo[,
2053 $flags]]
2054 This is an interface to the Linux pidfd_send_signal system call. The
2055 default for $siginfo is "undef" and the default for $flags is 0.
2056
2057 Returns the system call status. If the syscall is missing, fails
2058 with "ENOSYS".
2059
2060 When specified, $siginfo must be a reference to a hash with one or
2061 more of the following members:
2062
2063 code - the "si_code" member
2064 pid - the "si_pid" member
2065 uid - the "si_uid" member
2066 value_int - the "si_value.sival_int" member
2067 value_ptr - the "si_value.sival_ptr" member, specified as an integer
2068
2069 Example: send a SIGKILL to the specified process.
2070
2071 my $status = IO::AIO::pidfd_send_signal $pidfh, 9, undef
2072 and die "pidfd_send_signal: $!\n";
2073
2074 Example: send a SIGKILL to the specified process with extra data.
2075
2076 my $status = IO::AIO::pidfd_send_signal $pidfh, 9, { code => -1, value_int => 7 }
2077 and die "pidfd_send_signal: $!\n";
2078
2079 $fh = IO::AIO::pidfd_getfd $pidfh, $targetfd[, $flags]
2080 This is an interface to the Linux pidfd_getfd system call. The
2081 default for $flags is 0.
2082
2083 On success, returns a dup'ed copy of the target file descriptor
2084 (specified as an integer) returned (that is already set to
2085 close-on-exec), otherwise returns "undef". If the syscall is
2086 missing, fails with "ENOSYS".
2087
2088 Example: get a copy of standard error of another process and print
2089 soemthing to it.
2090
2091 my $errfh = IO::AIO::pidfd_getfd $pidfh, 2
2092 or die "pidfd_getfd: $!\n";
2093 print $errfh "stderr\n";
2094
2095 $fh = IO::AIO::eventfd [$initval, [$flags]]
2096 This is a direct interface to the Linux eventfd(2) system call. The
2097 (unhelpful) defaults for $initval and $flags are 0 for both.
2098
2099 On success, the new eventfd filehandle is returned, otherwise
2100 returns "undef". If the eventfd syscall is missing, fails with
2101 "ENOSYS".
2102
2103 Please refer to eventfd(2) for more info on this call.
2104
2105 The following symbol flag values are available:
2106 "IO::AIO::EFD_CLOEXEC", "IO::AIO::EFD_NONBLOCK" and
2107 "IO::AIO::EFD_SEMAPHORE" (Linux 2.6.30).
2108
2109 Example: create a new eventfd filehandle:
2110
2111 $fh = IO::AIO::eventfd 0, IO::AIO::EFD_CLOEXEC
2112 or die "eventfd: $!\n";
2113
2114 $fh = IO::AIO::timerfd_create $clockid[, $flags]
2115 This is a direct interface to the Linux timerfd_create(2) system
2116 call. The (unhelpful) default for $flags is 0, but your default
2117 should be "IO::AIO::TFD_CLOEXEC".
2118
2119 On success, the new timerfd filehandle is returned, otherwise
2120 returns "undef". If the timerfd_create syscall is missing, fails
2121 with "ENOSYS".
2122
2123 Please refer to timerfd_create(2) for more info on this call.
2124
2125 The following $clockid values are available:
2126 "IO::AIO::CLOCK_REALTIME", "IO::AIO::CLOCK_MONOTONIC"
2127 "IO::AIO::CLOCK_CLOCK_BOOTTIME" (Linux 3.15)
2128 "IO::AIO::CLOCK_CLOCK_REALTIME_ALARM" (Linux 3.11) and
2129 "IO::AIO::CLOCK_CLOCK_BOOTTIME_ALARM" (Linux 3.11).
2130
2131 The following $flags values are available (Linux 2.6.27):
2132 "IO::AIO::TFD_NONBLOCK" and "IO::AIO::TFD_CLOEXEC".
2133
2134 Example: create a new timerfd and set it to one-second repeated
2135 alarms, then wait for two alarms:
2136
2137 my $fh = IO::AIO::timerfd_create IO::AIO::CLOCK_BOOTTIME, IO::AIO::TFD_CLOEXEC
2138 or die "timerfd_create: $!\n";
2139
2140 defined IO::AIO::timerfd_settime $fh, 0, 1, 1
2141 or die "timerfd_settime: $!\n";
2142
2143 for (1..2) {
2144 8 == sysread $fh, my $buf, 8
2145 or die "timerfd read failure\n";
2146
2147 printf "number of expirations (likely 1): %d\n",
2148 unpack "Q", $buf;
2149 }
2150
2151 ($cur_interval, $cur_value) = IO::AIO::timerfd_settime $fh, $flags,
2152 $new_interval, $nbw_value
2153 This is a direct interface to the Linux timerfd_settime(2) system
2154 call. Please refer to its manpage for more info on this call.
2155
2156 The new itimerspec is specified using two (possibly fractional)
2157 second values, $new_interval and $new_value).
2158
2159 On success, the current interval and value are returned (as per
2160 "timerfd_gettime"). On failure, the empty list is returned.
2161
2162 The following $flags values are available:
2163 "IO::AIO::TFD_TIMER_ABSTIME" and "IO::AIO::TFD_TIMER_CANCEL_ON_SET".
2164
2165 See "IO::AIO::timerfd_create" for a full example.
2166
2167 ($cur_interval, $cur_value) = IO::AIO::timerfd_gettime $fh
2168 This is a direct interface to the Linux timerfd_gettime(2) system
2169 call. Please refer to its manpage for more info on this call.
2170
2171 On success, returns the current values of interval and value for the
2172 given timerfd (as potentially fractional second values). On failure,
2173 the empty list is returned.
1314 2174
1315EVENT LOOP INTEGRATION 2175EVENT LOOP INTEGRATION
1316 It is recommended to use AnyEvent::AIO to integrate IO::AIO 2176 It is recommended to use AnyEvent::AIO to integrate IO::AIO
1317 automatically into many event loops: 2177 automatically into many event loops:
1318 2178
… …
1341 # Danga::Socket integration 2201 # Danga::Socket integration
1342 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => 2202 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
1343 \&IO::AIO::poll_cb); 2203 \&IO::AIO::poll_cb);
1344 2204
1345 FORK BEHAVIOUR 2205 FORK BEHAVIOUR
1346 This module should do "the right thing" when the process using it forks: 2206 Usage of pthreads in a program changes the semantics of fork
2207 considerably. Specifically, only async-safe functions can be called
2208 after fork. Perl doesn't know about this, so in general, you cannot call
2209 fork with defined behaviour in perl if pthreads are involved. IO::AIO
2210 uses pthreads, so this applies, but many other extensions and (for
2211 inexplicable reasons) perl itself often is linked against pthreads, so
2212 this limitation applies to quite a lot of perls.
1347 2213
1348 Before the fork, IO::AIO enters a quiescent state where no requests can 2214 This module no longer tries to fight your OS, or POSIX. That means
1349 be added in other threads and no results will be processed. After the 2215 IO::AIO only works in the process that loaded it. Forking is fully
1350 fork the parent simply leaves the quiescent state and continues 2216 supported, but using IO::AIO in the child is not.
1351 request/result processing, while the child frees the request/result
1352 queue (so that the requests started before the fork will only be handled
1353 in the parent). Threads will be started on demand until the limit set in
1354 the parent process has been reached again.
1355 2217
1356 In short: the parent will, after a short pause, continue as if fork had 2218 You might get around by not *using* IO::AIO before (or after) forking.
1357 not been called, while the child will act as if IO::AIO has not been 2219 You could also try to call the IO::AIO::reinit function in the child:
1358 used yet. 2220
2221 IO::AIO::reinit
2222 Abandons all current requests and I/O threads and simply
2223 reinitialises all data structures. This is not an operation
2224 supported by any standards, but happens to work on GNU/Linux and
2225 some newer BSD systems.
2226
2227 The only reasonable use for this function is to call it after
2228 forking, if "IO::AIO" was used in the parent. Calling it while
2229 IO::AIO is active in the process will result in undefined behaviour.
2230 Calling it at any time will also result in any undefined (by POSIX)
2231 behaviour.
2232
2233 LINUX-SPECIFIC CALLS
2234 When a call is documented as "linux-specific" then this means it
2235 originated on GNU/Linux. "IO::AIO" will usually try to autodetect the
2236 availability and compatibility of such calls regardless of the platform
2237 it is compiled on, so platforms such as FreeBSD which often implement
2238 these calls will work. When in doubt, call them and see if they fail wth
2239 "ENOSYS".
1359 2240
1360 MEMORY USAGE 2241 MEMORY USAGE
1361 Per-request usage: 2242 Per-request usage:
1362 2243
1363 Each aio request uses - depending on your architecture - around 100-200 2244 Each aio request uses - depending on your architecture - around 100-200
… …
1374 In the execution phase, some aio requests require more memory for 2255 In the execution phase, some aio requests require more memory for
1375 temporary buffers, and each thread requires a stack and other data 2256 temporary buffers, and each thread requires a stack and other data
1376 structures (usually around 16k-128k, depending on the OS). 2257 structures (usually around 16k-128k, depending on the OS).
1377 2258
1378KNOWN BUGS 2259KNOWN BUGS
1379 Known bugs will be fixed in the next release. 2260 Known bugs will be fixed in the next release :)
2261
2262KNOWN ISSUES
2263 Calls that try to "import" foreign memory areas (such as "IO::AIO::mmap"
2264 or "IO::AIO::aio_slurp") do not work with generic lvalues, such as
2265 non-created hash slots or other scalars I didn't think of. It's best to
2266 avoid such and either use scalar variables or making sure that the
2267 scalar exists (e.g. by storing "undef") and isn't "funny" (e.g. tied).
2268
2269 I am not sure anything can be done about this, so this is considered a
2270 known issue, rather than a bug.
1380 2271
1381SEE ALSO 2272SEE ALSO
1382 AnyEvent::AIO for easy integration into event loops, Coro::AIO for a 2273 AnyEvent::AIO for easy integration into event loops, Coro::AIO for a
1383 more natural syntax. 2274 more natural syntax and IO::FDPass for file descriptor passing.
1384 2275
1385AUTHOR 2276AUTHOR
1386 Marc Lehmann <schmorp@schmorp.de> 2277 Marc Lehmann <schmorp@schmorp.de>
1387 http://home.schmorp.de/ 2278 http://home.schmorp.de/
1388 2279

Diff Legend

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