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

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.183 by root, Sun Sep 12 03:40:05 2010 UTC vs.
Revision 1.267 by root, Tue Aug 9 11:39:17 2016 UTC

1=head1 NAME 1=head1 NAME
2 2
3IO::AIO - Asynchronous Input/Output 3IO::AIO - Asynchronous/Advanced Input/Output
4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use IO::AIO; 7 use IO::AIO;
8 8
58not well-supported or restricted (GNU/Linux doesn't allow them on normal 58not well-supported or restricted (GNU/Linux doesn't allow them on normal
59files currently, for example), and they would only support aio_read and 59files currently, for example), and they would only support aio_read and
60aio_write, so the remaining functionality would have to be implemented 60aio_write, so the remaining functionality would have to be implemented
61using threads anyway. 61using threads anyway.
62 62
63In addition to asynchronous I/O, this module also exports some rather
64arcane interfaces, such as C<madvise> or linux's C<splice> system call,
65which is why the C<A> in C<AIO> can also mean I<advanced>.
66
63Although the module will work in the presence of other (Perl-) threads, 67Although the module will work in the presence of other (Perl-) threads,
64it is currently not reentrant in any way, so use appropriate locking 68it is currently not reentrant in any way, so use appropriate locking
65yourself, always call C<poll_cb> from within the same thread, or never 69yourself, always call C<poll_cb> from within the same thread, or never
66call C<poll_cb> (or other C<aio_> functions) recursively. 70call C<poll_cb> (or other C<aio_> functions) recursively.
67 71
68=head2 EXAMPLE 72=head2 EXAMPLE
69 73
70This is a simple example that uses the EV module and loads 74This is a simple example that uses the EV module and loads
71F</etc/passwd> asynchronously: 75F</etc/passwd> asynchronously:
72 76
73 use Fcntl;
74 use EV; 77 use EV;
75 use IO::AIO; 78 use IO::AIO;
76 79
77 # register the IO::AIO callback with EV 80 # register the IO::AIO callback with EV
78 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; 81 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
95 98
96 # file contents now in $contents 99 # file contents now in $contents
97 print $contents; 100 print $contents;
98 101
99 # exit event loop and program 102 # exit event loop and program
100 EV::unloop; 103 EV::break;
101 }; 104 };
102 }; 105 };
103 106
104 # possibly queue up other requests, or open GUI windows, 107 # possibly queue up other requests, or open GUI windows,
105 # check for sockets etc. etc. 108 # check for sockets etc. etc.
106 109
107 # process events as long as there are some: 110 # process events as long as there are some:
108 EV::loop; 111 EV::run;
109 112
110=head1 REQUEST ANATOMY AND LIFETIME 113=head1 REQUEST ANATOMY AND LIFETIME
111 114
112Every C<aio_*> function creates a request. which is a C data structure not 115Every C<aio_*> function creates a request. which is a C data structure not
113directly visible to Perl. 116directly visible to Perl.
168use common::sense; 171use common::sense;
169 172
170use base 'Exporter'; 173use base 'Exporter';
171 174
172BEGIN { 175BEGIN {
173 our $VERSION = '3.65'; 176 our $VERSION = 4.34;
174 177
175 our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close 178 our @AIO_REQ = qw(aio_sendfile aio_seek aio_read aio_write aio_open aio_close
176 aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx 179 aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
177 aio_scandir aio_symlink aio_readlink aio_sync aio_fsync 180 aio_scandir aio_symlink aio_readlink aio_realpath aio_fcntl aio_ioctl
178 aio_fdatasync aio_sync_file_range aio_pathsync aio_readahead 181 aio_sync aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range
182 aio_pathsync aio_readahead aio_fiemap aio_allocate
179 aio_rename aio_link aio_move aio_copy aio_group 183 aio_rename aio_link aio_move aio_copy aio_group
180 aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown 184 aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
181 aio_chmod aio_utime aio_truncate 185 aio_chmod aio_utime aio_truncate
182 aio_msync aio_mtouch aio_mlock aio_mlockall 186 aio_msync aio_mtouch aio_mlock aio_mlockall
183 aio_statvfs); 187 aio_statvfs
188 aio_wd);
184 189
185 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice)); 190 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
186 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush 191 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
187 min_parallel max_parallel max_idle 192 min_parallel max_parallel max_idle idle_timeout
188 nreqs nready npending nthreads 193 nreqs nready npending nthreads
189 max_poll_time max_poll_reqs 194 max_poll_time max_poll_reqs
190 sendfile fadvise madvise 195 sendfile fadvise madvise
191 mmap munmap munlock munlockall); 196 mmap munmap munlock munlockall);
192 197
200 205
201=head1 FUNCTIONS 206=head1 FUNCTIONS
202 207
203=head2 QUICK OVERVIEW 208=head2 QUICK OVERVIEW
204 209
205This section simply lists the prototypes of the most important functions 210This section simply lists the prototypes most of the functions for
206for quick reference. See the following sections for function-by-function 211quick reference. See the following sections for function-by-function
207documentation. 212documentation.
208 213
214 aio_wd $pathname, $callback->($wd)
209 aio_open $pathname, $flags, $mode, $callback->($fh) 215 aio_open $pathname, $flags, $mode, $callback->($fh)
210 aio_close $fh, $callback->($status) 216 aio_close $fh, $callback->($status)
217 aio_seek $fh,$offset,$whence, $callback->($offs)
211 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 218 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
212 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 219 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
213 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) 220 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
214 aio_readahead $fh,$offset,$length, $callback->($retval) 221 aio_readahead $fh,$offset,$length, $callback->($retval)
215 aio_stat $fh_or_path, $callback->($status) 222 aio_stat $fh_or_path, $callback->($status)
216 aio_lstat $fh, $callback->($status) 223 aio_lstat $fh, $callback->($status)
217 aio_statvfs $fh_or_path, $callback->($statvfs) 224 aio_statvfs $fh_or_path, $callback->($statvfs)
218 aio_utime $fh_or_path, $atime, $mtime, $callback->($status) 225 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
219 aio_chown $fh_or_path, $uid, $gid, $callback->($status) 226 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
227 aio_chmod $fh_or_path, $mode, $callback->($status)
220 aio_truncate $fh_or_path, $offset, $callback->($status) 228 aio_truncate $fh_or_path, $offset, $callback->($status)
221 aio_chmod $fh_or_path, $mode, $callback->($status) 229 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
230 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
222 aio_unlink $pathname, $callback->($status) 231 aio_unlink $pathname, $callback->($status)
223 aio_mknod $path, $mode, $dev, $callback->($status) 232 aio_mknod $pathname, $mode, $dev, $callback->($status)
224 aio_link $srcpath, $dstpath, $callback->($status) 233 aio_link $srcpath, $dstpath, $callback->($status)
225 aio_symlink $srcpath, $dstpath, $callback->($status) 234 aio_symlink $srcpath, $dstpath, $callback->($status)
226 aio_readlink $path, $callback->($link) 235 aio_readlink $pathname, $callback->($link)
236 aio_realpath $pathname, $callback->($path)
227 aio_rename $srcpath, $dstpath, $callback->($status) 237 aio_rename $srcpath, $dstpath, $callback->($status)
228 aio_mkdir $pathname, $mode, $callback->($status) 238 aio_mkdir $pathname, $mode, $callback->($status)
229 aio_rmdir $pathname, $callback->($status) 239 aio_rmdir $pathname, $callback->($status)
230 aio_readdir $pathname, $callback->($entries) 240 aio_readdir $pathname, $callback->($entries)
231 aio_readdirx $pathname, $flags, $callback->($entries, $flags) 241 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
232 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST 242 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
233 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN 243 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
244 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
234 aio_load $path, $data, $callback->($status) 245 aio_load $pathname, $data, $callback->($status)
235 aio_copy $srcpath, $dstpath, $callback->($status) 246 aio_copy $srcpath, $dstpath, $callback->($status)
236 aio_move $srcpath, $dstpath, $callback->($status) 247 aio_move $srcpath, $dstpath, $callback->($status)
237 aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
238 aio_rmtree $path, $callback->($status) 248 aio_rmtree $pathname, $callback->($status)
249 aio_fcntl $fh, $cmd, $arg, $callback->($status)
250 aio_ioctl $fh, $request, $buf, $callback->($status)
239 aio_sync $callback->($status) 251 aio_sync $callback->($status)
252 aio_syncfs $fh, $callback->($status)
240 aio_fsync $fh, $callback->($status) 253 aio_fsync $fh, $callback->($status)
241 aio_fdatasync $fh, $callback->($status) 254 aio_fdatasync $fh, $callback->($status)
242 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) 255 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
243 aio_pathsync $path, $callback->($status) 256 aio_pathsync $pathname, $callback->($status)
244 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) 257 aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
245 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status) 258 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
246 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) 259 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
247 aio_mlockall $flags, $callback->($status) 260 aio_mlockall $flags, $callback->($status)
248 aio_group $callback->(...) 261 aio_group $callback->(...)
258 IO::AIO::max_poll_reqs $nreqs 271 IO::AIO::max_poll_reqs $nreqs
259 IO::AIO::max_poll_time $seconds 272 IO::AIO::max_poll_time $seconds
260 IO::AIO::min_parallel $nthreads 273 IO::AIO::min_parallel $nthreads
261 IO::AIO::max_parallel $nthreads 274 IO::AIO::max_parallel $nthreads
262 IO::AIO::max_idle $nthreads 275 IO::AIO::max_idle $nthreads
276 IO::AIO::idle_timeout $seconds
263 IO::AIO::max_outstanding $maxreqs 277 IO::AIO::max_outstanding $maxreqs
264 IO::AIO::nreqs 278 IO::AIO::nreqs
265 IO::AIO::nready 279 IO::AIO::nready
266 IO::AIO::npending 280 IO::AIO::npending
267 281
268 IO::AIO::sendfile $ofh, $ifh, $offset, $count 282 IO::AIO::sendfile $ofh, $ifh, $offset, $count
269 IO::AIO::fadvise $fh, $offset, $len, $advice 283 IO::AIO::fadvise $fh, $offset, $len, $advice
284 IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
285 IO::AIO::munmap $scalar
286 IO::AIO::madvise $scalar, $offset, $length, $advice
287 IO::AIO::mprotect $scalar, $offset, $length, $protect
270 IO::AIO::munlock $scalar, $offset = 0, $length = undef 288 IO::AIO::munlock $scalar, $offset = 0, $length = undef
271 IO::AIO::munlockall 289 IO::AIO::munlockall
272 290
273=head2 AIO REQUEST FUNCTIONS 291=head2 API NOTES
274 292
275All the C<aio_*> calls are more or less thin wrappers around the syscall 293All the C<aio_*> calls are more or less thin wrappers around the syscall
276with the same name (sans C<aio_>). The arguments are similar or identical, 294with the same name (sans C<aio_>). The arguments are similar or identical,
277and they all accept an additional (and optional) C<$callback> argument 295and they all accept an additional (and optional) C<$callback> argument
278which must be a code reference. This code reference will get called with 296which must be a code reference. This code reference will be called after
279the syscall return code (e.g. most syscalls return C<-1> on error, unlike 297the syscall has been executed in an asynchronous fashion. The results
280perl, which usually delivers "false") as its sole argument after the given 298of the request will be passed as arguments to the callback (and, if an
281syscall has been executed asynchronously. 299error occured, in C<$!>) - for most requests the syscall return code (e.g.
300most syscalls return C<-1> on error, unlike perl, which usually delivers
301"false").
302
303Some requests (such as C<aio_readdir>) pass the actual results and
304communicate failures by passing C<undef>.
282 305
283All functions expecting a filehandle keep a copy of the filehandle 306All functions expecting a filehandle keep a copy of the filehandle
284internally until the request has finished. 307internally until the request has finished.
285 308
286All functions return request objects of type L<IO::AIO::REQ> that allow 309All functions return request objects of type L<IO::AIO::REQ> that allow
287further manipulation of those requests while they are in-flight. 310further manipulation of those requests while they are in-flight.
288 311
289The pathnames you pass to these routines I<must> be absolute and 312The pathnames you pass to these routines I<should> be absolute. The
290encoded as octets. The reason for the former is that at the time the 313reason for this is that at the time the request is being executed, the
291request is being executed, the current working directory could have 314current working directory could have changed. Alternatively, you can
292changed. Alternatively, you can make sure that you never change the 315make sure that you never change the current working directory anywhere
293current working directory anywhere in the program and then use relative 316in the program and then use relative paths. You can also take advantage
294paths. 317of IO::AIOs working directory abstraction, that lets you specify paths
318relative to some previously-opened "working directory object" - see the
319description of the C<IO::AIO::WD> class later in this document.
295 320
296To encode pathnames as octets, either make sure you either: a) always pass 321To encode pathnames as octets, either make sure you either: a) always pass
297in filenames you got from outside (command line, readdir etc.) without 322in filenames you got from outside (command line, readdir etc.) without
298tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode 323tinkering, b) are in your native filesystem encoding, c) use the Encode
299your pathnames to the locale (or other) encoding in effect in the user 324module and encode your pathnames to the locale (or other) encoding in
300environment, d) use Glib::filename_from_unicode on unicode filenames or e) 325effect in the user environment, d) use Glib::filename_from_unicode on
301use something else to ensure your scalar has the correct contents. 326unicode filenames or e) use something else to ensure your scalar has the
327correct contents.
302 328
303This works, btw. independent of the internal UTF-8 bit, which IO::AIO 329This works, btw. independent of the internal UTF-8 bit, which IO::AIO
304handles correctly whether it is set or not. 330handles correctly whether it is set or not.
331
332=head2 AIO REQUEST FUNCTIONS
305 333
306=over 4 334=over 4
307 335
308=item $prev_pri = aioreq_pri [$pri] 336=item $prev_pri = aioreq_pri [$pri]
309 337
339 367
340 368
341=item aio_open $pathname, $flags, $mode, $callback->($fh) 369=item aio_open $pathname, $flags, $mode, $callback->($fh)
342 370
343Asynchronously open or create a file and call the callback with a newly 371Asynchronously open or create a file and call the callback with a newly
344created filehandle for the file. 372created filehandle for the file (or C<undef> in case of an error).
345 373
346The pathname passed to C<aio_open> must be absolute. See API NOTES, above, 374The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
347for an explanation. 375for an explanation.
348 376
349The C<$flags> argument is a bitmask. See the C<Fcntl> module for a 377The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
365 } else { 393 } else {
366 die "open failed: $!\n"; 394 die "open failed: $!\n";
367 } 395 }
368 }; 396 };
369 397
398In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
399C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
400following POSIX and non-POSIX constants are available (missing ones on
401your system are, as usual, C<0>):
402
403C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
404C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
405C<O_RSYNC>, C<O_SYNC>, C<O_PATH>, C<O_TMPFILE>, and C<O_TTY_INIT>.
406
370 407
371=item aio_close $fh, $callback->($status) 408=item aio_close $fh, $callback->($status)
372 409
373Asynchronously close a file and call the callback with the result 410Asynchronously close a file and call the callback with the result
374code. 411code.
383Or in other words: the file descriptor will be closed, but it will not be 420Or in other words: the file descriptor will be closed, but it will not be
384free for reuse until the perl filehandle is closed. 421free for reuse until the perl filehandle is closed.
385 422
386=cut 423=cut
387 424
425=item aio_seek $fh, $offset, $whence, $callback->($offs)
426
427Seeks the filehandle to the new C<$offset>, similarly to perl's
428C<sysseek>. The C<$whence> can use the traditional values (C<0> for
429C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
430C<IO::AIO::SEEK_END>).
431
432The resulting absolute offset will be passed to the callback, or C<-1> in
433case of an error.
434
435In theory, the C<$whence> constants could be different than the
436corresponding values from L<Fcntl>, but perl guarantees they are the same,
437so don't panic.
438
439As a GNU/Linux (and maybe Solaris) extension, also the constants
440C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
441could be found. No guarantees about suitability for use in C<aio_seek> or
442Perl's C<sysseek> can be made though, although I would naively assume they
443"just work".
444
388=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 445=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
389 446
390=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 447=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
391 448
392Reads or writes C<$length> bytes from or to the specified C<$fh> and 449Reads or writes C<$length> bytes from or to the specified C<$fh> and
393C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> 450C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> and
394and calls the callback without the actual number of bytes read (or -1 on 451calls the callback with the actual number of bytes transferred (or -1 on
395error, just like the syscall). 452error, just like the syscall).
396 453
397C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to 454C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
398offset plus the actual number of bytes read. 455offset plus the actual number of bytes read.
399 456
424 481
425Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts 482Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
426reading at byte offset C<$in_offset>, and starts writing at the current 483reading at byte offset C<$in_offset>, and starts writing at the current
427file offset of C<$out_fh>. Because of that, it is not safe to issue more 484file offset of C<$out_fh>. Because of that, it is not safe to issue more
428than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each 485than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
429other. 486other. The same C<$in_fh> works fine though, as this function does not
487move or use the file offset of C<$in_fh>.
430 488
489Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
490are written, and there is no way to find out how many more bytes have been
491read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
492number of bytes written to C<$out_fh>. Only if the result value equals
493C<$length> one can assume that C<$length> bytes have been read.
494
495Unlike with other C<aio_> functions, it makes a lot of sense to use
496C<aio_sendfile> on non-blocking sockets, as long as one end (typically
497the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
498the socket I/O will be non-blocking. Note, however, that you can run
499into a trap where C<aio_sendfile> reads some data with readahead, then
500fails to write all data, and when the socket is ready the next time, the
501data in the cache is already lost, forcing C<aio_sendfile> to again hit
502the disk. Explicit C<aio_read> + C<aio_write> let's you better control
503resource usage.
504
431This call tries to make use of a native C<sendfile> syscall to provide 505This call tries to make use of a native C<sendfile>-like syscall to
432zero-copy operation. For this to work, C<$out_fh> should refer to a 506provide zero-copy operation. For this to work, C<$out_fh> should refer to
433socket, and C<$in_fh> should refer to an mmap'able file. 507a socket, and C<$in_fh> should refer to an mmap'able file.
434 508
435If a native sendfile cannot be found or it fails with C<ENOSYS>, 509If a native sendfile cannot be found or it fails with C<ENOSYS>,
436C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or C<ENOTSOCK>, 510C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
437it will be emulated, so you can call C<aio_sendfile> on any type of 511C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
438filehandle regardless of the limitations of the operating system. 512type of filehandle regardless of the limitations of the operating system.
439 513
440Please note, however, that C<aio_sendfile> can read more bytes from 514As native sendfile syscalls (as practically any non-POSIX interface hacked
441C<$in_fh> than are written, and there is no way to find out how many 515together in a hurry to improve benchmark numbers) tend to be rather buggy
442bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only 516on many systems, this implementation tries to work around some known bugs
443provides the number of bytes written to C<$out_fh>. Only if the result 517in Linux and FreeBSD kernels (probably others, too), but that might fail,
444value equals C<$length> one can assume that C<$length> bytes have been 518so you really really should check the return value of C<aio_sendfile> -
445read. 519fewer bytes than expected might have been transferred.
446 520
447 521
448=item aio_readahead $fh,$offset,$length, $callback->($retval) 522=item aio_readahead $fh,$offset,$length, $callback->($retval)
449 523
450C<aio_readahead> populates the page cache with data from a file so that 524C<aio_readahead> populates the page cache with data from a file so that
454whole pages, so that offset is effectively rounded down to a page boundary 528whole pages, so that offset is effectively rounded down to a page boundary
455and bytes are read up to the next page boundary greater than or equal to 529and bytes are read up to the next page boundary greater than or equal to
456(off-set+length). C<aio_readahead> does not read beyond the end of the 530(off-set+length). C<aio_readahead> does not read beyond the end of the
457file. The current file offset of the file is left unchanged. 531file. The current file offset of the file is left unchanged.
458 532
459If that syscall doesn't exist (likely if your OS isn't Linux) it will be 533If that syscall doesn't exist (likely if your kernel isn't Linux) it will
460emulated by simply reading the data, which would have a similar effect. 534be emulated by simply reading the data, which would have a similar effect.
461 535
462 536
463=item aio_stat $fh_or_path, $callback->($status) 537=item aio_stat $fh_or_path, $callback->($status)
464 538
465=item aio_lstat $fh, $callback->($status) 539=item aio_lstat $fh, $callback->($status)
472for an explanation. 546for an explanation.
473 547
474Currently, the stats are always 64-bit-stats, i.e. instead of returning an 548Currently, the stats are always 64-bit-stats, i.e. instead of returning an
475error when stat'ing a large file, the results will be silently truncated 549error when stat'ing a large file, the results will be silently truncated
476unless perl itself is compiled with large file support. 550unless perl itself is compiled with large file support.
551
552To help interpret the mode and dev/rdev stat values, IO::AIO offers the
553following constants and functions (if not implemented, the constants will
554be C<0> and the functions will either C<croak> or fall back on traditional
555behaviour).
556
557C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
558C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
559C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
477 560
478Example: Print the length of F</etc/passwd>: 561Example: Print the length of F</etc/passwd>:
479 562
480 aio_stat "/etc/passwd", sub { 563 aio_stat "/etc/passwd", sub {
481 $_[0] and die "stat failed: $!"; 564 $_[0] and die "stat failed: $!";
525 namemax => 255, 608 namemax => 255,
526 frsize => 1024, 609 frsize => 1024,
527 fsid => 1810 610 fsid => 1810
528 } 611 }
529 612
613Here is a (likely partial - send me updates!) list of fsid values used by
614Linux - it is safe to hardcode these when C<$^O> is C<linux>:
615
616 0x0000adf5 adfs
617 0x0000adff affs
618 0x5346414f afs
619 0x09041934 anon-inode filesystem
620 0x00000187 autofs
621 0x42465331 befs
622 0x1badface bfs
623 0x42494e4d binfmt_misc
624 0x9123683e btrfs
625 0x0027e0eb cgroupfs
626 0xff534d42 cifs
627 0x73757245 coda
628 0x012ff7b7 coh
629 0x28cd3d45 cramfs
630 0x453dcd28 cramfs-wend (wrong endianness)
631 0x64626720 debugfs
632 0x00001373 devfs
633 0x00001cd1 devpts
634 0x0000f15f ecryptfs
635 0x00414a53 efs
636 0x0000137d ext
637 0x0000ef53 ext2/ext3/ext4
638 0x0000ef51 ext2
639 0xf2f52010 f2fs
640 0x00004006 fat
641 0x65735546 fuseblk
642 0x65735543 fusectl
643 0x0bad1dea futexfs
644 0x01161970 gfs2
645 0x47504653 gpfs
646 0x00004244 hfs
647 0xf995e849 hpfs
648 0x00c0ffee hostfs
649 0x958458f6 hugetlbfs
650 0x2bad1dea inotifyfs
651 0x00009660 isofs
652 0x000072b6 jffs2
653 0x3153464a jfs
654 0x6b414653 k-afs
655 0x0bd00bd0 lustre
656 0x0000137f minix
657 0x0000138f minix 30 char names
658 0x00002468 minix v2
659 0x00002478 minix v2 30 char names
660 0x00004d5a minix v3
661 0x19800202 mqueue
662 0x00004d44 msdos
663 0x0000564c novell
664 0x00006969 nfs
665 0x6e667364 nfsd
666 0x00003434 nilfs
667 0x5346544e ntfs
668 0x00009fa1 openprom
669 0x7461636F ocfs2
670 0x00009fa0 proc
671 0x6165676c pstorefs
672 0x0000002f qnx4
673 0x68191122 qnx6
674 0x858458f6 ramfs
675 0x52654973 reiserfs
676 0x00007275 romfs
677 0x67596969 rpc_pipefs
678 0x73636673 securityfs
679 0xf97cff8c selinux
680 0x0000517b smb
681 0x534f434b sockfs
682 0x73717368 squashfs
683 0x62656572 sysfs
684 0x012ff7b6 sysv2
685 0x012ff7b5 sysv4
686 0x01021994 tmpfs
687 0x15013346 udf
688 0x00011954 ufs
689 0x54190100 ufs byteswapped
690 0x00009fa2 usbdevfs
691 0x01021997 v9fs
692 0xa501fcf5 vxfs
693 0xabba1974 xenfs
694 0x012ff7b4 xenix
695 0x58465342 xfs
696 0x012fd16d xia
530 697
531=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status) 698=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
532 699
533Works like perl's C<utime> function (including the special case of $atime 700Works like perl's C<utime> function (including the special case of $atime
534and $mtime being undef). Fractional times are supported if the underlying 701and $mtime being undef). Fractional times are supported if the underlying
562=item aio_truncate $fh_or_path, $offset, $callback->($status) 729=item aio_truncate $fh_or_path, $offset, $callback->($status)
563 730
564Works like truncate(2) or ftruncate(2). 731Works like truncate(2) or ftruncate(2).
565 732
566 733
734=item aio_allocate $fh, $mode, $offset, $len, $callback->($status)
735
736Allocates or frees disk space according to the C<$mode> argument. See the
737linux C<fallocate> documentation for details.
738
739C<$mode> is usually C<0> or C<IO::AIO::FALLOC_FL_KEEP_SIZE> to allocate
740space, or C<IO::AIO::FALLOC_FL_PUNCH_HOLE | IO::AIO::FALLOC_FL_KEEP_SIZE>,
741to deallocate a file range.
742
743IO::AIO also supports C<FALLOC_FL_COLLAPSE_RANGE>, to remove a range
744(without leaving a hole) and C<FALLOC_FL_ZERO_RANGE>, to zero a range (see
745your L<fallocate(2)> manpage).
746
747The file system block size used by C<fallocate> is presumably the
748C<f_bsize> returned by C<statvfs>.
749
750If C<fallocate> isn't available or cannot be emulated (currently no
751emulation will be attempted), passes C<-1> and sets C<$!> to C<ENOSYS>.
752
753
567=item aio_chmod $fh_or_path, $mode, $callback->($status) 754=item aio_chmod $fh_or_path, $mode, $callback->($status)
568 755
569Works like perl's C<chmod> function. 756Works like perl's C<chmod> function.
570 757
571 758
573 760
574Asynchronously unlink (delete) a file and call the callback with the 761Asynchronously unlink (delete) a file and call the callback with the
575result code. 762result code.
576 763
577 764
578=item aio_mknod $path, $mode, $dev, $callback->($status) 765=item aio_mknod $pathname, $mode, $dev, $callback->($status)
579 766
580[EXPERIMENTAL] 767[EXPERIMENTAL]
581 768
582Asynchronously create a device node (or fifo). See mknod(2). 769Asynchronously create a device node (or fifo). See mknod(2).
583 770
584The only (POSIX-) portable way of calling this function is: 771The only (POSIX-) portable way of calling this function is:
585 772
586 aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ... 773 aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
587 774
775See C<aio_stat> for info about some potentially helpful extra constants
776and functions.
588 777
589=item aio_link $srcpath, $dstpath, $callback->($status) 778=item aio_link $srcpath, $dstpath, $callback->($status)
590 779
591Asynchronously create a new link to the existing object at C<$srcpath> at 780Asynchronously create a new link to the existing object at C<$srcpath> at
592the path C<$dstpath> and call the callback with the result code. 781the path C<$dstpath> and call the callback with the result code.
596 785
597Asynchronously create a new symbolic link to the existing object at C<$srcpath> at 786Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
598the path C<$dstpath> and call the callback with the result code. 787the path C<$dstpath> and call the callback with the result code.
599 788
600 789
601=item aio_readlink $path, $callback->($link) 790=item aio_readlink $pathname, $callback->($link)
602 791
603Asynchronously read the symlink specified by C<$path> and pass it to 792Asynchronously read the symlink specified by C<$path> and pass it to
604the callback. If an error occurs, nothing or undef gets passed to the 793the callback. If an error occurs, nothing or undef gets passed to the
605callback. 794callback.
606 795
607 796
797=item aio_realpath $pathname, $callback->($path)
798
799Asynchronously make the path absolute and resolve any symlinks in
800C<$path>. The resulting path only consists of directories (same as
801L<Cwd::realpath>).
802
803This request can be used to get the absolute path of the current working
804directory by passing it a path of F<.> (a single dot).
805
806
608=item aio_rename $srcpath, $dstpath, $callback->($status) 807=item aio_rename $srcpath, $dstpath, $callback->($status)
609 808
610Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as 809Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
611rename(2) and call the callback with the result code. 810rename(2) and call the callback with the result code.
811
812On systems that support the AIO::WD working directory abstraction
813natively, the case C<[$wd, "."]> as C<$srcpath> is specialcased - instead
814of failing, C<rename> is called on the absolute path of C<$wd>.
612 815
613 816
614=item aio_mkdir $pathname, $mode, $callback->($status) 817=item aio_mkdir $pathname, $mode, $callback->($status)
615 818
616Asynchronously mkdir (create) a directory and call the callback with 819Asynchronously mkdir (create) a directory and call the callback with
621=item aio_rmdir $pathname, $callback->($status) 824=item aio_rmdir $pathname, $callback->($status)
622 825
623Asynchronously rmdir (delete) a directory and call the callback with the 826Asynchronously rmdir (delete) a directory and call the callback with the
624result code. 827result code.
625 828
829On systems that support the AIO::WD working directory abstraction
830natively, the case C<[$wd, "."]> is specialcased - instead of failing,
831C<rmdir> is called on the absolute path of C<$wd>.
832
626 833
627=item aio_readdir $pathname, $callback->($entries) 834=item aio_readdir $pathname, $callback->($entries)
628 835
629Unlike the POSIX call of the same name, C<aio_readdir> reads an entire 836Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
630directory (i.e. opendir + readdir + closedir). The entries will not be 837directory (i.e. opendir + readdir + closedir). The entries will not be
634array-ref with the filenames. 841array-ref with the filenames.
635 842
636 843
637=item aio_readdirx $pathname, $flags, $callback->($entries, $flags) 844=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
638 845
639Quite similar to C<aio_readdir>, but the C<$flags> argument allows to tune 846Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
640behaviour and output format. In case of an error, C<$entries> will be 847tune behaviour and output format. In case of an error, C<$entries> will be
641C<undef>. 848C<undef>.
642 849
643The flags are a combination of the following constants, ORed together (the 850The flags are a combination of the following constants, ORed together (the
644flags will also be passed to the callback, possibly modified): 851flags will also be passed to the callback, possibly modified):
645 852
646=over 4 853=over 4
647 854
648=item IO::AIO::READDIR_DENTS 855=item IO::AIO::READDIR_DENTS
649 856
650When this flag is off, then the callback gets an arrayref with of names 857When this flag is off, then the callback gets an arrayref consisting of
651only (as with C<aio_readdir>), otherwise it gets an arrayref with 858names only (as with C<aio_readdir>), otherwise it gets an arrayref with
652C<[$name, $type, $inode]> arrayrefs, each describing a single directory 859C<[$name, $type, $inode]> arrayrefs, each describing a single directory
653entry in more detail. 860entry in more detail.
654 861
655C<$name> is the name of the entry. 862C<$name> is the name of the entry.
656 863
669systems that do not deliver the inode information. 876systems that do not deliver the inode information.
670 877
671=item IO::AIO::READDIR_DIRS_FIRST 878=item IO::AIO::READDIR_DIRS_FIRST
672 879
673When this flag is set, then the names will be returned in an order where 880When this flag is set, then the names will be returned in an order where
674likely directories come first. This is useful when you need to quickly 881likely directories come first, in optimal stat order. This is useful when
675find directories, or you want to find all directories while avoiding to 882you need to quickly find directories, or you want to find all directories
676stat() each entry. 883while avoiding to stat() each entry.
677 884
678If the system returns type information in readdir, then this is used 885If the system returns type information in readdir, then this is used
679to find directories directly. Otherwise, likely directories are files 886to find directories directly. Otherwise, likely directories are names
680beginning with ".", or otherwise files with no dots, of which files with 887beginning with ".", or otherwise names with no dots, of which names with
681short names are tried first. 888short names are tried first.
682 889
683=item IO::AIO::READDIR_STAT_ORDER 890=item IO::AIO::READDIR_STAT_ORDER
684 891
685When this flag is set, then the names will be returned in an order 892When this flag is set, then the names will be returned in an order
692 899
693=item IO::AIO::READDIR_FOUND_UNKNOWN 900=item IO::AIO::READDIR_FOUND_UNKNOWN
694 901
695This flag should not be set when calling C<aio_readdirx>. Instead, it 902This flag should not be set when calling C<aio_readdirx>. Instead, it
696is being set by C<aio_readdirx>, when any of the C<$type>'s found were 903is being set by C<aio_readdirx>, when any of the C<$type>'s found were
697C<IO::AIO::DT_UNKNOWN>. The absense of this flag therefore indicates that all 904C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
698C<$type>'s are known, which can be used to speed up some algorithms. 905C<$type>'s are known, which can be used to speed up some algorithms.
699 906
700=back 907=back
701 908
702 909
703=item aio_load $path, $data, $callback->($status) 910=item aio_load $pathname, $data, $callback->($status)
704 911
705This is a composite request that tries to fully load the given file into 912This is a composite request that tries to fully load the given file into
706memory. Status is the same as with aio_read. 913memory. Status is the same as with aio_read.
707 914
708=cut 915=cut
830 if ($_[0] && $! == EXDEV) { 1037 if ($_[0] && $! == EXDEV) {
831 aioreq_pri $pri; 1038 aioreq_pri $pri;
832 add $grp aio_copy $src, $dst, sub { 1039 add $grp aio_copy $src, $dst, sub {
833 $grp->result ($_[0]); 1040 $grp->result ($_[0]);
834 1041
835 if (!$_[0]) { 1042 unless ($_[0]) {
836 aioreq_pri $pri; 1043 aioreq_pri $pri;
837 add $grp aio_unlink $src; 1044 add $grp aio_unlink $src;
838 } 1045 }
839 }; 1046 };
840 } else { 1047 } else {
843 }; 1050 };
844 1051
845 $grp 1052 $grp
846} 1053}
847 1054
848=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 1055=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
849 1056
850Scans a directory (similar to C<aio_readdir>) but additionally tries to 1057Scans a directory (similar to C<aio_readdir>) but additionally tries to
851efficiently separate the entries of directory C<$path> into two sets of 1058efficiently separate the entries of directory C<$path> into two sets of
852names, directories you can recurse into (directories), and ones you cannot 1059names, directories you can recurse into (directories), and ones you cannot
853recurse into (everything else, including symlinks to directories). 1060recurse into (everything else, including symlinks to directories).
884Then entries will be sorted into likely directories a non-initial dot 1091Then entries will be sorted into likely directories a non-initial dot
885currently) and likely non-directories (see C<aio_readdirx>). Then every 1092currently) and likely non-directories (see C<aio_readdirx>). Then every
886entry plus an appended C</.> will be C<stat>'ed, likely directories first, 1093entry plus an appended C</.> will be C<stat>'ed, likely directories first,
887in order of their inode numbers. If that succeeds, it assumes that the 1094in order of their inode numbers. If that succeeds, it assumes that the
888entry is a directory or a symlink to directory (which will be checked 1095entry is a directory or a symlink to directory (which will be checked
889seperately). This is often faster than stat'ing the entry itself because 1096separately). This is often faster than stat'ing the entry itself because
890filesystems might detect the type of the entry without reading the inode 1097filesystems might detect the type of the entry without reading the inode
891data (e.g. ext2fs filetype feature), even on systems that cannot return 1098data (e.g. ext2fs filetype feature), even on systems that cannot return
892the filetype information on readdir. 1099the filetype information on readdir.
893 1100
894If the known number of directories (link count - 2) has been reached, the 1101If the known number of directories (link count - 2) has been reached, the
910 1117
911 my $grp = aio_group $cb; 1118 my $grp = aio_group $cb;
912 1119
913 $maxreq = 4 if $maxreq <= 0; 1120 $maxreq = 4 if $maxreq <= 0;
914 1121
915 # stat once 1122 # get a wd object
916 aioreq_pri $pri; 1123 aioreq_pri $pri;
917 add $grp aio_stat $path, sub { 1124 add $grp aio_wd $path, sub {
1125 $_[0]
918 return $grp->result () if $_[0]; 1126 or return $grp->result ();
919 my $now = time;
920 my $hash1 = join ":", (stat _)[0,1,3,7,9];
921 1127
922 # read the directory entries 1128 my $wd = [shift, "."];
1129
1130 # stat once
923 aioreq_pri $pri; 1131 aioreq_pri $pri;
924 add $grp aio_readdirx $path, READDIR_DIRS_FIRST, sub { 1132 add $grp aio_stat $wd, sub {
925 my $entries = shift
926 or return $grp->result (); 1133 return $grp->result () if $_[0];
1134 my $now = time;
1135 my $hash1 = join ":", (stat _)[0,1,3,7,9];
927 1136
928 # stat the dir another time 1137 # read the directory entries
929 aioreq_pri $pri; 1138 aioreq_pri $pri;
1139 add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
1140 my $entries = shift
1141 or return $grp->result ();
1142
1143 # stat the dir another time
1144 aioreq_pri $pri;
930 add $grp aio_stat $path, sub { 1145 add $grp aio_stat $wd, sub {
931 my $hash2 = join ":", (stat _)[0,1,3,7,9]; 1146 my $hash2 = join ":", (stat _)[0,1,3,7,9];
932 1147
933 my $ndirs; 1148 my $ndirs;
934 1149
935 # take the slow route if anything looks fishy 1150 # take the slow route if anything looks fishy
936 if ($hash1 ne $hash2 or (stat _)[9] == $now) { 1151 if ($hash1 ne $hash2 or (stat _)[9] == $now) {
937 $ndirs = -1; 1152 $ndirs = -1;
938 } else { 1153 } else {
939 # if nlink == 2, we are finished 1154 # if nlink == 2, we are finished
940 # for non-posix-fs's, we rely on nlink < 2 1155 # for non-posix-fs's, we rely on nlink < 2
941 $ndirs = (stat _)[3] - 2 1156 $ndirs = (stat _)[3] - 2
942 or return $grp->result ([], $entries); 1157 or return $grp->result ([], $entries);
943 } 1158 }
944 1159
945 my (@dirs, @nondirs); 1160 my (@dirs, @nondirs);
946 1161
947 my $statgrp = add $grp aio_group sub { 1162 my $statgrp = add $grp aio_group sub {
948 $grp->result (\@dirs, \@nondirs); 1163 $grp->result (\@dirs, \@nondirs);
949 }; 1164 };
950 1165
951 limit $statgrp $maxreq; 1166 limit $statgrp $maxreq;
952 feed $statgrp sub { 1167 feed $statgrp sub {
953 return unless @$entries; 1168 return unless @$entries;
954 my $entry = shift @$entries; 1169 my $entry = shift @$entries;
955 1170
956 aioreq_pri $pri; 1171 aioreq_pri $pri;
1172 $wd->[1] = "$entry/.";
957 add $statgrp aio_stat "$path/$entry/.", sub { 1173 add $statgrp aio_stat $wd, sub {
958 if ($_[0] < 0) { 1174 if ($_[0] < 0) {
959 push @nondirs, $entry; 1175 push @nondirs, $entry;
960 } else { 1176 } else {
961 # need to check for real directory 1177 # need to check for real directory
962 aioreq_pri $pri; 1178 aioreq_pri $pri;
1179 $wd->[1] = $entry;
963 add $statgrp aio_lstat "$path/$entry", sub { 1180 add $statgrp aio_lstat $wd, sub {
964 if (-d _) { 1181 if (-d _) {
965 push @dirs, $entry; 1182 push @dirs, $entry;
966 1183
967 unless (--$ndirs) { 1184 unless (--$ndirs) {
968 push @nondirs, @$entries; 1185 push @nondirs, @$entries;
969 feed $statgrp; 1186 feed $statgrp;
1187 }
1188 } else {
1189 push @nondirs, $entry;
970 } 1190 }
971 } else {
972 push @nondirs, $entry;
973 } 1191 }
974 } 1192 }
975 } 1193 };
976 }; 1194 };
977 }; 1195 };
978 }; 1196 };
979 }; 1197 };
980 }; 1198 };
981 1199
982 $grp 1200 $grp
983} 1201}
984 1202
985=item aio_rmtree $path, $callback->($status) 1203=item aio_rmtree $pathname, $callback->($status)
986 1204
987Delete a directory tree starting (and including) C<$path>, return the 1205Delete a directory tree starting (and including) C<$path>, return the
988status of the final C<rmdir> only. This is a composite request that 1206status of the final C<rmdir> only. This is a composite request that
989uses C<aio_scandir> to recurse into and rmdir directories, and unlink 1207uses C<aio_scandir> to recurse into and rmdir directories, and unlink
990everything else. 1208everything else.
991 1209
992=cut 1210=cut
993 1211
1015 }; 1233 };
1016 1234
1017 $grp 1235 $grp
1018} 1236}
1019 1237
1238=item aio_fcntl $fh, $cmd, $arg, $callback->($status)
1239
1240=item aio_ioctl $fh, $request, $buf, $callback->($status)
1241
1242These work just like the C<fcntl> and C<ioctl> built-in functions, except
1243they execute asynchronously and pass the return value to the callback.
1244
1245Both calls can be used for a lot of things, some of which make more sense
1246to run asynchronously in their own thread, while some others make less
1247sense. For example, calls that block waiting for external events, such
1248as locking, will also lock down an I/O thread while it is waiting, which
1249can deadlock the whole I/O system. At the same time, there might be no
1250alternative to using a thread to wait.
1251
1252So in general, you should only use these calls for things that do
1253(filesystem) I/O, not for things that wait for other events (network,
1254other processes), although if you are careful and know what you are doing,
1255you still can.
1256
1257The following constants are available (missing ones are, as usual C<0>):
1258
1259C<FIFREEZE>, C<FITHAW>, C<FITRIM>, C<FICLONE>, C<FICLONERANGE>, C<FIDEDUPERANGE>.
1260
1261C<FS_IOC_GETFLAGS>, C<FS_IOC_SETFLAGS>, C<FS_IOC_GETVERSION>, C<FS_IOC_SETVERSION>,
1262C<FS_IOC_FIEMAP>.
1263
1264C<FS_IOC_FSGETXATTR>, C<FS_IOC_FSSETXATTR>, C<FS_IOC_SET_ENCRYPTION_POLICY>,
1265C<FS_IOC_GET_ENCRYPTION_PWSALT>, C<FS_IOC_GET_ENCRYPTION_POLICY>, C<FS_KEY_DESCRIPTOR_SIZE>.
1266
1267C<FS_SECRM_FL>, C<FS_UNRM_FL>, C<FS_COMPR_FL>, C<FS_SYNC_FL>, C<FS_IMMUTABLE_FL>,
1268C<FS_APPEND_FL>, C<FS_NODUMP_FL>, C<FS_NOATIME_FL>, C<FS_DIRTY_FL>,
1269C<FS_COMPRBLK_FL>, C<FS_NOCOMP_FL>, C<FS_ENCRYPT_FL>, C<FS_BTREE_FL>,
1270C<FS_INDEX_FL>, C<FS_JOURNAL_DATA_FL>, C<FS_NOTAIL_FL>, C<FS_DIRSYNC_FL>, C<FS_TOPDIR_FL>,
1271C<FS_FL_USER_MODIFIABLE>.
1272
1273C<FS_XFLAG_REALTIME>, C<FS_XFLAG_PREALLOC>, C<FS_XFLAG_IMMUTABLE>, C<FS_XFLAG_APPEND>,
1274C<FS_XFLAG_SYNC>, C<FS_XFLAG_NOATIME>, C<FS_XFLAG_NODUMP>, C<FS_XFLAG_RTINHERIT>,
1275C<FS_XFLAG_PROJINHERIT>, C<FS_XFLAG_NOSYMLINKS>, C<FS_XFLAG_EXTSIZE>, C<FS_XFLAG_EXTSZINHERIT>,
1276C<FS_XFLAG_NODEFRAG>, C<FS_XFLAG_FILESTREAM>, C<FS_XFLAG_DAX>, C<FS_XFLAG_HASATTR>,
1277
1020=item aio_sync $callback->($status) 1278=item aio_sync $callback->($status)
1021 1279
1022Asynchronously call sync and call the callback when finished. 1280Asynchronously call sync and call the callback when finished.
1023 1281
1024=item aio_fsync $fh, $callback->($status) 1282=item aio_fsync $fh, $callback->($status)
1031Asynchronously call fdatasync on the given filehandle and call the 1289Asynchronously call fdatasync on the given filehandle and call the
1032callback with the fdatasync result code. 1290callback with the fdatasync result code.
1033 1291
1034If this call isn't available because your OS lacks it or it couldn't be 1292If this call isn't available because your OS lacks it or it couldn't be
1035detected, it will be emulated by calling C<fsync> instead. 1293detected, it will be emulated by calling C<fsync> instead.
1294
1295=item aio_syncfs $fh, $callback->($status)
1296
1297Asynchronously call the syncfs syscall to sync the filesystem associated
1298to the given filehandle and call the callback with the syncfs result
1299code. If syncfs is not available, calls sync(), but returns C<-1> and sets
1300errno to C<ENOSYS> nevertheless.
1036 1301
1037=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status) 1302=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
1038 1303
1039Sync the data portion of the file specified by C<$offset> and C<$length> 1304Sync the data portion of the file specified by C<$offset> and C<$length>
1040to disk (but NOT the metadata), by calling the Linux-specific 1305to disk (but NOT the metadata), by calling the Linux-specific
1044C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>, 1309C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
1045C<IO::AIO::SYNC_FILE_RANGE_WRITE> and 1310C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
1046C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range 1311C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
1047manpage for details. 1312manpage for details.
1048 1313
1049=item aio_pathsync $path, $callback->($status) 1314=item aio_pathsync $pathname, $callback->($status)
1050 1315
1051This request tries to open, fsync and close the given path. This is a 1316This request tries to open, fsync and close the given path. This is a
1052composite request intended to sync directories after directory operations 1317composite request intended to sync directories after directory operations
1053(E.g. rename). This might not work on all operating systems or have any 1318(E.g. rename). This might not work on all operating systems or have any
1054specific effect, but usually it makes sure that directory changes get 1319specific effect, but usually it makes sure that directory changes get
1106 1371
1107This is a rather advanced IO::AIO call, which works best on mmap(2)ed 1372This is a rather advanced IO::AIO call, which works best on mmap(2)ed
1108scalars. 1373scalars.
1109 1374
1110It touches (reads or writes) all memory pages in the specified 1375It touches (reads or writes) all memory pages in the specified
1111range inside the scalar. All caveats and parameters are the same 1376range inside the scalar. All caveats and parameters are the same
1112as for C<aio_msync>, above, except for flags, which must be either 1377as for C<aio_msync>, above, except for flags, which must be either
1113C<0> (which reads all pages and ensures they are instantiated) or 1378C<0> (which reads all pages and ensures they are instantiated) or
1114C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and 1379C<IO::AIO::MT_MODIFY>, which modifies the memory pages (by reading and
1115writing an octet from it, which dirties the page). 1380writing an octet from it, which dirties the page).
1116 1381
1117=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status) 1382=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
1118 1383
1119This is a rather advanced IO::AIO call, which works best on mmap(2)ed 1384This is a rather advanced IO::AIO call, which works best on mmap(2)ed
1150documented under L<MISCELLANEOUS FUNCTIONS>. 1415documented under L<MISCELLANEOUS FUNCTIONS>.
1151 1416
1152Example: asynchronously lock all current and future pages into memory. 1417Example: asynchronously lock all current and future pages into memory.
1153 1418
1154 aio_mlockall IO::AIO::MCL_FUTURE; 1419 aio_mlockall IO::AIO::MCL_FUTURE;
1420
1421=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
1422
1423Queries the extents of the given file (by calling the Linux C<FIEMAP>
1424ioctl, see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If
1425the ioctl is not available on your OS, then this request will fail with
1426C<ENOSYS>.
1427
1428C<$start> is the starting offset to query extents for, C<$length> is the
1429size of the range to query - if it is C<undef>, then the whole file will
1430be queried.
1431
1432C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
1433C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
1434exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
1435the data portion.
1436
1437C<$count> is the maximum number of extent records to return. If it is
1438C<undef>, then IO::AIO queries all extents of the range. As a very special
1439case, if it is C<0>, then the callback receives the number of extents
1440instead of the extents themselves (which is unreliable, see below).
1441
1442If an error occurs, the callback receives no arguments. The special
1443C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
1444
1445Otherwise, the callback receives an array reference with extent
1446structures. Each extent structure is an array reference itself, with the
1447following members:
1448
1449 [$logical, $physical, $length, $flags]
1450
1451Flags is any combination of the following flag values (typically either C<0>
1452or C<IO::AIO::FIEMAP_EXTENT_LAST> (1)):
1453
1454C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
1455C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
1456C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
1457C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
1458C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
1459C<IO::AIO::FIEMAP_EXTENT_SHARED>.
1460
1461At the time of this writing (Linux 3.2), this requets is unreliable unless
1462C<$count> is C<undef>, as the kernel has all sorts of bugs preventing
1463it to return all extents of a range for files with large number of
1464extents. The code works around all these issues if C<$count> is undef.
1155 1465
1156=item aio_group $callback->(...) 1466=item aio_group $callback->(...)
1157 1467
1158This is a very special aio request: Instead of doing something, it is a 1468This is a very special aio request: Instead of doing something, it is a
1159container for other aio requests, which is useful if you want to bundle 1469container for other aio requests, which is useful if you want to bundle
1196like sleep and file handle readable/writable, the overhead this creates is 1506like sleep and file handle readable/writable, the overhead this creates is
1197immense (it blocks a thread for a long time) so do not use this function 1507immense (it blocks a thread for a long time) so do not use this function
1198except to put your application under artificial I/O pressure. 1508except to put your application under artificial I/O pressure.
1199 1509
1200=back 1510=back
1511
1512
1513=head2 IO::AIO::WD - multiple working directories
1514
1515Your process only has one current working directory, which is used by all
1516threads. This makes it hard to use relative paths (some other component
1517could call C<chdir> at any time, and it is hard to control when the path
1518will be used by IO::AIO).
1519
1520One solution for this is to always use absolute paths. This usually works,
1521but can be quite slow (the kernel has to walk the whole path on every
1522access), and can also be a hassle to implement.
1523
1524Newer POSIX systems have a number of functions (openat, fdopendir,
1525futimensat and so on) that make it possible to specify working directories
1526per operation.
1527
1528For portability, and because the clowns who "designed", or shall I write,
1529perpetrated this new interface were obviously half-drunk, this abstraction
1530cannot be perfect, though.
1531
1532IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
1533object. This object stores the canonicalised, absolute version of the
1534path, and on systems that allow it, also a directory file descriptor.
1535
1536Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
1537or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
1538object and a pathname instead (or the IO::AIO::WD object alone, which
1539gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
1540IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
1541to that IO::AIO::WD object.
1542
1543For example, to get a wd object for F</etc> and then stat F<passwd>
1544inside, you would write:
1545
1546 aio_wd "/etc", sub {
1547 my $etcdir = shift;
1548
1549 # although $etcdir can be undef on error, there is generally no reason
1550 # to check for errors here, as aio_stat will fail with ENOENT
1551 # when $etcdir is undef.
1552
1553 aio_stat [$etcdir, "passwd"], sub {
1554 # yay
1555 };
1556 };
1557
1558The fact that C<aio_wd> is a request and not a normal function shows that
1559creating an IO::AIO::WD object is itself a potentially blocking operation,
1560which is why it is done asynchronously.
1561
1562To stat the directory obtained with C<aio_wd> above, one could write
1563either of the following three request calls:
1564
1565 aio_lstat "/etc" , sub { ... # pathname as normal string
1566 aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
1567 aio_lstat $wd , sub { ... # shorthand for the previous
1568
1569As with normal pathnames, IO::AIO keeps a copy of the working directory
1570object and the pathname string, so you could write the following without
1571causing any issues due to C<$path> getting reused:
1572
1573 my $path = [$wd, undef];
1574
1575 for my $name (qw(abc def ghi)) {
1576 $path->[1] = $name;
1577 aio_stat $path, sub {
1578 # ...
1579 };
1580 }
1581
1582There are some caveats: when directories get renamed (or deleted), the
1583pathname string doesn't change, so will point to the new directory (or
1584nowhere at all), while the directory fd, if available on the system,
1585will still point to the original directory. Most functions accepting a
1586pathname will use the directory fd on newer systems, and the string on
1587older systems. Some functions (such as realpath) will always rely on the
1588string form of the pathname.
1589
1590So this functionality is mainly useful to get some protection against
1591C<chdir>, to easily get an absolute path out of a relative path for future
1592reference, and to speed up doing many operations in the same directory
1593(e.g. when stat'ing all files in a directory).
1594
1595The following functions implement this working directory abstraction:
1596
1597=over 4
1598
1599=item aio_wd $pathname, $callback->($wd)
1600
1601Asynchonously canonicalise the given pathname and convert it to an
1602IO::AIO::WD object representing it. If possible and supported on the
1603system, also open a directory fd to speed up pathname resolution relative
1604to this working directory.
1605
1606If something goes wrong, then C<undef> is passwd to the callback instead
1607of a working directory object and C<$!> is set appropriately. Since
1608passing C<undef> as working directory component of a pathname fails the
1609request with C<ENOENT>, there is often no need for error checking in the
1610C<aio_wd> callback, as future requests using the value will fail in the
1611expected way.
1612
1613=item IO::AIO::CWD
1614
1615This is a compiletime constant (object) that represents the process
1616current working directory.
1617
1618Specifying this object as working directory object for a pathname is as if
1619the pathname would be specified directly, without a directory object. For
1620example, these calls are functionally identical:
1621
1622 aio_stat "somefile", sub { ... };
1623 aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
1624
1625=back
1626
1627To recover the path associated with an IO::AIO::WD object, you can use
1628C<aio_realpath>:
1629
1630 aio_realpath $wd, sub {
1631 warn "path is $_[0]\n";
1632 };
1633
1634Currently, C<aio_statvfs> always, and C<aio_rename> and C<aio_rmdir>
1635sometimes, fall back to using an absolue path.
1201 1636
1202=head2 IO::AIO::REQ CLASS 1637=head2 IO::AIO::REQ CLASS
1203 1638
1204All non-aggregate C<aio_*> functions return an object of this class when 1639All non-aggregate C<aio_*> functions return an object of this class when
1205called in non-void context. 1640called in non-void context.
1323 1758
1324Sets a feeder/generator on this group: every group can have an attached 1759Sets a feeder/generator on this group: every group can have an attached
1325generator that generates requests if idle. The idea behind this is that, 1760generator that generates requests if idle. The idea behind this is that,
1326although you could just queue as many requests as you want in a group, 1761although you could just queue as many requests as you want in a group,
1327this might starve other requests for a potentially long time. For example, 1762this might starve other requests for a potentially long time. For example,
1328C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests, 1763C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
1329delaying any later requests for a long time. 1764requests, delaying any later requests for a long time.
1330 1765
1331To avoid this, and allow incremental generation of requests, you can 1766To avoid this, and allow incremental generation of requests, you can
1332instead a group and set a feeder on it that generates those requests. The 1767instead a group and set a feeder on it that generates those requests. The
1333feed callback will be called whenever there are few enough (see C<limit>, 1768feed callback will be called whenever there are few enough (see C<limit>,
1334below) requests active in the group itself and is expected to queue more 1769below) requests active in the group itself and is expected to queue more
1383 1818
1384See C<poll_cb> for an example. 1819See C<poll_cb> for an example.
1385 1820
1386=item IO::AIO::poll_cb 1821=item IO::AIO::poll_cb
1387 1822
1388Process some outstanding events on the result pipe. You have to call this 1823Process some requests that have reached the result phase (i.e. they have
1389regularly. Returns C<0> if all events could be processed, or C<-1> if it 1824been executed but the results are not yet reported). You have to call
1390returned earlier for whatever reason. Returns immediately when no events 1825this "regularly" to finish outstanding requests.
1391are outstanding. The amount of events processed depends on the settings of
1392C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
1393 1826
1827Returns C<0> if all events could be processed (or there were no
1828events to process), or C<-1> if it returned earlier for whatever
1829reason. Returns immediately when no events are outstanding. The amount
1830of events processed depends on the settings of C<IO::AIO::max_poll_req>,
1831C<IO::AIO::max_poll_time> and C<IO::AIO::max_outstanding>.
1832
1394If not all requests were processed for whatever reason, the filehandle 1833If not all requests were processed for whatever reason, the poll file
1395will still be ready when C<poll_cb> returns, so normally you don't have to 1834descriptor will still be ready when C<poll_cb> returns, so normally you
1396do anything special to have it called later. 1835don't have to do anything special to have it called later.
1836
1837Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
1838ready, it can be beneficial to call this function from loops which submit
1839a lot of requests, to make sure the results get processed when they become
1840available and not just when the loop is finished and the event loop takes
1841over again. This function returns very fast when there are no outstanding
1842requests.
1397 1843
1398Example: Install an Event watcher that automatically calls 1844Example: Install an Event watcher that automatically calls
1399IO::AIO::poll_cb with high priority (more examples can be found in the 1845IO::AIO::poll_cb with high priority (more examples can be found in the
1400SYNOPSIS section, at the top of this document): 1846SYNOPSIS section, at the top of this document):
1401 1847
1403 poll => 'r', async => 1, 1849 poll => 'r', async => 1,
1404 cb => \&IO::AIO::poll_cb); 1850 cb => \&IO::AIO::poll_cb);
1405 1851
1406=item IO::AIO::poll_wait 1852=item IO::AIO::poll_wait
1407 1853
1408If there are any outstanding requests and none of them in the result 1854Wait until either at least one request is in the result phase or no
1409phase, wait till the result filehandle becomes ready for reading (simply 1855requests are outstanding anymore.
1410does a C<select> on the filehandle. This is useful if you want to 1856
1411synchronously wait for some requests to finish). 1857This is useful if you want to synchronously wait for some requests to
1858become ready, without actually handling them.
1412 1859
1413See C<nreqs> for an example. 1860See C<nreqs> for an example.
1414 1861
1415=item IO::AIO::poll 1862=item IO::AIO::poll
1416 1863
1503 1950
1504Under normal circumstances you don't need to call this function. 1951Under normal circumstances you don't need to call this function.
1505 1952
1506=item IO::AIO::max_idle $nthreads 1953=item IO::AIO::max_idle $nthreads
1507 1954
1508Limit the number of threads (default: 4) that are allowed to idle (i.e., 1955Limit the number of threads (default: 4) that are allowed to idle
1509threads that did not get a request to process within 10 seconds). That 1956(i.e., threads that did not get a request to process within the idle
1510means if a thread becomes idle while C<$nthreads> other threads are also 1957timeout (default: 10 seconds). That means if a thread becomes idle while
1511idle, it will free its resources and exit. 1958C<$nthreads> other threads are also idle, it will free its resources and
1959exit.
1512 1960
1513This is useful when you allow a large number of threads (e.g. 100 or 1000) 1961This is useful when you allow a large number of threads (e.g. 100 or 1000)
1514to allow for extremely high load situations, but want to free resources 1962to allow for extremely high load situations, but want to free resources
1515under normal circumstances (1000 threads can easily consume 30MB of RAM). 1963under normal circumstances (1000 threads can easily consume 30MB of RAM).
1516 1964
1517The default is probably ok in most situations, especially if thread 1965The default is probably ok in most situations, especially if thread
1518creation is fast. If thread creation is very slow on your system you might 1966creation is fast. If thread creation is very slow on your system you might
1519want to use larger values. 1967want to use larger values.
1520 1968
1969=item IO::AIO::idle_timeout $seconds
1970
1971Sets the minimum idle timeout (default 10) after which worker threads are
1972allowed to exit. SEe C<IO::AIO::max_idle>.
1973
1521=item IO::AIO::max_outstanding $maxreqs 1974=item IO::AIO::max_outstanding $maxreqs
1975
1976Sets the maximum number of outstanding requests to C<$nreqs>. If
1977you do queue up more than this number of requests, the next call to
1978C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
1979C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
1980longer exceeded.
1981
1982In other words, this setting does not enforce a queue limit, but can be
1983used to make poll functions block if the limit is exceeded.
1522 1984
1523This is a very bad function to use in interactive programs because it 1985This is a very bad function to use in interactive programs because it
1524blocks, and a bad way to reduce concurrency because it is inexact: Better 1986blocks, and a bad way to reduce concurrency because it is inexact: Better
1525use an C<aio_group> together with a feed callback. 1987use an C<aio_group> together with a feed callback.
1526 1988
1527Sets the maximum number of outstanding requests to C<$nreqs>. If you 1989Its main use is in scripts without an event loop - when you want to stat
1528do queue up more than this number of requests, the next call to the 1990a lot of files, you can write somehting like this:
1529C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
1530function will block until the limit is no longer exceeded.
1531 1991
1532The default value is very large, so there is no practical limit on the 1992 IO::AIO::max_outstanding 32;
1533number of outstanding requests.
1534 1993
1535You can still queue as many requests as you want. Therefore, 1994 for my $path (...) {
1536C<max_outstanding> is mainly useful in simple scripts (with low values) or 1995 aio_stat $path , ...;
1537as a stop gap to shield against fatal memory overflow (with large values). 1996 IO::AIO::poll_cb;
1997 }
1998
1999 IO::AIO::flush;
2000
2001The call to C<poll_cb> inside the loop will normally return instantly, but
2002as soon as more thna C<32> reqeusts are in-flight, it will block until
2003some requests have been handled. This keeps the loop from pushing a large
2004number of C<aio_stat> requests onto the queue.
2005
2006The default value for C<max_outstanding> is very large, so there is no
2007practical limit on the number of outstanding requests.
1538 2008
1539=back 2009=back
1540 2010
1541=head3 STATISTICAL INFORMATION 2011=head3 STATISTICAL INFORMATION
1542 2012
1564 2034
1565=back 2035=back
1566 2036
1567=head3 MISCELLANEOUS FUNCTIONS 2037=head3 MISCELLANEOUS FUNCTIONS
1568 2038
1569IO::AIO implements some functions that might be useful, but are not 2039IO::AIO implements some functions that are useful when you want to use
1570asynchronous. 2040some "Advanced I/O" function not available to in Perl, without going the
2041"Asynchronous I/O" route. Many of these have an asynchronous C<aio_*>
2042counterpart.
1571 2043
1572=over 4 2044=over 4
1573 2045
1574=item IO::AIO::sendfile $ofh, $ifh, $offset, $count 2046=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
1575 2047
1580 2052
1581Returns the number of bytes copied, or C<-1> on error. 2053Returns the number of bytes copied, or C<-1> on error.
1582 2054
1583=item IO::AIO::fadvise $fh, $offset, $len, $advice 2055=item IO::AIO::fadvise $fh, $offset, $len, $advice
1584 2056
1585Simply calls the C<posix_fadvise> function (see it's 2057Simply calls the C<posix_fadvise> function (see its
1586manpage for details). The following advice constants are 2058manpage for details). The following advice constants are
1587avaiable: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>, 2059available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
1588C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>, 2060C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
1589C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>. 2061C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
1590 2062
1591On systems that do not implement C<posix_fadvise>, this function returns 2063On systems that do not implement C<posix_fadvise>, this function returns
1592ENOSYS, otherwise the return value of C<posix_fadvise>. 2064ENOSYS, otherwise the return value of C<posix_fadvise>.
1593 2065
2066=item IO::AIO::madvise $scalar, $offset, $len, $advice
2067
2068Simply calls the C<posix_madvise> function (see its
2069manpage for details). The following advice constants are
2070available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
2071C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>,
2072C<IO::AIO::MADV_FREE>.
2073
2074On systems that do not implement C<posix_madvise>, this function returns
2075ENOSYS, otherwise the return value of C<posix_madvise>.
2076
2077=item IO::AIO::mprotect $scalar, $offset, $len, $protect
2078
2079Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
2080$scalar (see its manpage for details). The following protect
2081constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
2082C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
2083
2084On systems that do not implement C<mprotect>, this function returns
2085ENOSYS, otherwise the return value of C<mprotect>.
2086
1594=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset] 2087=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
1595 2088
1596Memory-maps a file (or anonymous memory range) and attaches it to the 2089Memory-maps a file (or anonymous memory range) and attaches it to the
1597given C<$scalar>, which will act like a string scalar. 2090given C<$scalar>, which will act like a string scalar. Returns true on
2091success, and false otherwise.
1598 2092
1599The only operations allowed on the scalar are C<substr>/C<vec> that don't 2093The only operations allowed on the scalar are C<substr>/C<vec> that don't
1600change the string length, and most read-only operations such as copying it 2094change the string length, and most read-only operations such as copying it
1601or searching it with regexes and so on. 2095or searching it with regexes and so on.
1602 2096
1613filesize. 2107filesize.
1614 2108
1615C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>, 2109C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
1616C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>, 2110C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
1617 2111
1618C<$flags> can be a combination of C<IO::AIO::MAP_SHARED> or 2112C<$flags> can be a combination of
1619C<IO::AIO::MAP_PRIVATE>, or a number of system-specific flags (when 2113C<IO::AIO::MAP_SHARED> or
1620not available, the are defined as 0): C<IO::AIO::MAP_ANONYMOUS> 2114C<IO::AIO::MAP_PRIVATE>,
2115or a number of system-specific flags (when not available, the are C<0>):
1621(which is set to C<MAP_ANON> if your system only provides this 2116C<IO::AIO::MAP_ANONYMOUS> (which is set to C<MAP_ANON> if your system only provides this constant),
1622constant), C<IO::AIO::MAP_HUGETLB>, C<IO::AIO::MAP_LOCKED>, 2117C<IO::AIO::MAP_LOCKED>,
1623C<IO::AIO::MAP_NORESERVE>, C<IO::AIO::MAP_POPULATE> or 2118C<IO::AIO::MAP_NORESERVE>,
2119C<IO::AIO::MAP_POPULATE>,
1624C<IO::AIO::MAP_NONBLOCK> 2120C<IO::AIO::MAP_NONBLOCK>,
2121C<IO::AIO::MAP_FIXED>,
2122C<IO::AIO::MAP_GROWSDOWN>,
2123C<IO::AIO::MAP_32BIT>,
2124C<IO::AIO::MAP_HUGETLB> or
2125C<IO::AIO::MAP_STACK>.
1625 2126
1626If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed. 2127If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
1627 2128
1628C<$offset> is the offset from the start of the file - it generally must be 2129C<$offset> is the offset from the start of the file - it generally must be
1629a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>. 2130a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
1654 2155
1655Calls the C<munlockall> function. 2156Calls the C<munlockall> function.
1656 2157
1657On systems that do not implement C<munlockall>, this function returns 2158On systems that do not implement C<munlockall>, this function returns
1658ENOSYS, otherwise the return value of C<munlockall>. 2159ENOSYS, otherwise the return value of C<munlockall>.
2160
2161=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
2162
2163Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
2164C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
2165should be the file offset.
2166
2167C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
2168silently corrupt the data in this case.
2169
2170The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
2171C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
2172C<IO::AIO::SPLICE_F_GIFT>.
2173
2174See the C<splice(2)> manpage for details.
2175
2176=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
2177
2178Calls the GNU/Linux C<tee(2)> syscall, see its manpage and the
2179description for C<IO::AIO::splice> above for details.
2180
2181=item $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
2182
2183Attempts to query or change the pipe buffer size. Obviously works only
2184on pipes, and currently works only on GNU/Linux systems, and fails with
2185C<-1>/C<ENOSYS> everywhere else. If anybody knows how to influence pipe buffer
2186size on other systems, drop me a note.
2187
2188=item ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
2189
2190This is a direct interface to the Linux L<pipe2(2)> system call. If
2191C<$flags> is missing or C<0>, then this should be the same as a call to
2192perl's built-in C<pipe> function and create a new pipe, and works on
2193systems that lack the pipe2 syscall. On win32, this case invokes C<_pipe
2194(..., 4096, O_BINARY)>.
2195
2196If C<$flags> is non-zero, it tries to invoke the pipe2 system call with
2197the given flags (Linux 2.6.27, glibc 2.9).
2198
2199On success, the read and write file handles are returned.
2200
2201On error, nothing will be returned. If the pipe2 syscall is missing and
2202C<$flags> is non-zero, fails with C<ENOSYS>.
2203
2204Please refer to L<pipe2(2)> for more info on the C<$flags>, but at the
2205time of this writing, C<IO::AIO::O_CLOEXEC>, C<IO::AIO::O_NONBLOCK> and
2206C<IO::AIO::O_DIRECT> (Linux 3.4, for packet-based pipes) were supported.
1659 2207
1660=back 2208=back
1661 2209
1662=cut 2210=cut
1663 2211
1698 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => 2246 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
1699 \&IO::AIO::poll_cb); 2247 \&IO::AIO::poll_cb);
1700 2248
1701=head2 FORK BEHAVIOUR 2249=head2 FORK BEHAVIOUR
1702 2250
1703This module should do "the right thing" when the process using it forks: 2251Usage of pthreads in a program changes the semantics of fork
2252considerably. Specifically, only async-safe functions can be called after
2253fork. Perl doesn't know about this, so in general, you cannot call fork
2254with defined behaviour in perl if pthreads are involved. IO::AIO uses
2255pthreads, so this applies, but many other extensions and (for inexplicable
2256reasons) perl itself often is linked against pthreads, so this limitation
2257applies to quite a lot of perls.
1704 2258
1705Before the fork, IO::AIO enters a quiescent state where no requests 2259This module no longer tries to fight your OS, or POSIX. That means IO::AIO
1706can be added in other threads and no results will be processed. After 2260only works in the process that loaded it. Forking is fully supported, but
1707the fork the parent simply leaves the quiescent state and continues 2261using IO::AIO in the child is not.
1708request/result processing, while the child frees the request/result queue
1709(so that the requests started before the fork will only be handled in the
1710parent). Threads will be started on demand until the limit set in the
1711parent process has been reached again.
1712 2262
1713In short: the parent will, after a short pause, continue as if fork had 2263You might get around by not I<using> IO::AIO before (or after)
1714not been called, while the child will act as if IO::AIO has not been used 2264forking. You could also try to call the L<IO::AIO::reinit> function in the
1715yet. 2265child:
2266
2267=over 4
2268
2269=item IO::AIO::reinit
2270
2271Abandons all current requests and I/O threads and simply reinitialises all
2272data structures. This is not an operation supported by any standards, but
2273happens to work on GNU/Linux and some newer BSD systems.
2274
2275The only reasonable use for this function is to call it after forking, if
2276C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
2277the process will result in undefined behaviour. Calling it at any time
2278will also result in any undefined (by POSIX) behaviour.
2279
2280=back
1716 2281
1717=head2 MEMORY USAGE 2282=head2 MEMORY USAGE
1718 2283
1719Per-request usage: 2284Per-request usage:
1720 2285

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines