ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.62 by root, Sun Oct 22 21:13:47 2006 UTC vs.
Revision 1.175 by root, Sun Jan 10 20:37:33 2010 UTC

…		…
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", O_RDONLY, 0, sub {
10	my ($fh) = @_;	10	my $fh = shift
		11	or die "/etc/passwd: $!";
11	...	12	...
12	};	13	};
13		14
14	aio_unlink "/tmp/file", sub { };	15	aio_unlink "/tmp/file", sub { };
15		16
…		…
18	};	19	};
19		20
20	# version 2+ has request and group objects	21	# version 2+ has request and group objects
21	use IO::AIO 2;	22	use IO::AIO 2;
22		23
		24	aioreq_pri 4; # give next request a very high priority
23	my $req = aio_unlink "/tmp/file", sub { };	25	my $req = aio_unlink "/tmp/file", sub { };
24	$req->cancel; # cancel request if still in queue	26	$req->cancel; # cancel request if still in queue
25		27
26	my $grp = aio_group sub { print "all stats done\n" };	28	my $grp = aio_group sub { print "all stats done\n" };
27	add $grp aio_stat "..." for ...;	29	add $grp aio_stat "..." for ...;
28		30
29	# AnyEvent integration
30	open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
31	my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
32
33	# Event integration
34	Event->io (fd => IO::AIO::poll_fileno,
35	poll => 'r',
36	cb => \&IO::AIO::poll_cb);
37
38	# Glib/Gtk2 integration
39	add_watch Glib::IO IO::AIO::poll_fileno,
40	in => sub { IO::AIO::poll_cb; 1 };
41
42	# Tk integration
43	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
44	readable => \&IO::AIO::poll_cb);
45
46	# Danga::Socket integration
47	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
48	\&IO::AIO::poll_cb);
49
50	=head1 DESCRIPTION	31	=head1 DESCRIPTION
51		32
52	This module implements asynchronous I/O using whatever means your	33	This module implements asynchronous I/O using whatever means your
53	operating system supports.	34	operating system supports. It is implemented as an interface to C<libeio>
		35	(L<http://software.schmorp.de/pkg/libeio.html>).
54		36
		37	Asynchronous means that operations that can normally block your program
		38	(e.g. reading from disk) will be done asynchronously: the operation
		39	will still block, but you can do something else in the meantime. This
		40	is extremely useful for programs that need to stay interactive even
		41	when doing heavy I/O (GUI programs, high performance network servers
		42	etc.), but can also be used to easily do operations in parallel that are
		43	normally done sequentially, e.g. stat'ing many files, which is much faster
		44	on a RAID volume or over NFS when you do a number of stat operations
		45	concurrently.
		46
		47	While most of this works on all types of file descriptors (for
		48	example sockets), using these functions on file descriptors that
		49	support nonblocking operation (again, sockets, pipes etc.) is
		50	very inefficient. Use an event loop for that (such as the L<EV>
		51	module): IO::AIO will naturally fit into such an event loop itself.
		52
55	Currently, a number of threads are started that execute your read/writes	53	In this version, a number of threads are started that execute your
56	and signal their completion. You don't need thread support in your libc or	54	requests and signal their completion. You don't need thread support
57	perl, and the threads created by this module will not be visible to the	55	in perl, and the threads created by this module will not be visible
58	pthreads library. In the future, this module might make use of the native	56	to perl. In the future, this module might make use of the native aio
59	aio functions available on many operating systems. However, they are often	57	functions available on many operating systems. However, they are often
60	not well-supported (Linux doesn't allow them on normal files currently,	58	not well-supported or restricted (GNU/Linux doesn't allow them on normal
61	for example), and they would only support aio_read and aio_write, so the	59	files currently, for example), and they would only support aio_read and
62	remaining functionality would have to be implemented using threads anyway.	60	aio_write, so the remaining functionality would have to be implemented
		61	using threads anyway.
63		62
64	Although the module will work with in the presence of other threads, it is	63	Although the module will work in the presence of other (Perl-) threads,
65	currently not reentrant, so use appropriate locking yourself, always call	64	it is currently not reentrant in any way, so use appropriate locking
66	C<poll_cb> from within the same thread, or never call C<poll_cb> (or other	65	yourself, always call C<poll_cb> from within the same thread, or never
67	C<aio_> functions) recursively.	66	call C<poll_cb> (or other C<aio_> functions) recursively.
		67
		68	=head2 EXAMPLE
		69
		70	This is a simple example that uses the EV module and loads
		71	F</etc/passwd> asynchronously:
		72
		73	use Fcntl;
		74	use EV;
		75	use IO::AIO;
		76
		77	# register the IO::AIO callback with EV
		78	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		79
		80	# queue the request to open /etc/passwd
		81	aio_open "/etc/passwd", O_RDONLY, 0, sub {
		82	my $fh = shift
		83	or die "error while opening: $!";
		84
		85	# stat'ing filehandles is generally non-blocking
		86	my $size = -s $fh;
		87
		88	# queue a request to read the file
		89	my $contents;
		90	aio_read $fh, 0, $size, $contents, 0, sub {
		91	$_[0] == $size
		92	or die "short read: $!";
		93
		94	close $fh;
		95
		96	# file contents now in $contents
		97	print $contents;
		98
		99	# exit event loop and program
		100	EV::unloop;
		101	};
		102	};
		103
		104	# possibly queue up other requests, or open GUI windows,
		105	# check for sockets etc. etc.
		106
		107	# process events as long as there are some:
		108	EV::loop;
		109
		110	=head1 REQUEST ANATOMY AND LIFETIME
		111
		112	Every C<aio_*> function creates a request. which is a C data structure not
		113	directly visible to Perl.
		114
		115	If called in non-void context, every request function returns a Perl
		116	object representing the request. In void context, nothing is returned,
		117	which saves a bit of memory.
		118
		119	The perl object is a fairly standard ref-to-hash object. The hash contents
		120	are not used by IO::AIO so you are free to store anything you like in it.
		121
		122	During their existance, aio requests travel through the following states,
		123	in order:
		124
		125	=over 4
		126
		127	=item ready
		128
		129	Immediately after a request is created it is put into the ready state,
		130	waiting for a thread to execute it.
		131
		132	=item execute
		133
		134	A thread has accepted the request for processing and is currently
		135	executing it (e.g. blocking in read).
		136
		137	=item pending
		138
		139	The request has been executed and is waiting for result processing.
		140
		141	While request submission and execution is fully asynchronous, result
		142	processing is not and relies on the perl interpreter calling C<poll_cb>
		143	(or another function with the same effect).
		144
		145	=item result
		146
		147	The request results are processed synchronously by C<poll_cb>.
		148
		149	The C<poll_cb> function will process all outstanding aio requests by
		150	calling their callbacks, freeing memory associated with them and managing
		151	any groups they are contained in.
		152
		153	=item done
		154
		155	Request has reached the end of its lifetime and holds no resources anymore
		156	(except possibly for the Perl object, but its connection to the actual
		157	aio request is severed and calling its methods will either do nothing or
		158	result in a runtime error).
		159
		160	=back
68		161
69	=cut	162	=cut
70		163
71	package IO::AIO;	164	package IO::AIO;
72		165
73	no warnings;	166	use Carp ();
74	use strict 'vars';	167
		168	use common::sense;
75		169
76	use base 'Exporter';	170	use base 'Exporter';
77		171
78	BEGIN {	172	BEGIN {
79	our $VERSION = '2.0';	173	our $VERSION = '3.5';
80		174
81	our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat	175	our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close
82	aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink	176	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
83	aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move	177	aio_scandir aio_symlink aio_readlink aio_sync aio_fsync
84	aio_group);	178	aio_fdatasync aio_sync_file_range aio_pathsync aio_readahead
85	our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);	179	aio_rename aio_link aio_move aio_copy aio_group
		180	aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
		181	aio_chmod aio_utime aio_truncate
		182	aio_msync aio_mtouch aio_statvfs);
		183
		184	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		185	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		186	min_parallel max_parallel max_idle
		187	nreqs nready npending nthreads
		188	max_poll_time max_poll_reqs
		189	sendfile fadvise);
		190
		191	push @AIO_REQ, qw(aio_busy); # not exported
86		192
87	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';	193	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
88		194
89	require XSLoader;	195	require XSLoader;
90	XSLoader::load ("IO::AIO", $VERSION);	196	XSLoader::load ("IO::AIO", $VERSION);
91	}	197	}
92		198
93	=head1 FUNCTIONS	199	=head1 FUNCTIONS
94		200
		201	=head2 QUICK OVERVIEW
		202
		203	This section simply lists the prototypes of the most important functions
		204	for quick reference. See the following sections for function-by-function
		205	documentation.
		206
		207	aio_open $pathname, $flags, $mode, $callback->($fh)
		208	aio_close $fh, $callback->($status)
		209	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		210	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		211	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		212	aio_readahead $fh,$offset,$length, $callback->($retval)
		213	aio_stat $fh_or_path, $callback->($status)
		214	aio_lstat $fh, $callback->($status)
		215	aio_statvfs $fh_or_path, $callback->($statvfs)
		216	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		217	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		218	aio_truncate $fh_or_path, $offset, $callback->($status)
		219	aio_chmod $fh_or_path, $mode, $callback->($status)
		220	aio_unlink $pathname, $callback->($status)
		221	aio_mknod $path, $mode, $dev, $callback->($status)
		222	aio_link $srcpath, $dstpath, $callback->($status)
		223	aio_symlink $srcpath, $dstpath, $callback->($status)
		224	aio_readlink $path, $callback->($link)
		225	aio_rename $srcpath, $dstpath, $callback->($status)
		226	aio_mkdir $pathname, $mode, $callback->($status)
		227	aio_rmdir $pathname, $callback->($status)
		228	aio_readdir $pathname, $callback->($entries)
		229	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		230	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		231	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		232	aio_load $path, $data, $callback->($status)
		233	aio_copy $srcpath, $dstpath, $callback->($status)
		234	aio_move $srcpath, $dstpath, $callback->($status)
		235	aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
		236	aio_rmtree $path, $callback->($status)
		237	aio_sync $callback->($status)
		238	aio_fsync $fh, $callback->($status)
		239	aio_fdatasync $fh, $callback->($status)
		240	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		241	aio_pathsync $path, $callback->($status)
		242	aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		243	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		244	aio_group $callback->(...)
		245	aio_nop $callback->()
		246
		247	$prev_pri = aioreq_pri [$pri]
		248	aioreq_nice $pri_adjust
		249
		250	IO::AIO::poll_wait
		251	IO::AIO::poll_cb
		252	IO::AIO::poll
		253	IO::AIO::flush
		254	IO::AIO::max_poll_reqs $nreqs
		255	IO::AIO::max_poll_time $seconds
		256	IO::AIO::min_parallel $nthreads
		257	IO::AIO::max_parallel $nthreads
		258	IO::AIO::max_idle $nthreads
		259	IO::AIO::max_outstanding $maxreqs
		260	IO::AIO::nreqs
		261	IO::AIO::nready
		262	IO::AIO::npending
		263
		264	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		265	IO::AIO::fadvise $fh, $offset, $len, $advice
		266	IO::AIO::mlockall $flags
		267	IO::AIO::munlockall
		268
95	=head2 AIO FUNCTIONS	269	=head2 AIO REQUEST FUNCTIONS
96		270
97	All the C<aio_*> calls are more or less thin wrappers around the syscall	271	All the C<aio_*> calls are more or less thin wrappers around the syscall
98	with the same name (sans C<aio_>). The arguments are similar or identical,	272	with the same name (sans C<aio_>). The arguments are similar or identical,
99	and they all accept an additional (and optional) C<$callback> argument	273	and they all accept an additional (and optional) C<$callback> argument
100	which must be a code reference. This code reference will get called with	274	which must be a code reference. This code reference will get called with
101	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	275	the syscall return code (e.g. most syscalls return C<-1> on error, unlike
102	perl, which usually delivers "false") as it's sole argument when the given	276	perl, which usually delivers "false") as its sole argument after the given
103	syscall has been executed asynchronously.	277	syscall has been executed asynchronously.
104		278
105	All functions expecting a filehandle keep a copy of the filehandle	279	All functions expecting a filehandle keep a copy of the filehandle
106	internally until the request has finished.	280	internally until the request has finished.
107		281
108	All requests return objects of type L<IO::AIO::REQ> that allow further	282	All functions return request objects of type L<IO::AIO::REQ> that allow
109	manipulation of those requests while they are in-flight.	283	further manipulation of those requests while they are in-flight.
110		284
111	The pathnames you pass to these routines I<must> be absolute and	285	The pathnames you pass to these routines I<must> be absolute and
112	encoded in byte form. The reason for the former is that at the time the	286	encoded as octets. The reason for the former is that at the time the
113	request is being executed, the current working directory could have	287	request is being executed, the current working directory could have
114	changed. Alternatively, you can make sure that you never change the	288	changed. Alternatively, you can make sure that you never change the
115	current working directory.	289	current working directory anywhere in the program and then use relative
		290	paths.
116		291
117	To encode pathnames to byte form, either make sure you either: a)	292	To encode pathnames as octets, either make sure you either: a) always pass
118	always pass in filenames you got from outside (command line, readdir	293	in filenames you got from outside (command line, readdir etc.) without
119	etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode	294	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
120	your pathnames to the locale (or other) encoding in effect in the user	295	your pathnames to the locale (or other) encoding in effect in the user
121	environment, d) use Glib::filename_from_unicode on unicode filenames or e)	296	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
122	use something else.	297	use something else to ensure your scalar has the correct contents.
		298
		299	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		300	handles correctly whether it is set or not.
123		301
124	=over 4	302	=over 4
		303
		304	=item $prev_pri = aioreq_pri [$pri]
		305
		306	Returns the priority value that would be used for the next request and, if
		307	C<$pri> is given, sets the priority for the next aio request.
		308
		309	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		310	and C<4>, respectively. Requests with higher priority will be serviced
		311	first.
		312
		313	The priority will be reset to C<0> after each call to one of the C<aio_*>
		314	functions.
		315
		316	Example: open a file with low priority, then read something from it with
		317	higher priority so the read request is serviced before other low priority
		318	open requests (potentially spamming the cache):
		319
		320	aioreq_pri -3;
		321	aio_open ..., sub {
		322	return unless $_[0];
		323
		324	aioreq_pri -2;
		325	aio_read $_[0], ..., sub {
		326	...
		327	};
		328	};
		329
		330
		331	=item aioreq_nice $pri_adjust
		332
		333	Similar to C<aioreq_pri>, but subtracts the given value from the current
		334	priority, so the effect is cumulative.
		335
125		336
126	=item aio_open $pathname, $flags, $mode, $callback->($fh)	337	=item aio_open $pathname, $flags, $mode, $callback->($fh)
127		338
128	Asynchronously open or create a file and call the callback with a newly	339	Asynchronously open or create a file and call the callback with a newly
129	created filehandle for the file.	340	created filehandle for the file.
…		…
135	list. They are the same as used by C<sysopen>.	346	list. They are the same as used by C<sysopen>.
136		347
137	Likewise, C<$mode> specifies the mode of the newly created file, if it	348	Likewise, C<$mode> specifies the mode of the newly created file, if it
138	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,	349	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
139	except that it is mandatory (i.e. use C<0> if you don't create new files,	350	except that it is mandatory (i.e. use C<0> if you don't create new files,
140	and C<0666> or C<0777> if you do).	351	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		352	by the umask in effect then the request is being executed, so better never
		353	change the umask.
141		354
142	Example:	355	Example:
143		356
144	aio_open "/etc/passwd", O_RDONLY, 0, sub {	357	aio_open "/etc/passwd", O_RDONLY, 0, sub {
145	if ($_[0]) {	358	if ($_[0]) {
…		…
148	} else {	361	} else {
149	die "open failed: $!\n";	362	die "open failed: $!\n";
150	}	363	}
151	};	364	};
152		365
		366
153	=item aio_close $fh, $callback->($status)	367	=item aio_close $fh, $callback->($status)
154		368
155	Asynchronously close a file and call the callback with the result	369	Asynchronously close a file and call the callback with the result
156	code. I<WARNING:> although accepted, you should not pass in a perl	370	code.
157	filehandle here, as perl will likely close the file descriptor another
158	time when the filehandle is destroyed. Normally, you can safely call perls
159	C<close> or just let filehandles go out of scope.
160		371
161	This is supposed to be a bug in the API, so that might change. It's	372	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
162	therefore best to avoid this function.	373	closing the file descriptor associated with the filehandle itself.
		374
		375	Therefore, C<aio_close> will not close the filehandle - instead it will
		376	use dup2 to overwrite the file descriptor with the write-end of a pipe
		377	(the pipe fd will be created on demand and will be cached).
		378
		379	Or in other words: the file descriptor will be closed, but it will not be
		380	free for reuse until the perl filehandle is closed.
		381
		382	=cut
163		383
164	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	384	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
165		385
166	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	386	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
167		387
168	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	388	Reads or writes C<$length> bytes from or to the specified C<$fh> and
169	into the scalar given by C<data> and offset C<dataoffset> and calls the	389	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset>
170	callback without the actual number of bytes read (or -1 on error, just	390	and calls the callback without the actual number of bytes read (or -1 on
171	like the syscall).	391	error, just like the syscall).
		392
		393	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		394	offset plus the actual number of bytes read.
		395
		396	If C<$offset> is undefined, then the current file descriptor offset will
		397	be used (and updated), otherwise the file descriptor offset will not be
		398	changed by these calls.
		399
		400	If C<$length> is undefined in C<aio_write>, use the remaining length of
		401	C<$data>.
		402
		403	If C<$dataoffset> is less than zero, it will be counted from the end of
		404	C<$data>.
172		405
173	The C<$data> scalar I<MUST NOT> be modified in any way while the request	406	The C<$data> scalar I<MUST NOT> be modified in any way while the request
174	is outstanding. Modifying it can result in segfaults or WW3 (if the	407	is outstanding. Modifying it can result in segfaults or World War III (if
175	necessary/optional hardware is installed).	408	the necessary/optional hardware is installed).
176		409
177	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	410	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
178	offset C<0> within the scalar:	411	offset C<0> within the scalar:
179		412
180	aio_read $fh, 7, 15, $buffer, 0, sub {	413	aio_read $fh, 7, 15, $buffer, 0, sub {
181	$_[0] > 0 or die "read error: $!";	414	$_[0] > 0 or die "read error: $!";
182	print "read $_[0] bytes: <$buffer>\n";	415	print "read $_[0] bytes: <$buffer>\n";
183	};	416	};
184		417
185	=item aio_move $srcpath, $dstpath, $callback->($status)
186
187	[EXPERIMENTAL due to internal aio_group use]
188
189	Try to move the I<file> (directories not supported as either source or
190	destination) from C<$srcpath> to C<$dstpath> and call the callback with
191	the C<0> (error) or C<-1> ok.
192
193	This is a composite request that tries to rename(2) the file first. If
194	rename files with C<EXDEV>, it creates the destination file with mode 0200
195	and copies the contents of the source file into it using C<aio_sendfile>,
196	followed by restoring atime, mtime, access mode and uid/gid, in that
197	order, and unlinking the C<$srcpath>.
198
199	If an error occurs, the partial destination file will be unlinked, if
200	possible, except when setting atime, mtime, access mode and uid/gid, where
201	errors are being ignored.
202
203	=cut
204
205	sub aio_move($$$) {
206	my ($src, $dst, $cb) = @_;
207
208	my $grp = aio_group $cb;
209
210	add $grp aio_rename $src, $dst, sub {
211	if ($_[0] && $! == EXDEV) {
212	add $grp aio_open $src, O_RDONLY, 0, sub {
213	if (my $src_fh = $_[0]) {
214	my @stat = stat $src_fh;
215
216	add $grp aio_open $dst, O_WRONLY, 0200, sub {
217	if (my $dst_fh = $_[0]) {
218	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
219	close $src_fh;
220
221	if ($_[0] == $stat[7]) {
222	utime $stat[8], $stat[9], $dst;
223	chmod $stat[2] & 07777, $dst_fh;
224	chown $stat[4], $stat[5], $dst_fh;
225	close $dst_fh;
226
227	add $grp aio_unlink $src, sub {
228	$grp->result ($_[0]);
229	};
230	} else {
231	my $errno = $!;
232	add $grp aio_unlink $dst, sub {
233	$! = $errno;
234	$grp->result (-1);
235	};
236	}
237	};
238	} else {
239	$grp->result (-1);
240	}
241	},
242
243	} else {
244	$grp->result (-1);
245	}
246	};
247	} else {
248	$grp->result ($_[0]);
249	}
250	};
251
252	$grp
253	}
254		418
255	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)	419	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
256		420
257	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts	421	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
258	reading at byte offset C<$in_offset>, and starts writing at the current	422	reading at byte offset C<$in_offset>, and starts writing at the current
…		…
262		426
263	This call tries to make use of a native C<sendfile> syscall to provide	427	This call tries to make use of a native C<sendfile> syscall to provide
264	zero-copy operation. For this to work, C<$out_fh> should refer to a	428	zero-copy operation. For this to work, C<$out_fh> should refer to a
265	socket, and C<$in_fh> should refer to mmap'able file.	429	socket, and C<$in_fh> should refer to mmap'able file.
266		430
267	If the native sendfile call fails or is not implemented, it will be	431	If a native sendfile cannot be found or it fails with C<ENOSYS>,
		432	C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or C<ENOTSOCK>,
268	emulated, so you can call C<aio_sendfile> on any type of filehandle	433	it will be emulated, so you can call C<aio_sendfile> on any type of
269	regardless of the limitations of the operating system.	434	filehandle regardless of the limitations of the operating system.
270		435
271	Please note, however, that C<aio_sendfile> can read more bytes from	436	Please note, however, that C<aio_sendfile> can read more bytes from
272	C<$in_fh> than are written, and there is no way to find out how many	437	C<$in_fh> than are written, and there is no way to find out how many
273	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only	438	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
274	provides the number of bytes written to C<$out_fh>. Only if the result	439	provides the number of bytes written to C<$out_fh>. Only if the result
275	value equals C<$length> one can assume that C<$length> bytes have been	440	value equals C<$length> one can assume that C<$length> bytes have been
276	read.	441	read.
		442
277		443
278	=item aio_readahead $fh,$offset,$length, $callback->($retval)	444	=item aio_readahead $fh,$offset,$length, $callback->($retval)
279		445
280	C<aio_readahead> populates the page cache with data from a file so that	446	C<aio_readahead> populates the page cache with data from a file so that
281	subsequent reads from that file will not block on disk I/O. The C<$offset>	447	subsequent reads from that file will not block on disk I/O. The C<$offset>
…		…
287	file. The current file offset of the file is left unchanged.	453	file. The current file offset of the file is left unchanged.
288		454
289	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	455	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
290	emulated by simply reading the data, which would have a similar effect.	456	emulated by simply reading the data, which would have a similar effect.
291		457
		458
292	=item aio_stat $fh_or_path, $callback->($status)	459	=item aio_stat $fh_or_path, $callback->($status)
293		460
294	=item aio_lstat $fh, $callback->($status)	461	=item aio_lstat $fh, $callback->($status)
295		462
296	Works like perl's C<stat> or C<lstat> in void context. The callback will	463	Works like perl's C<stat> or C<lstat> in void context. The callback will
…		…
309	aio_stat "/etc/passwd", sub {	476	aio_stat "/etc/passwd", sub {
310	$_[0] and die "stat failed: $!";	477	$_[0] and die "stat failed: $!";
311	print "size is ", -s _, "\n";	478	print "size is ", -s _, "\n";
312	};	479	};
313		480
		481
		482	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		483
		484	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		485	whether a file handle or path was passed.
		486
		487	On success, the callback is passed a hash reference with the following
		488	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		489	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		490	is passed.
		491
		492	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		493	C<ST_NOSUID>.
		494
		495	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		496	their correct value when available, or to C<0> on systems that do
		497	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		498	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		499	C<ST_NODIRATIME> and C<ST_RELATIME>.
		500
		501	Example: stat C</wd> and dump out the data if successful.
		502
		503	aio_statvfs "/wd", sub {
		504	my $f = $_[0]
		505	or die "statvfs: $!";
		506
		507	use Data::Dumper;
		508	say Dumper $f;
		509	};
		510
		511	# result:
		512	{
		513	bsize => 1024,
		514	bfree => 4333064312,
		515	blocks => 10253828096,
		516	files => 2050765568,
		517	flag => 4096,
		518	favail => 2042092649,
		519	bavail => 4333064312,
		520	ffree => 2042092649,
		521	namemax => 255,
		522	frsize => 1024,
		523	fsid => 1810
		524	}
		525
		526
		527	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		528
		529	Works like perl's C<utime> function (including the special case of $atime
		530	and $mtime being undef). Fractional times are supported if the underlying
		531	syscalls support them.
		532
		533	When called with a pathname, uses utimes(2) if available, otherwise
		534	utime(2). If called on a file descriptor, uses futimes(2) if available,
		535	otherwise returns ENOSYS, so this is not portable.
		536
		537	Examples:
		538
		539	# set atime and mtime to current time (basically touch(1)):
		540	aio_utime "path", undef, undef;
		541	# set atime to current time and mtime to beginning of the epoch:
		542	aio_utime "path", time, undef; # undef==0
		543
		544
		545	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		546
		547	Works like perl's C<chown> function, except that C<undef> for either $uid
		548	or $gid is being interpreted as "do not change" (but -1 can also be used).
		549
		550	Examples:
		551
		552	# same as "chown root path" in the shell:
		553	aio_chown "path", 0, -1;
		554	# same as above:
		555	aio_chown "path", 0, undef;
		556
		557
		558	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		559
		560	Works like truncate(2) or ftruncate(2).
		561
		562
		563	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		564
		565	Works like perl's C<chmod> function.
		566
		567
314	=item aio_unlink $pathname, $callback->($status)	568	=item aio_unlink $pathname, $callback->($status)
315		569
316	Asynchronously unlink (delete) a file and call the callback with the	570	Asynchronously unlink (delete) a file and call the callback with the
317	result code.	571	result code.
318		572
		573
		574	=item aio_mknod $path, $mode, $dev, $callback->($status)
		575
		576	[EXPERIMENTAL]
		577
		578	Asynchronously create a device node (or fifo). See mknod(2).
		579
		580	The only (POSIX-) portable way of calling this function is:
		581
		582	aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		583
		584
319	=item aio_link $srcpath, $dstpath, $callback->($status)	585	=item aio_link $srcpath, $dstpath, $callback->($status)
320		586
321	Asynchronously create a new link to the existing object at C<$srcpath> at	587	Asynchronously create a new link to the existing object at C<$srcpath> at
322	the path C<$dstpath> and call the callback with the result code.	588	the path C<$dstpath> and call the callback with the result code.
323		589
		590
324	=item aio_symlink $srcpath, $dstpath, $callback->($status)	591	=item aio_symlink $srcpath, $dstpath, $callback->($status)
325		592
326	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at	593	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
327	the path C<$dstpath> and call the callback with the result code.	594	the path C<$dstpath> and call the callback with the result code.
328		595
		596
		597	=item aio_readlink $path, $callback->($link)
		598
		599	Asynchronously read the symlink specified by C<$path> and pass it to
		600	the callback. If an error occurs, nothing or undef gets passed to the
		601	callback.
		602
		603
329	=item aio_rename $srcpath, $dstpath, $callback->($status)	604	=item aio_rename $srcpath, $dstpath, $callback->($status)
330		605
331	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as	606	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
332	rename(2) and call the callback with the result code.	607	rename(2) and call the callback with the result code.
333		608
		609
		610	=item aio_mkdir $pathname, $mode, $callback->($status)
		611
		612	Asynchronously mkdir (create) a directory and call the callback with
		613	the result code. C<$mode> will be modified by the umask at the time the
		614	request is executed, so do not change your umask.
		615
		616
334	=item aio_rmdir $pathname, $callback->($status)	617	=item aio_rmdir $pathname, $callback->($status)
335		618
336	Asynchronously rmdir (delete) a directory and call the callback with the	619	Asynchronously rmdir (delete) a directory and call the callback with the
337	result code.	620	result code.
		621
338		622
339	=item aio_readdir $pathname, $callback->($entries)	623	=item aio_readdir $pathname, $callback->($entries)
340		624
341	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire	625	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
342	directory (i.e. opendir + readdir + closedir). The entries will not be	626	directory (i.e. opendir + readdir + closedir). The entries will not be
343	sorted, and will B<NOT> include the C<.> and C<..> entries.	627	sorted, and will B<NOT> include the C<.> and C<..> entries.
344		628
345	The callback a single argument which is either C<undef> or an array-ref	629	The callback is passed a single argument which is either C<undef> or an
346	with the filenames.	630	array-ref with the filenames.
		631
		632
		633	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		634
		635	Quite similar to C<aio_readdir>, but the C<$flags> argument allows to tune
		636	behaviour and output format. In case of an error, C<$entries> will be
		637	C<undef>.
		638
		639	The flags are a combination of the following constants, ORed together (the
		640	flags will also be passed to the callback, possibly modified):
		641
		642	=over 4
		643
		644	=item IO::AIO::READDIR_DENTS
		645
		646	When this flag is off, then the callback gets an arrayref with of names
		647	only (as with C<aio_readdir>), otherwise it gets an arrayref with
		648	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		649	entry in more detail.
		650
		651	C<$name> is the name of the entry.
		652
		653	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		654
		655	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		656	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		657	C<IO::AIO::DT_WHT>.
		658
		659	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		660	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		661	scalars are read-only: you can not modify them.
		662
		663	C<$inode> is the inode number (which might not be exact on systems with 64
		664	bit inode numbers and 32 bit perls). This field has unspecified content on
		665	systems that do not deliver the inode information.
		666
		667	=item IO::AIO::READDIR_DIRS_FIRST
		668
		669	When this flag is set, then the names will be returned in an order where
		670	likely directories come first. This is useful when you need to quickly
		671	find directories, or you want to find all directories while avoiding to
		672	stat() each entry.
		673
		674	If the system returns type information in readdir, then this is used
		675	to find directories directly. Otherwise, likely directories are files
		676	beginning with ".", or otherwise files with no dots, of which files with
		677	short names are tried first.
		678
		679	=item IO::AIO::READDIR_STAT_ORDER
		680
		681	When this flag is set, then the names will be returned in an order
		682	suitable for stat()'ing each one. That is, when you plan to stat()
		683	all files in the given directory, then the returned order will likely
		684	be fastest.
		685
		686	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		687	the likely dirs come first, resulting in a less optimal stat order.
		688
		689	=item IO::AIO::READDIR_FOUND_UNKNOWN
		690
		691	This flag should not be set when calling C<aio_readdirx>. Instead, it
		692	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		693	C<IO::AIO::DT_UNKNOWN>. The absense of this flag therefore indicates that all
		694	C<$type>'s are known, which can be used to speed up some algorithms.
		695
		696	=back
		697
		698
		699	=item aio_load $path, $data, $callback->($status)
		700
		701	This is a composite request that tries to fully load the given file into
		702	memory. Status is the same as with aio_read.
		703
		704	=cut
		705
		706	sub aio_load($$;$) {
		707	my ($path, undef, $cb) = @_;
		708	my $data = \$_[1];
		709
		710	my $pri = aioreq_pri;
		711	my $grp = aio_group $cb;
		712
		713	aioreq_pri $pri;
		714	add $grp aio_open $path, O_RDONLY, 0, sub {
		715	my $fh = shift
		716	or return $grp->result (-1);
		717
		718	aioreq_pri $pri;
		719	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		720	$grp->result ($_[0]);
		721	};
		722	};
		723
		724	$grp
		725	}
		726
		727	=item aio_copy $srcpath, $dstpath, $callback->($status)
		728
		729	Try to copy the I<file> (directories not supported as either source or
		730	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		731	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		732
		733	This is a composite request that creates the destination file with
		734	mode 0200 and copies the contents of the source file into it using
		735	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		736	uid/gid, in that order.
		737
		738	If an error occurs, the partial destination file will be unlinked, if
		739	possible, except when setting atime, mtime, access mode and uid/gid, where
		740	errors are being ignored.
		741
		742	=cut
		743
		744	sub aio_copy($$;$) {
		745	my ($src, $dst, $cb) = @_;
		746
		747	my $pri = aioreq_pri;
		748	my $grp = aio_group $cb;
		749
		750	aioreq_pri $pri;
		751	add $grp aio_open $src, O_RDONLY, 0, sub {
		752	if (my $src_fh = $_[0]) {
		753	my @stat = stat $src_fh; # hmm, might block over nfs?
		754
		755	aioreq_pri $pri;
		756	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		757	if (my $dst_fh = $_[0]) {
		758	aioreq_pri $pri;
		759	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		760	if ($_[0] == $stat[7]) {
		761	$grp->result (0);
		762	close $src_fh;
		763
		764	my $ch = sub {
		765	aioreq_pri $pri;
		766	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		767	aioreq_pri $pri;
		768	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		769	aioreq_pri $pri;
		770	add $grp aio_close $dst_fh;
		771	}
		772	};
		773	};
		774
		775	aioreq_pri $pri;
		776	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		777	if ($_[0] < 0 && $! == ENOSYS) {
		778	aioreq_pri $pri;
		779	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		780	} else {
		781	$ch->();
		782	}
		783	};
		784	} else {
		785	$grp->result (-1);
		786	close $src_fh;
		787	close $dst_fh;
		788
		789	aioreq $pri;
		790	add $grp aio_unlink $dst;
		791	}
		792	};
		793	} else {
		794	$grp->result (-1);
		795	}
		796	},
		797
		798	} else {
		799	$grp->result (-1);
		800	}
		801	};
		802
		803	$grp
		804	}
		805
		806	=item aio_move $srcpath, $dstpath, $callback->($status)
		807
		808	Try to move the I<file> (directories not supported as either source or
		809	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		810	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		811
		812	This is a composite request that tries to rename(2) the file first; if
		813	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		814	that is successful, unlinks the C<$srcpath>.
		815
		816	=cut
		817
		818	sub aio_move($$;$) {
		819	my ($src, $dst, $cb) = @_;
		820
		821	my $pri = aioreq_pri;
		822	my $grp = aio_group $cb;
		823
		824	aioreq_pri $pri;
		825	add $grp aio_rename $src, $dst, sub {
		826	if ($_[0] && $! == EXDEV) {
		827	aioreq_pri $pri;
		828	add $grp aio_copy $src, $dst, sub {
		829	$grp->result ($_[0]);
		830
		831	if (!$_[0]) {
		832	aioreq_pri $pri;
		833	add $grp aio_unlink $src;
		834	}
		835	};
		836	} else {
		837	$grp->result ($_[0]);
		838	}
		839	};
		840
		841	$grp
		842	}
347		843
348	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)	844	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
349		845
350	[EXPERIMENTAL due to internal aio_group use]
351
352	Scans a directory (similar to C<aio_readdir>) but additionally tries to	846	Scans a directory (similar to C<aio_readdir>) but additionally tries to
353	separate the entries of directory C<$path> into two sets of names, ones	847	efficiently separate the entries of directory C<$path> into two sets of
354	you can recurse into (directories or links to them), and ones you cannot	848	names, directories you can recurse into (directories), and ones you cannot
355	recurse into (everything else).	849	recurse into (everything else, including symlinks to directories).
356		850
357	C<aio_scandir> is a composite request that creates of many sub requests_	851	C<aio_scandir> is a composite request that creates of many sub requests_
358	C<$maxreq> specifies the maximum number of outstanding aio requests that	852	C<$maxreq> specifies the maximum number of outstanding aio requests that
359	this function generates. If it is C<< <= 0 >>, then a suitable default	853	this function generates. If it is C<< <= 0 >>, then a suitable default
360	will be chosen (currently 6).	854	will be chosen (currently 4).
361		855
362	On error, the callback is called without arguments, otherwise it receives	856	On error, the callback is called without arguments, otherwise it receives
363	two array-refs with path-relative entry names.	857	two array-refs with path-relative entry names.
364		858
365	Example:	859	Example:
…		…
372		866
373	Implementation notes.	867	Implementation notes.
374		868
375	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.	869	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
376		870
		871	If readdir returns file type information, then this is used directly to
		872	find directories.
		873
377	After reading the directory, the modification time, size etc. of the	874	Otherwise, after reading the directory, the modification time, size etc.
378	directory before and after the readdir is checked, and if they match (and	875	of the directory before and after the readdir is checked, and if they
379	isn't the current time), the link count will be used to decide how many	876	match (and isn't the current time), the link count will be used to decide
380	entries are directories (if >= 2). Otherwise, no knowledge of the number	877	how many entries are directories (if >= 2). Otherwise, no knowledge of the
381	of subdirectories will be assumed.	878	number of subdirectories will be assumed.
382		879
383	Then entries will be sorted into likely directories (everything without	880	Then entries will be sorted into likely directories a non-initial dot
384	a non-initial dot currently) and likely non-directories (everything	881	currently) and likely non-directories (see C<aio_readdirx>). Then every
385	else). Then every entry plus an appended C</.> will be C<stat>'ed,	882	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
386	likely directories first. If that succeeds, it assumes that the entry	883	in order of their inode numbers. If that succeeds, it assumes that the
387	is a directory or a symlink to directory (which will be checked	884	entry is a directory or a symlink to directory (which will be checked
388	seperately). This is often faster than stat'ing the entry itself because	885	seperately). This is often faster than stat'ing the entry itself because
389	filesystems might detect the type of the entry without reading the inode	886	filesystems might detect the type of the entry without reading the inode
390	data (e.g. ext2fs filetype feature).	887	data (e.g. ext2fs filetype feature), even on systems that cannot return
		888	the filetype information on readdir.
391		889
392	If the known number of directories (link count - 2) has been reached, the	890	If the known number of directories (link count - 2) has been reached, the
393	rest of the entries is assumed to be non-directories.	891	rest of the entries is assumed to be non-directories.
394		892
395	This only works with certainty on POSIX (= UNIX) filesystems, which	893	This only works with certainty on POSIX (= UNIX) filesystems, which
…		…
399	as those tend to return 0 or 1 as link counts, which disables the	897	as those tend to return 0 or 1 as link counts, which disables the
400	directory counting heuristic.	898	directory counting heuristic.
401		899
402	=cut	900	=cut
403		901
404	sub aio_scandir($$$) {	902	sub aio_scandir($$;$) {
405	my ($path, $maxreq, $cb) = @_;	903	my ($path, $maxreq, $cb) = @_;
406		904
		905	my $pri = aioreq_pri;
		906
407	my $grp = aio_group $cb;	907	my $grp = aio_group $cb;
408		908
409	$maxreq = 6 if $maxreq <= 0;	909	$maxreq = 4 if $maxreq <= 0;
410		910
411	# stat once	911	# stat once
		912	aioreq_pri $pri;
412	add $grp aio_stat $path, sub {	913	add $grp aio_stat $path, sub {
413	return $grp->result () if $_[0];	914	return $grp->result () if $_[0];
414	my $now = time;	915	my $now = time;
415	my $hash1 = join ":", (stat _)[0,1,3,7,9];	916	my $hash1 = join ":", (stat _)[0,1,3,7,9];
416		917
417	# read the directory entries	918	# read the directory entries
		919	aioreq_pri $pri;
418	add $grp aio_readdir $path, sub {	920	add $grp aio_readdirx $path, READDIR_DIRS_FIRST, sub {
419	my $entries = shift	921	my $entries = shift
420	or return $grp->result ();	922	or return $grp->result ();
421		923
422	# stat the dir another time	924	# stat the dir another time
		925	aioreq_pri $pri;
423	add $grp aio_stat $path, sub {	926	add $grp aio_stat $path, sub {
424	my $hash2 = join ":", (stat _)[0,1,3,7,9];	927	my $hash2 = join ":", (stat _)[0,1,3,7,9];
425		928
426	my $ndirs;	929	my $ndirs;
427		930
428	# take the slow route if anything looks fishy	931	# take the slow route if anything looks fishy
429	if ($hash1 ne $hash2 or (stat _)[9] == $now) {	932	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
430	$ndirs = -1;	933	$ndirs = -1;
431	} else {	934	} else {
432	# if nlink == 2, we are finished	935	# if nlink == 2, we are finished
433	# on non-posix-fs's, we rely on nlink < 2	936	# for non-posix-fs's, we rely on nlink < 2
434	$ndirs = (stat _)[3] - 2	937	$ndirs = (stat _)[3] - 2
435	or return $grp->result ([], $entries);	938	or return $grp->result ([], $entries);
436	}	939	}
437		940
438	# sort into likely dirs and likely nondirs
439	# dirs == files without ".", short entries first
440	$entries = [map $_->[0],
441	sort { $b->[1] cmp $a->[1] }
442	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
443	@$entries];
444
445	my (@dirs, @nondirs);	941	my (@dirs, @nondirs);
446		942
447	my ($statcb, $schedcb);
448	my $nreq = 0;
449
450	my $statgrp = add $grp aio_group;	943	my $statgrp = add $grp aio_group sub {
		944	$grp->result (\@dirs, \@nondirs);
		945	};
451		946
452	$schedcb = sub {	947	limit $statgrp $maxreq;
453	if (@$entries) {	948	feed $statgrp sub {
454	if ($nreq < $maxreq) {	949	return unless @$entries;
455	my $ent = pop @$entries;	950	my $entry = shift @$entries;
		951
		952	aioreq_pri $pri;
		953	add $statgrp aio_stat "$path/$entry/.", sub {
		954	if ($_[0] < 0) {
		955	push @nondirs, $entry;
		956	} else {
		957	# need to check for real directory
		958	aioreq_pri $pri;
		959	add $statgrp aio_lstat "$path/$entry", sub {
		960	if (-d _) {
		961	push @dirs, $entry;
		962
		963	unless (--$ndirs) {
		964	push @nondirs, @$entries;
		965	feed $statgrp;
		966	}
		967	} else {
		968	push @nondirs, $entry;
		969	}
456	$nreq++;	970	}
457	add $statgrp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) };
458	}	971	}
459	} elsif (!$nreq) {
460	# finished
461	$statgrp->cancel;
462	undef $statcb;
463	undef $schedcb;
464	$grp->result (\@dirs, \@nondirs);
465	}	972	};
466	};	973	};
467	$statcb = sub {
468	my ($status, $entry) = @_;
469
470	if ($status < 0) {
471	$nreq--;
472	push @nondirs, $entry;
473	&$schedcb;
474	} else {
475	# need to check for real directory
476	add $grp aio_lstat "$path/$entry", sub {
477	$nreq--;
478
479	if (-d _) {
480	push @dirs, $entry;
481
482	if (!--$ndirs) {
483	push @nondirs, @$entries;
484	$entries = [];
485	}
486	} else {
487	push @nondirs, $entry;
488	}
489
490	&$schedcb;
491	}
492	}
493	};
494
495	&$schedcb while @$entries && $nreq < $maxreq;
496	};	974	};
497	};	975	};
498	};	976	};
499		977
500	$grp	978	$grp
501	}	979	}
502		980
		981	=item aio_rmtree $path, $callback->($status)
		982
		983	Delete a directory tree starting (and including) C<$path>, return the
		984	status of the final C<rmdir> only. This is a composite request that
		985	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		986	everything else.
		987
		988	=cut
		989
		990	sub aio_rmtree;
		991	sub aio_rmtree($;$) {
		992	my ($path, $cb) = @_;
		993
		994	my $pri = aioreq_pri;
		995	my $grp = aio_group $cb;
		996
		997	aioreq_pri $pri;
		998	add $grp aio_scandir $path, 0, sub {
		999	my ($dirs, $nondirs) = @_;
		1000
		1001	my $dirgrp = aio_group sub {
		1002	add $grp aio_rmdir $path, sub {
		1003	$grp->result ($_[0]);
		1004	};
		1005	};
		1006
		1007	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1008	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1009
		1010	add $grp $dirgrp;
		1011	};
		1012
		1013	$grp
		1014	}
		1015
		1016	=item aio_sync $callback->($status)
		1017
		1018	Asynchronously call sync and call the callback when finished.
		1019
503	=item aio_fsync $fh, $callback->($status)	1020	=item aio_fsync $fh, $callback->($status)
504		1021
505	Asynchronously call fsync on the given filehandle and call the callback	1022	Asynchronously call fsync on the given filehandle and call the callback
506	with the fsync result code.	1023	with the fsync result code.
507		1024
…		…
511	callback with the fdatasync result code.	1028	callback with the fdatasync result code.
512		1029
513	If this call isn't available because your OS lacks it or it couldn't be	1030	If this call isn't available because your OS lacks it or it couldn't be
514	detected, it will be emulated by calling C<fsync> instead.	1031	detected, it will be emulated by calling C<fsync> instead.
515		1032
		1033	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1034
		1035	Sync the data portion of the file specified by C<$offset> and C<$length>
		1036	to disk (but NOT the metadata), by calling the Linux-specific
		1037	sync_file_range call. If sync_file_range is not available or it returns
		1038	ENOSYS, then fdatasync or fsync is being substituted.
		1039
		1040	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1041	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1042	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1043	manpage for details.
		1044
		1045	=item aio_pathsync $path, $callback->($status)
		1046
		1047	This request tries to open, fsync and close the given path. This is a
		1048	composite request intended to sync directories after directory operations
		1049	(E.g. rename). This might not work on all operating systems or have any
		1050	specific effect, but usually it makes sure that directory changes get
		1051	written to disc. It works for anything that can be opened for read-only,
		1052	not just directories.
		1053
		1054	Future versions of this function might fall back to other methods when
		1055	C<fsync> on the directory fails (such as calling C<sync>).
		1056
		1057	Passes C<0> when everything went ok, and C<-1> on error.
		1058
		1059	=cut
		1060
		1061	sub aio_pathsync($;$) {
		1062	my ($path, $cb) = @_;
		1063
		1064	my $pri = aioreq_pri;
		1065	my $grp = aio_group $cb;
		1066
		1067	aioreq_pri $pri;
		1068	add $grp aio_open $path, O_RDONLY, 0, sub {
		1069	my ($fh) = @_;
		1070	if ($fh) {
		1071	aioreq_pri $pri;
		1072	add $grp aio_fsync $fh, sub {
		1073	$grp->result ($_[0]);
		1074
		1075	aioreq_pri $pri;
		1076	add $grp aio_close $fh;
		1077	};
		1078	} else {
		1079	$grp->result (-1);
		1080	}
		1081	};
		1082
		1083	$grp
		1084	}
		1085
		1086	=item aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1087
		1088	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1089	scalars (see the L<Sys::Mmap> or L<Mmap> modules for details on this, note
		1090	that the scalar must only be modified in-place while an aio operation is
		1091	pending on it).
		1092
		1093	It calls the C<msync> function of your OS, if available, with the memory
		1094	area starting at C<$offset> in the string and ending C<$length> bytes
		1095	later. If C<$length> is negative, counts from the end, and if C<$length>
		1096	is C<undef>, then it goes till the end of the string. The flags can be
		1097	a combination of C<IO::AIO::MS_ASYNC>, C<IO::AIO::MS_INVALIDATE> and
		1098	C<IO::AIO::MS_SYNC>.
		1099
		1100	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1101
		1102	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1103	scalars.
		1104
		1105	It touches (reads or writes) all memory pages in the specified
		1106	range inside the scalar. All caveats and parameters are the same
		1107	as for C<aio_msync>, above, except for flags, which must be either
		1108	C<0> (which reads all pages and ensures they are instantiated) or
		1109	C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and
		1110	writing an octet from it, which dirties the page).
		1111
516	=item aio_group $callback->(...)	1112	=item aio_group $callback->(...)
517
518	[EXPERIMENTAL]
519		1113
520	This is a very special aio request: Instead of doing something, it is a	1114	This is a very special aio request: Instead of doing something, it is a
521	container for other aio requests, which is useful if you want to bundle	1115	container for other aio requests, which is useful if you want to bundle
522	many requests into a single, composite, request.	1116	many requests into a single, composite, request with a definite callback
		1117	and the ability to cancel the whole request with its subrequests.
523		1118
524	Returns an object of class L<IO::AIO::GRP>. See its documentation below	1119	Returns an object of class L<IO::AIO::GRP>. See its documentation below
525	for more info.	1120	for more info.
526		1121
527	Example:	1122	Example:
…		…
533	add $grp	1128	add $grp
534	(aio_stat ...),	1129	(aio_stat ...),
535	(aio_stat ...),	1130	(aio_stat ...),
536	...;	1131	...;
537		1132
		1133	=item aio_nop $callback->()
		1134
		1135	This is a special request - it does nothing in itself and is only used for
		1136	side effects, such as when you want to add a dummy request to a group so
		1137	that finishing the requests in the group depends on executing the given
		1138	code.
		1139
		1140	While this request does nothing, it still goes through the execution
		1141	phase and still requires a worker thread. Thus, the callback will not
		1142	be executed immediately but only after other requests in the queue have
		1143	entered their execution phase. This can be used to measure request
		1144	latency.
		1145
538	=item IO::AIO::aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED*	1146	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
539		1147
540	Mainly used for debugging and benchmarking, this aio request puts one of	1148	Mainly used for debugging and benchmarking, this aio request puts one of
541	the request workers to sleep for the given time.	1149	the request workers to sleep for the given time.
542		1150
543	While it is theoretically handy to have simple I/O scheduling requests	1151	While it is theoretically handy to have simple I/O scheduling requests
544	like sleep and file handle readable/writable, the overhead this creates	1152	like sleep and file handle readable/writable, the overhead this creates is
545	is immense, so do not use this function except to put your application	1153	immense (it blocks a thread for a long time) so do not use this function
546	under artificial I/O pressure.	1154	except to put your application under artificial I/O pressure.
547		1155
548	=back	1156	=back
549		1157
550	=head2 IO::AIO::REQ CLASS	1158	=head2 IO::AIO::REQ CLASS
551		1159
552	All non-aggregate C<aio_*> functions return an object of this class when	1160	All non-aggregate C<aio_*> functions return an object of this class when
553	called in non-void context.	1161	called in non-void context.
554		1162
555	A request always moves through the following five states in its lifetime,
556	in order: B<ready> (request has been created, but has not been executed
557	yet), B<execute> (request is currently being executed), B<pending>
558	(request has been executed but callback has not been called yet),
559	B<result> (results are being processed synchronously, includes calling the
560	callback) and B<done> (request has reached the end of its lifetime and
561	holds no resources anymore).
562
563	=over 4	1163	=over 4
564		1164
565	=item $req->cancel	1165	=item cancel $req
566		1166
567	Cancels the request, if possible. Has the effect of skipping execution	1167	Cancels the request, if possible. Has the effect of skipping execution
568	when entering the B<execute> state and skipping calling the callback when	1168	when entering the B<execute> state and skipping calling the callback when
569	entering the the B<result> state, but will leave the request otherwise	1169	entering the the B<result> state, but will leave the request otherwise
570	untouched. That means that requests that currently execute will not be	1170	untouched (with the exception of readdir). That means that requests that
571	stopped and resources held by the request will not be freed prematurely.	1171	currently execute will not be stopped and resources held by the request
		1172	will not be freed prematurely.
		1173
		1174	=item cb $req $callback->(...)
		1175
		1176	Replace (or simply set) the callback registered to the request.
572		1177
573	=back	1178	=back
574		1179
575	=head2 IO::AIO::GRP CLASS	1180	=head2 IO::AIO::GRP CLASS
576		1181
…		…
616	=item * They can also can also be added to other IO::AIO::GRP objects.	1221	=item * They can also can also be added to other IO::AIO::GRP objects.
617		1222
618	=item * You must not add requests to a group from within the group callback (or	1223	=item * You must not add requests to a group from within the group callback (or
619	any later time).	1224	any later time).
620		1225
621	=item * This does not harmonise well with C<max_outstanding>, so best do
622	not combine C<aio_group> with it. Groups and feeders are recommended for
623	this kind of concurrency-limiting.
624
625	=back	1226	=back
626		1227
627	Their lifetime, simplified, looks like this: when they are empty, they	1228	Their lifetime, simplified, looks like this: when they are empty, they
628	will finish very quickly. If they contain only requests that are in the	1229	will finish very quickly. If they contain only requests that are in the
629	C<done> state, they will also finish. Otherwise they will continue to	1230	C<done> state, they will also finish. Otherwise they will continue to
630	exist.	1231	exist.
631		1232
632	That means after creating a group you have some time to add requests. And	1233	That means after creating a group you have some time to add requests
633	in the callbacks of those requests, you can add further requests to the	1234	(precisely before the callback has been invoked, which is only done within
634	group. And only when all those requests have finished will the the group	1235	the C<poll_cb>). And in the callbacks of those requests, you can add
635	itself finish.	1236	further requests to the group. And only when all those requests have
		1237	finished will the the group itself finish.
636		1238
637	=over 4	1239	=over 4
638		1240
		1241	=item add $grp ...
		1242
639	=item $grp->add (...)	1243	=item $grp->add (...)
640
641	=item add $grp ...
642		1244
643	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can	1245	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
644	be added, including other groups, as long as you do not create circular	1246	be added, including other groups, as long as you do not create circular
645	dependencies.	1247	dependencies.
646		1248
647	Returns all its arguments.	1249	Returns all its arguments.
648		1250
		1251	=item $grp->cancel_subs
		1252
		1253	Cancel all subrequests and clears any feeder, but not the group request
		1254	itself. Useful when you queued a lot of events but got a result early.
		1255
		1256	The group request will finish normally (you cannot add requests to the
		1257	group).
		1258
649	=item $grp->result (...)	1259	=item $grp->result (...)
650		1260
651	Set the result value(s) that will be passed to the group callback when all	1261	Set the result value(s) that will be passed to the group callback when all
652	subrequests have finished. By default, no argument will be passed.	1262	subrequests have finished and set the groups errno to the current value
		1263	of errno (just like calling C<errno> without an error number). By default,
		1264	no argument will be passed and errno is zero.
653		1265
654	=item $grp->set_feeder ($callback->($grp))	1266	=item $grp->errno ([$errno])
655		1267
656	[VERY EXPERIMENTAL]	1268	Sets the group errno value to C<$errno>, or the current value of errno
		1269	when the argument is missing.
		1270
		1271	Every aio request has an associated errno value that is restored when
		1272	the callback is invoked. This method lets you change this value from its
		1273	default (0).
		1274
		1275	Calling C<result> will also set errno, so make sure you either set C<$!>
		1276	before the call to C<result>, or call c<errno> after it.
		1277
		1278	=item feed $grp $callback->($grp)
657		1279
658	Sets a feeder/generator on this group: every group can have an attached	1280	Sets a feeder/generator on this group: every group can have an attached
659	generator that generates requests if idle. The idea behind this is that,	1281	generator that generates requests if idle. The idea behind this is that,
660	although you could just queue as many requests as you want in a group,	1282	although you could just queue as many requests as you want in a group,
661	this might starve other requests for a potentially long time. For	1283	this might starve other requests for a potentially long time. For example,
662	example, C<aio_scandir> might generate hundreds of thousands C<aio_stat>	1284	C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests,
663	requests, delaying any later requests for a long time.	1285	delaying any later requests for a long time.
664		1286
665	To avoid this, and allow incremental generation of requests, you can	1287	To avoid this, and allow incremental generation of requests, you can
666	instead a group and set a feeder on it that generates those requests. The	1288	instead a group and set a feeder on it that generates those requests. The
667	feeder will be called whenever there are few enough (see C<feeder_limit>,	1289	feed callback will be called whenever there are few enough (see C<limit>,
668	below) requests active in the group itself and is expected to queue more	1290	below) requests active in the group itself and is expected to queue more
669	requests.	1291	requests.
670		1292
671	The feeder can queue as many requests as it likes (i.e. C<add> does not	1293	The feed callback can queue as many requests as it likes (i.e. C<add> does
672	impose any limits).	1294	not impose any limits).
673		1295
674	If the feeder does not queue more requests when called, it will be	1296	If the feed does not queue more requests when called, it will be
675	automatically removed from the group.	1297	automatically removed from the group.
676		1298
677	If the feeder limit is C<0>, it will be set to C<2> automatically.	1299	If the feed limit is C<0> when this method is called, it will be set to
		1300	C<2> automatically.
678		1301
679	Example:	1302	Example:
680		1303
681	# stat all files in @files, but only ever use four aio requests concurrently:	1304	# stat all files in @files, but only ever use four aio requests concurrently:
682		1305
683	my $grp = aio_group sub { print "finished\n" };	1306	my $grp = aio_group sub { print "finished\n" };
684	$grp->feeder_limit (4);	1307	limit $grp 4;
685	$grp->set_feeder (sub {	1308	feed $grp sub {
686	my $file = pop @files	1309	my $file = pop @files
687	or return;	1310	or return;
688		1311
689	add $grp aio_stat $file, sub { ... };	1312	add $grp aio_stat $file, sub { ... };
690	});	1313	};
691		1314
692	=item $grp->feeder_limit ($num)	1315	=item limit $grp $num
693		1316
694	Sets the feeder limit for the group: The feeder will be called whenever	1317	Sets the feeder limit for the group: The feeder will be called whenever
695	the group contains less than this many requests.	1318	the group contains less than this many requests.
696		1319
697	Setting the limit to C<0> will pause the feeding process.	1320	Setting the limit to C<0> will pause the feeding process.
698		1321
		1322	The default value for the limit is C<0>, but note that setting a feeder
		1323	automatically bumps it up to C<2>.
		1324
699	=back	1325	=back
700		1326
701	=head2 SUPPORT FUNCTIONS	1327	=head2 SUPPORT FUNCTIONS
702		1328
		1329	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1330
703	=over 4	1331	=over 4
704		1332
705	=item $fileno = IO::AIO::poll_fileno	1333	=item $fileno = IO::AIO::poll_fileno
706		1334
707	Return the I<request result pipe file descriptor>. This filehandle must be	1335	Return the I<request result pipe file descriptor>. This filehandle must be
708	polled for reading by some mechanism outside this module (e.g. Event or	1336	polled for reading by some mechanism outside this module (e.g. EV, Glib,
709	select, see below or the SYNOPSIS). If the pipe becomes readable you have	1337	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
710	to call C<poll_cb> to check the results.	1338	you have to call C<poll_cb> to check the results.
711		1339
712	See C<poll_cb> for an example.	1340	See C<poll_cb> for an example.
713		1341
714	=item IO::AIO::poll_cb	1342	=item IO::AIO::poll_cb
715		1343
716	Process all outstanding events on the result pipe. You have to call this	1344	Process some outstanding events on the result pipe. You have to call this
717	regularly. Returns the number of events processed. Returns immediately	1345	regularly. Returns C<0> if all events could be processed, or C<-1> if it
718	when no events are outstanding.	1346	returned earlier for whatever reason. Returns immediately when no events
		1347	are outstanding. The amount of events processed depends on the settings of
		1348	C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
		1349
		1350	If not all requests were processed for whatever reason, the filehandle
		1351	will still be ready when C<poll_cb> returns, so normally you don't have to
		1352	do anything special to have it called later.
719		1353
720	Example: Install an Event watcher that automatically calls	1354	Example: Install an Event watcher that automatically calls
721	IO::AIO::poll_cb with high priority:	1355	IO::AIO::poll_cb with high priority (more examples can be found in the
		1356	SYNOPSIS section, at the top of this document):
722		1357
723	Event->io (fd => IO::AIO::poll_fileno,	1358	Event->io (fd => IO::AIO::poll_fileno,
724	poll => 'r', async => 1,	1359	poll => 'r', async => 1,
725	cb => \&IO::AIO::poll_cb);	1360	cb => \&IO::AIO::poll_cb);
726		1361
727	=item IO::AIO::poll_wait	1362	=item IO::AIO::poll_wait
728		1363
		1364	If there are any outstanding requests and none of them in the result
729	Wait till the result filehandle becomes ready for reading (simply does a	1365	phase, wait till the result filehandle becomes ready for reading (simply
730	C<select> on the filehandle. This is useful if you want to synchronously wait	1366	does a C<select> on the filehandle. This is useful if you want to
731	for some requests to finish).	1367	synchronously wait for some requests to finish).
732		1368
733	See C<nreqs> for an example.	1369	See C<nreqs> for an example.
734		1370
		1371	=item IO::AIO::poll
		1372
		1373	Waits until some requests have been handled.
		1374
		1375	Returns the number of requests processed, but is otherwise strictly
		1376	equivalent to:
		1377
		1378	IO::AIO::poll_wait, IO::AIO::poll_cb
		1379
735	=item IO::AIO::nreqs	1380	=item IO::AIO::flush
736		1381
737	Returns the number of requests currently outstanding (i.e. for which their	1382	Wait till all outstanding AIO requests have been handled.
738	callback has not been invoked yet).
739		1383
740	Example: wait till there are no outstanding requests anymore:	1384	Strictly equivalent to:
741		1385
742	IO::AIO::poll_wait, IO::AIO::poll_cb	1386	IO::AIO::poll_wait, IO::AIO::poll_cb
743	while IO::AIO::nreqs;	1387	while IO::AIO::nreqs;
744		1388
745	=item IO::AIO::flush	1389	=item IO::AIO::max_poll_reqs $nreqs
746		1390
747	Wait till all outstanding AIO requests have been handled.	1391	=item IO::AIO::max_poll_time $seconds
748		1392
749	Strictly equivalent to:	1393	These set the maximum number of requests (default C<0>, meaning infinity)
		1394	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1395	the maximum amount of time (default C<0>, meaning infinity) spent in
		1396	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1397	of time C<poll_cb> is allowed to use).
750		1398
751	IO::AIO::poll_wait, IO::AIO::poll_cb	1399	Setting C<max_poll_time> to a non-zero value creates an overhead of one
752	while IO::AIO::nreqs;	1400	syscall per request processed, which is not normally a problem unless your
		1401	callbacks are really really fast or your OS is really really slow (I am
		1402	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
753		1403
754	=item IO::AIO::poll	1404	Setting these is useful if you want to ensure some level of
		1405	interactiveness when perl is not fast enough to process all requests in
		1406	time.
755		1407
756	Waits until some requests have been handled.	1408	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
757		1409
758	Strictly equivalent to:	1410	Example: Install an Event watcher that automatically calls
		1411	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1412	program get the CPU sometimes even under high AIO load.
759		1413
760	IO::AIO::poll_wait, IO::AIO::poll_cb	1414	# try not to spend much more than 0.1s in poll_cb
761	if IO::AIO::nreqs;	1415	IO::AIO::max_poll_time 0.1;
		1416
		1417	# use a low priority so other tasks have priority
		1418	Event->io (fd => IO::AIO::poll_fileno,
		1419	poll => 'r', nice => 1,
		1420	cb => &IO::AIO::poll_cb);
		1421
		1422	=back
		1423
		1424	=head3 CONTROLLING THE NUMBER OF THREADS
		1425
		1426	=over
762		1427
763	=item IO::AIO::min_parallel $nthreads	1428	=item IO::AIO::min_parallel $nthreads
764		1429
765	Set the minimum number of AIO threads to C<$nthreads>. The current	1430	Set the minimum number of AIO threads to C<$nthreads>. The current
766	default is C<8>, which means eight asynchronous operations can execute	1431	default is C<8>, which means eight asynchronous operations can execute
767	concurrently at any one time (the number of outstanding requests,	1432	concurrently at any one time (the number of outstanding requests,
768	however, is unlimited).	1433	however, is unlimited).
769		1434
770	IO::AIO starts threads only on demand, when an AIO request is queued and	1435	IO::AIO starts threads only on demand, when an AIO request is queued and
771	no free thread exists.	1436	no free thread exists. Please note that queueing up a hundred requests can
		1437	create demand for a hundred threads, even if it turns out that everything
		1438	is in the cache and could have been processed faster by a single thread.
772		1439
773	It is recommended to keep the number of threads relatively low, as some	1440	It is recommended to keep the number of threads relatively low, as some
774	Linux kernel versions will scale negatively with the number of threads	1441	Linux kernel versions will scale negatively with the number of threads
775	(higher parallelity => MUCH higher latency). With current Linux 2.6	1442	(higher parallelity => MUCH higher latency). With current Linux 2.6
776	versions, 4-32 threads should be fine.	1443	versions, 4-32 threads should be fine.
…		…
790	This module automatically runs C<max_parallel 0> at program end, to ensure	1457	This module automatically runs C<max_parallel 0> at program end, to ensure
791	that all threads are killed and that there are no outstanding requests.	1458	that all threads are killed and that there are no outstanding requests.
792		1459
793	Under normal circumstances you don't need to call this function.	1460	Under normal circumstances you don't need to call this function.
794		1461
		1462	=item IO::AIO::max_idle $nthreads
		1463
		1464	Limit the number of threads (default: 4) that are allowed to idle (i.e.,
		1465	threads that did not get a request to process within 10 seconds). That
		1466	means if a thread becomes idle while C<$nthreads> other threads are also
		1467	idle, it will free its resources and exit.
		1468
		1469	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1470	to allow for extremely high load situations, but want to free resources
		1471	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1472
		1473	The default is probably ok in most situations, especially if thread
		1474	creation is fast. If thread creation is very slow on your system you might
		1475	want to use larger values.
		1476
795	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1477	=item IO::AIO::max_outstanding $maxreqs
796		1478
797	[DEPRECATED]	1479	This is a very bad function to use in interactive programs because it
		1480	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1481	use an C<aio_group> together with a feed callback.
798		1482
799	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1483	Sets the maximum number of outstanding requests to C<$nreqs>. If you
800	try to queue up more than this number of requests, the caller will block until	1484	do queue up more than this number of requests, the next call to the
801	some requests have been handled.	1485	C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
		1486	function will block until the limit is no longer exceeded.
802		1487
803	The default is very large, so normally there is no practical limit. If you	1488	The default value is very large, so there is no practical limit on the
804	queue up many requests in a loop it often improves speed if you set	1489	number of outstanding requests.
805	this to a relatively low number, such as C<100>.
806		1490
807	This function does not work well together with C<aio_group>'s, and their	1491	You can still queue as many requests as you want. Therefore,
808	feeder interface is better suited to limiting concurrency, so do not use	1492	C<max_outstanding> is mainly useful in simple scripts (with low values) or
809	this function.	1493	as a stop gap to shield against fatal memory overflow (with large values).
810
811	Under normal circumstances you don't need to call this function.
812		1494
813	=back	1495	=back
814		1496
		1497	=head3 STATISTICAL INFORMATION
		1498
		1499	=over
		1500
		1501	=item IO::AIO::nreqs
		1502
		1503	Returns the number of requests currently in the ready, execute or pending
		1504	states (i.e. for which their callback has not been invoked yet).
		1505
		1506	Example: wait till there are no outstanding requests anymore:
		1507
		1508	IO::AIO::poll_wait, IO::AIO::poll_cb
		1509	while IO::AIO::nreqs;
		1510
		1511	=item IO::AIO::nready
		1512
		1513	Returns the number of requests currently in the ready state (not yet
		1514	executed).
		1515
		1516	=item IO::AIO::npending
		1517
		1518	Returns the number of requests currently in the pending state (executed,
		1519	but not yet processed by poll_cb).
		1520
		1521	=back
		1522
		1523	=head3 MISCELLANEOUS FUNCTIONS
		1524
		1525	IO::AIO implements some functions that might be useful, but are not
		1526	asynchronous.
		1527
		1528	=over 4
		1529
		1530	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		1531
		1532	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		1533	but is blocking (this makes most sense if you know the input data is
		1534	likely cached already and the output filehandle is set to non-blocking
		1535	operations).
		1536
		1537	Returns the number of bytes copied, or C<-1> on error.
		1538
		1539	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		1540
		1541	Simply calls the C<posix_fadvise> function (see it's
		1542	manpage for details). The following advice constants are
		1543	avaiable: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		1544	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		1545	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		1546
		1547	On systems that do not implement C<posix_fadvise>, this function returns
		1548	ENOSYS, otherwise the return value of C<posix_fadvise>.
		1549
		1550	=item IO::AIO::mlockall $flags
		1551
		1552	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1553	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL__FUTURE>).
		1554
		1555	On systems that do not implement C<mlockall>, this function returns
		1556	ENOSYS, otherwise the return value of C<mlockall>.
		1557
		1558	=item IO::AIO::munlockall
		1559
		1560	Calls the C<munlockall> function.
		1561
		1562	On systems that do not implement C<munlockall>, this function returns
		1563	ENOSYS, otherwise the return value of C<munlockall>.
		1564
		1565	=back
		1566
815	=cut	1567	=cut
816		1568
817	# support function to convert a fd into a perl filehandle
818	sub _fd2fh {
819	return undef if $_[0] < 0;
820
821	# try to generate nice filehandles
822	my $sym = "IO::AIO::fd#$_[0]";
823	local *$sym;
824
825	open *$sym, "+<&=$_[0]" # usually works under any unix
826	or open *$sym, "<&=$_[0]" # cygwin needs this
827	or open *$sym, ">&=$_[0]" # or this
828	or return undef;
829
830	*$sym
831	}
832
833	min_parallel 8;	1569	min_parallel 8;
834		1570
835	END {	1571	END { flush }
836	max_parallel 0;
837	}
838		1572
839	1;	1573	1;
		1574
		1575	=head1 EVENT LOOP INTEGRATION
		1576
		1577	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		1578	automatically into many event loops:
		1579
		1580	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		1581	use AnyEvent::AIO;
		1582
		1583	You can also integrate IO::AIO manually into many event loops, here are
		1584	some examples of how to do this:
		1585
		1586	# EV integration
		1587	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		1588
		1589	# Event integration
		1590	Event->io (fd => IO::AIO::poll_fileno,
		1591	poll => 'r',
		1592	cb => \&IO::AIO::poll_cb);
		1593
		1594	# Glib/Gtk2 integration
		1595	add_watch Glib::IO IO::AIO::poll_fileno,
		1596	in => sub { IO::AIO::poll_cb; 1 };
		1597
		1598	# Tk integration
		1599	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		1600	readable => \&IO::AIO::poll_cb);
		1601
		1602	# Danga::Socket integration
		1603	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		1604	\&IO::AIO::poll_cb);
840		1605
841	=head2 FORK BEHAVIOUR	1606	=head2 FORK BEHAVIOUR
842		1607
843	This module should do "the right thing" when the process using it forks:	1608	This module should do "the right thing" when the process using it forks:
844		1609
845	Before the fork, IO::AIO enters a quiescent state where no requests	1610	Before the fork, IO::AIO enters a quiescent state where no requests
846	can be added in other threads and no results will be processed. After	1611	can be added in other threads and no results will be processed. After
847	the fork the parent simply leaves the quiescent state and continues	1612	the fork the parent simply leaves the quiescent state and continues
848	request/result processing, while the child clears the request/result	1613	request/result processing, while the child frees the request/result queue
849	queue (so the requests started before the fork will only be handled in	1614	(so that the requests started before the fork will only be handled in the
850	the parent). Threads will be started on demand until the limit ste in the	1615	parent). Threads will be started on demand until the limit set in the
851	parent process has been reached again.	1616	parent process has been reached again.
852		1617
853	In short: the parent will, after a short pause, continue as if fork had	1618	In short: the parent will, after a short pause, continue as if fork had
854	not been called, while the child will act as if IO::AIO has not been used	1619	not been called, while the child will act as if IO::AIO has not been used
855	yet.	1620	yet.
856		1621
857	=head2 MEMORY USAGE	1622	=head2 MEMORY USAGE
858		1623
		1624	Per-request usage:
		1625
859	Each aio request uses - depending on your architecture - around 128 bytes	1626	Each aio request uses - depending on your architecture - around 100-200
860	of memory. In addition, stat requests need a stat buffer (possibly a few	1627	bytes of memory. In addition, stat requests need a stat buffer (possibly
861	hundred bytes). Perl scalars and other data passed into aio requests will	1628	a few hundred bytes), readdir requires a result buffer and so on. Perl
862	also be locked.	1629	scalars and other data passed into aio requests will also be locked and
		1630	will consume memory till the request has entered the done state.
863		1631
864	This is now awfully much, so queuing lots of requests is not usually a	1632	This is not awfully much, so queuing lots of requests is not usually a
865	problem.	1633	problem.
866		1634
867	Each thread needs a stack area which is usually around 16k, sometimes much	1635	Per-thread usage:
868	larger, depending on the OS.	1636
		1637	In the execution phase, some aio requests require more memory for
		1638	temporary buffers, and each thread requires a stack and other data
		1639	structures (usually around 16k-128k, depending on the OS).
		1640
		1641	=head1 KNOWN BUGS
		1642
		1643	Known bugs will be fixed in the next release.
869		1644
870	=head1 SEE ALSO	1645	=head1 SEE ALSO
871		1646
872	L<Coro>, L<Linux::AIO> (obsolete).	1647	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		1648	more natural syntax.
873		1649
874	=head1 AUTHOR	1650	=head1 AUTHOR
875		1651
876	Marc Lehmann <schmorp@schmorp.de>	1652	Marc Lehmann <schmorp@schmorp.de>
877	http://home.schmorp.de/	1653	http://home.schmorp.de/

Diff Legend

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