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 1.136 by root, Tue Sep 30 14:07:59 2008 UTC vs.
Revision 1.296 by root, Sun Aug 26 03:17:35 2018 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
9 aio_open "/etc/passwd", O_RDONLY, 0, sub { 9 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
10 my $fh = shift 10 my $fh = shift
11 or die "/etc/passwd: $!"; 11 or die "/etc/passwd: $!";
12 ... 12 ...
13 }; 13 };
14 14
26 $req->cancel; # cancel request if still in queue 26 $req->cancel; # cancel request if still in queue
27 27
28 my $grp = aio_group sub { print "all stats done\n" }; 28 my $grp = aio_group sub { print "all stats done\n" };
29 add $grp aio_stat "..." for ...; 29 add $grp aio_stat "..." for ...;
30 30
31 # AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
32 use AnyEvent::AIO;
33
34 # EV integration
35 my $w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
36
37 # Event integration
38 Event->io (fd => IO::AIO::poll_fileno,
39 poll => 'r',
40 cb => \&IO::AIO::poll_cb);
41
42 # Glib/Gtk2 integration
43 add_watch Glib::IO IO::AIO::poll_fileno,
44 in => sub { IO::AIO::poll_cb; 1 };
45
46 # Tk integration
47 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
48 readable => \&IO::AIO::poll_cb);
49
50 # Danga::Socket integration
51 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
52 \&IO::AIO::poll_cb);
53
54=head1 DESCRIPTION 31=head1 DESCRIPTION
55 32
56This module implements asynchronous I/O using whatever means your 33This module implements asynchronous I/O using whatever means your
57operating system supports. 34operating system supports. It is implemented as an interface to C<libeio>
35(L<http://software.schmorp.de/pkg/libeio.html>).
58 36
59Asynchronous means that operations that can normally block your program 37Asynchronous means that operations that can normally block your program
60(e.g. reading from disk) will be done asynchronously: the operation 38(e.g. reading from disk) will be done asynchronously: the operation
61will still block, but you can do something else in the meantime. This 39will still block, but you can do something else in the meantime. This
62is extremely useful for programs that need to stay interactive even 40is extremely useful for programs that need to stay interactive even
66on a RAID volume or over NFS when you do a number of stat operations 44on a RAID volume or over NFS when you do a number of stat operations
67concurrently. 45concurrently.
68 46
69While most of this works on all types of file descriptors (for 47While most of this works on all types of file descriptors (for
70example sockets), using these functions on file descriptors that 48example sockets), using these functions on file descriptors that
71support nonblocking operation (again, sockets, pipes etc.) is very 49support nonblocking operation (again, sockets, pipes etc.) is
72inefficient. Use an event loop for that (such as the L<Event|Event> 50very inefficient. Use an event loop for that (such as the L<EV>
73module): IO::AIO will naturally fit into such an event loop itself. 51module): IO::AIO will naturally fit into such an event loop itself.
74 52
75In this version, a number of threads are started that execute your 53In this version, a number of threads are started that execute your
76requests and signal their completion. You don't need thread support 54requests and signal their completion. You don't need thread support
77in perl, and the threads created by this module will not be visible 55in perl, and the threads created by this module will not be visible
80not 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
81files currently, for example), and they would only support aio_read and 59files currently, for example), and they would only support aio_read and
82aio_write, so the remaining functionality would have to be implemented 60aio_write, so the remaining functionality would have to be implemented
83using threads anyway. 61using threads anyway.
84 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
85Although the module will work in the presence of other (Perl-) threads, 67Although the module will work in the presence of other (Perl-) threads,
86it is currently not reentrant in any way, so use appropriate locking 68it is currently not reentrant in any way, so use appropriate locking
87yourself, 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
88call C<poll_cb> (or other C<aio_> functions) recursively. 70call C<poll_cb> (or other C<aio_> functions) recursively.
89 71
90=head2 EXAMPLE 72=head2 EXAMPLE
91 73
92This is a simple example that uses the Event module and loads 74This is a simple example that uses the EV module and loads
93F</etc/passwd> asynchronously: 75F</etc/passwd> asynchronously:
94 76
95 use Fcntl;
96 use Event; 77 use EV;
97 use IO::AIO; 78 use IO::AIO;
98 79
99 # register the IO::AIO callback with Event 80 # register the IO::AIO callback with EV
100 Event->io (fd => IO::AIO::poll_fileno, 81 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
101 poll => 'r',
102 cb => \&IO::AIO::poll_cb);
103 82
104 # queue the request to open /etc/passwd 83 # queue the request to open /etc/passwd
105 aio_open "/etc/passwd", O_RDONLY, 0, sub { 84 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
106 my $fh = shift 85 my $fh = shift
107 or die "error while opening: $!"; 86 or die "error while opening: $!";
108 87
109 # stat'ing filehandles is generally non-blocking 88 # stat'ing filehandles is generally non-blocking
110 my $size = -s $fh; 89 my $size = -s $fh;
119 98
120 # file contents now in $contents 99 # file contents now in $contents
121 print $contents; 100 print $contents;
122 101
123 # exit event loop and program 102 # exit event loop and program
124 Event::unloop; 103 EV::break;
125 }; 104 };
126 }; 105 };
127 106
128 # possibly queue up other requests, or open GUI windows, 107 # possibly queue up other requests, or open GUI windows,
129 # check for sockets etc. etc. 108 # check for sockets etc. etc.
130 109
131 # process events as long as there are some: 110 # process events as long as there are some:
132 Event::loop; 111 EV::run;
133 112
134=head1 REQUEST ANATOMY AND LIFETIME 113=head1 REQUEST ANATOMY AND LIFETIME
135 114
136Every 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
137directly visible to Perl. 116directly visible to Perl.
187 166
188package IO::AIO; 167package IO::AIO;
189 168
190use Carp (); 169use Carp ();
191 170
192no warnings; 171use common::sense;
193use strict 'vars';
194 172
195use base 'Exporter'; 173use base 'Exporter';
196 174
197BEGIN { 175BEGIN {
198 our $VERSION = '3.1'; 176 our $VERSION = 4.6;
199 177
200 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
201 aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir 179 aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
202 aio_scandir aio_symlink aio_readlink aio_sync aio_fsync 180 aio_scandir aio_symlink aio_readlink aio_realpath aio_fcntl aio_ioctl
203 aio_fdatasync 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
204 aio_rename aio_link aio_move aio_copy aio_group 183 aio_rename aio_rename2 aio_link aio_move aio_copy aio_group
205 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
206 aio_chmod aio_utime aio_truncate); 185 aio_chmod aio_utime aio_truncate
186 aio_msync aio_mtouch aio_mlock aio_mlockall
187 aio_statvfs
188 aio_slurp
189 aio_wd);
207 190
208 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice)); 191 our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
209 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush 192 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
210 min_parallel max_parallel max_idle 193 min_parallel max_parallel max_idle idle_timeout
211 nreqs nready npending nthreads 194 nreqs nready npending nthreads
212 max_poll_time max_poll_reqs); 195 max_poll_time max_poll_reqs
196 sendfile fadvise madvise
197 mmap munmap mremap munlock munlockall);
198
199 push @AIO_REQ, qw(aio_busy); # not exported
213 200
214 @IO::AIO::GRP::ISA = 'IO::AIO::REQ'; 201 @IO::AIO::GRP::ISA = 'IO::AIO::REQ';
215 202
216 require XSLoader; 203 require XSLoader;
217 XSLoader::load ("IO::AIO", $VERSION); 204 XSLoader::load ("IO::AIO", $VERSION);
218} 205}
219 206
220=head1 FUNCTIONS 207=head1 FUNCTIONS
221 208
222=head2 AIO REQUEST FUNCTIONS 209=head2 QUICK OVERVIEW
210
211This section simply lists the prototypes most of the functions for
212quick reference. See the following sections for function-by-function
213documentation.
214
215 aio_wd $pathname, $callback->($wd)
216 aio_open $pathname, $flags, $mode, $callback->($fh)
217 aio_close $fh, $callback->($status)
218 aio_seek $fh,$offset,$whence, $callback->($offs)
219 aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
220 aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
221 aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
222 aio_readahead $fh,$offset,$length, $callback->($retval)
223 aio_stat $fh_or_path, $callback->($status)
224 aio_lstat $fh, $callback->($status)
225 aio_statvfs $fh_or_path, $callback->($statvfs)
226 aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
227 aio_chown $fh_or_path, $uid, $gid, $callback->($status)
228 aio_chmod $fh_or_path, $mode, $callback->($status)
229 aio_truncate $fh_or_path, $offset, $callback->($status)
230 aio_allocate $fh, $mode, $offset, $len, $callback->($status)
231 aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
232 aio_unlink $pathname, $callback->($status)
233 aio_mknod $pathname, $mode, $dev, $callback->($status)
234 aio_link $srcpath, $dstpath, $callback->($status)
235 aio_symlink $srcpath, $dstpath, $callback->($status)
236 aio_readlink $pathname, $callback->($link)
237 aio_realpath $pathname, $callback->($path)
238 aio_rename $srcpath, $dstpath, $callback->($status)
239 aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
240 aio_mkdir $pathname, $mode, $callback->($status)
241 aio_rmdir $pathname, $callback->($status)
242 aio_readdir $pathname, $callback->($entries)
243 aio_readdirx $pathname, $flags, $callback->($entries, $flags)
244 IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
245 IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
246 aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
247 aio_load $pathname, $data, $callback->($status)
248 aio_copy $srcpath, $dstpath, $callback->($status)
249 aio_move $srcpath, $dstpath, $callback->($status)
250 aio_rmtree $pathname, $callback->($status)
251 aio_fcntl $fh, $cmd, $arg, $callback->($status)
252 aio_ioctl $fh, $request, $buf, $callback->($status)
253 aio_sync $callback->($status)
254 aio_syncfs $fh, $callback->($status)
255 aio_fsync $fh, $callback->($status)
256 aio_fdatasync $fh, $callback->($status)
257 aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
258 aio_pathsync $pathname, $callback->($status)
259 aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
260 aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
261 aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
262 aio_mlockall $flags, $callback->($status)
263 aio_group $callback->(...)
264 aio_nop $callback->()
265
266 $prev_pri = aioreq_pri [$pri]
267 aioreq_nice $pri_adjust
268
269 IO::AIO::poll_wait
270 IO::AIO::poll_cb
271 IO::AIO::poll
272 IO::AIO::flush
273 IO::AIO::max_poll_reqs $nreqs
274 IO::AIO::max_poll_time $seconds
275 IO::AIO::min_parallel $nthreads
276 IO::AIO::max_parallel $nthreads
277 IO::AIO::max_idle $nthreads
278 IO::AIO::idle_timeout $seconds
279 IO::AIO::max_outstanding $maxreqs
280 IO::AIO::nreqs
281 IO::AIO::nready
282 IO::AIO::npending
283 $nfd = IO::AIO::get_fdlimit [EXPERIMENTAL]
284 IO::AIO::min_fdlimit $nfd [EXPERIMENTAL]
285
286 IO::AIO::sendfile $ofh, $ifh, $offset, $count
287 IO::AIO::fadvise $fh, $offset, $len, $advice
288 IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
289 IO::AIO::munmap $scalar
290 IO::AIO::mremap $scalar, $new_length, $flags[, $new_address]
291 IO::AIO::madvise $scalar, $offset, $length, $advice
292 IO::AIO::mprotect $scalar, $offset, $length, $protect
293 IO::AIO::munlock $scalar, $offset = 0, $length = undef
294 IO::AIO::munlockall
295
296=head2 API NOTES
223 297
224All the C<aio_*> calls are more or less thin wrappers around the syscall 298All the C<aio_*> calls are more or less thin wrappers around the syscall
225with the same name (sans C<aio_>). The arguments are similar or identical, 299with the same name (sans C<aio_>). The arguments are similar or identical,
226and they all accept an additional (and optional) C<$callback> argument 300and they all accept an additional (and optional) C<$callback> argument
227which must be a code reference. This code reference will get called with 301which must be a code reference. This code reference will be called after
228the syscall return code (e.g. most syscalls return C<-1> on error, unlike 302the syscall has been executed in an asynchronous fashion. The results
229perl, which usually delivers "false") as its sole argument after the given 303of the request will be passed as arguments to the callback (and, if an
230syscall has been executed asynchronously. 304error occured, in C<$!>) - for most requests the syscall return code (e.g.
305most syscalls return C<-1> on error, unlike perl, which usually delivers
306"false").
307
308Some requests (such as C<aio_readdir>) pass the actual results and
309communicate failures by passing C<undef>.
231 310
232All functions expecting a filehandle keep a copy of the filehandle 311All functions expecting a filehandle keep a copy of the filehandle
233internally until the request has finished. 312internally until the request has finished.
234 313
235All functions return request objects of type L<IO::AIO::REQ> that allow 314All functions return request objects of type L<IO::AIO::REQ> that allow
236further manipulation of those requests while they are in-flight. 315further manipulation of those requests while they are in-flight.
237 316
238The pathnames you pass to these routines I<must> be absolute and 317The pathnames you pass to these routines I<should> be absolute. The
239encoded as octets. The reason for the former is that at the time the 318reason for this is that at the time the request is being executed, the
240request is being executed, the current working directory could have 319current working directory could have changed. Alternatively, you can
241changed. Alternatively, you can make sure that you never change the 320make sure that you never change the current working directory anywhere
242current working directory anywhere in the program and then use relative 321in the program and then use relative paths. You can also take advantage
243paths. 322of IO::AIOs working directory abstraction, that lets you specify paths
323relative to some previously-opened "working directory object" - see the
324description of the C<IO::AIO::WD> class later in this document.
244 325
245To encode pathnames as octets, either make sure you either: a) always pass 326To encode pathnames as octets, either make sure you either: a) always pass
246in filenames you got from outside (command line, readdir etc.) without 327in filenames you got from outside (command line, readdir etc.) without
247tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode 328tinkering, b) are in your native filesystem encoding, c) use the Encode
248your pathnames to the locale (or other) encoding in effect in the user 329module and encode your pathnames to the locale (or other) encoding in
249environment, d) use Glib::filename_from_unicode on unicode filenames or e) 330effect in the user environment, d) use Glib::filename_from_unicode on
250use something else to ensure your scalar has the correct contents. 331unicode filenames or e) use something else to ensure your scalar has the
332correct contents.
251 333
252This works, btw. independent of the internal UTF-8 bit, which IO::AIO 334This works, btw. independent of the internal UTF-8 bit, which IO::AIO
253handles correctly whether it is set or not. 335handles correctly whether it is set or not.
336
337=head2 AIO REQUEST FUNCTIONS
254 338
255=over 4 339=over 4
256 340
257=item $prev_pri = aioreq_pri [$pri] 341=item $prev_pri = aioreq_pri [$pri]
258 342
288 372
289 373
290=item aio_open $pathname, $flags, $mode, $callback->($fh) 374=item aio_open $pathname, $flags, $mode, $callback->($fh)
291 375
292Asynchronously open or create a file and call the callback with a newly 376Asynchronously open or create a file and call the callback with a newly
293created filehandle for the file. 377created filehandle for the file (or C<undef> in case of an error).
294 378
295The pathname passed to C<aio_open> must be absolute. See API NOTES, above, 379The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
296for an explanation. 380for an explanation.
297 381
298The C<$flags> argument is a bitmask. See the C<Fcntl> module for a 382The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
305by the umask in effect then the request is being executed, so better never 389by the umask in effect then the request is being executed, so better never
306change the umask. 390change the umask.
307 391
308Example: 392Example:
309 393
310 aio_open "/etc/passwd", O_RDONLY, 0, sub { 394 aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
311 if ($_[0]) { 395 if ($_[0]) {
312 print "open successful, fh is $_[0]\n"; 396 print "open successful, fh is $_[0]\n";
313 ... 397 ...
314 } else { 398 } else {
315 die "open failed: $!\n"; 399 die "open failed: $!\n";
316 } 400 }
317 }; 401 };
318 402
403In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
404C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
405following POSIX and non-POSIX constants are available (missing ones on
406your system are, as usual, C<0>):
407
408C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
409C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
410C<O_RSYNC>, C<O_SYNC>, C<O_PATH>, C<O_TMPFILE>, C<O_TTY_INIT> and C<O_ACCMODE>.
411
319 412
320=item aio_close $fh, $callback->($status) 413=item aio_close $fh, $callback->($status)
321 414
322Asynchronously close a file and call the callback with the result 415Asynchronously close a file and call the callback with the result
323code. 416code.
332Or in other words: the file descriptor will be closed, but it will not be 425Or in other words: the file descriptor will be closed, but it will not be
333free for reuse until the perl filehandle is closed. 426free for reuse until the perl filehandle is closed.
334 427
335=cut 428=cut
336 429
430=item aio_seek $fh, $offset, $whence, $callback->($offs)
431
432Seeks the filehandle to the new C<$offset>, similarly to perl's
433C<sysseek>. The C<$whence> can use the traditional values (C<0> for
434C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
435C<IO::AIO::SEEK_END>).
436
437The resulting absolute offset will be passed to the callback, or C<-1> in
438case of an error.
439
440In theory, the C<$whence> constants could be different than the
441corresponding values from L<Fcntl>, but perl guarantees they are the same,
442so don't panic.
443
444As a GNU/Linux (and maybe Solaris) extension, also the constants
445C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
446could be found. No guarantees about suitability for use in C<aio_seek> or
447Perl's C<sysseek> can be made though, although I would naively assume they
448"just work".
449
337=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 450=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
338 451
339=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) 452=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
340 453
341Reads or writes C<$length> bytes from the specified C<$fh> and C<$offset> 454Reads or writes C<$length> bytes from or to the specified C<$fh> and
342into the scalar given by C<$data> and offset C<$dataoffset> and calls the 455C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> and
343callback without the actual number of bytes read (or -1 on error, just 456calls the callback with the actual number of bytes transferred (or -1 on
344like the syscall). 457error, just like the syscall).
458
459C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
460offset plus the actual number of bytes read.
345 461
346If C<$offset> is undefined, then the current file descriptor offset will 462If C<$offset> is undefined, then the current file descriptor offset will
347be used (and updated), otherwise the file descriptor offset will not be 463be used (and updated), otherwise the file descriptor offset will not be
348changed by these calls. 464changed by these calls.
349 465
350If C<$length> is undefined in C<aio_write>, use the remaining length of C<$data>. 466If C<$length> is undefined in C<aio_write>, use the remaining length of
467C<$data>.
351 468
352If C<$dataoffset> is less than zero, it will be counted from the end of 469If C<$dataoffset> is less than zero, it will be counted from the end of
353C<$data>. 470C<$data>.
354 471
355The C<$data> scalar I<MUST NOT> be modified in any way while the request 472The C<$data> scalar I<MUST NOT> be modified in any way while the request
369 486
370Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts 487Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
371reading at byte offset C<$in_offset>, and starts writing at the current 488reading at byte offset C<$in_offset>, and starts writing at the current
372file offset of C<$out_fh>. Because of that, it is not safe to issue more 489file offset of C<$out_fh>. Because of that, it is not safe to issue more
373than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each 490than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
374other. 491other. The same C<$in_fh> works fine though, as this function does not
492move or use the file offset of C<$in_fh>.
375 493
494Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
495are written, and there is no way to find out how many more bytes have been
496read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
497number of bytes written to C<$out_fh>. Only if the result value equals
498C<$length> one can assume that C<$length> bytes have been read.
499
500Unlike with other C<aio_> functions, it makes a lot of sense to use
501C<aio_sendfile> on non-blocking sockets, as long as one end (typically
502the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
503the socket I/O will be non-blocking. Note, however, that you can run
504into a trap where C<aio_sendfile> reads some data with readahead, then
505fails to write all data, and when the socket is ready the next time, the
506data in the cache is already lost, forcing C<aio_sendfile> to again hit
507the disk. Explicit C<aio_read> + C<aio_write> let's you better control
508resource usage.
509
376This call tries to make use of a native C<sendfile> syscall to provide 510This call tries to make use of a native C<sendfile>-like syscall to
377zero-copy operation. For this to work, C<$out_fh> should refer to a 511provide zero-copy operation. For this to work, C<$out_fh> should refer to
378socket, and C<$in_fh> should refer to mmap'able file. 512a socket, and C<$in_fh> should refer to an mmap'able file.
379 513
380If the native sendfile call fails or is not implemented, it will be 514If a native sendfile cannot be found or it fails with C<ENOSYS>,
381emulated, so you can call C<aio_sendfile> on any type of filehandle 515C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
516C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
382regardless of the limitations of the operating system. 517type of filehandle regardless of the limitations of the operating system.
383 518
384Please note, however, that C<aio_sendfile> can read more bytes from 519As native sendfile syscalls (as practically any non-POSIX interface hacked
385C<$in_fh> than are written, and there is no way to find out how many 520together in a hurry to improve benchmark numbers) tend to be rather buggy
386bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only 521on many systems, this implementation tries to work around some known bugs
387provides the number of bytes written to C<$out_fh>. Only if the result 522in Linux and FreeBSD kernels (probably others, too), but that might fail,
388value equals C<$length> one can assume that C<$length> bytes have been 523so you really really should check the return value of C<aio_sendfile> -
389read. 524fewer bytes than expected might have been transferred.
390 525
391 526
392=item aio_readahead $fh,$offset,$length, $callback->($retval) 527=item aio_readahead $fh,$offset,$length, $callback->($retval)
393 528
394C<aio_readahead> populates the page cache with data from a file so that 529C<aio_readahead> populates the page cache with data from a file so that
398whole pages, so that offset is effectively rounded down to a page boundary 533whole pages, so that offset is effectively rounded down to a page boundary
399and bytes are read up to the next page boundary greater than or equal to 534and bytes are read up to the next page boundary greater than or equal to
400(off-set+length). C<aio_readahead> does not read beyond the end of the 535(off-set+length). C<aio_readahead> does not read beyond the end of the
401file. The current file offset of the file is left unchanged. 536file. The current file offset of the file is left unchanged.
402 537
403If that syscall doesn't exist (likely if your OS isn't Linux) it will be 538If that syscall doesn't exist (likely if your kernel isn't Linux) it will
404emulated by simply reading the data, which would have a similar effect. 539be emulated by simply reading the data, which would have a similar effect.
405 540
406 541
407=item aio_stat $fh_or_path, $callback->($status) 542=item aio_stat $fh_or_path, $callback->($status)
408 543
409=item aio_lstat $fh, $callback->($status) 544=item aio_lstat $fh, $callback->($status)
410 545
411Works like perl's C<stat> or C<lstat> in void context. The callback will 546Works almost exactly like perl's C<stat> or C<lstat> in void context. The
412be called after the stat and the results will be available using C<stat _> 547callback will be called after the stat and the results will be available
413or C<-s _> etc... 548using C<stat _> or C<-s _> and other tests (with the exception of C<-B>
549and C<-T>).
414 550
415The pathname passed to C<aio_stat> must be absolute. See API NOTES, above, 551The pathname passed to C<aio_stat> must be absolute. See API NOTES, above,
416for an explanation. 552for an explanation.
417 553
418Currently, the stats are always 64-bit-stats, i.e. instead of returning an 554Currently, the stats are always 64-bit-stats, i.e. instead of returning an
419error when stat'ing a large file, the results will be silently truncated 555error when stat'ing a large file, the results will be silently truncated
420unless perl itself is compiled with large file support. 556unless perl itself is compiled with large file support.
557
558To help interpret the mode and dev/rdev stat values, IO::AIO offers the
559following constants and functions (if not implemented, the constants will
560be C<0> and the functions will either C<croak> or fall back on traditional
561behaviour).
562
563C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
564C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
565C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
566
567To access higher resolution stat timestamps, see L<SUBSECOND STAT TIME
568ACCESS>.
421 569
422Example: Print the length of F</etc/passwd>: 570Example: Print the length of F</etc/passwd>:
423 571
424 aio_stat "/etc/passwd", sub { 572 aio_stat "/etc/passwd", sub {
425 $_[0] and die "stat failed: $!"; 573 $_[0] and die "stat failed: $!";
426 print "size is ", -s _, "\n"; 574 print "size is ", -s _, "\n";
427 }; 575 };
428 576
429 577
578=item aio_statvfs $fh_or_path, $callback->($statvfs)
579
580Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
581whether a file handle or path was passed.
582
583On success, the callback is passed a hash reference with the following
584members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
585C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
586is passed.
587
588The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
589C<ST_NOSUID>.
590
591The following non-POSIX IO::AIO::ST_* flag masks are defined to
592their correct value when available, or to C<0> on systems that do
593not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
594C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
595C<ST_NODIRATIME> and C<ST_RELATIME>.
596
597Example: stat C</wd> and dump out the data if successful.
598
599 aio_statvfs "/wd", sub {
600 my $f = $_[0]
601 or die "statvfs: $!";
602
603 use Data::Dumper;
604 say Dumper $f;
605 };
606
607 # result:
608 {
609 bsize => 1024,
610 bfree => 4333064312,
611 blocks => 10253828096,
612 files => 2050765568,
613 flag => 4096,
614 favail => 2042092649,
615 bavail => 4333064312,
616 ffree => 2042092649,
617 namemax => 255,
618 frsize => 1024,
619 fsid => 1810
620 }
621
430=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status) 622=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
431 623
432Works like perl's C<utime> function (including the special case of $atime 624Works like perl's C<utime> function (including the special case of $atime
433and $mtime being undef). Fractional times are supported if the underlying 625and $mtime being undef). Fractional times are supported if the underlying
434syscalls support them. 626syscalls support them.
435 627
436When called with a pathname, uses utimes(2) if available, otherwise 628When called with a pathname, uses utimensat(2) or utimes(2) if available,
437utime(2). If called on a file descriptor, uses futimes(2) if available, 629otherwise utime(2). If called on a file descriptor, uses futimens(2)
438otherwise returns ENOSYS, so this is not portable. 630or futimes(2) if available, otherwise returns ENOSYS, so this is not
631portable.
439 632
440Examples: 633Examples:
441 634
442 # set atime and mtime to current time (basically touch(1)): 635 # set atime and mtime to current time (basically touch(1)):
443 aio_utime "path", undef, undef; 636 aio_utime "path", undef, undef;
461=item aio_truncate $fh_or_path, $offset, $callback->($status) 654=item aio_truncate $fh_or_path, $offset, $callback->($status)
462 655
463Works like truncate(2) or ftruncate(2). 656Works like truncate(2) or ftruncate(2).
464 657
465 658
659=item aio_allocate $fh, $mode, $offset, $len, $callback->($status)
660
661Allocates or frees disk space according to the C<$mode> argument. See the
662linux C<fallocate> documentation for details.
663
664C<$mode> is usually C<0> or C<IO::AIO::FALLOC_FL_KEEP_SIZE> to allocate
665space, or C<IO::AIO::FALLOC_FL_PUNCH_HOLE | IO::AIO::FALLOC_FL_KEEP_SIZE>,
666to deallocate a file range.
667
668IO::AIO also supports C<FALLOC_FL_COLLAPSE_RANGE>, to remove a range
669(without leaving a hole), C<FALLOC_FL_ZERO_RANGE>, to zero a range,
670C<FALLOC_FL_INSERT_RANGE> to insert a range and C<FALLOC_FL_UNSHARE_RANGE>
671to unshare shared blocks (see your L<fallocate(2)> manpage).
672
673The file system block size used by C<fallocate> is presumably the
674C<f_bsize> returned by C<statvfs>, but different filesystems and filetypes
675can dictate other limitations.
676
677If C<fallocate> isn't available or cannot be emulated (currently no
678emulation will be attempted), passes C<-1> and sets C<$!> to C<ENOSYS>.
679
680
466=item aio_chmod $fh_or_path, $mode, $callback->($status) 681=item aio_chmod $fh_or_path, $mode, $callback->($status)
467 682
468Works like perl's C<chmod> function. 683Works like perl's C<chmod> function.
469 684
470 685
472 687
473Asynchronously unlink (delete) a file and call the callback with the 688Asynchronously unlink (delete) a file and call the callback with the
474result code. 689result code.
475 690
476 691
477=item aio_mknod $path, $mode, $dev, $callback->($status) 692=item aio_mknod $pathname, $mode, $dev, $callback->($status)
478 693
479[EXPERIMENTAL] 694[EXPERIMENTAL]
480 695
481Asynchronously create a device node (or fifo). See mknod(2). 696Asynchronously create a device node (or fifo). See mknod(2).
482 697
483The only (POSIX-) portable way of calling this function is: 698The only (POSIX-) portable way of calling this function is:
484 699
485 aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ... 700 aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
486 701
702See C<aio_stat> for info about some potentially helpful extra constants
703and functions.
487 704
488=item aio_link $srcpath, $dstpath, $callback->($status) 705=item aio_link $srcpath, $dstpath, $callback->($status)
489 706
490Asynchronously create a new link to the existing object at C<$srcpath> at 707Asynchronously create a new link to the existing object at C<$srcpath> at
491the path C<$dstpath> and call the callback with the result code. 708the path C<$dstpath> and call the callback with the result code.
495 712
496Asynchronously create a new symbolic link to the existing object at C<$srcpath> at 713Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
497the path C<$dstpath> and call the callback with the result code. 714the path C<$dstpath> and call the callback with the result code.
498 715
499 716
500=item aio_readlink $path, $callback->($link) 717=item aio_readlink $pathname, $callback->($link)
501 718
502Asynchronously read the symlink specified by C<$path> and pass it to 719Asynchronously read the symlink specified by C<$path> and pass it to
503the callback. If an error occurs, nothing or undef gets passed to the 720the callback. If an error occurs, nothing or undef gets passed to the
504callback. 721callback.
505 722
506 723
724=item aio_realpath $pathname, $callback->($path)
725
726Asynchronously make the path absolute and resolve any symlinks in
727C<$path>. The resulting path only consists of directories (same as
728L<Cwd::realpath>).
729
730This request can be used to get the absolute path of the current working
731directory by passing it a path of F<.> (a single dot).
732
733
507=item aio_rename $srcpath, $dstpath, $callback->($status) 734=item aio_rename $srcpath, $dstpath, $callback->($status)
508 735
509Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as 736Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
510rename(2) and call the callback with the result code. 737rename(2) and call the callback with the result code.
738
739On systems that support the AIO::WD working directory abstraction
740natively, the case C<[$wd, "."]> as C<$srcpath> is specialcased - instead
741of failing, C<rename> is called on the absolute path of C<$wd>.
742
743
744=item aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
745
746Basically a version of C<aio_rename> with an additional C<$flags>
747argument. Calling this with C<$flags=0> is the same as calling
748C<aio_rename>.
749
750Non-zero flags are currently only supported on GNU/Linux systems that
751support renameat2. Other systems fail with C<ENOSYS> in this case.
752
753The following constants are available (missing ones are, as usual C<0>),
754see renameat2(2) for details:
755
756C<IO::AIO::RENAME_NOREPLACE>, C<IO::AIO::RENAME_EXCHANGE>
757and C<IO::AIO::RENAME_WHITEOUT>.
511 758
512 759
513=item aio_mkdir $pathname, $mode, $callback->($status) 760=item aio_mkdir $pathname, $mode, $callback->($status)
514 761
515Asynchronously mkdir (create) a directory and call the callback with 762Asynchronously mkdir (create) a directory and call the callback with
520=item aio_rmdir $pathname, $callback->($status) 767=item aio_rmdir $pathname, $callback->($status)
521 768
522Asynchronously rmdir (delete) a directory and call the callback with the 769Asynchronously rmdir (delete) a directory and call the callback with the
523result code. 770result code.
524 771
772On systems that support the AIO::WD working directory abstraction
773natively, the case C<[$wd, "."]> is specialcased - instead of failing,
774C<rmdir> is called on the absolute path of C<$wd>.
775
525 776
526=item aio_readdir $pathname, $callback->($entries) 777=item aio_readdir $pathname, $callback->($entries)
527 778
528Unlike the POSIX call of the same name, C<aio_readdir> reads an entire 779Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
529directory (i.e. opendir + readdir + closedir). The entries will not be 780directory (i.e. opendir + readdir + closedir). The entries will not be
530sorted, and will B<NOT> include the C<.> and C<..> entries. 781sorted, and will B<NOT> include the C<.> and C<..> entries.
531 782
532The callback a single argument which is either C<undef> or an array-ref 783The callback is passed a single argument which is either C<undef> or an
533with the filenames. 784array-ref with the filenames.
534 785
535 786
787=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
788
789Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
790tune behaviour and output format. In case of an error, C<$entries> will be
791C<undef>.
792
793The flags are a combination of the following constants, ORed together (the
794flags will also be passed to the callback, possibly modified):
795
796=over 4
797
798=item IO::AIO::READDIR_DENTS
799
800Normally the callback gets an arrayref consisting of names only (as
801with C<aio_readdir>). If this flag is set, then the callback gets an
802arrayref with C<[$name, $type, $inode]> arrayrefs, each describing a
803single directory entry in more detail:
804
805C<$name> is the name of the entry.
806
807C<$type> is one of the C<IO::AIO::DT_xxx> constants:
808
809C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
810C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
811C<IO::AIO::DT_WHT>.
812
813C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need
814to know, you have to run stat yourself. Also, for speed/memory reasons,
815the C<$type> scalars are read-only: you must not modify them.
816
817C<$inode> is the inode number (which might not be exact on systems with 64
818bit inode numbers and 32 bit perls). This field has unspecified content on
819systems that do not deliver the inode information.
820
821=item IO::AIO::READDIR_DIRS_FIRST
822
823When this flag is set, then the names will be returned in an order where
824likely directories come first, in optimal stat order. This is useful when
825you need to quickly find directories, or you want to find all directories
826while avoiding to stat() each entry.
827
828If the system returns type information in readdir, then this is used
829to find directories directly. Otherwise, likely directories are names
830beginning with ".", or otherwise names with no dots, of which names with
831short names are tried first.
832
833=item IO::AIO::READDIR_STAT_ORDER
834
835When this flag is set, then the names will be returned in an order
836suitable for stat()'ing each one. That is, when you plan to stat() most or
837all files in the given directory, then the returned order will likely be
838faster.
839
840If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified,
841then the likely dirs come first, resulting in a less optimal stat order
842for stat'ing all entries, but likely a more optimal order for finding
843subdirectories.
844
845=item IO::AIO::READDIR_FOUND_UNKNOWN
846
847This flag should not be set when calling C<aio_readdirx>. Instead, it
848is being set by C<aio_readdirx>, when any of the C<$type>'s found were
849C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
850C<$type>'s are known, which can be used to speed up some algorithms.
851
852=back
853
854
855=item aio_slurp $pathname, $offset, $length, $data, $callback->($status)
856
857Opens, reads and closes the given file. The data is put into C<$data>,
858which is resized as required.
859
860If C<$offset> is negative, then it is counted from the end of the file.
861
862If C<$length> is zero, then the remaining length of the file is
863used. Also, in this case, the same limitations to modifying C<$data> apply
864as when IO::AIO::mmap is used, i.e. it must only be modified in-place
865with C<substr>. If the size of the file is known, specifying a non-zero
866C<$length> results in a performance advantage.
867
868This request is similar to the older C<aio_load> request, but since it is
869a single request, it might be more efficient to use.
870
871Example: load F</etc/passwd> into C<$passwd>.
872
873 my $passwd;
874 aio_slurp "/etc/passwd", 0, 0, $passwd, sub {
875 $_[0] >= 0
876 or die "/etc/passwd: $!\n";
877
878 printf "/etc/passwd is %d bytes long, and contains:\n", length $passwd;
879 print $passwd;
880 };
881 IO::AIO::flush;
882
883
536=item aio_load $path, $data, $callback->($status) 884=item aio_load $pathname, $data, $callback->($status)
537 885
538This is a composite request that tries to fully load the given file into 886This is a composite request that tries to fully load the given file into
539memory. Status is the same as with aio_read. 887memory. Status is the same as with aio_read.
888
889Using C<aio_slurp> might be more efficient, as it is a single request.
540 890
541=cut 891=cut
542 892
543sub aio_load($$;$) { 893sub aio_load($$;$) {
544 my ($path, undef, $cb) = @_; 894 my ($path, undef, $cb) = @_;
563 913
564=item aio_copy $srcpath, $dstpath, $callback->($status) 914=item aio_copy $srcpath, $dstpath, $callback->($status)
565 915
566Try to copy the I<file> (directories not supported as either source or 916Try to copy the I<file> (directories not supported as either source or
567destination) from C<$srcpath> to C<$dstpath> and call the callback with 917destination) from C<$srcpath> to C<$dstpath> and call the callback with
568the C<0> (error) or C<-1> ok. 918a status of C<0> (ok) or C<-1> (error, see C<$!>).
919
920Existing destination files will be truncated.
569 921
570This is a composite request that creates the destination file with 922This is a composite request that creates the destination file with
571mode 0200 and copies the contents of the source file into it using 923mode 0200 and copies the contents of the source file into it using
572C<aio_sendfile>, followed by restoring atime, mtime, access mode and 924C<aio_sendfile>, followed by restoring atime, mtime, access mode and
573uid/gid, in that order. 925uid/gid, in that order.
585 my $grp = aio_group $cb; 937 my $grp = aio_group $cb;
586 938
587 aioreq_pri $pri; 939 aioreq_pri $pri;
588 add $grp aio_open $src, O_RDONLY, 0, sub { 940 add $grp aio_open $src, O_RDONLY, 0, sub {
589 if (my $src_fh = $_[0]) { 941 if (my $src_fh = $_[0]) {
590 my @stat = stat $src_fh; 942 my @stat = stat $src_fh; # hmm, might block over nfs?
591 943
592 aioreq_pri $pri; 944 aioreq_pri $pri;
593 add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub { 945 add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
594 if (my $dst_fh = $_[0]) { 946 if (my $dst_fh = $_[0]) {
595 aioreq_pri $pri; 947 aioreq_pri $pri;
596 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub { 948 add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
597 if ($_[0] == $stat[7]) { 949 if ($_[0] == $stat[7]) {
598 $grp->result (0); 950 $grp->result (0);
599 close $src_fh; 951 close $src_fh;
600 952
601 # those should not normally block. should. should. 953 my $ch = sub {
602 utime $stat[8], $stat[9], $dst; 954 aioreq_pri $pri;
603 chmod $stat[2] & 07777, $dst_fh; 955 add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
604 chown $stat[4], $stat[5], $dst_fh; 956 aioreq_pri $pri;
957 add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
958 aioreq_pri $pri;
959 add $grp aio_close $dst_fh;
960 }
961 };
962 };
605 963
606 aioreq_pri $pri; 964 aioreq_pri $pri;
607 add $grp aio_close $dst_fh; 965 add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
966 if ($_[0] < 0 && $! == ENOSYS) {
967 aioreq_pri $pri;
968 add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
969 } else {
970 $ch->();
971 }
972 };
608 } else { 973 } else {
609 $grp->result (-1); 974 $grp->result (-1);
610 close $src_fh; 975 close $src_fh;
611 close $dst_fh; 976 close $dst_fh;
612 977
629 994
630=item aio_move $srcpath, $dstpath, $callback->($status) 995=item aio_move $srcpath, $dstpath, $callback->($status)
631 996
632Try to move the I<file> (directories not supported as either source or 997Try to move the I<file> (directories not supported as either source or
633destination) from C<$srcpath> to C<$dstpath> and call the callback with 998destination) from C<$srcpath> to C<$dstpath> and call the callback with
634the C<0> (error) or C<-1> ok. 999a status of C<0> (ok) or C<-1> (error, see C<$!>).
635 1000
636This is a composite request that tries to rename(2) the file first. If 1001This is a composite request that tries to rename(2) the file first; if
637rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if 1002rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
638that is successful, unlinking the C<$srcpath>. 1003that is successful, unlinks the C<$srcpath>.
639 1004
640=cut 1005=cut
641 1006
642sub aio_move($$;$) { 1007sub aio_move($$;$) {
643 my ($src, $dst, $cb) = @_; 1008 my ($src, $dst, $cb) = @_;
650 if ($_[0] && $! == EXDEV) { 1015 if ($_[0] && $! == EXDEV) {
651 aioreq_pri $pri; 1016 aioreq_pri $pri;
652 add $grp aio_copy $src, $dst, sub { 1017 add $grp aio_copy $src, $dst, sub {
653 $grp->result ($_[0]); 1018 $grp->result ($_[0]);
654 1019
655 if (!$_[0]) { 1020 unless ($_[0]) {
656 aioreq_pri $pri; 1021 aioreq_pri $pri;
657 add $grp aio_unlink $src; 1022 add $grp aio_unlink $src;
658 } 1023 }
659 }; 1024 };
660 } else { 1025 } else {
663 }; 1028 };
664 1029
665 $grp 1030 $grp
666} 1031}
667 1032
668=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) 1033=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
669 1034
670Scans a directory (similar to C<aio_readdir>) but additionally tries to 1035Scans a directory (similar to C<aio_readdir>) but additionally tries to
671efficiently separate the entries of directory C<$path> into two sets of 1036efficiently separate the entries of directory C<$path> into two sets of
672names, directories you can recurse into (directories), and ones you cannot 1037names, directories you can recurse into (directories), and ones you cannot
673recurse into (everything else, including symlinks to directories). 1038recurse into (everything else, including symlinks to directories).
674 1039
675C<aio_scandir> is a composite request that creates of many sub requests_ 1040C<aio_scandir> is a composite request that generates many sub requests.
676C<$maxreq> specifies the maximum number of outstanding aio requests that 1041C<$maxreq> specifies the maximum number of outstanding aio requests that
677this function generates. If it is C<< <= 0 >>, then a suitable default 1042this function generates. If it is C<< <= 0 >>, then a suitable default
678will be chosen (currently 4). 1043will be chosen (currently 4).
679 1044
680On error, the callback is called without arguments, otherwise it receives 1045On error, the callback is called without arguments, otherwise it receives
690 1055
691Implementation notes. 1056Implementation notes.
692 1057
693The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. 1058The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
694 1059
1060If readdir returns file type information, then this is used directly to
1061find directories.
1062
695After reading the directory, the modification time, size etc. of the 1063Otherwise, after reading the directory, the modification time, size etc.
696directory before and after the readdir is checked, and if they match (and 1064of the directory before and after the readdir is checked, and if they
697isn't the current time), the link count will be used to decide how many 1065match (and isn't the current time), the link count will be used to decide
698entries are directories (if >= 2). Otherwise, no knowledge of the number 1066how many entries are directories (if >= 2). Otherwise, no knowledge of the
699of subdirectories will be assumed. 1067number of subdirectories will be assumed.
700 1068
701Then entries will be sorted into likely directories (everything without 1069Then entries will be sorted into likely directories a non-initial dot
702a non-initial dot currently) and likely non-directories (everything 1070currently) and likely non-directories (see C<aio_readdirx>). Then every
703else). Then every entry plus an appended C</.> will be C<stat>'ed, 1071entry plus an appended C</.> will be C<stat>'ed, likely directories first,
704likely directories first. If that succeeds, it assumes that the entry 1072in order of their inode numbers. If that succeeds, it assumes that the
705is a directory or a symlink to directory (which will be checked 1073entry is a directory or a symlink to directory (which will be checked
706seperately). This is often faster than stat'ing the entry itself because 1074separately). This is often faster than stat'ing the entry itself because
707filesystems might detect the type of the entry without reading the inode 1075filesystems might detect the type of the entry without reading the inode
708data (e.g. ext2fs filetype feature). 1076data (e.g. ext2fs filetype feature), even on systems that cannot return
1077the filetype information on readdir.
709 1078
710If the known number of directories (link count - 2) has been reached, the 1079If the known number of directories (link count - 2) has been reached, the
711rest of the entries is assumed to be non-directories. 1080rest of the entries is assumed to be non-directories.
712 1081
713This only works with certainty on POSIX (= UNIX) filesystems, which 1082This only works with certainty on POSIX (= UNIX) filesystems, which
726 1095
727 my $grp = aio_group $cb; 1096 my $grp = aio_group $cb;
728 1097
729 $maxreq = 4 if $maxreq <= 0; 1098 $maxreq = 4 if $maxreq <= 0;
730 1099
731 # stat once 1100 # get a wd object
732 aioreq_pri $pri; 1101 aioreq_pri $pri;
733 add $grp aio_stat $path, sub { 1102 add $grp aio_wd $path, sub {
1103 $_[0]
734 return $grp->result () if $_[0]; 1104 or return $grp->result ();
735 my $now = time;
736 my $hash1 = join ":", (stat _)[0,1,3,7,9];
737 1105
738 # read the directory entries 1106 my $wd = [shift, "."];
1107
1108 # stat once
739 aioreq_pri $pri; 1109 aioreq_pri $pri;
740 add $grp aio_readdir $path, sub { 1110 add $grp aio_stat $wd, sub {
741 my $entries = shift
742 or return $grp->result (); 1111 return $grp->result () if $_[0];
1112 my $now = time;
1113 my $hash1 = join ":", (stat _)[0,1,3,7,9];
743 1114
744 # stat the dir another time 1115 # read the directory entries
745 aioreq_pri $pri; 1116 aioreq_pri $pri;
1117 add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
1118 my $entries = shift
1119 or return $grp->result ();
1120
1121 # stat the dir another time
1122 aioreq_pri $pri;
746 add $grp aio_stat $path, sub { 1123 add $grp aio_stat $wd, sub {
747 my $hash2 = join ":", (stat _)[0,1,3,7,9]; 1124 my $hash2 = join ":", (stat _)[0,1,3,7,9];
748 1125
749 my $ndirs; 1126 my $ndirs;
750 1127
751 # take the slow route if anything looks fishy 1128 # take the slow route if anything looks fishy
752 if ($hash1 ne $hash2 or (stat _)[9] == $now) { 1129 if ($hash1 ne $hash2 or (stat _)[9] == $now) {
753 $ndirs = -1; 1130 $ndirs = -1;
754 } else { 1131 } else {
755 # if nlink == 2, we are finished 1132 # if nlink == 2, we are finished
756 # on non-posix-fs's, we rely on nlink < 2 1133 # for non-posix-fs's, we rely on nlink < 2
757 $ndirs = (stat _)[3] - 2 1134 $ndirs = (stat _)[3] - 2
758 or return $grp->result ([], $entries); 1135 or return $grp->result ([], $entries);
759 } 1136 }
760 1137
761 # sort into likely dirs and likely nondirs
762 # dirs == files without ".", short entries first
763 $entries = [map $_->[0],
764 sort { $b->[1] cmp $a->[1] }
765 map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
766 @$entries];
767
768 my (@dirs, @nondirs); 1138 my (@dirs, @nondirs);
769 1139
770 my $statgrp = add $grp aio_group sub { 1140 my $statgrp = add $grp aio_group sub {
771 $grp->result (\@dirs, \@nondirs); 1141 $grp->result (\@dirs, \@nondirs);
772 }; 1142 };
773 1143
774 limit $statgrp $maxreq; 1144 limit $statgrp $maxreq;
775 feed $statgrp sub { 1145 feed $statgrp sub {
776 return unless @$entries; 1146 return unless @$entries;
777 my $entry = pop @$entries; 1147 my $entry = shift @$entries;
778 1148
779 aioreq_pri $pri; 1149 aioreq_pri $pri;
1150 $wd->[1] = "$entry/.";
780 add $statgrp aio_stat "$path/$entry/.", sub { 1151 add $statgrp aio_stat $wd, sub {
781 if ($_[0] < 0) { 1152 if ($_[0] < 0) {
782 push @nondirs, $entry; 1153 push @nondirs, $entry;
783 } else { 1154 } else {
784 # need to check for real directory 1155 # need to check for real directory
785 aioreq_pri $pri; 1156 aioreq_pri $pri;
1157 $wd->[1] = $entry;
786 add $statgrp aio_lstat "$path/$entry", sub { 1158 add $statgrp aio_lstat $wd, sub {
787 if (-d _) { 1159 if (-d _) {
788 push @dirs, $entry; 1160 push @dirs, $entry;
789 1161
790 unless (--$ndirs) { 1162 unless (--$ndirs) {
791 push @nondirs, @$entries; 1163 push @nondirs, @$entries;
792 feed $statgrp; 1164 feed $statgrp;
1165 }
1166 } else {
1167 push @nondirs, $entry;
793 } 1168 }
794 } else {
795 push @nondirs, $entry;
796 } 1169 }
797 } 1170 }
798 } 1171 };
799 }; 1172 };
800 }; 1173 };
801 }; 1174 };
802 }; 1175 };
803 }; 1176 };
804 1177
805 $grp 1178 $grp
806} 1179}
807 1180
808=item aio_rmtree $path, $callback->($status) 1181=item aio_rmtree $pathname, $callback->($status)
809 1182
810Delete a directory tree starting (and including) C<$path>, return the 1183Delete a directory tree starting (and including) C<$path>, return the
811status of the final C<rmdir> only. This is a composite request that 1184status of the final C<rmdir> only. This is a composite request that
812uses C<aio_scandir> to recurse into and rmdir directories, and unlink 1185uses C<aio_scandir> to recurse into and rmdir directories, and unlink
813everything else. 1186everything else.
814 1187
815=cut 1188=cut
816 1189
838 }; 1211 };
839 1212
840 $grp 1213 $grp
841} 1214}
842 1215
1216=item aio_fcntl $fh, $cmd, $arg, $callback->($status)
1217
1218=item aio_ioctl $fh, $request, $buf, $callback->($status)
1219
1220These work just like the C<fcntl> and C<ioctl> built-in functions, except
1221they execute asynchronously and pass the return value to the callback.
1222
1223Both calls can be used for a lot of things, some of which make more sense
1224to run asynchronously in their own thread, while some others make less
1225sense. For example, calls that block waiting for external events, such
1226as locking, will also lock down an I/O thread while it is waiting, which
1227can deadlock the whole I/O system. At the same time, there might be no
1228alternative to using a thread to wait.
1229
1230So in general, you should only use these calls for things that do
1231(filesystem) I/O, not for things that wait for other events (network,
1232other processes), although if you are careful and know what you are doing,
1233you still can.
1234
1235The following constants are available (missing ones are, as usual C<0>):
1236
1237C<F_DUPFD_CLOEXEC>,
1238
1239C<F_OFD_GETLK>, C<F_OFD_SETLK>, C<F_OFD_GETLKW>,
1240
1241C<FIFREEZE>, C<FITHAW>, C<FITRIM>, C<FICLONE>, C<FICLONERANGE>, C<FIDEDUPERANGE>.
1242
1243C<FS_IOC_GETFLAGS>, C<FS_IOC_SETFLAGS>, C<FS_IOC_GETVERSION>, C<FS_IOC_SETVERSION>,
1244C<FS_IOC_FIEMAP>.
1245
1246C<FS_IOC_FSGETXATTR>, C<FS_IOC_FSSETXATTR>, C<FS_IOC_SET_ENCRYPTION_POLICY>,
1247C<FS_IOC_GET_ENCRYPTION_PWSALT>, C<FS_IOC_GET_ENCRYPTION_POLICY>, C<FS_KEY_DESCRIPTOR_SIZE>.
1248
1249C<FS_SECRM_FL>, C<FS_UNRM_FL>, C<FS_COMPR_FL>, C<FS_SYNC_FL>, C<FS_IMMUTABLE_FL>,
1250C<FS_APPEND_FL>, C<FS_NODUMP_FL>, C<FS_NOATIME_FL>, C<FS_DIRTY_FL>,
1251C<FS_COMPRBLK_FL>, C<FS_NOCOMP_FL>, C<FS_ENCRYPT_FL>, C<FS_BTREE_FL>,
1252C<FS_INDEX_FL>, C<FS_JOURNAL_DATA_FL>, C<FS_NOTAIL_FL>, C<FS_DIRSYNC_FL>, C<FS_TOPDIR_FL>,
1253C<FS_FL_USER_MODIFIABLE>.
1254
1255C<FS_XFLAG_REALTIME>, C<FS_XFLAG_PREALLOC>, C<FS_XFLAG_IMMUTABLE>, C<FS_XFLAG_APPEND>,
1256C<FS_XFLAG_SYNC>, C<FS_XFLAG_NOATIME>, C<FS_XFLAG_NODUMP>, C<FS_XFLAG_RTINHERIT>,
1257C<FS_XFLAG_PROJINHERIT>, C<FS_XFLAG_NOSYMLINKS>, C<FS_XFLAG_EXTSIZE>, C<FS_XFLAG_EXTSZINHERIT>,
1258C<FS_XFLAG_NODEFRAG>, C<FS_XFLAG_FILESTREAM>, C<FS_XFLAG_DAX>, C<FS_XFLAG_HASATTR>,
1259
843=item aio_sync $callback->($status) 1260=item aio_sync $callback->($status)
844 1261
845Asynchronously call sync and call the callback when finished. 1262Asynchronously call sync and call the callback when finished.
846 1263
847=item aio_fsync $fh, $callback->($status) 1264=item aio_fsync $fh, $callback->($status)
855callback with the fdatasync result code. 1272callback with the fdatasync result code.
856 1273
857If this call isn't available because your OS lacks it or it couldn't be 1274If this call isn't available because your OS lacks it or it couldn't be
858detected, it will be emulated by calling C<fsync> instead. 1275detected, it will be emulated by calling C<fsync> instead.
859 1276
1277=item aio_syncfs $fh, $callback->($status)
1278
1279Asynchronously call the syncfs syscall to sync the filesystem associated
1280to the given filehandle and call the callback with the syncfs result
1281code. If syncfs is not available, calls sync(), but returns C<-1> and sets
1282errno to C<ENOSYS> nevertheless.
1283
1284=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
1285
1286Sync the data portion of the file specified by C<$offset> and C<$length>
1287to disk (but NOT the metadata), by calling the Linux-specific
1288sync_file_range call. If sync_file_range is not available or it returns
1289ENOSYS, then fdatasync or fsync is being substituted.
1290
1291C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
1292C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
1293C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
1294manpage for details.
1295
860=item aio_pathsync $path, $callback->($status) 1296=item aio_pathsync $pathname, $callback->($status)
861 1297
862This request tries to open, fsync and close the given path. This is a 1298This request tries to open, fsync and close the given path. This is a
863composite request intended to sync directories after directory operations 1299composite request intended to sync directories after directory operations
864(E.g. rename). This might not work on all operating systems or have any 1300(E.g. rename). This might not work on all operating systems or have any
865specific effect, but usually it makes sure that directory changes get 1301specific effect, but usually it makes sure that directory changes get
866written to disc. It works for anything that can be opened for read-only, 1302written to disc. It works for anything that can be opened for read-only,
867not just directories. 1303not just directories.
1304
1305Future versions of this function might fall back to other methods when
1306C<fsync> on the directory fails (such as calling C<sync>).
868 1307
869Passes C<0> when everything went ok, and C<-1> on error. 1308Passes C<0> when everything went ok, and C<-1> on error.
870 1309
871=cut 1310=cut
872 1311
893 }; 1332 };
894 1333
895 $grp 1334 $grp
896} 1335}
897 1336
1337=item aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
1338
1339This is a rather advanced IO::AIO call, which only works on mmap(2)ed
1340scalars (see the C<IO::AIO::mmap> function, although it also works on data
1341scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
1342scalar must only be modified in-place while an aio operation is pending on
1343it).
1344
1345It calls the C<msync> function of your OS, if available, with the memory
1346area starting at C<$offset> in the string and ending C<$length> bytes
1347later. If C<$length> is negative, counts from the end, and if C<$length>
1348is C<undef>, then it goes till the end of the string. The flags can be
1349either C<IO::AIO::MS_ASYNC> or C<IO::AIO::MS_SYNC>, plus an optional
1350C<IO::AIO::MS_INVALIDATE>.
1351
1352=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
1353
1354This is a rather advanced IO::AIO call, which works best on mmap(2)ed
1355scalars.
1356
1357It touches (reads or writes) all memory pages in the specified
1358range inside the scalar. All caveats and parameters are the same
1359as for C<aio_msync>, above, except for flags, which must be either
1360C<0> (which reads all pages and ensures they are instantiated) or
1361C<IO::AIO::MT_MODIFY>, which modifies the memory pages (by reading and
1362writing an octet from it, which dirties the page).
1363
1364=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
1365
1366This is a rather advanced IO::AIO call, which works best on mmap(2)ed
1367scalars.
1368
1369It reads in all the pages of the underlying storage into memory (if any)
1370and locks them, so they are not getting swapped/paged out or removed.
1371
1372If C<$length> is undefined, then the scalar will be locked till the end.
1373
1374On systems that do not implement C<mlock>, this function returns C<-1>
1375and sets errno to C<ENOSYS>.
1376
1377Note that the corresponding C<munlock> is synchronous and is
1378documented under L<MISCELLANEOUS FUNCTIONS>.
1379
1380Example: open a file, mmap and mlock it - both will be undone when
1381C<$data> gets destroyed.
1382
1383 open my $fh, "<", $path or die "$path: $!";
1384 my $data;
1385 IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
1386 aio_mlock $data; # mlock in background
1387
1388=item aio_mlockall $flags, $callback->($status)
1389
1390Calls the C<mlockall> function with the given C<$flags> (a combination of
1391C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
1392
1393On systems that do not implement C<mlockall>, this function returns C<-1>
1394and sets errno to C<ENOSYS>.
1395
1396Note that the corresponding C<munlockall> is synchronous and is
1397documented under L<MISCELLANEOUS FUNCTIONS>.
1398
1399Example: asynchronously lock all current and future pages into memory.
1400
1401 aio_mlockall IO::AIO::MCL_FUTURE;
1402
1403=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
1404
1405Queries the extents of the given file (by calling the Linux C<FIEMAP>
1406ioctl, see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If
1407the ioctl is not available on your OS, then this request will fail with
1408C<ENOSYS>.
1409
1410C<$start> is the starting offset to query extents for, C<$length> is the
1411size of the range to query - if it is C<undef>, then the whole file will
1412be queried.
1413
1414C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
1415C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
1416exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
1417the data portion.
1418
1419C<$count> is the maximum number of extent records to return. If it is
1420C<undef>, then IO::AIO queries all extents of the range. As a very special
1421case, if it is C<0>, then the callback receives the number of extents
1422instead of the extents themselves (which is unreliable, see below).
1423
1424If an error occurs, the callback receives no arguments. The special
1425C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
1426
1427Otherwise, the callback receives an array reference with extent
1428structures. Each extent structure is an array reference itself, with the
1429following members:
1430
1431 [$logical, $physical, $length, $flags]
1432
1433Flags is any combination of the following flag values (typically either C<0>
1434or C<IO::AIO::FIEMAP_EXTENT_LAST> (1)):
1435
1436C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
1437C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
1438C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
1439C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
1440C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
1441C<IO::AIO::FIEMAP_EXTENT_SHARED>.
1442
1443At the time of this writing (Linux 3.2), this request is unreliable unless
1444C<$count> is C<undef>, as the kernel has all sorts of bugs preventing
1445it to return all extents of a range for files with a large number of
1446extents. The code (only) works around all these issues if C<$count> is
1447C<undef>.
1448
898=item aio_group $callback->(...) 1449=item aio_group $callback->(...)
899 1450
900This is a very special aio request: Instead of doing something, it is a 1451This is a very special aio request: Instead of doing something, it is a
901container for other aio requests, which is useful if you want to bundle 1452container for other aio requests, which is useful if you want to bundle
902many requests into a single, composite, request with a definite callback 1453many requests into a single, composite, request with a definite callback
939immense (it blocks a thread for a long time) so do not use this function 1490immense (it blocks a thread for a long time) so do not use this function
940except to put your application under artificial I/O pressure. 1491except to put your application under artificial I/O pressure.
941 1492
942=back 1493=back
943 1494
1495
1496=head2 IO::AIO::WD - multiple working directories
1497
1498Your process only has one current working directory, which is used by all
1499threads. This makes it hard to use relative paths (some other component
1500could call C<chdir> at any time, and it is hard to control when the path
1501will be used by IO::AIO).
1502
1503One solution for this is to always use absolute paths. This usually works,
1504but can be quite slow (the kernel has to walk the whole path on every
1505access), and can also be a hassle to implement.
1506
1507Newer POSIX systems have a number of functions (openat, fdopendir,
1508futimensat and so on) that make it possible to specify working directories
1509per operation.
1510
1511For portability, and because the clowns who "designed", or shall I write,
1512perpetrated this new interface were obviously half-drunk, this abstraction
1513cannot be perfect, though.
1514
1515IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
1516object. This object stores the canonicalised, absolute version of the
1517path, and on systems that allow it, also a directory file descriptor.
1518
1519Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
1520or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
1521object and a pathname instead (or the IO::AIO::WD object alone, which
1522gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
1523IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
1524to that IO::AIO::WD object.
1525
1526For example, to get a wd object for F</etc> and then stat F<passwd>
1527inside, you would write:
1528
1529 aio_wd "/etc", sub {
1530 my $etcdir = shift;
1531
1532 # although $etcdir can be undef on error, there is generally no reason
1533 # to check for errors here, as aio_stat will fail with ENOENT
1534 # when $etcdir is undef.
1535
1536 aio_stat [$etcdir, "passwd"], sub {
1537 # yay
1538 };
1539 };
1540
1541The fact that C<aio_wd> is a request and not a normal function shows that
1542creating an IO::AIO::WD object is itself a potentially blocking operation,
1543which is why it is done asynchronously.
1544
1545To stat the directory obtained with C<aio_wd> above, one could write
1546either of the following three request calls:
1547
1548 aio_lstat "/etc" , sub { ... # pathname as normal string
1549 aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
1550 aio_lstat $wd , sub { ... # shorthand for the previous
1551
1552As with normal pathnames, IO::AIO keeps a copy of the working directory
1553object and the pathname string, so you could write the following without
1554causing any issues due to C<$path> getting reused:
1555
1556 my $path = [$wd, undef];
1557
1558 for my $name (qw(abc def ghi)) {
1559 $path->[1] = $name;
1560 aio_stat $path, sub {
1561 # ...
1562 };
1563 }
1564
1565There are some caveats: when directories get renamed (or deleted), the
1566pathname string doesn't change, so will point to the new directory (or
1567nowhere at all), while the directory fd, if available on the system,
1568will still point to the original directory. Most functions accepting a
1569pathname will use the directory fd on newer systems, and the string on
1570older systems. Some functions (such as C<aio_realpath>) will always rely on
1571the string form of the pathname.
1572
1573So this functionality is mainly useful to get some protection against
1574C<chdir>, to easily get an absolute path out of a relative path for future
1575reference, and to speed up doing many operations in the same directory
1576(e.g. when stat'ing all files in a directory).
1577
1578The following functions implement this working directory abstraction:
1579
1580=over 4
1581
1582=item aio_wd $pathname, $callback->($wd)
1583
1584Asynchonously canonicalise the given pathname and convert it to an
1585IO::AIO::WD object representing it. If possible and supported on the
1586system, also open a directory fd to speed up pathname resolution relative
1587to this working directory.
1588
1589If something goes wrong, then C<undef> is passwd to the callback instead
1590of a working directory object and C<$!> is set appropriately. Since
1591passing C<undef> as working directory component of a pathname fails the
1592request with C<ENOENT>, there is often no need for error checking in the
1593C<aio_wd> callback, as future requests using the value will fail in the
1594expected way.
1595
1596=item IO::AIO::CWD
1597
1598This is a compiletime constant (object) that represents the process
1599current working directory.
1600
1601Specifying this object as working directory object for a pathname is as if
1602the pathname would be specified directly, without a directory object. For
1603example, these calls are functionally identical:
1604
1605 aio_stat "somefile", sub { ... };
1606 aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
1607
1608=back
1609
1610To recover the path associated with an IO::AIO::WD object, you can use
1611C<aio_realpath>:
1612
1613 aio_realpath $wd, sub {
1614 warn "path is $_[0]\n";
1615 };
1616
1617Currently, C<aio_statvfs> always, and C<aio_rename> and C<aio_rmdir>
1618sometimes, fall back to using an absolue path.
1619
944=head2 IO::AIO::REQ CLASS 1620=head2 IO::AIO::REQ CLASS
945 1621
946All non-aggregate C<aio_*> functions return an object of this class when 1622All non-aggregate C<aio_*> functions return an object of this class when
947called in non-void context. 1623called in non-void context.
948 1624
951=item cancel $req 1627=item cancel $req
952 1628
953Cancels the request, if possible. Has the effect of skipping execution 1629Cancels the request, if possible. Has the effect of skipping execution
954when entering the B<execute> state and skipping calling the callback when 1630when entering the B<execute> state and skipping calling the callback when
955entering the the B<result> state, but will leave the request otherwise 1631entering the the B<result> state, but will leave the request otherwise
956untouched. That means that requests that currently execute will not be 1632untouched (with the exception of readdir). That means that requests that
957stopped and resources held by the request will not be freed prematurely. 1633currently execute will not be stopped and resources held by the request
1634will not be freed prematurely.
958 1635
959=item cb $req $callback->(...) 1636=item cb $req $callback->(...)
960 1637
961Replace (or simply set) the callback registered to the request. 1638Replace (or simply set) the callback registered to the request.
962 1639
1036=item $grp->cancel_subs 1713=item $grp->cancel_subs
1037 1714
1038Cancel all subrequests and clears any feeder, but not the group request 1715Cancel all subrequests and clears any feeder, but not the group request
1039itself. Useful when you queued a lot of events but got a result early. 1716itself. Useful when you queued a lot of events but got a result early.
1040 1717
1718The group request will finish normally (you cannot add requests to the
1719group).
1720
1041=item $grp->result (...) 1721=item $grp->result (...)
1042 1722
1043Set the result value(s) that will be passed to the group callback when all 1723Set the result value(s) that will be passed to the group callback when all
1044subrequests have finished and set the groups errno to the current value 1724subrequests have finished and set the groups errno to the current value
1045of errno (just like calling C<errno> without an error number). By default, 1725of errno (just like calling C<errno> without an error number). By default,
1060=item feed $grp $callback->($grp) 1740=item feed $grp $callback->($grp)
1061 1741
1062Sets a feeder/generator on this group: every group can have an attached 1742Sets a feeder/generator on this group: every group can have an attached
1063generator that generates requests if idle. The idea behind this is that, 1743generator that generates requests if idle. The idea behind this is that,
1064although you could just queue as many requests as you want in a group, 1744although you could just queue as many requests as you want in a group,
1065this might starve other requests for a potentially long time. For 1745this might starve other requests for a potentially long time. For example,
1066example, C<aio_scandir> might generate hundreds of thousands C<aio_stat> 1746C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
1067requests, delaying any later requests for a long time. 1747requests, delaying any later requests for a long time.
1068 1748
1069To avoid this, and allow incremental generation of requests, you can 1749To avoid this, and allow incremental generation of requests, you can
1070instead a group and set a feeder on it that generates those requests. The 1750instead a group and set a feeder on it that generates those requests. The
1071feed callback will be called whenever there are few enough (see C<limit>, 1751feed callback will be called whenever there are few enough (see C<limit>,
1076not impose any limits). 1756not impose any limits).
1077 1757
1078If the feed does not queue more requests when called, it will be 1758If the feed does not queue more requests when called, it will be
1079automatically removed from the group. 1759automatically removed from the group.
1080 1760
1081If the feed limit is C<0>, it will be set to C<2> automatically. 1761If the feed limit is C<0> when this method is called, it will be set to
1762C<2> automatically.
1082 1763
1083Example: 1764Example:
1084 1765
1085 # stat all files in @files, but only ever use four aio requests concurrently: 1766 # stat all files in @files, but only ever use four aio requests concurrently:
1086 1767
1098Sets the feeder limit for the group: The feeder will be called whenever 1779Sets the feeder limit for the group: The feeder will be called whenever
1099the group contains less than this many requests. 1780the group contains less than this many requests.
1100 1781
1101Setting the limit to C<0> will pause the feeding process. 1782Setting the limit to C<0> will pause the feeding process.
1102 1783
1784The default value for the limit is C<0>, but note that setting a feeder
1785automatically bumps it up to C<2>.
1786
1103=back 1787=back
1104 1788
1789
1105=head2 SUPPORT FUNCTIONS 1790=head2 SUPPORT FUNCTIONS
1106 1791
1107=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION 1792=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
1108 1793
1109=over 4 1794=over 4
1110 1795
1111=item $fileno = IO::AIO::poll_fileno 1796=item $fileno = IO::AIO::poll_fileno
1112 1797
1113Return the I<request result pipe file descriptor>. This filehandle must be 1798Return the I<request result pipe file descriptor>. This filehandle must be
1114polled for reading by some mechanism outside this module (e.g. Event or 1799polled for reading by some mechanism outside this module (e.g. EV, Glib,
1115select, see below or the SYNOPSIS). If the pipe becomes readable you have 1800select and so on, see below or the SYNOPSIS). If the pipe becomes readable
1116to call C<poll_cb> to check the results. 1801you have to call C<poll_cb> to check the results.
1117 1802
1118See C<poll_cb> for an example. 1803See C<poll_cb> for an example.
1119 1804
1120=item IO::AIO::poll_cb 1805=item IO::AIO::poll_cb
1121 1806
1122Process some outstanding events on the result pipe. You have to call this 1807Process some requests that have reached the result phase (i.e. they have
1123regularly. Returns C<0> if all events could be processed, or C<-1> if it 1808been executed but the results are not yet reported). You have to call
1124returned earlier for whatever reason. Returns immediately when no events 1809this "regularly" to finish outstanding requests.
1125are outstanding. The amount of events processed depends on the settings of
1126C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
1127 1810
1811Returns C<0> if all events could be processed (or there were no
1812events to process), or C<-1> if it returned earlier for whatever
1813reason. Returns immediately when no events are outstanding. The amount
1814of events processed depends on the settings of C<IO::AIO::max_poll_req>,
1815C<IO::AIO::max_poll_time> and C<IO::AIO::max_outstanding>.
1816
1128If not all requests were processed for whatever reason, the filehandle 1817If not all requests were processed for whatever reason, the poll file
1129will still be ready when C<poll_cb> returns, so normally you don't have to 1818descriptor will still be ready when C<poll_cb> returns, so normally you
1130do anything special to have it called later. 1819don't have to do anything special to have it called later.
1820
1821Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
1822ready, it can be beneficial to call this function from loops which submit
1823a lot of requests, to make sure the results get processed when they become
1824available and not just when the loop is finished and the event loop takes
1825over again. This function returns very fast when there are no outstanding
1826requests.
1131 1827
1132Example: Install an Event watcher that automatically calls 1828Example: Install an Event watcher that automatically calls
1133IO::AIO::poll_cb with high priority: 1829IO::AIO::poll_cb with high priority (more examples can be found in the
1830SYNOPSIS section, at the top of this document):
1134 1831
1135 Event->io (fd => IO::AIO::poll_fileno, 1832 Event->io (fd => IO::AIO::poll_fileno,
1136 poll => 'r', async => 1, 1833 poll => 'r', async => 1,
1137 cb => \&IO::AIO::poll_cb); 1834 cb => \&IO::AIO::poll_cb);
1835
1836=item IO::AIO::poll_wait
1837
1838Wait until either at least one request is in the result phase or no
1839requests are outstanding anymore.
1840
1841This is useful if you want to synchronously wait for some requests to
1842become ready, without actually handling them.
1843
1844See C<nreqs> for an example.
1845
1846=item IO::AIO::poll
1847
1848Waits until some requests have been handled.
1849
1850Returns the number of requests processed, but is otherwise strictly
1851equivalent to:
1852
1853 IO::AIO::poll_wait, IO::AIO::poll_cb
1854
1855=item IO::AIO::flush
1856
1857Wait till all outstanding AIO requests have been handled.
1858
1859Strictly equivalent to:
1860
1861 IO::AIO::poll_wait, IO::AIO::poll_cb
1862 while IO::AIO::nreqs;
1863
1864This function can be useful at program aborts, to make sure outstanding
1865I/O has been done (C<IO::AIO> uses an C<END> block which already calls
1866this function on normal exits), or when you are merely using C<IO::AIO>
1867for its more advanced functions, rather than for async I/O, e.g.:
1868
1869 my ($dirs, $nondirs);
1870 IO::AIO::aio_scandir "/tmp", 0, sub { ($dirs, $nondirs) = @_ };
1871 IO::AIO::flush;
1872 # $dirs, $nondirs are now set
1138 1873
1139=item IO::AIO::max_poll_reqs $nreqs 1874=item IO::AIO::max_poll_reqs $nreqs
1140 1875
1141=item IO::AIO::max_poll_time $seconds 1876=item IO::AIO::max_poll_time $seconds
1142 1877
1167 # use a low priority so other tasks have priority 1902 # use a low priority so other tasks have priority
1168 Event->io (fd => IO::AIO::poll_fileno, 1903 Event->io (fd => IO::AIO::poll_fileno,
1169 poll => 'r', nice => 1, 1904 poll => 'r', nice => 1,
1170 cb => &IO::AIO::poll_cb); 1905 cb => &IO::AIO::poll_cb);
1171 1906
1172=item IO::AIO::poll_wait
1173
1174If there are any outstanding requests and none of them in the result
1175phase, wait till the result filehandle becomes ready for reading (simply
1176does a C<select> on the filehandle. This is useful if you want to
1177synchronously wait for some requests to finish).
1178
1179See C<nreqs> for an example.
1180
1181=item IO::AIO::poll
1182
1183Waits until some requests have been handled.
1184
1185Returns the number of requests processed, but is otherwise strictly
1186equivalent to:
1187
1188 IO::AIO::poll_wait, IO::AIO::poll_cb
1189
1190=item IO::AIO::flush
1191
1192Wait till all outstanding AIO requests have been handled.
1193
1194Strictly equivalent to:
1195
1196 IO::AIO::poll_wait, IO::AIO::poll_cb
1197 while IO::AIO::nreqs;
1198
1199=back 1907=back
1908
1200 1909
1201=head3 CONTROLLING THE NUMBER OF THREADS 1910=head3 CONTROLLING THE NUMBER OF THREADS
1202 1911
1203=over 1912=over
1204 1913
1236 1945
1237Under normal circumstances you don't need to call this function. 1946Under normal circumstances you don't need to call this function.
1238 1947
1239=item IO::AIO::max_idle $nthreads 1948=item IO::AIO::max_idle $nthreads
1240 1949
1241Limit the number of threads (default: 4) that are allowed to idle (i.e., 1950Limit the number of threads (default: 4) that are allowed to idle
1242threads that did not get a request to process within 10 seconds). That 1951(i.e., threads that did not get a request to process within the idle
1243means if a thread becomes idle while C<$nthreads> other threads are also 1952timeout (default: 10 seconds). That means if a thread becomes idle while
1244idle, it will free its resources and exit. 1953C<$nthreads> other threads are also idle, it will free its resources and
1954exit.
1245 1955
1246This is useful when you allow a large number of threads (e.g. 100 or 1000) 1956This is useful when you allow a large number of threads (e.g. 100 or 1000)
1247to allow for extremely high load situations, but want to free resources 1957to allow for extremely high load situations, but want to free resources
1248under normal circumstances (1000 threads can easily consume 30MB of RAM). 1958under normal circumstances (1000 threads can easily consume 30MB of RAM).
1249 1959
1250The default is probably ok in most situations, especially if thread 1960The default is probably ok in most situations, especially if thread
1251creation is fast. If thread creation is very slow on your system you might 1961creation is fast. If thread creation is very slow on your system you might
1252want to use larger values. 1962want to use larger values.
1253 1963
1964=item IO::AIO::idle_timeout $seconds
1965
1966Sets the minimum idle timeout (default 10) after which worker threads are
1967allowed to exit. SEe C<IO::AIO::max_idle>.
1968
1254=item IO::AIO::max_outstanding $maxreqs 1969=item IO::AIO::max_outstanding $maxreqs
1970
1971Sets the maximum number of outstanding requests to C<$nreqs>. If
1972you do queue up more than this number of requests, the next call to
1973C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
1974C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
1975longer exceeded.
1976
1977In other words, this setting does not enforce a queue limit, but can be
1978used to make poll functions block if the limit is exceeded.
1255 1979
1256This is a very bad function to use in interactive programs because it 1980This is a very bad function to use in interactive programs because it
1257blocks, and a bad way to reduce concurrency because it is inexact: Better 1981blocks, and a bad way to reduce concurrency because it is inexact: Better
1258use an C<aio_group> together with a feed callback. 1982use an C<aio_group> together with a feed callback.
1259 1983
1260Sets the maximum number of outstanding requests to C<$nreqs>. If you 1984Its main use is in scripts without an event loop - when you want to stat
1261do queue up more than this number of requests, the next call to the 1985a lot of files, you can write something like this:
1262C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
1263function will block until the limit is no longer exceeded.
1264 1986
1265The default value is very large, so there is no practical limit on the 1987 IO::AIO::max_outstanding 32;
1266number of outstanding requests.
1267 1988
1268You can still queue as many requests as you want. Therefore, 1989 for my $path (...) {
1269C<max_outstanding> is mainly useful in simple scripts (with low values) or 1990 aio_stat $path , ...;
1270as a stop gap to shield against fatal memory overflow (with large values). 1991 IO::AIO::poll_cb;
1992 }
1993
1994 IO::AIO::flush;
1995
1996The call to C<poll_cb> inside the loop will normally return instantly, but
1997as soon as more thna C<32> reqeusts are in-flight, it will block until
1998some requests have been handled. This keeps the loop from pushing a large
1999number of C<aio_stat> requests onto the queue.
2000
2001The default value for C<max_outstanding> is very large, so there is no
2002practical limit on the number of outstanding requests.
1271 2003
1272=back 2004=back
2005
1273 2006
1274=head3 STATISTICAL INFORMATION 2007=head3 STATISTICAL INFORMATION
1275 2008
1276=over 2009=over
1277 2010
1295Returns the number of requests currently in the pending state (executed, 2028Returns the number of requests currently in the pending state (executed,
1296but not yet processed by poll_cb). 2029but not yet processed by poll_cb).
1297 2030
1298=back 2031=back
1299 2032
2033
2034=head3 SUBSECOND STAT TIME ACCESS
2035
2036Both C<aio_stat>/C<aio_lstat> and perl's C<stat>/C<lstat> functions can
2037generally find access/modification and change times with subsecond time
2038accuracy of the system supports it, but perl's built-in functions only
2039return the integer part.
2040
2041The following functions return the timestamps of the most recent
2042stat with subsecond precision on most systems and work both after
2043C<aio_stat>/C<aio_lstat> and perl's C<stat>/C<lstat> calls. Their return
2044value is only meaningful after a successful C<stat>/C<lstat> call, or
2045during/after a successful C<aio_stat>/C<aio_lstat> callback.
2046
2047This is similar to the L<Time::HiRes> C<stat> functions, but can return
2048full resolution without rounding and work with standard perl C<stat>,
2049alleviating the need to call the special C<Time::HiRes> functions, which
2050do not act like their perl counterparts.
2051
2052On operating systems or file systems where subsecond time resolution is
2053not supported or could not be detected, a fractional part of C<0> is
2054returned, so it is always safe to call these functions.
2055
2056=over 4
2057
2058=item $seconds = IO::AIO::st_atime, IO::AIO::st_mtime, IO::AIO::st_ctime, IO::AIO::st_btime
2059
2060Return the access, modication, change or birth time, respectively,
2061including fractional part. Due to the limited precision of floating point,
2062the accuracy on most platforms is only a bit better than milliseconds
2063for times around now - see the I<nsec> function family, below, for full
2064accuracy.
2065
2066File birth time is only available when the OS and perl support it (on
2067FreeBSD and NetBSD at the time of this writing, although support is
2068adaptive, so if your OS/perl gains support, IO::AIO can take avdantage of
2069it). On systems where it isn't available, C<0> is currently returned, but
2070this might change to C<undef> in a future version.
2071
2072=item ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtime
2073
2074Returns access, modification, change and birth time all in one go, and
2075maybe more times in the future version.
2076
2077=item $nanoseconds = IO::AIO::st_atimensec, IO::AIO::st_mtimensec, IO::AIO::st_ctimensec, IO::AIO::st_btimensec
2078
2079Return the fractional access, modifcation, change or birth time, in nanoseconds,
2080as an integer in the range C<0> to C<999999999>.
2081
2082Note that no accessors are provided for access, modification and
2083change times - you need to get those from C<stat _> if required (C<int
2084IO::AIO::st_atime> and so on will I<not> generally give you the correct
2085value).
2086
2087=item $seconds = IO::AIO::st_btimesec
2088
2089The (integral) seconds part of the file birth time, if available.
2090
2091=item ($atime, $mtime, $ctime, $btime, ...) = IO::AIO::st_xtimensec
2092
2093Like the functions above, but returns all four times in one go (and maybe
2094more in future versions).
2095
2096=item $counter = IO::AIO::st_gen
2097
2098Returns the generation counter (in practice this is just a random number)
2099of the file. This is only available on platforms which have this member in
2100their C<struct stat> (most BSDs at the time of this writing) and generally
2101only to the root usert. If unsupported, C<0> is returned, but this might
2102change to C<undef> in a future version.
2103
2104=back
2105
2106Example: print the high resolution modification time of F</etc>, using
2107C<stat>, and C<IO::AIO::aio_stat>.
2108
2109 if (stat "/etc") {
2110 printf "stat(/etc) mtime: %f\n", IO::AIO::st_mtime;
2111 }
2112
2113 IO::AIO::aio_stat "/etc", sub {
2114 $_[0]
2115 and return;
2116
2117 printf "aio_stat(/etc) mtime: %d.%09d\n", (stat _)[9], IO::AIO::st_mtimensec;
2118 };
2119
2120 IO::AIO::flush;
2121
2122Output of the awbove on my system, showing reduced and full accuracy:
2123
2124 stat(/etc) mtime: 1534043702.020808
2125 aio_stat(/etc) mtime: 1534043702.020807792
2126
2127
2128=head3 MISCELLANEOUS FUNCTIONS
2129
2130IO::AIO implements some functions that are useful when you want to use
2131some "Advanced I/O" function not available to in Perl, without going the
2132"Asynchronous I/O" route. Many of these have an asynchronous C<aio_*>
2133counterpart.
2134
2135=over 4
2136
2137=item $numfd = IO::AIO::get_fdlimit
2138
2139This function is I<EXPERIMENTAL> and subject to change.
2140
2141Tries to find the current file descriptor limit and returns it, or
2142C<undef> and sets C<$!> in case of an error. The limit is one larger than
2143the highest valid file descriptor number.
2144
2145=item IO::AIO::min_fdlimit [$numfd]
2146
2147This function is I<EXPERIMENTAL> and subject to change.
2148
2149Try to increase the current file descriptor limit(s) to at least C<$numfd>
2150by changing the soft or hard file descriptor resource limit. If C<$numfd>
2151is missing, it will try to set a very high limit, although this is not
2152recommended when you know the actual minimum that you require.
2153
2154If the limit cannot be raised enough, the function makes a best-effort
2155attempt to increase the limit as much as possible, using various
2156tricks, while still failing. You can query the resulting limit using
2157C<IO::AIO::get_fdlimit>.
2158
2159If an error occurs, returns C<undef> and sets C<$!>, otherwise returns
2160true.
2161
2162=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
2163
2164Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
2165but is blocking (this makes most sense if you know the input data is
2166likely cached already and the output filehandle is set to non-blocking
2167operations).
2168
2169Returns the number of bytes copied, or C<-1> on error.
2170
2171=item IO::AIO::fadvise $fh, $offset, $len, $advice
2172
2173Simply calls the C<posix_fadvise> function (see its
2174manpage for details). The following advice constants are
2175available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
2176C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
2177C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
2178
2179On systems that do not implement C<posix_fadvise>, this function returns
2180ENOSYS, otherwise the return value of C<posix_fadvise>.
2181
2182=item IO::AIO::madvise $scalar, $offset, $len, $advice
2183
2184Simply calls the C<posix_madvise> function (see its
2185manpage for details). The following advice constants are
2186available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
2187C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>,
2188C<IO::AIO::MADV_DONTNEED>.
2189
2190If C<$offset> is negative, counts from the end. If C<$length> is negative,
2191the remaining length of the C<$scalar> is used. If possible, C<$length>
2192will be reduced to fit into the C<$scalar>.
2193
2194On systems that do not implement C<posix_madvise>, this function returns
2195ENOSYS, otherwise the return value of C<posix_madvise>.
2196
2197=item IO::AIO::mprotect $scalar, $offset, $len, $protect
2198
2199Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
2200$scalar (see its manpage for details). The following protect
2201constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
2202C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
2203
2204If C<$offset> is negative, counts from the end. If C<$length> is negative,
2205the remaining length of the C<$scalar> is used. If possible, C<$length>
2206will be reduced to fit into the C<$scalar>.
2207
2208On systems that do not implement C<mprotect>, this function returns
2209ENOSYS, otherwise the return value of C<mprotect>.
2210
2211=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
2212
2213Memory-maps a file (or anonymous memory range) and attaches it to the
2214given C<$scalar>, which will act like a string scalar. Returns true on
2215success, and false otherwise.
2216
2217The scalar must exist, but its contents do not matter - this means you
2218cannot use a nonexistant array or hash element. When in doubt, C<undef>
2219the scalar first.
2220
2221The only operations allowed on the mmapped scalar are C<substr>/C<vec>,
2222which don't change the string length, and most read-only operations such
2223as copying it or searching it with regexes and so on.
2224
2225Anything else is unsafe and will, at best, result in memory leaks.
2226
2227The memory map associated with the C<$scalar> is automatically removed
2228when the C<$scalar> is undef'd or destroyed, or when the C<IO::AIO::mmap>
2229or C<IO::AIO::munmap> functions are called on it.
2230
2231This calls the C<mmap>(2) function internally. See your system's manual
2232page for details on the C<$length>, C<$prot> and C<$flags> parameters.
2233
2234The C<$length> must be larger than zero and smaller than the actual
2235filesize.
2236
2237C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
2238C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
2239
2240C<$flags> can be a combination of
2241C<IO::AIO::MAP_SHARED> or
2242C<IO::AIO::MAP_PRIVATE>,
2243or a number of system-specific flags (when not available, the are C<0>):
2244C<IO::AIO::MAP_ANONYMOUS> (which is set to C<MAP_ANON> if your system only provides this constant),
2245C<IO::AIO::MAP_LOCKED>,
2246C<IO::AIO::MAP_NORESERVE>,
2247C<IO::AIO::MAP_POPULATE>,
2248C<IO::AIO::MAP_NONBLOCK>,
2249C<IO::AIO::MAP_FIXED>,
2250C<IO::AIO::MAP_GROWSDOWN>,
2251C<IO::AIO::MAP_32BIT>,
2252C<IO::AIO::MAP_HUGETLB> or
2253C<IO::AIO::MAP_STACK>.
2254
2255If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
2256
2257C<$offset> is the offset from the start of the file - it generally must be
2258a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
2259
2260Example:
2261
2262 use Digest::MD5;
2263 use IO::AIO;
2264
2265 open my $fh, "<verybigfile"
2266 or die "$!";
2267
2268 IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
2269 or die "verybigfile: $!";
2270
2271 my $fast_md5 = md5 $data;
2272
2273=item IO::AIO::munmap $scalar
2274
2275Removes a previous mmap and undefines the C<$scalar>.
2276
2277=item IO::AIO::mremap $scalar, $new_length, $flags = MREMAP_MAYMOVE[, $new_address = 0]
2278
2279Calls the Linux-specific mremap(2) system call. The C<$scalar> must have
2280been mapped by C<IO::AIO::mmap>, and C<$flags> must currently either be
2281C<0> or C<IO::AIO::MREMAP_MAYMOVE>.
2282
2283Returns true if successful, and false otherwise. If the underlying mmapped
2284region has changed address, then the true value has the numerical value
2285C<1>, otherwise it has the numerical value C<0>:
2286
2287 my $success = IO::AIO::mremap $mmapped, 8192, IO::AIO::MREMAP_MAYMOVE
2288 or die "mremap: $!";
2289
2290 if ($success*1) {
2291 warn "scalar has chanegd address in memory\n";
2292 }
2293
2294C<IO::AIO::MREMAP_FIXED> and the C<$new_address> argument are currently
2295implemented, but not supported and might go away in a future version.
2296
2297On systems where this call is not supported or is not emulated, this call
2298returns falls and sets C<$!> to C<ENOSYS>.
2299
2300=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
2301
2302Calls the C<munlock> function, undoing the effects of a previous
2303C<aio_mlock> call (see its description for details).
2304
2305=item IO::AIO::munlockall
2306
2307Calls the C<munlockall> function.
2308
2309On systems that do not implement C<munlockall>, this function returns
2310ENOSYS, otherwise the return value of C<munlockall>.
2311
2312=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
2313
2314Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
2315C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
2316should be the file offset.
2317
2318C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
2319silently corrupt the data in this case.
2320
2321The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
2322C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
2323C<IO::AIO::SPLICE_F_GIFT>.
2324
2325See the C<splice(2)> manpage for details.
2326
2327=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
2328
2329Calls the GNU/Linux C<tee(2)> syscall, see its manpage and the
2330description for C<IO::AIO::splice> above for details.
2331
2332=item $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
2333
2334Attempts to query or change the pipe buffer size. Obviously works only
2335on pipes, and currently works only on GNU/Linux systems, and fails with
2336C<-1>/C<ENOSYS> everywhere else. If anybody knows how to influence pipe buffer
2337size on other systems, drop me a note.
2338
2339=item ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
2340
2341This is a direct interface to the Linux L<pipe2(2)> system call. If
2342C<$flags> is missing or C<0>, then this should be the same as a call to
2343perl's built-in C<pipe> function and create a new pipe, and works on
2344systems that lack the pipe2 syscall. On win32, this case invokes C<_pipe
2345(..., 4096, O_BINARY)>.
2346
2347If C<$flags> is non-zero, it tries to invoke the pipe2 system call with
2348the given flags (Linux 2.6.27, glibc 2.9).
2349
2350On success, the read and write file handles are returned.
2351
2352On error, nothing will be returned. If the pipe2 syscall is missing and
2353C<$flags> is non-zero, fails with C<ENOSYS>.
2354
2355Please refer to L<pipe2(2)> for more info on the C<$flags>, but at the
2356time of this writing, C<IO::AIO::O_CLOEXEC>, C<IO::AIO::O_NONBLOCK> and
2357C<IO::AIO::O_DIRECT> (Linux 3.4, for packet-based pipes) were supported.
2358
2359Example: create a pipe race-free w.r.t. threads and fork:
2360
2361 my ($rfh, $wfh) = IO::AIO::pipe2 IO::AIO::O_CLOEXEC
2362 or die "pipe2: $!\n";
2363
2364=item $fh = IO::AIO::eventfd [$initval, [$flags]]
2365
2366This is a direct interface to the Linux L<eventfd(2)> system call. The
2367(unhelpful) defaults for C<$initval> and C<$flags> are C<0> for both.
2368
2369On success, the new eventfd filehandle is returned, otherwise returns
2370C<undef>. If the eventfd syscall is missing, fails with C<ENOSYS>.
2371
2372Please refer to L<eventfd(2)> for more info on this call.
2373
2374The following symbol flag values are available: C<IO::AIO::EFD_CLOEXEC>,
2375C<IO::AIO::EFD_NONBLOCK> and C<IO::AIO::EFD_SEMAPHORE> (Linux 2.6.30).
2376
2377Example: create a new eventfd filehandle:
2378
2379 $fh = IO::AIO::eventfd 0, IO::AIO::O_CLOEXEC
2380 or die "eventfd: $!\n";
2381
2382=item $fh = IO::AIO::timerfd_create $clockid[, $flags]
2383
2384This is a direct interface to the Linux L<timerfd_create(2)> system call. The
2385(unhelpful) default for C<$flags> is C<0>.
2386
2387On success, the new timerfd filehandle is returned, otherwise returns
2388C<undef>. If the eventfd syscall is missing, fails with C<ENOSYS>.
2389
2390Please refer to L<timerfd_create(2)> for more info on this call.
2391
2392The following C<$clockid> values are
2393available: C<IO::AIO::CLOCK_REALTIME>, C<IO::AIO::CLOCK_MONOTONIC>
2394C<IO::AIO::CLOCK_CLOCK_BOOTTIME> (Linux 3.15)
2395C<IO::AIO::CLOCK_CLOCK_REALTIME_ALARM> (Linux 3.11) and
2396C<IO::AIO::CLOCK_CLOCK_BOOTTIME_ALARM> (Linux 3.11).
2397
2398The following C<$flags> values are available (Linux
23992.6.27): C<IO::AIO::TFD_NONBLOCK> and C<IO::AIO::TFD_CLOEXEC>.
2400
2401Example: create a new timerfd and set it to one-second repeated alarms,
2402then wait for two alarms:
2403
2404 my $fh = IO::AIO::timerfd_create IO::AIO::CLOCK_BOOTTIME, IO::AIO::TFD_CLOEXEC
2405 or die "timerfd_create: $!\n";
2406
2407 defined IO::AIO::timerfd_settime $fh, 0, 1, 1
2408 or die "timerfd_settime: $!\n";
2409
2410 for (1..2) {
2411 8 == sysread $fh, my $buf, 8
2412 or die "timerfd read failure\n";
2413
2414 printf "number of expirations (likely 1): %d\n",
2415 unpack "Q", $buf;
2416 }
2417
2418=item ($cur_interval, $cur_value) = IO::AIO::timerfd_settime $fh, $flags, $new_interval, $nbw_value
2419
2420This is a direct interface to the Linux L<timerfd_settime(2)> system
2421call. Please refer to its manpage for more info on this call.
2422
2423The new itimerspec is specified using two (possibly fractional) second
2424values, C<$new_interval> and C<$new_value>).
2425
2426On success, the current interval and value are returned (as per
2427C<timerfd_gettime>). On failure, the empty list is returned.
2428
2429The following C<$flags> values are
2430available: C<IO::AIO::TFD_TIMER_ABSTIME> and
2431C<IO::AIO::TFD_TIMER_CANCEL_ON_SET>.
2432
2433See C<IO::AIO::timerfd_create> for a full example.
2434
2435=item ($cur_interval, $cur_value) = IO::AIO::timerfd_gettime $fh
2436
2437This is a direct interface to the Linux L<timerfd_gettime(2)> system
2438call. Please refer to its manpage for more info on this call.
2439
2440On success, returns the current values of interval and value for the given
2441timerfd (as potentially fractional second values). On failure, the empty
2442list is returned.
2443
2444=back
2445
1300=cut 2446=cut
1301 2447
1302min_parallel 8; 2448min_parallel 8;
1303 2449
1304END { flush } 2450END { flush }
1305 2451
13061; 24521;
1307 2453
2454=head1 EVENT LOOP INTEGRATION
2455
2456It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
2457automatically into many event loops:
2458
2459 # AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
2460 use AnyEvent::AIO;
2461
2462You can also integrate IO::AIO manually into many event loops, here are
2463some examples of how to do this:
2464
2465 # EV integration
2466 my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
2467
2468 # Event integration
2469 Event->io (fd => IO::AIO::poll_fileno,
2470 poll => 'r',
2471 cb => \&IO::AIO::poll_cb);
2472
2473 # Glib/Gtk2 integration
2474 add_watch Glib::IO IO::AIO::poll_fileno,
2475 in => sub { IO::AIO::poll_cb; 1 };
2476
2477 # Tk integration
2478 Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
2479 readable => \&IO::AIO::poll_cb);
2480
2481 # Danga::Socket integration
2482 Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
2483 \&IO::AIO::poll_cb);
2484
1308=head2 FORK BEHAVIOUR 2485=head2 FORK BEHAVIOUR
1309 2486
1310This module should do "the right thing" when the process using it forks: 2487Usage of pthreads in a program changes the semantics of fork
2488considerably. Specifically, only async-safe functions can be called after
2489fork. Perl doesn't know about this, so in general, you cannot call fork
2490with defined behaviour in perl if pthreads are involved. IO::AIO uses
2491pthreads, so this applies, but many other extensions and (for inexplicable
2492reasons) perl itself often is linked against pthreads, so this limitation
2493applies to quite a lot of perls.
1311 2494
1312Before the fork, IO::AIO enters a quiescent state where no requests 2495This module no longer tries to fight your OS, or POSIX. That means IO::AIO
1313can be added in other threads and no results will be processed. After 2496only works in the process that loaded it. Forking is fully supported, but
1314the fork the parent simply leaves the quiescent state and continues 2497using IO::AIO in the child is not.
1315request/result processing, while the child frees the request/result queue
1316(so that the requests started before the fork will only be handled in the
1317parent). Threads will be started on demand until the limit set in the
1318parent process has been reached again.
1319 2498
1320In short: the parent will, after a short pause, continue as if fork had 2499You might get around by not I<using> IO::AIO before (or after)
1321not been called, while the child will act as if IO::AIO has not been used 2500forking. You could also try to call the L<IO::AIO::reinit> function in the
1322yet. 2501child:
2502
2503=over 4
2504
2505=item IO::AIO::reinit
2506
2507Abandons all current requests and I/O threads and simply reinitialises all
2508data structures. This is not an operation supported by any standards, but
2509happens to work on GNU/Linux and some newer BSD systems.
2510
2511The only reasonable use for this function is to call it after forking, if
2512C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
2513the process will result in undefined behaviour. Calling it at any time
2514will also result in any undefined (by POSIX) behaviour.
2515
2516=back
2517
2518=head2 LINUX-SPECIFIC CALLS
2519
2520When a call is documented as "linux-specific" then this means it
2521originated on GNU/Linux. C<IO::AIO> will usually try to autodetect the
2522availability and compatibility of such calls regardless of the platform
2523it is compiled on, so platforms such as FreeBSD which often implement
2524these calls will work. When in doubt, call them and see if they fail wth
2525C<ENOSYS>.
1323 2526
1324=head2 MEMORY USAGE 2527=head2 MEMORY USAGE
1325 2528
1326Per-request usage: 2529Per-request usage:
1327 2530
1340temporary buffers, and each thread requires a stack and other data 2543temporary buffers, and each thread requires a stack and other data
1341structures (usually around 16k-128k, depending on the OS). 2544structures (usually around 16k-128k, depending on the OS).
1342 2545
1343=head1 KNOWN BUGS 2546=head1 KNOWN BUGS
1344 2547
1345Known bugs will be fixed in the next release. 2548Known bugs will be fixed in the next release :)
2549
2550=head1 KNOWN ISSUES
2551
2552Calls that try to "import" foreign memory areas (such as C<IO::AIO::mmap>
2553or C<IO::AIO::aio_slurp>) do not work with generic lvalues, such as
2554non-created hash slots or other scalars I didn't think of. It's best to
2555avoid such and either use scalar variables or making sure that the scalar
2556exists (e.g. by storing C<undef>) and isn't "funny" (e.g. tied).
2557
2558I am not sure anything can be done about this, so this is considered a
2559known issue, rather than a bug.
1346 2560
1347=head1 SEE ALSO 2561=head1 SEE ALSO
1348 2562
1349L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a 2563L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
1350more natural syntax. 2564more natural syntax.

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines