ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.39 by root, Sun Aug 28 11:05:50 2005 UTC vs.
Revision 1.228 by root, Sun Jun 17 17:07:25 2012 UTC

…		…
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) = @_;	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
16	aio_read $fh, 30000, 1024, $buffer, 0, sub {	17	aio_read $fh, 30000, 1024, $buffer, 0, sub {
17	$_[0] > 0 or die "read error: $!";	18	$_[0] > 0 or die "read error: $!";
18	};	19	};
19		20
20	# Event	21	# version 2+ has request and group objects
21	Event->io (fd => IO::AIO::poll_fileno,	22	use IO::AIO 2;
22	poll => 'r',
23	cb => \&IO::AIO::poll_cb);
24		23
25	# Glib/Gtk2	24	aioreq_pri 4; # give next request a very high priority
26	add_watch Glib::IO IO::AIO::poll_fileno,	25	my $req = aio_unlink "/tmp/file", sub { };
27	in => sub { IO::AIO::poll_cb; 1 };	26	$req->cancel; # cancel request if still in queue
28		27
29	# Tk	28	my $grp = aio_group sub { print "all stats done\n" };
30	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",	29	add $grp aio_stat "..." for ...;
31	readable => \&IO::AIO::poll_cb);
32
33	# Danga::Socket
34	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
35	\&IO::AIO::poll_cb);
36
37		30
38	=head1 DESCRIPTION	31	=head1 DESCRIPTION
39		32
40	This module implements asynchronous I/O using whatever means your	33	This module implements asynchronous I/O using whatever means your
41	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>).
42		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
43	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
44	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
45	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
46	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
47	aio functions available on many operating systems. However, they are often	57	functions available on many operating systems. However, they are often
48	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
49	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
50	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.
51		62
52	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,
53	currently not reentrant, so use appropriate locking yourself, always call	64	it is currently not reentrant in any way, so use appropriate locking
54	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
55	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", IO::AIO::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
56		161
57	=cut	162	=cut
58		163
59	package IO::AIO;	164	package IO::AIO;
60		165
61	no warnings;	166	use Carp ();
		167
		168	use common::sense;
62		169
63	use base 'Exporter';	170	use base 'Exporter';
64		171
65	use Fcntl ();
66
67	BEGIN {	172	BEGIN {
68	$VERSION = 1.6;	173	our $VERSION = '4.15';
69		174
70	@EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat	175	our @AIO_REQ = qw(aio_sendfile aio_seek aio_read aio_write aio_open aio_close
71	aio_lstat aio_unlink aio_rmdir aio_readdir aio_symlink	176	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
72	aio_fsync aio_fdatasync aio_readahead);	177	aio_scandir aio_symlink aio_readlink aio_realpath aio_sync
73	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel	178	aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range aio_fallocate
74	max_outstanding nreqs);	179	aio_pathsync aio_readahead aio_fiemap
		180	aio_rename aio_link aio_move aio_copy aio_group
		181	aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
		182	aio_chmod aio_utime aio_truncate
		183	aio_msync aio_mtouch aio_mlock aio_mlockall
		184	aio_statvfs
		185	aio_wd);
		186
		187	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		188	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		189	min_parallel max_parallel max_idle idle_timeout
		190	nreqs nready npending nthreads
		191	max_poll_time max_poll_reqs
		192	sendfile fadvise madvise
		193	mmap munmap munlock munlockall);
		194
		195	push @AIO_REQ, qw(aio_busy); # not exported
		196
		197	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
75		198
76	require XSLoader;	199	require XSLoader;
77	XSLoader::load IO::AIO, $VERSION;	200	XSLoader::load ("IO::AIO", $VERSION);
78	}	201	}
79		202
80	=head1 FUNCTIONS	203	=head1 FUNCTIONS
81		204
82	=head2 AIO FUNCTIONS	205	=head2 QUICK OVERVIEW
		206
		207	This section simply lists the prototypes of the most important functions
		208	for quick reference. See the following sections for function-by-function
		209	documentation.
		210
		211	aio_wd $pathname, $callback->($wd)
		212	aio_open $pathname, $flags, $mode, $callback->($fh)
		213	aio_close $fh, $callback->($status)
		214	aio_seek $fh,$offset,$whence, $callback->($offs)
		215	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		216	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		217	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		218	aio_readahead $fh,$offset,$length, $callback->($retval)
		219	aio_stat $fh_or_path, $callback->($status)
		220	aio_lstat $fh, $callback->($status)
		221	aio_statvfs $fh_or_path, $callback->($statvfs)
		222	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		223	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		224	aio_chmod $fh_or_path, $mode, $callback->($status)
		225	aio_truncate $fh_or_path, $offset, $callback->($status)
		226	aio_unlink $pathname, $callback->($status)
		227	aio_mknod $pathname, $mode, $dev, $callback->($status)
		228	aio_link $srcpath, $dstpath, $callback->($status)
		229	aio_symlink $srcpath, $dstpath, $callback->($status)
		230	aio_readlink $pathname, $callback->($link)
		231	aio_realpath $pathname, $callback->($link)
		232	aio_rename $srcpath, $dstpath, $callback->($status)
		233	aio_mkdir $pathname, $mode, $callback->($status)
		234	aio_rmdir $pathname, $callback->($status)
		235	aio_readdir $pathname, $callback->($entries)
		236	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		237	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		238	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		239	aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		240	aio_load $pathname, $data, $callback->($status)
		241	aio_copy $srcpath, $dstpath, $callback->($status)
		242	aio_move $srcpath, $dstpath, $callback->($status)
		243	aio_rmtree $pathname, $callback->($status)
		244	aio_sync $callback->($status)
		245	aio_syncfs $fh, $callback->($status)
		246	aio_fsync $fh, $callback->($status)
		247	aio_fdatasync $fh, $callback->($status)
		248	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		249	aio_pathsync $pathname, $callback->($status)
		250	aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		251	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		252	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		253	aio_mlockall $flags, $callback->($status)
		254	aio_group $callback->(...)
		255	aio_nop $callback->()
		256
		257	$prev_pri = aioreq_pri [$pri]
		258	aioreq_nice $pri_adjust
		259
		260	IO::AIO::poll_wait
		261	IO::AIO::poll_cb
		262	IO::AIO::poll
		263	IO::AIO::flush
		264	IO::AIO::max_poll_reqs $nreqs
		265	IO::AIO::max_poll_time $seconds
		266	IO::AIO::min_parallel $nthreads
		267	IO::AIO::max_parallel $nthreads
		268	IO::AIO::max_idle $nthreads
		269	IO::AIO::idle_timeout $seconds
		270	IO::AIO::max_outstanding $maxreqs
		271	IO::AIO::nreqs
		272	IO::AIO::nready
		273	IO::AIO::npending
		274
		275	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		276	IO::AIO::fadvise $fh, $offset, $len, $advice
		277	IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
		278	IO::AIO::munmap $scalar
		279	IO::AIO::madvise $scalar, $offset, $length, $advice
		280	IO::AIO::mprotect $scalar, $offset, $length, $protect
		281	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		282	IO::AIO::munlockall
		283
		284	=head2 API NOTES
83		285
84	All the C<aio_*> calls are more or less thin wrappers around the syscall	286	All the C<aio_*> calls are more or less thin wrappers around the syscall
85	with the same name (sans C<aio_>). The arguments are similar or identical,	287	with the same name (sans C<aio_>). The arguments are similar or identical,
86	and they all accept an additional (and optional) C<$callback> argument	288	and they all accept an additional (and optional) C<$callback> argument
87	which must be a code reference. This code reference will get called with	289	which must be a code reference. This code reference will be called after
88	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	290	the syscall has been executed in an asynchronous fashion. The results
89	perl, which usually delivers "false") as it's sole argument when the given	291	of the request will be passed as arguments to the callback (and, if an
90	syscall has been executed asynchronously.	292	error occured, in C<$!>) - for most requests the syscall return code (e.g.
		293	most syscalls return C<-1> on error, unlike perl, which usually delivers
		294	"false").
		295
		296	Some requests (such as C<aio_readdir>) pass the actual results and
		297	communicate failures by passing C<undef>.
91		298
92	All functions expecting a filehandle keep a copy of the filehandle	299	All functions expecting a filehandle keep a copy of the filehandle
93	internally until the request has finished.	300	internally until the request has finished.
94		301
		302	All functions return request objects of type L<IO::AIO::REQ> that allow
		303	further manipulation of those requests while they are in-flight.
		304
95	The pathnames you pass to these routines I<must> be absolute and	305	The pathnames you pass to these routines I<should> be absolute. The
96	encoded in byte form. The reason for the former is that at the time the	306	reason for this is that at the time the request is being executed, the
97	request is being executed, the current working directory could have	307	current working directory could have changed. Alternatively, you can
98	changed. Alternatively, you can make sure that you never change the	308	make sure that you never change the current working directory anywhere
99	current working directory.	309	in the program and then use relative paths. You can also take advantage
		310	of IO::AIOs working directory abstraction, that lets you specify paths
		311	relative to some previously-opened "working directory object" - see the
		312	description of the C<IO::AIO::WD> class later in this document.
100		313
101	To encode pathnames to byte form, either make sure you either: a)	314	To encode pathnames as octets, either make sure you either: a) always pass
102	always pass in filenames you got from outside (command line, readdir	315	in filenames you got from outside (command line, readdir etc.) without
103	etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode	316	tinkering, b) are in your native filesystem encoding, c) use the Encode
104	your pathnames to the locale (or other) encoding in effect in the user	317	module and encode your pathnames to the locale (or other) encoding in
105	environment, d) use Glib::filename_from_unicode on unicode filenames or e)	318	effect in the user environment, d) use Glib::filename_from_unicode on
106	use something else.	319	unicode filenames or e) use something else to ensure your scalar has the
		320	correct contents.
		321
		322	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		323	handles correctly whether it is set or not.
		324
		325	=head2 AIO REQUEST FUNCTIONS
107		326
108	=over 4	327	=over 4
109		328
		329	=item $prev_pri = aioreq_pri [$pri]
		330
		331	Returns the priority value that would be used for the next request and, if
		332	C<$pri> is given, sets the priority for the next aio request.
		333
		334	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		335	and C<4>, respectively. Requests with higher priority will be serviced
		336	first.
		337
		338	The priority will be reset to C<0> after each call to one of the C<aio_*>
		339	functions.
		340
		341	Example: open a file with low priority, then read something from it with
		342	higher priority so the read request is serviced before other low priority
		343	open requests (potentially spamming the cache):
		344
		345	aioreq_pri -3;
		346	aio_open ..., sub {
		347	return unless $_[0];
		348
		349	aioreq_pri -2;
		350	aio_read $_[0], ..., sub {
		351	...
		352	};
		353	};
		354
		355
		356	=item aioreq_nice $pri_adjust
		357
		358	Similar to C<aioreq_pri>, but subtracts the given value from the current
		359	priority, so the effect is cumulative.
		360
		361
110	=item aio_open $pathname, $flags, $mode, $callback	362	=item aio_open $pathname, $flags, $mode, $callback->($fh)
111		363
112	Asynchronously open or create a file and call the callback with a newly	364	Asynchronously open or create a file and call the callback with a newly
113	created filehandle for the file.	365	created filehandle for the file.
114		366
115	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	367	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
…		…
119	list. They are the same as used by C<sysopen>.	371	list. They are the same as used by C<sysopen>.
120		372
121	Likewise, C<$mode> specifies the mode of the newly created file, if it	373	Likewise, C<$mode> specifies the mode of the newly created file, if it
122	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,	374	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
123	except that it is mandatory (i.e. use C<0> if you don't create new files,	375	except that it is mandatory (i.e. use C<0> if you don't create new files,
124	and C<0666> or C<0777> if you do).	376	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		377	by the umask in effect then the request is being executed, so better never
		378	change the umask.
125		379
126	Example:	380	Example:
127		381
128	aio_open "/etc/passwd", O_RDONLY, 0, sub {	382	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
129	if ($_[0]) {	383	if ($_[0]) {
130	print "open successful, fh is $_[0]\n";	384	print "open successful, fh is $_[0]\n";
131	...	385	...
132	} else {	386	} else {
133	die "open failed: $!\n";	387	die "open failed: $!\n";
134	}	388	}
135	};	389	};
136		390
		391	In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
		392	C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
		393	following POSIX and non-POSIX constants are available (missing ones on
		394	your system are, as usual, C<0>):
		395
		396	C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
		397	C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
		398	C<O_RSYNC>, C<O_SYNC> and C<O_TTY_INIT>.
		399
		400
137	=item aio_close $fh, $callback	401	=item aio_close $fh, $callback->($status)
138		402
139	Asynchronously close a file and call the callback with the result	403	Asynchronously close a file and call the callback with the result
140	code. I<WARNING:> although accepted, you should not pass in a perl	404	code.
141	filehandle here, as perl will likely close the file descriptor another
142	time when the filehandle is destroyed. Normally, you can safely call perls
143	C<close> or just let filehandles go out of scope.
144		405
145	This is supposed to be a bug in the API, so that might change. It's	406	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
146	therefore best to avoid this function.	407	closing the file descriptor associated with the filehandle itself.
147		408
		409	Therefore, C<aio_close> will not close the filehandle - instead it will
		410	use dup2 to overwrite the file descriptor with the write-end of a pipe
		411	(the pipe fd will be created on demand and will be cached).
		412
		413	Or in other words: the file descriptor will be closed, but it will not be
		414	free for reuse until the perl filehandle is closed.
		415
		416	=cut
		417
		418	=item aio_seek $fh, $offset, $whence, $callback->($offs)
		419
		420	Seeks the filehandle to the new C<$offset>, similarly to perl's
		421	C<sysseek>. The C<$whence> can use the traditional values (C<0> for
		422	C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
		423	C<IO::AIO::SEEK_END>).
		424
		425	The resulting absolute offset will be passed to the callback, or C<-1> in
		426	case of an error.
		427
		428	In theory, the C<$whence> constants could be different than the
		429	corresponding values from L<Fcntl>, but perl guarantees they are the same,
		430	so don't panic.
		431
		432	As a GNU/Linux (and maybe Solaris) extension, also the constants
		433	C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
		434	could be found. No guarantees about suitability for use in C<aio_seek> or
		435	Perl's C<sysseek> can be made though, although I would naively assume they
		436	"just work".
		437
148	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	438	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
149		439
150	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	440	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
151		441
152	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	442	Reads or writes C<$length> bytes from or to the specified C<$fh> and
153	into the scalar given by C<data> and offset C<dataoffset> and calls the	443	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset>
154	callback without the actual number of bytes read (or -1 on error, just	444	and calls the callback without the actual number of bytes read (or -1 on
155	like the syscall).	445	error, just like the syscall).
		446
		447	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		448	offset plus the actual number of bytes read.
		449
		450	If C<$offset> is undefined, then the current file descriptor offset will
		451	be used (and updated), otherwise the file descriptor offset will not be
		452	changed by these calls.
		453
		454	If C<$length> is undefined in C<aio_write>, use the remaining length of
		455	C<$data>.
		456
		457	If C<$dataoffset> is less than zero, it will be counted from the end of
		458	C<$data>.
156		459
157	The C<$data> scalar I<MUST NOT> be modified in any way while the request	460	The C<$data> scalar I<MUST NOT> be modified in any way while the request
158	is outstanding. Modifying it can result in segfaults or WW3 (if the	461	is outstanding. Modifying it can result in segfaults or World War III (if
159	necessary/optional hardware is installed).	462	the necessary/optional hardware is installed).
160		463
161	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	464	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
162	offset C<0> within the scalar:	465	offset C<0> within the scalar:
163		466
164	aio_read $fh, 7, 15, $buffer, 0, sub {	467	aio_read $fh, 7, 15, $buffer, 0, sub {
165	$_[0] > 0 or die "read error: $!";	468	$_[0] > 0 or die "read error: $!";
166	print "read $_[0] bytes: <$buffer>\n";	469	print "read $_[0] bytes: <$buffer>\n";
167	};	470	};
168		471
		472
169	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback	473	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
170		474
171	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts	475	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
172	reading at byte offset C<$in_offset>, and starts writing at the current	476	reading at byte offset C<$in_offset>, and starts writing at the current
173	file offset of C<$out_fh>. Because of that, it is not safe to issue more	477	file offset of C<$out_fh>. Because of that, it is not safe to issue more
174	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each	478	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
175	other.	479	other. The same C<$in_fh> works fine though, as this function does not
		480	move or use the file offset of C<$in_fh>.
176		481
		482	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		483	are written, and there is no way to find out how many more bytes have been
		484	read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
		485	number of bytes written to C<$out_fh>. Only if the result value equals
		486	C<$length> one can assume that C<$length> bytes have been read.
		487
		488	Unlike with other C<aio_> functions, it makes a lot of sense to use
		489	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		490	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		491	the socket I/O will be non-blocking. Note, however, that you can run
		492	into a trap where C<aio_sendfile> reads some data with readahead, then
		493	fails to write all data, and when the socket is ready the next time, the
		494	data in the cache is already lost, forcing C<aio_sendfile> to again hit
		495	the disk. Explicit C<aio_read> + C<aio_write> let's you better control
		496	resource usage.
		497
177	This call tries to make use of a native C<sendfile> syscall to provide	498	This call tries to make use of a native C<sendfile>-like syscall to
178	zero-copy operation. For this to work, C<$out_fh> should refer to a	499	provide zero-copy operation. For this to work, C<$out_fh> should refer to
179	socket, and C<$in_fh> should refer to mmap'able file.	500	a socket, and C<$in_fh> should refer to an mmap'able file.
180		501
181	If the native sendfile call fails or is not implemented, it will be	502	If a native sendfile cannot be found or it fails with C<ENOSYS>,
182	emulated, so you can call C<aio_sendfile> on any type of filehandle	503	C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
		504	C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
183	regardless of the limitations of the operating system.	505	type of filehandle regardless of the limitations of the operating system.
184		506
185	Please note, however, that C<aio_sendfile> can read more bytes from	507	As native sendfile syscalls (as practically any non-POSIX interface hacked
186	C<$in_fh> than are written, and there is no way to find out how many	508	together in a hurry to improve benchmark numbers) tend to be rather buggy
187	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only	509	on many systems, this implementation tries to work around some known bugs
188	provides the number of bytes written to C<$out_fh>. Only if the result	510	in Linux and FreeBSD kernels (probably others, too), but that might fail,
189	value equals C<$length> one can assume that C<$length> bytes have been	511	so you really really should check the return value of C<aio_sendfile> -
190	read.	512	fewre bytes than expected might have been transferred.
191		513
		514
192	=item aio_readahead $fh,$offset,$length, $callback	515	=item aio_readahead $fh,$offset,$length, $callback->($retval)
193		516
194	C<aio_readahead> populates the page cache with data from a file so that	517	C<aio_readahead> populates the page cache with data from a file so that
195	subsequent reads from that file will not block on disk I/O. The C<$offset>	518	subsequent reads from that file will not block on disk I/O. The C<$offset>
196	argument specifies the starting point from which data is to be read and	519	argument specifies the starting point from which data is to be read and
197	C<$length> specifies the number of bytes to be read. I/O is performed in	520	C<$length> specifies the number of bytes to be read. I/O is performed in
…		…
201	file. The current file offset of the file is left unchanged.	524	file. The current file offset of the file is left unchanged.
202		525
203	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	526	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
204	emulated by simply reading the data, which would have a similar effect.	527	emulated by simply reading the data, which would have a similar effect.
205		528
		529
206	=item aio_stat $fh_or_path, $callback	530	=item aio_stat $fh_or_path, $callback->($status)
207		531
208	=item aio_lstat $fh, $callback	532	=item aio_lstat $fh, $callback->($status)
209		533
210	Works like perl's C<stat> or C<lstat> in void context. The callback will	534	Works like perl's C<stat> or C<lstat> in void context. The callback will
211	be called after the stat and the results will be available using C<stat _>	535	be called after the stat and the results will be available using C<stat _>
212	or C<-s _> etc...	536	or C<-s _> etc...
213		537
…		…
215	for an explanation.	539	for an explanation.
216		540
217	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	541	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
218	error when stat'ing a large file, the results will be silently truncated	542	error when stat'ing a large file, the results will be silently truncated
219	unless perl itself is compiled with large file support.	543	unless perl itself is compiled with large file support.
		544
		545	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		546	following constants and functions (if not implemented, the constants will
		547	be C<0> and the functions will either C<croak> or fall back on traditional
		548	behaviour).
		549
		550	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		551	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		552	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
220		553
221	Example: Print the length of F</etc/passwd>:	554	Example: Print the length of F</etc/passwd>:
222		555
223	aio_stat "/etc/passwd", sub {	556	aio_stat "/etc/passwd", sub {
224	$_[0] and die "stat failed: $!";	557	$_[0] and die "stat failed: $!";
225	print "size is ", -s _, "\n";	558	print "size is ", -s _, "\n";
226	};	559	};
227		560
		561
		562	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		563
		564	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		565	whether a file handle or path was passed.
		566
		567	On success, the callback is passed a hash reference with the following
		568	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		569	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		570	is passed.
		571
		572	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		573	C<ST_NOSUID>.
		574
		575	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		576	their correct value when available, or to C<0> on systems that do
		577	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		578	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		579	C<ST_NODIRATIME> and C<ST_RELATIME>.
		580
		581	Example: stat C</wd> and dump out the data if successful.
		582
		583	aio_statvfs "/wd", sub {
		584	my $f = $_[0]
		585	or die "statvfs: $!";
		586
		587	use Data::Dumper;
		588	say Dumper $f;
		589	};
		590
		591	# result:
		592	{
		593	bsize => 1024,
		594	bfree => 4333064312,
		595	blocks => 10253828096,
		596	files => 2050765568,
		597	flag => 4096,
		598	favail => 2042092649,
		599	bavail => 4333064312,
		600	ffree => 2042092649,
		601	namemax => 255,
		602	frsize => 1024,
		603	fsid => 1810
		604	}
		605
		606
		607	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		608
		609	Works like perl's C<utime> function (including the special case of $atime
		610	and $mtime being undef). Fractional times are supported if the underlying
		611	syscalls support them.
		612
		613	When called with a pathname, uses utimes(2) if available, otherwise
		614	utime(2). If called on a file descriptor, uses futimes(2) if available,
		615	otherwise returns ENOSYS, so this is not portable.
		616
		617	Examples:
		618
		619	# set atime and mtime to current time (basically touch(1)):
		620	aio_utime "path", undef, undef;
		621	# set atime to current time and mtime to beginning of the epoch:
		622	aio_utime "path", time, undef; # undef==0
		623
		624
		625	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		626
		627	Works like perl's C<chown> function, except that C<undef> for either $uid
		628	or $gid is being interpreted as "do not change" (but -1 can also be used).
		629
		630	Examples:
		631
		632	# same as "chown root path" in the shell:
		633	aio_chown "path", 0, -1;
		634	# same as above:
		635	aio_chown "path", 0, undef;
		636
		637
		638	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		639
		640	Works like truncate(2) or ftruncate(2).
		641
		642
		643	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		644
		645	Works like perl's C<chmod> function.
		646
		647
228	=item aio_unlink $pathname, $callback	648	=item aio_unlink $pathname, $callback->($status)
229		649
230	Asynchronously unlink (delete) a file and call the callback with the	650	Asynchronously unlink (delete) a file and call the callback with the
231	result code.	651	result code.
232		652
		653
		654	=item aio_mknod $pathname, $mode, $dev, $callback->($status)
		655
		656	[EXPERIMENTAL]
		657
		658	Asynchronously create a device node (or fifo). See mknod(2).
		659
		660	The only (POSIX-) portable way of calling this function is:
		661
		662	aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		663
		664	See C<aio_stat> for info about some potentially helpful extra constants
		665	and functions.
		666
		667	=item aio_link $srcpath, $dstpath, $callback->($status)
		668
		669	Asynchronously create a new link to the existing object at C<$srcpath> at
		670	the path C<$dstpath> and call the callback with the result code.
		671
		672
		673	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		674
		675	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		676	the path C<$dstpath> and call the callback with the result code.
		677
		678
		679	=item aio_readlink $pathname, $callback->($link)
		680
		681	Asynchronously read the symlink specified by C<$path> and pass it to
		682	the callback. If an error occurs, nothing or undef gets passed to the
		683	callback.
		684
		685
		686	=item aio_realpath $pathname, $callback->($path)
		687
		688	Asynchronously make the path absolute and resolve any symlinks in
		689	C<$path>. The resulting path only consists of directories (Same as
		690	L<Cwd::realpath>).
		691
		692	This request can be used to get the absolute path of the current working
		693	directory by passing it a path of F<.> (a single dot).
		694
		695
		696	=item aio_rename $srcpath, $dstpath, $callback->($status)
		697
		698	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		699	rename(2) and call the callback with the result code.
		700
		701
		702	=item aio_mkdir $pathname, $mode, $callback->($status)
		703
		704	Asynchronously mkdir (create) a directory and call the callback with
		705	the result code. C<$mode> will be modified by the umask at the time the
		706	request is executed, so do not change your umask.
		707
		708
233	=item aio_rmdir $pathname, $callback	709	=item aio_rmdir $pathname, $callback->($status)
234		710
235	Asynchronously rmdir (delete) a directory and call the callback with the	711	Asynchronously rmdir (delete) a directory and call the callback with the
236	result code.	712	result code.
237		713
		714
238	=item aio_readdir $pathname $callback	715	=item aio_readdir $pathname, $callback->($entries)
239		716
240	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire	717	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
241	directory (i.e. opendir + readdir + closedir). The entries will not be	718	directory (i.e. opendir + readdir + closedir). The entries will not be
242	sorted, and will B<NOT> include the C<.> and C<..> entries.	719	sorted, and will B<NOT> include the C<.> and C<..> entries.
243		720
244	The callback a single argument which is either C<undef> or an array-ref	721	The callback is passed a single argument which is either C<undef> or an
245	with the filenames.	722	array-ref with the filenames.
246		723
		724
		725	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		726
		727	Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
		728	tune behaviour and output format. In case of an error, C<$entries> will be
		729	C<undef>.
		730
		731	The flags are a combination of the following constants, ORed together (the
		732	flags will also be passed to the callback, possibly modified):
		733
		734	=over 4
		735
		736	=item IO::AIO::READDIR_DENTS
		737
		738	When this flag is off, then the callback gets an arrayref consisting of
		739	names only (as with C<aio_readdir>), otherwise it gets an arrayref with
		740	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		741	entry in more detail.
		742
		743	C<$name> is the name of the entry.
		744
		745	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		746
		747	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		748	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		749	C<IO::AIO::DT_WHT>.
		750
		751	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		752	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		753	scalars are read-only: you can not modify them.
		754
		755	C<$inode> is the inode number (which might not be exact on systems with 64
		756	bit inode numbers and 32 bit perls). This field has unspecified content on
		757	systems that do not deliver the inode information.
		758
		759	=item IO::AIO::READDIR_DIRS_FIRST
		760
		761	When this flag is set, then the names will be returned in an order where
		762	likely directories come first, in optimal stat order. This is useful when
		763	you need to quickly find directories, or you want to find all directories
		764	while avoiding to stat() each entry.
		765
		766	If the system returns type information in readdir, then this is used
		767	to find directories directly. Otherwise, likely directories are names
		768	beginning with ".", or otherwise names with no dots, of which names with
		769	short names are tried first.
		770
		771	=item IO::AIO::READDIR_STAT_ORDER
		772
		773	When this flag is set, then the names will be returned in an order
		774	suitable for stat()'ing each one. That is, when you plan to stat()
		775	all files in the given directory, then the returned order will likely
		776	be fastest.
		777
		778	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		779	the likely dirs come first, resulting in a less optimal stat order.
		780
		781	=item IO::AIO::READDIR_FOUND_UNKNOWN
		782
		783	This flag should not be set when calling C<aio_readdirx>. Instead, it
		784	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		785	C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
		786	C<$type>'s are known, which can be used to speed up some algorithms.
		787
		788	=back
		789
		790
		791	=item aio_load $pathname, $data, $callback->($status)
		792
		793	This is a composite request that tries to fully load the given file into
		794	memory. Status is the same as with aio_read.
		795
		796	=cut
		797
		798	sub aio_load($$;$) {
		799	my ($path, undef, $cb) = @_;
		800	my $data = \$_[1];
		801
		802	my $pri = aioreq_pri;
		803	my $grp = aio_group $cb;
		804
		805	aioreq_pri $pri;
		806	add $grp aio_open $path, O_RDONLY, 0, sub {
		807	my $fh = shift
		808	or return $grp->result (-1);
		809
		810	aioreq_pri $pri;
		811	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		812	$grp->result ($_[0]);
		813	};
		814	};
		815
		816	$grp
		817	}
		818
		819	=item aio_copy $srcpath, $dstpath, $callback->($status)
		820
		821	Try to copy the I<file> (directories not supported as either source or
		822	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		823	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		824
		825	This is a composite request that creates the destination file with
		826	mode 0200 and copies the contents of the source file into it using
		827	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		828	uid/gid, in that order.
		829
		830	If an error occurs, the partial destination file will be unlinked, if
		831	possible, except when setting atime, mtime, access mode and uid/gid, where
		832	errors are being ignored.
		833
		834	=cut
		835
		836	sub aio_copy($$;$) {
		837	my ($src, $dst, $cb) = @_;
		838
		839	my $pri = aioreq_pri;
		840	my $grp = aio_group $cb;
		841
		842	aioreq_pri $pri;
		843	add $grp aio_open $src, O_RDONLY, 0, sub {
		844	if (my $src_fh = $_[0]) {
		845	my @stat = stat $src_fh; # hmm, might block over nfs?
		846
		847	aioreq_pri $pri;
		848	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		849	if (my $dst_fh = $_[0]) {
		850	aioreq_pri $pri;
		851	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		852	if ($_[0] == $stat[7]) {
		853	$grp->result (0);
		854	close $src_fh;
		855
		856	my $ch = sub {
		857	aioreq_pri $pri;
		858	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		859	aioreq_pri $pri;
		860	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		861	aioreq_pri $pri;
		862	add $grp aio_close $dst_fh;
		863	}
		864	};
		865	};
		866
		867	aioreq_pri $pri;
		868	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		869	if ($_[0] < 0 && $! == ENOSYS) {
		870	aioreq_pri $pri;
		871	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		872	} else {
		873	$ch->();
		874	}
		875	};
		876	} else {
		877	$grp->result (-1);
		878	close $src_fh;
		879	close $dst_fh;
		880
		881	aioreq $pri;
		882	add $grp aio_unlink $dst;
		883	}
		884	};
		885	} else {
		886	$grp->result (-1);
		887	}
		888	},
		889
		890	} else {
		891	$grp->result (-1);
		892	}
		893	};
		894
		895	$grp
		896	}
		897
		898	=item aio_move $srcpath, $dstpath, $callback->($status)
		899
		900	Try to move the I<file> (directories not supported as either source or
		901	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		902	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		903
		904	This is a composite request that tries to rename(2) the file first; if
		905	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		906	that is successful, unlinks the C<$srcpath>.
		907
		908	=cut
		909
		910	sub aio_move($$;$) {
		911	my ($src, $dst, $cb) = @_;
		912
		913	my $pri = aioreq_pri;
		914	my $grp = aio_group $cb;
		915
		916	aioreq_pri $pri;
		917	add $grp aio_rename $src, $dst, sub {
		918	if ($_[0] && $! == EXDEV) {
		919	aioreq_pri $pri;
		920	add $grp aio_copy $src, $dst, sub {
		921	$grp->result ($_[0]);
		922
		923	unless ($_[0]) {
		924	aioreq_pri $pri;
		925	add $grp aio_unlink $src;
		926	}
		927	};
		928	} else {
		929	$grp->result ($_[0]);
		930	}
		931	};
		932
		933	$grp
		934	}
		935
		936	=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		937
		938	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		939	efficiently separate the entries of directory C<$path> into two sets of
		940	names, directories you can recurse into (directories), and ones you cannot
		941	recurse into (everything else, including symlinks to directories).
		942
		943	C<aio_scandir> is a composite request that creates of many sub requests_
		944	C<$maxreq> specifies the maximum number of outstanding aio requests that
		945	this function generates. If it is C<< <= 0 >>, then a suitable default
		946	will be chosen (currently 4).
		947
		948	On error, the callback is called without arguments, otherwise it receives
		949	two array-refs with path-relative entry names.
		950
		951	Example:
		952
		953	aio_scandir $dir, 0, sub {
		954	my ($dirs, $nondirs) = @_;
		955	print "real directories: @$dirs\n";
		956	print "everything else: @$nondirs\n";
		957	};
		958
		959	Implementation notes.
		960
		961	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		962
		963	If readdir returns file type information, then this is used directly to
		964	find directories.
		965
		966	Otherwise, after reading the directory, the modification time, size etc.
		967	of the directory before and after the readdir is checked, and if they
		968	match (and isn't the current time), the link count will be used to decide
		969	how many entries are directories (if >= 2). Otherwise, no knowledge of the
		970	number of subdirectories will be assumed.
		971
		972	Then entries will be sorted into likely directories a non-initial dot
		973	currently) and likely non-directories (see C<aio_readdirx>). Then every
		974	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
		975	in order of their inode numbers. If that succeeds, it assumes that the
		976	entry is a directory or a symlink to directory (which will be checked
		977	separately). This is often faster than stat'ing the entry itself because
		978	filesystems might detect the type of the entry without reading the inode
		979	data (e.g. ext2fs filetype feature), even on systems that cannot return
		980	the filetype information on readdir.
		981
		982	If the known number of directories (link count - 2) has been reached, the
		983	rest of the entries is assumed to be non-directories.
		984
		985	This only works with certainty on POSIX (= UNIX) filesystems, which
		986	fortunately are the vast majority of filesystems around.
		987
		988	It will also likely work on non-POSIX filesystems with reduced efficiency
		989	as those tend to return 0 or 1 as link counts, which disables the
		990	directory counting heuristic.
		991
		992	=cut
		993
		994	sub aio_scandir($$;$) {
		995	my ($path, $maxreq, $cb) = @_;
		996
		997	my $pri = aioreq_pri;
		998
		999	my $grp = aio_group $cb;
		1000
		1001	$maxreq = 4 if $maxreq <= 0;
		1002
		1003	# get a wd object
		1004	aioreq_pri $pri;
		1005	add $grp aio_wd $path, sub {
		1006	$_[0]
		1007	or return $grp->result ();
		1008
		1009	my $wd = [shift, "."];
		1010
		1011	# stat once
		1012	aioreq_pri $pri;
		1013	add $grp aio_stat $wd, sub {
		1014	return $grp->result () if $_[0];
		1015	my $now = time;
		1016	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		1017
		1018	# read the directory entries
		1019	aioreq_pri $pri;
		1020	add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
		1021	my $entries = shift
		1022	or return $grp->result ();
		1023
		1024	# stat the dir another time
		1025	aioreq_pri $pri;
		1026	add $grp aio_stat $wd, sub {
		1027	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		1028
		1029	my $ndirs;
		1030
		1031	# take the slow route if anything looks fishy
		1032	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		1033	$ndirs = -1;
		1034	} else {
		1035	# if nlink == 2, we are finished
		1036	# for non-posix-fs's, we rely on nlink < 2
		1037	$ndirs = (stat _)[3] - 2
		1038	or return $grp->result ([], $entries);
		1039	}
		1040
		1041	my (@dirs, @nondirs);
		1042
		1043	my $statgrp = add $grp aio_group sub {
		1044	$grp->result (\@dirs, \@nondirs);
		1045	};
		1046
		1047	limit $statgrp $maxreq;
		1048	feed $statgrp sub {
		1049	return unless @$entries;
		1050	my $entry = shift @$entries;
		1051
		1052	aioreq_pri $pri;
		1053	$wd->[1] = "$entry/.";
		1054	add $statgrp aio_stat $wd, sub {
		1055	if ($_[0] < 0) {
		1056	push @nondirs, $entry;
		1057	} else {
		1058	# need to check for real directory
		1059	aioreq_pri $pri;
		1060	$wd->[1] = $entry;
		1061	add $statgrp aio_lstat $wd, sub {
		1062	if (-d _) {
		1063	push @dirs, $entry;
		1064
		1065	unless (--$ndirs) {
		1066	push @nondirs, @$entries;
		1067	feed $statgrp;
		1068	}
		1069	} else {
		1070	push @nondirs, $entry;
		1071	}
		1072	}
		1073	}
		1074	};
		1075	};
		1076	};
		1077	};
		1078	};
		1079	};
		1080
		1081	$grp
		1082	}
		1083
		1084	=item aio_rmtree $pathname, $callback->($status)
		1085
		1086	Delete a directory tree starting (and including) C<$path>, return the
		1087	status of the final C<rmdir> only. This is a composite request that
		1088	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1089	everything else.
		1090
		1091	=cut
		1092
		1093	sub aio_rmtree;
		1094	sub aio_rmtree($;$) {
		1095	my ($path, $cb) = @_;
		1096
		1097	my $pri = aioreq_pri;
		1098	my $grp = aio_group $cb;
		1099
		1100	aioreq_pri $pri;
		1101	add $grp aio_scandir $path, 0, sub {
		1102	my ($dirs, $nondirs) = @_;
		1103
		1104	my $dirgrp = aio_group sub {
		1105	add $grp aio_rmdir $path, sub {
		1106	$grp->result ($_[0]);
		1107	};
		1108	};
		1109
		1110	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1111	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1112
		1113	add $grp $dirgrp;
		1114	};
		1115
		1116	$grp
		1117	}
		1118
		1119	=item aio_sync $callback->($status)
		1120
		1121	Asynchronously call sync and call the callback when finished.
		1122
247	=item aio_fsync $fh, $callback	1123	=item aio_fsync $fh, $callback->($status)
248		1124
249	Asynchronously call fsync on the given filehandle and call the callback	1125	Asynchronously call fsync on the given filehandle and call the callback
250	with the fsync result code.	1126	with the fsync result code.
251		1127
252	=item aio_fdatasync $fh, $callback	1128	=item aio_fdatasync $fh, $callback->($status)
253		1129
254	Asynchronously call fdatasync on the given filehandle and call the	1130	Asynchronously call fdatasync on the given filehandle and call the
255	callback with the fdatasync result code.	1131	callback with the fdatasync result code.
256		1132
257	If this call isn't available because your OS lacks it or it couldn't be	1133	If this call isn't available because your OS lacks it or it couldn't be
258	detected, it will be emulated by calling C<fsync> instead.	1134	detected, it will be emulated by calling C<fsync> instead.
259		1135
		1136	=item aio_syncfs $fh, $callback->($status)
		1137
		1138	Asynchronously call the syncfs syscall to sync the filesystem associated
		1139	to the given filehandle and call the callback with the syncfs result
		1140	code. If syncfs is not available, calls sync(), but returns C<-1> and sets
		1141	errno to C<ENOSYS> nevertheless.
		1142
		1143	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1144
		1145	Sync the data portion of the file specified by C<$offset> and C<$length>
		1146	to disk (but NOT the metadata), by calling the Linux-specific
		1147	sync_file_range call. If sync_file_range is not available or it returns
		1148	ENOSYS, then fdatasync or fsync is being substituted.
		1149
		1150	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1151	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1152	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1153	manpage for details.
		1154
		1155	=item aio_pathsync $pathname, $callback->($status)
		1156
		1157	This request tries to open, fsync and close the given path. This is a
		1158	composite request intended to sync directories after directory operations
		1159	(E.g. rename). This might not work on all operating systems or have any
		1160	specific effect, but usually it makes sure that directory changes get
		1161	written to disc. It works for anything that can be opened for read-only,
		1162	not just directories.
		1163
		1164	Future versions of this function might fall back to other methods when
		1165	C<fsync> on the directory fails (such as calling C<sync>).
		1166
		1167	Passes C<0> when everything went ok, and C<-1> on error.
		1168
		1169	=cut
		1170
		1171	sub aio_pathsync($;$) {
		1172	my ($path, $cb) = @_;
		1173
		1174	my $pri = aioreq_pri;
		1175	my $grp = aio_group $cb;
		1176
		1177	aioreq_pri $pri;
		1178	add $grp aio_open $path, O_RDONLY, 0, sub {
		1179	my ($fh) = @_;
		1180	if ($fh) {
		1181	aioreq_pri $pri;
		1182	add $grp aio_fsync $fh, sub {
		1183	$grp->result ($_[0]);
		1184
		1185	aioreq_pri $pri;
		1186	add $grp aio_close $fh;
		1187	};
		1188	} else {
		1189	$grp->result (-1);
		1190	}
		1191	};
		1192
		1193	$grp
		1194	}
		1195
		1196	=item aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1197
		1198	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1199	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1200	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1201	scalar must only be modified in-place while an aio operation is pending on
		1202	it).
		1203
		1204	It calls the C<msync> function of your OS, if available, with the memory
		1205	area starting at C<$offset> in the string and ending C<$length> bytes
		1206	later. If C<$length> is negative, counts from the end, and if C<$length>
		1207	is C<undef>, then it goes till the end of the string. The flags can be
		1208	a combination of C<IO::AIO::MS_ASYNC>, C<IO::AIO::MS_INVALIDATE> and
		1209	C<IO::AIO::MS_SYNC>.
		1210
		1211	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1212
		1213	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1214	scalars.
		1215
		1216	It touches (reads or writes) all memory pages in the specified
		1217	range inside the scalar. All caveats and parameters are the same
		1218	as for C<aio_msync>, above, except for flags, which must be either
		1219	C<0> (which reads all pages and ensures they are instantiated) or
		1220	C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and
		1221	writing an octet from it, which dirties the page).
		1222
		1223	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1224
		1225	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1226	scalars.
		1227
		1228	It reads in all the pages of the underlying storage into memory (if any)
		1229	and locks them, so they are not getting swapped/paged out or removed.
		1230
		1231	If C<$length> is undefined, then the scalar will be locked till the end.
		1232
		1233	On systems that do not implement C<mlock>, this function returns C<-1>
		1234	and sets errno to C<ENOSYS>.
		1235
		1236	Note that the corresponding C<munlock> is synchronous and is
		1237	documented under L<MISCELLANEOUS FUNCTIONS>.
		1238
		1239	Example: open a file, mmap and mlock it - both will be undone when
		1240	C<$data> gets destroyed.
		1241
		1242	open my $fh, "<", $path or die "$path: $!";
		1243	my $data;
		1244	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1245	aio_mlock $data; # mlock in background
		1246
		1247	=item aio_mlockall $flags, $callback->($status)
		1248
		1249	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1250	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1251
		1252	On systems that do not implement C<mlockall>, this function returns C<-1>
		1253	and sets errno to C<ENOSYS>.
		1254
		1255	Note that the corresponding C<munlockall> is synchronous and is
		1256	documented under L<MISCELLANEOUS FUNCTIONS>.
		1257
		1258	Example: asynchronously lock all current and future pages into memory.
		1259
		1260	aio_mlockall IO::AIO::MCL_FUTURE;
		1261
		1262	=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		1263
		1264	Queries the extents of the given file (by calling the Linux FIEMAP ioctl,
		1265	see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If the
		1266	C<ioctl> is not available on your OS, then this rquiest will fail with
		1267	C<ENOSYS>.
		1268
		1269	C<$start> is the starting offset to query extents for, C<$length> is the
		1270	size of the range to query - if it is C<undef>, then the whole file will
		1271	be queried.
		1272
		1273	C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
		1274	C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
		1275	exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
		1276	the data portion.
		1277
		1278	C<$count> is the maximum number of extent records to return. If it is
		1279	C<undef>, then IO::AIO queries all extents of the file. As a very special
		1280	case, if it is C<0>, then the callback receives the number of extents
		1281	instead of the extents themselves.
		1282
		1283	If an error occurs, the callback receives no arguments. The special
		1284	C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
		1285
		1286	Otherwise, the callback receives an array reference with extent
		1287	structures. Each extent structure is an array reference itself, with the
		1288	following members:
		1289
		1290	[$logical, $physical, $length, $flags]
		1291
		1292	Flags is any combination of the following flag values (typically either C<0>
		1293	or C<IO::AIO::FIEMAP_EXTENT_LAST>):
		1294
		1295	C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
		1296	C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
		1297	C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
		1298	C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
		1299	C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
		1300	C<IO::AIO::FIEMAP_EXTENT_SHARED>.
		1301
		1302	=item aio_group $callback->(...)
		1303
		1304	This is a very special aio request: Instead of doing something, it is a
		1305	container for other aio requests, which is useful if you want to bundle
		1306	many requests into a single, composite, request with a definite callback
		1307	and the ability to cancel the whole request with its subrequests.
		1308
		1309	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		1310	for more info.
		1311
		1312	Example:
		1313
		1314	my $grp = aio_group sub {
		1315	print "all stats done\n";
		1316	};
		1317
		1318	add $grp
		1319	(aio_stat ...),
		1320	(aio_stat ...),
		1321	...;
		1322
		1323	=item aio_nop $callback->()
		1324
		1325	This is a special request - it does nothing in itself and is only used for
		1326	side effects, such as when you want to add a dummy request to a group so
		1327	that finishing the requests in the group depends on executing the given
		1328	code.
		1329
		1330	While this request does nothing, it still goes through the execution
		1331	phase and still requires a worker thread. Thus, the callback will not
		1332	be executed immediately but only after other requests in the queue have
		1333	entered their execution phase. This can be used to measure request
		1334	latency.
		1335
		1336	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		1337
		1338	Mainly used for debugging and benchmarking, this aio request puts one of
		1339	the request workers to sleep for the given time.
		1340
		1341	While it is theoretically handy to have simple I/O scheduling requests
		1342	like sleep and file handle readable/writable, the overhead this creates is
		1343	immense (it blocks a thread for a long time) so do not use this function
		1344	except to put your application under artificial I/O pressure.
		1345
260	=back	1346	=back
261		1347
		1348
		1349	=head2 IO::AIO::WD - multiple working directories
		1350
		1351	Your process only has one current working directory, which is used by all
		1352	threads. This makes it hard to use relative paths (some other component
		1353	could call C<chdir> at any time, and it is hard to control when the path
		1354	will be used by IO::AIO).
		1355
		1356	One solution for this is to always use absolute paths. This usually works,
		1357	but can be quite slow (the kernel has to walk the whole path on every
		1358	access), and can also be a hassle to implement.
		1359
		1360	Newer POSIX systems have a number of functions (openat, fdopendir,
		1361	futimensat and so on) that make it possible to specify working directories
		1362	per operation.
		1363
		1364	For portability, and because the clowns who "designed", or shall I write,
		1365	perpetrated this new interface were obviously half-drunk, this abstraction
		1366	cannot be perfect, though.
		1367
		1368	IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
		1369	object. This object stores the canonicalised, absolute version of the
		1370	path, and on systems that allow it, also a directory file descriptor.
		1371
		1372	Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
		1373	or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
		1374	object and a pathname instead (or the IO::AIO::WD object alone, which
		1375	gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
		1376	IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
		1377	to that IO::AIO::WD object.
		1378
		1379	For example, to get a wd object for F</etc> and then stat F<passwd>
		1380	inside, you would write:
		1381
		1382	aio_wd "/etc", sub {
		1383	my $etcdir = shift;
		1384
		1385	# although $etcdir can be undef on error, there is generally no reason
		1386	# to check for errors here, as aio_stat will fail with ENOENT
		1387	# when $etcdir is undef.
		1388
		1389	aio_stat [$etcdir, "passwd"], sub {
		1390	# yay
		1391	};
		1392	};
		1393
		1394	That C<aio_wd> is a request and not a normal function shows that creating
		1395	an IO::AIO::WD object is itself a potentially blocking operation, which is
		1396	why it is done asynchronously.
		1397
		1398	To stat the directory obtained with C<aio_wd> above, one could write
		1399	either of the following three request calls:
		1400
		1401	aio_lstat "/etc" , sub { ... # pathname as normal string
		1402	aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
		1403	aio_lstat $wd , sub { ... # shorthand for the previous
		1404
		1405	As with normal pathnames, IO::AIO keeps a copy of the working directory
		1406	object and the pathname string, so you could write the following without
		1407	causing any issues due to C<$path> getting reused:
		1408
		1409	my $path = [$wd, undef];
		1410
		1411	for my $name (qw(abc def ghi)) {
		1412	$path->[1] = $name;
		1413	aio_stat $path, sub {
		1414	# ...
		1415	};
		1416	}
		1417
		1418	There are some caveats: when directories get renamed (or deleted), the
		1419	pathname string doesn't change, so will point to the new directory (or
		1420	nowhere at all), while the directory fd, if available on the system,
		1421	will still point to the original directory. Most functions accepting a
		1422	pathname will use the directory fd on newer systems, and the string on
		1423	older systems. Some functions (such as realpath) will always rely on the
		1424	string form of the pathname.
		1425
		1426	So this fucntionality is mainly useful to get some protection against
		1427	C<chdir>, to easily get an absolute path out of a relative path for future
		1428	reference, and to speed up doing many operations in the same directory
		1429	(e.g. when stat'ing all files in a directory).
		1430
		1431	The following functions implement this working directory abstraction:
		1432
		1433	=over 4
		1434
		1435	=item aio_wd $pathname, $callback->($wd)
		1436
		1437	Asynchonously canonicalise the given pathname and convert it to an
		1438	IO::AIO::WD object representing it. If possible and supported on the
		1439	system, also open a directory fd to speed up pathname resolution relative
		1440	to this working directory.
		1441
		1442	If something goes wrong, then C<undef> is passwd to the callback instead
		1443	of a working directory object and C<$!> is set appropriately. Since
		1444	passing C<undef> as working directory component of a pathname fails the
		1445	request with C<ENOENT>, there is often no need for error checking in the
		1446	C<aio_wd> callback, as future requests using the value will fail in the
		1447	expected way.
		1448
		1449	If this call isn't available because your OS lacks it or it couldn't be
		1450	detected, it will be emulated by calling C<fsync> instead.
		1451
		1452	=item IO::AIO::CWD
		1453
		1454	This is a compiletime constant (object) that represents the process
		1455	current working directory.
		1456
		1457	Specifying this object as working directory object for a pathname is as
		1458	if the pathname would be specified directly, without a directory object,
		1459	e.g., these calls are functionally identical:
		1460
		1461	aio_stat "somefile", sub { ... };
		1462	aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
		1463
		1464	=back
		1465
		1466
		1467	=head2 IO::AIO::REQ CLASS
		1468
		1469	All non-aggregate C<aio_*> functions return an object of this class when
		1470	called in non-void context.
		1471
		1472	=over 4
		1473
		1474	=item cancel $req
		1475
		1476	Cancels the request, if possible. Has the effect of skipping execution
		1477	when entering the B<execute> state and skipping calling the callback when
		1478	entering the the B<result> state, but will leave the request otherwise
		1479	untouched (with the exception of readdir). That means that requests that
		1480	currently execute will not be stopped and resources held by the request
		1481	will not be freed prematurely.
		1482
		1483	=item cb $req $callback->(...)
		1484
		1485	Replace (or simply set) the callback registered to the request.
		1486
		1487	=back
		1488
		1489	=head2 IO::AIO::GRP CLASS
		1490
		1491	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		1492	objects of this class, too.
		1493
		1494	A IO::AIO::GRP object is a special request that can contain multiple other
		1495	aio requests.
		1496
		1497	You create one by calling the C<aio_group> constructing function with a
		1498	callback that will be called when all contained requests have entered the
		1499	C<done> state:
		1500
		1501	my $grp = aio_group sub {
		1502	print "all requests are done\n";
		1503	};
		1504
		1505	You add requests by calling the C<add> method with one or more
		1506	C<IO::AIO::REQ> objects:
		1507
		1508	$grp->add (aio_unlink "...");
		1509
		1510	add $grp aio_stat "...", sub {
		1511	$_[0] or return $grp->result ("error");
		1512
		1513	# add another request dynamically, if first succeeded
		1514	add $grp aio_open "...", sub {
		1515	$grp->result ("ok");
		1516	};
		1517	};
		1518
		1519	This makes it very easy to create composite requests (see the source of
		1520	C<aio_move> for an application) that work and feel like simple requests.
		1521
		1522	=over 4
		1523
		1524	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		1525	C<IO::AIO::poll_cb>, just like any other request.
		1526
		1527	=item * They can be canceled like any other request. Canceling will cancel not
		1528	only the request itself, but also all requests it contains.
		1529
		1530	=item * They can also can also be added to other IO::AIO::GRP objects.
		1531
		1532	=item * You must not add requests to a group from within the group callback (or
		1533	any later time).
		1534
		1535	=back
		1536
		1537	Their lifetime, simplified, looks like this: when they are empty, they
		1538	will finish very quickly. If they contain only requests that are in the
		1539	C<done> state, they will also finish. Otherwise they will continue to
		1540	exist.
		1541
		1542	That means after creating a group you have some time to add requests
		1543	(precisely before the callback has been invoked, which is only done within
		1544	the C<poll_cb>). And in the callbacks of those requests, you can add
		1545	further requests to the group. And only when all those requests have
		1546	finished will the the group itself finish.
		1547
		1548	=over 4
		1549
		1550	=item add $grp ...
		1551
		1552	=item $grp->add (...)
		1553
		1554	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1555	be added, including other groups, as long as you do not create circular
		1556	dependencies.
		1557
		1558	Returns all its arguments.
		1559
		1560	=item $grp->cancel_subs
		1561
		1562	Cancel all subrequests and clears any feeder, but not the group request
		1563	itself. Useful when you queued a lot of events but got a result early.
		1564
		1565	The group request will finish normally (you cannot add requests to the
		1566	group).
		1567
		1568	=item $grp->result (...)
		1569
		1570	Set the result value(s) that will be passed to the group callback when all
		1571	subrequests have finished and set the groups errno to the current value
		1572	of errno (just like calling C<errno> without an error number). By default,
		1573	no argument will be passed and errno is zero.
		1574
		1575	=item $grp->errno ([$errno])
		1576
		1577	Sets the group errno value to C<$errno>, or the current value of errno
		1578	when the argument is missing.
		1579
		1580	Every aio request has an associated errno value that is restored when
		1581	the callback is invoked. This method lets you change this value from its
		1582	default (0).
		1583
		1584	Calling C<result> will also set errno, so make sure you either set C<$!>
		1585	before the call to C<result>, or call c<errno> after it.
		1586
		1587	=item feed $grp $callback->($grp)
		1588
		1589	Sets a feeder/generator on this group: every group can have an attached
		1590	generator that generates requests if idle. The idea behind this is that,
		1591	although you could just queue as many requests as you want in a group,
		1592	this might starve other requests for a potentially long time. For example,
		1593	C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
		1594	requests, delaying any later requests for a long time.
		1595
		1596	To avoid this, and allow incremental generation of requests, you can
		1597	instead a group and set a feeder on it that generates those requests. The
		1598	feed callback will be called whenever there are few enough (see C<limit>,
		1599	below) requests active in the group itself and is expected to queue more
		1600	requests.
		1601
		1602	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1603	not impose any limits).
		1604
		1605	If the feed does not queue more requests when called, it will be
		1606	automatically removed from the group.
		1607
		1608	If the feed limit is C<0> when this method is called, it will be set to
		1609	C<2> automatically.
		1610
		1611	Example:
		1612
		1613	# stat all files in @files, but only ever use four aio requests concurrently:
		1614
		1615	my $grp = aio_group sub { print "finished\n" };
		1616	limit $grp 4;
		1617	feed $grp sub {
		1618	my $file = pop @files
		1619	or return;
		1620
		1621	add $grp aio_stat $file, sub { ... };
		1622	};
		1623
		1624	=item limit $grp $num
		1625
		1626	Sets the feeder limit for the group: The feeder will be called whenever
		1627	the group contains less than this many requests.
		1628
		1629	Setting the limit to C<0> will pause the feeding process.
		1630
		1631	The default value for the limit is C<0>, but note that setting a feeder
		1632	automatically bumps it up to C<2>.
		1633
		1634	=back
		1635
262	=head2 SUPPORT FUNCTIONS	1636	=head2 SUPPORT FUNCTIONS
263		1637
		1638	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1639
264	=over 4	1640	=over 4
265		1641
266	=item $fileno = IO::AIO::poll_fileno	1642	=item $fileno = IO::AIO::poll_fileno
267		1643
268	Return the I<request result pipe file descriptor>. This filehandle must be	1644	Return the I<request result pipe file descriptor>. This filehandle must be
269	polled for reading by some mechanism outside this module (e.g. Event or	1645	polled for reading by some mechanism outside this module (e.g. EV, Glib,
270	select, see below or the SYNOPSIS). If the pipe becomes readable you have	1646	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
271	to call C<poll_cb> to check the results.	1647	you have to call C<poll_cb> to check the results.
272		1648
273	See C<poll_cb> for an example.	1649	See C<poll_cb> for an example.
274		1650
275	=item IO::AIO::poll_cb	1651	=item IO::AIO::poll_cb
276		1652
277	Process all outstanding events on the result pipe. You have to call this	1653	Process some outstanding events on the result pipe. You have to call
278	regularly. Returns the number of events processed. Returns immediately	1654	this regularly. Returns C<0> if all events could be processed (or there
279	when no events are outstanding.	1655	were no events to process), or C<-1> if it returned earlier for whatever
		1656	reason. Returns immediately when no events are outstanding. The amount of
		1657	events processed depends on the settings of C<IO::AIO::max_poll_req> and
		1658	C<IO::AIO::max_poll_time>.
		1659
		1660	If not all requests were processed for whatever reason, the filehandle
		1661	will still be ready when C<poll_cb> returns, so normally you don't have to
		1662	do anything special to have it called later.
		1663
		1664	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1665	ready, it can be beneficial to call this function from loops which submit
		1666	a lot of requests, to make sure the results get processed when they become
		1667	available and not just when the loop is finished and the event loop takes
		1668	over again. This function returns very fast when there are no outstanding
		1669	requests.
280		1670
281	Example: Install an Event watcher that automatically calls	1671	Example: Install an Event watcher that automatically calls
282	IO::AIO::poll_cb with high priority:	1672	IO::AIO::poll_cb with high priority (more examples can be found in the
		1673	SYNOPSIS section, at the top of this document):
283		1674
284	Event->io (fd => IO::AIO::poll_fileno,	1675	Event->io (fd => IO::AIO::poll_fileno,
285	poll => 'r', async => 1,	1676	poll => 'r', async => 1,
286	cb => \&IO::AIO::poll_cb);	1677	cb => \&IO::AIO::poll_cb);
287		1678
288	=item IO::AIO::poll_wait	1679	=item IO::AIO::poll_wait
289		1680
		1681	If there are any outstanding requests and none of them in the result
290	Wait till the result filehandle becomes ready for reading (simply does a	1682	phase, wait till the result filehandle becomes ready for reading (simply
291	C<select> on the filehandle. This is useful if you want to synchronously wait	1683	does a C<select> on the filehandle. This is useful if you want to
292	for some requests to finish).	1684	synchronously wait for some requests to finish).
293		1685
294	See C<nreqs> for an example.	1686	See C<nreqs> for an example.
295		1687
		1688	=item IO::AIO::poll
		1689
		1690	Waits until some requests have been handled.
		1691
		1692	Returns the number of requests processed, but is otherwise strictly
		1693	equivalent to:
		1694
		1695	IO::AIO::poll_wait, IO::AIO::poll_cb
		1696
296	=item IO::AIO::nreqs	1697	=item IO::AIO::flush
297		1698
298	Returns the number of requests currently outstanding (i.e. for which their	1699	Wait till all outstanding AIO requests have been handled.
299	callback has not been invoked yet).
300		1700
301	Example: wait till there are no outstanding requests anymore:	1701	Strictly equivalent to:
302		1702
303	IO::AIO::poll_wait, IO::AIO::poll_cb	1703	IO::AIO::poll_wait, IO::AIO::poll_cb
304	while IO::AIO::nreqs;	1704	while IO::AIO::nreqs;
305		1705
306	=item IO::AIO::flush	1706	=item IO::AIO::max_poll_reqs $nreqs
307		1707
308	Wait till all outstanding AIO requests have been handled.	1708	=item IO::AIO::max_poll_time $seconds
309		1709
310	Strictly equivalent to:	1710	These set the maximum number of requests (default C<0>, meaning infinity)
		1711	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1712	the maximum amount of time (default C<0>, meaning infinity) spent in
		1713	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1714	of time C<poll_cb> is allowed to use).
311		1715
312	IO::AIO::poll_wait, IO::AIO::poll_cb	1716	Setting C<max_poll_time> to a non-zero value creates an overhead of one
313	while IO::AIO::nreqs;	1717	syscall per request processed, which is not normally a problem unless your
		1718	callbacks are really really fast or your OS is really really slow (I am
		1719	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
314		1720
315	=item IO::AIO::poll	1721	Setting these is useful if you want to ensure some level of
		1722	interactiveness when perl is not fast enough to process all requests in
		1723	time.
316		1724
317	Waits until some requests have been handled.	1725	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
318		1726
319	Strictly equivalent to:	1727	Example: Install an Event watcher that automatically calls
		1728	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1729	program get the CPU sometimes even under high AIO load.
320		1730
321	IO::AIO::poll_wait, IO::AIO::poll_cb	1731	# try not to spend much more than 0.1s in poll_cb
322	if IO::AIO::nreqs;	1732	IO::AIO::max_poll_time 0.1;
		1733
		1734	# use a low priority so other tasks have priority
		1735	Event->io (fd => IO::AIO::poll_fileno,
		1736	poll => 'r', nice => 1,
		1737	cb => &IO::AIO::poll_cb);
		1738
		1739	=back
		1740
		1741	=head3 CONTROLLING THE NUMBER OF THREADS
		1742
		1743	=over
323		1744
324	=item IO::AIO::min_parallel $nthreads	1745	=item IO::AIO::min_parallel $nthreads
325		1746
326	Set the minimum number of AIO threads to C<$nthreads>. The current default	1747	Set the minimum number of AIO threads to C<$nthreads>. The current
327	is C<4>, which means four asynchronous operations can be done at one time	1748	default is C<8>, which means eight asynchronous operations can execute
328	(the number of outstanding operations, however, is unlimited).	1749	concurrently at any one time (the number of outstanding requests,
		1750	however, is unlimited).
329		1751
330	IO::AIO starts threads only on demand, when an AIO request is queued and	1752	IO::AIO starts threads only on demand, when an AIO request is queued and
331	no free thread exists.	1753	no free thread exists. Please note that queueing up a hundred requests can
		1754	create demand for a hundred threads, even if it turns out that everything
		1755	is in the cache and could have been processed faster by a single thread.
332		1756
333	It is recommended to keep the number of threads low, as some Linux	1757	It is recommended to keep the number of threads relatively low, as some
334	kernel versions will scale negatively with the number of threads (higher	1758	Linux kernel versions will scale negatively with the number of threads
335	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1759	(higher parallelity => MUCH higher latency). With current Linux 2.6
336	threads should be fine.	1760	versions, 4-32 threads should be fine.
337		1761
338	Under most circumstances you don't need to call this function, as the	1762	Under most circumstances you don't need to call this function, as the
339	module selects a default that is suitable for low to moderate load.	1763	module selects a default that is suitable for low to moderate load.
340		1764
341	=item IO::AIO::max_parallel $nthreads	1765	=item IO::AIO::max_parallel $nthreads
…		…
350	This module automatically runs C<max_parallel 0> at program end, to ensure	1774	This module automatically runs C<max_parallel 0> at program end, to ensure
351	that all threads are killed and that there are no outstanding requests.	1775	that all threads are killed and that there are no outstanding requests.
352		1776
353	Under normal circumstances you don't need to call this function.	1777	Under normal circumstances you don't need to call this function.
354		1778
		1779	=item IO::AIO::max_idle $nthreads
		1780
		1781	Limit the number of threads (default: 4) that are allowed to idle
		1782	(i.e., threads that did not get a request to process within the idle
		1783	timeout (default: 10 seconds). That means if a thread becomes idle while
		1784	C<$nthreads> other threads are also idle, it will free its resources and
		1785	exit.
		1786
		1787	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1788	to allow for extremely high load situations, but want to free resources
		1789	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1790
		1791	The default is probably ok in most situations, especially if thread
		1792	creation is fast. If thread creation is very slow on your system you might
		1793	want to use larger values.
		1794
		1795	=item IO::AIO::idle_timeout $seconds
		1796
		1797	Sets the minimum idle timeout (default 10) after which worker threads are
		1798	allowed to exit. SEe C<IO::AIO::max_idle>.
		1799
355	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1800	=item IO::AIO::max_outstanding $maxreqs
356		1801
357	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1802	Sets the maximum number of outstanding requests to C<$nreqs>. If
358	try to queue up more than this number of requests, the caller will block until	1803	you do queue up more than this number of requests, the next call to
359	some requests have been handled.	1804	C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
		1805	C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
		1806	longer exceeded.
360		1807
361	The default is very large, so normally there is no practical limit. If you	1808	In other words, this setting does not enforce a queue limit, but can be
362	queue up many requests in a loop it often improves speed if you set	1809	used to make poll functions block if the limit is exceeded.
363	this to a relatively low number, such as C<100>.
364		1810
365	Under normal circumstances you don't need to call this function.	1811	This is a very bad function to use in interactive programs because it
		1812	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1813	use an C<aio_group> together with a feed callback.
		1814
		1815	It's main use is in scripts without an event loop - when you want to stat
		1816	a lot of files, you can write somehting like this:
		1817
		1818	IO::AIO::max_outstanding 32;
		1819
		1820	for my $path (...) {
		1821	aio_stat $path , ...;
		1822	IO::AIO::poll_cb;
		1823	}
		1824
		1825	IO::AIO::flush;
		1826
		1827	The call to C<poll_cb> inside the loop will normally return instantly, but
		1828	as soon as more thna C<32> reqeusts are in-flight, it will block until
		1829	some requests have been handled. This keeps the loop from pushing a large
		1830	number of C<aio_stat> requests onto the queue.
		1831
		1832	The default value for C<max_outstanding> is very large, so there is no
		1833	practical limit on the number of outstanding requests.
366		1834
367	=back	1835	=back
368		1836
		1837	=head3 STATISTICAL INFORMATION
		1838
		1839	=over
		1840
		1841	=item IO::AIO::nreqs
		1842
		1843	Returns the number of requests currently in the ready, execute or pending
		1844	states (i.e. for which their callback has not been invoked yet).
		1845
		1846	Example: wait till there are no outstanding requests anymore:
		1847
		1848	IO::AIO::poll_wait, IO::AIO::poll_cb
		1849	while IO::AIO::nreqs;
		1850
		1851	=item IO::AIO::nready
		1852
		1853	Returns the number of requests currently in the ready state (not yet
		1854	executed).
		1855
		1856	=item IO::AIO::npending
		1857
		1858	Returns the number of requests currently in the pending state (executed,
		1859	but not yet processed by poll_cb).
		1860
		1861	=back
		1862
		1863	=head3 MISCELLANEOUS FUNCTIONS
		1864
		1865	IO::AIO implements some functions that might be useful, but are not
		1866	asynchronous.
		1867
		1868	=over 4
		1869
		1870	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		1871
		1872	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		1873	but is blocking (this makes most sense if you know the input data is
		1874	likely cached already and the output filehandle is set to non-blocking
		1875	operations).
		1876
		1877	Returns the number of bytes copied, or C<-1> on error.
		1878
		1879	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		1880
		1881	Simply calls the C<posix_fadvise> function (see its
		1882	manpage for details). The following advice constants are
		1883	available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		1884	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		1885	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		1886
		1887	On systems that do not implement C<posix_fadvise>, this function returns
		1888	ENOSYS, otherwise the return value of C<posix_fadvise>.
		1889
		1890	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		1891
		1892	Simply calls the C<posix_madvise> function (see its
		1893	manpage for details). The following advice constants are
		1894	available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		1895	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>.
		1896
		1897	On systems that do not implement C<posix_madvise>, this function returns
		1898	ENOSYS, otherwise the return value of C<posix_madvise>.
		1899
		1900	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		1901
		1902	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		1903	$scalar (see its manpage for details). The following protect
		1904	constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		1905	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		1906
		1907	On systems that do not implement C<mprotect>, this function returns
		1908	ENOSYS, otherwise the return value of C<mprotect>.
		1909
		1910	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		1911
		1912	Memory-maps a file (or anonymous memory range) and attaches it to the
		1913	given C<$scalar>, which will act like a string scalar. Returns true on
		1914	success, and false otherwise.
		1915
		1916	The only operations allowed on the scalar are C<substr>/C<vec> that don't
		1917	change the string length, and most read-only operations such as copying it
		1918	or searching it with regexes and so on.
		1919
		1920	Anything else is unsafe and will, at best, result in memory leaks.
		1921
		1922	The memory map associated with the C<$scalar> is automatically removed
		1923	when the C<$scalar> is destroyed, or when the C<IO::AIO::mmap> or
		1924	C<IO::AIO::munmap> functions are called.
		1925
		1926	This calls the C<mmap>(2) function internally. See your system's manual
		1927	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		1928
		1929	The C<$length> must be larger than zero and smaller than the actual
		1930	filesize.
		1931
		1932	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		1933	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		1934
		1935	C<$flags> can be a combination of C<IO::AIO::MAP_SHARED> or
		1936	C<IO::AIO::MAP_PRIVATE>, or a number of system-specific flags (when
		1937	not available, the are defined as 0): C<IO::AIO::MAP_ANONYMOUS>
		1938	(which is set to C<MAP_ANON> if your system only provides this
		1939	constant), C<IO::AIO::MAP_HUGETLB>, C<IO::AIO::MAP_LOCKED>,
		1940	C<IO::AIO::MAP_NORESERVE>, C<IO::AIO::MAP_POPULATE> or
		1941	C<IO::AIO::MAP_NONBLOCK>
		1942
		1943	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		1944
		1945	C<$offset> is the offset from the start of the file - it generally must be
		1946	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		1947
		1948	Example:
		1949
		1950	use Digest::MD5;
		1951	use IO::AIO;
		1952
		1953	open my $fh, "<verybigfile"
		1954	or die "$!";
		1955
		1956	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		1957	or die "verybigfile: $!";
		1958
		1959	my $fast_md5 = md5 $data;
		1960
		1961	=item IO::AIO::munmap $scalar
		1962
		1963	Removes a previous mmap and undefines the C<$scalar>.
		1964
		1965	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		1966
		1967	Calls the C<munlock> function, undoing the effects of a previous
		1968	C<aio_mlock> call (see its description for details).
		1969
		1970	=item IO::AIO::munlockall
		1971
		1972	Calls the C<munlockall> function.
		1973
		1974	On systems that do not implement C<munlockall>, this function returns
		1975	ENOSYS, otherwise the return value of C<munlockall>.
		1976
		1977	=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
		1978
		1979	Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
		1980	C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
		1981	should be the file offset.
		1982
		1983	C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
		1984	silently corrupt the data in this case.
		1985
		1986	The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
		1987	C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
		1988	C<IO::AIO::SPLICE_F_GIFT>.
		1989
		1990	See the C<splice(2)> manpage for details.
		1991
		1992	=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
		1993
		1994	Calls the GNU/Linux C<tee(2)> syscall, see it's manpage and the
		1995	description for C<IO::AIO::splice> above for details.
		1996
		1997	=back
		1998
369	=cut	1999	=cut
370		2000
371	# support function to convert a fd into a perl filehandle
372	sub _fd2fh {
373	return undef if $_[0] < 0;
374
375	# try to generate nice filehandles
376	my $sym = "IO::AIO::fd#$_[0]";
377	local *$sym;
378
379	open *$sym, "+<&=$_[0]" # usually works under any unix
380	or open *$sym, "<&=$_[0]" # cygwin needs this
381	or open *$sym, ">&=$_[0]" # or this
382	or return undef;
383
384	*$sym
385	}
386
387	min_parallel 4;	2001	min_parallel 8;
388		2002
389	END {	2003	END { flush }
390	max_parallel 0;
391	}
392		2004
393	1;	2005	1;
394		2006
		2007	=head1 EVENT LOOP INTEGRATION
		2008
		2009	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		2010	automatically into many event loops:
		2011
		2012	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		2013	use AnyEvent::AIO;
		2014
		2015	You can also integrate IO::AIO manually into many event loops, here are
		2016	some examples of how to do this:
		2017
		2018	# EV integration
		2019	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		2020
		2021	# Event integration
		2022	Event->io (fd => IO::AIO::poll_fileno,
		2023	poll => 'r',
		2024	cb => \&IO::AIO::poll_cb);
		2025
		2026	# Glib/Gtk2 integration
		2027	add_watch Glib::IO IO::AIO::poll_fileno,
		2028	in => sub { IO::AIO::poll_cb; 1 };
		2029
		2030	# Tk integration
		2031	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		2032	readable => \&IO::AIO::poll_cb);
		2033
		2034	# Danga::Socket integration
		2035	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		2036	\&IO::AIO::poll_cb);
		2037
395	=head2 FORK BEHAVIOUR	2038	=head2 FORK BEHAVIOUR
396		2039
397	Before the fork, IO::AIO enters a quiescent state where no requests	2040	Usage of pthreads in a program changes the semantics of fork
398	can be added in other threads and no results will be processed. After	2041	considerably. Specifically, only async-safe functions can be called after
399	the fork the parent simply leaves the quiescent state and continues	2042	fork. Perl doesn't know about this, so in general, you cannot call fork
400	request/result processing, while the child clears the request/result	2043	with defined behaviour in perl if pthreads are involved. IO::AIO uses
401	queue (so the requests started before the fork will only be handled in	2044	pthreads, so this applies, but many other extensions and (for inexplicable
402	the parent). Threats will be started on demand until the limit ste in the	2045	reasons) perl itself often is linked against pthreads, so this limitation
403	parent process has been reached again.	2046	applies to quite a lot of perls.
		2047
		2048	This module no longer tries to fight your OS, or POSIX. That means IO::AIO
		2049	only works in the process that loaded it. Forking is fully supported, but
		2050	using IO::AIO in the child is not.
		2051
		2052	You might get around by not I<using> IO::AIO before (or after)
		2053	forking. You could also try to call the L<IO::AIO::reinit> function in the
		2054	child:
		2055
		2056	=over 4
		2057
		2058	=item IO::AIO::reinit
		2059
		2060	Abandons all current requests and I/O threads and simply reinitialises all
		2061	data structures. This is not an operation supported by any standards, but
		2062	happens to work on GNU/Linux and some newer BSD systems.
		2063
		2064	The only reasonable use for this function is to call it after forking, if
		2065	C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
		2066	the process will result in undefined behaviour. Calling it at any time
		2067	will also result in any undefined (by POSIX) behaviour.
		2068
		2069	=back
		2070
		2071	=head2 MEMORY USAGE
		2072
		2073	Per-request usage:
		2074
		2075	Each aio request uses - depending on your architecture - around 100-200
		2076	bytes of memory. In addition, stat requests need a stat buffer (possibly
		2077	a few hundred bytes), readdir requires a result buffer and so on. Perl
		2078	scalars and other data passed into aio requests will also be locked and
		2079	will consume memory till the request has entered the done state.
		2080
		2081	This is not awfully much, so queuing lots of requests is not usually a
		2082	problem.
		2083
		2084	Per-thread usage:
		2085
		2086	In the execution phase, some aio requests require more memory for
		2087	temporary buffers, and each thread requires a stack and other data
		2088	structures (usually around 16k-128k, depending on the OS).
		2089
		2090	=head1 KNOWN BUGS
		2091
		2092	Known bugs will be fixed in the next release.
404		2093
405	=head1 SEE ALSO	2094	=head1 SEE ALSO
406		2095
407	L<Coro>, L<Linux::AIO>.	2096	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		2097	more natural syntax.
408		2098
409	=head1 AUTHOR	2099	=head1 AUTHOR
410		2100
411	Marc Lehmann <schmorp@schmorp.de>	2101	Marc Lehmann <schmorp@schmorp.de>
412	http://home.schmorp.de/	2102	http://home.schmorp.de/

Diff Legend

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