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 1.284 by root, Fri Mar 23 01:14:08 2018 UTC

1	=head1 NAME	1	=head1 NAME
2		2
3	IO::AIO - Asynchronous Input/Output	3	IO::AIO - Asynchronous/Advanced Input/Output
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use IO::AIO;	7	use IO::AIO;
8		8
9	aio_open "/etc/passwd", O_RDONLY, 0, sub {	9	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
10	my ($fh) = @_;	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
		63	In addition to asynchronous I/O, this module also exports some rather
		64	arcane interfaces, such as C<madvise> or linux's C<splice> system call,
		65	which is why the C<A> in C<AIO> can also mean I<advanced>.
		66
52	Although the module will work with in the presence of other threads, it is	67	Although the module will work in the presence of other (Perl-) threads,
53	currently not reentrant, so use appropriate locking yourself, always call	68	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	69	yourself, always call C<poll_cb> from within the same thread, or never
55	C<aio_> functions) recursively.	70	call C<poll_cb> (or other C<aio_> functions) recursively.
		71
		72	=head2 EXAMPLE
		73
		74	This is a simple example that uses the EV module and loads
		75	F</etc/passwd> asynchronously:
		76
		77	use EV;
		78	use IO::AIO;
		79
		80	# register the IO::AIO callback with EV
		81	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		82
		83	# queue the request to open /etc/passwd
		84	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
		85	my $fh = shift
		86	or die "error while opening: $!";
		87
		88	# stat'ing filehandles is generally non-blocking
		89	my $size = -s $fh;
		90
		91	# queue a request to read the file
		92	my $contents;
		93	aio_read $fh, 0, $size, $contents, 0, sub {
		94	$_[0] == $size
		95	or die "short read: $!";
		96
		97	close $fh;
		98
		99	# file contents now in $contents
		100	print $contents;
		101
		102	# exit event loop and program
		103	EV::break;
		104	};
		105	};
		106
		107	# possibly queue up other requests, or open GUI windows,
		108	# check for sockets etc. etc.
		109
		110	# process events as long as there are some:
		111	EV::run;
		112
		113	=head1 REQUEST ANATOMY AND LIFETIME
		114
		115	Every C<aio_*> function creates a request. which is a C data structure not
		116	directly visible to Perl.
		117
		118	If called in non-void context, every request function returns a Perl
		119	object representing the request. In void context, nothing is returned,
		120	which saves a bit of memory.
		121
		122	The perl object is a fairly standard ref-to-hash object. The hash contents
		123	are not used by IO::AIO so you are free to store anything you like in it.
		124
		125	During their existance, aio requests travel through the following states,
		126	in order:
		127
		128	=over 4
		129
		130	=item ready
		131
		132	Immediately after a request is created it is put into the ready state,
		133	waiting for a thread to execute it.
		134
		135	=item execute
		136
		137	A thread has accepted the request for processing and is currently
		138	executing it (e.g. blocking in read).
		139
		140	=item pending
		141
		142	The request has been executed and is waiting for result processing.
		143
		144	While request submission and execution is fully asynchronous, result
		145	processing is not and relies on the perl interpreter calling C<poll_cb>
		146	(or another function with the same effect).
		147
		148	=item result
		149
		150	The request results are processed synchronously by C<poll_cb>.
		151
		152	The C<poll_cb> function will process all outstanding aio requests by
		153	calling their callbacks, freeing memory associated with them and managing
		154	any groups they are contained in.
		155
		156	=item done
		157
		158	Request has reached the end of its lifetime and holds no resources anymore
		159	(except possibly for the Perl object, but its connection to the actual
		160	aio request is severed and calling its methods will either do nothing or
		161	result in a runtime error).
		162
		163	=back
56		164
57	=cut	165	=cut
58		166
59	package IO::AIO;	167	package IO::AIO;
60		168
61	no warnings;	169	use Carp ();
		170
		171	use common::sense;
62		172
63	use base 'Exporter';	173	use base 'Exporter';
64		174
65	use Fcntl ();
66
67	BEGIN {	175	BEGIN {
68	$VERSION = 1.6;	176	our $VERSION = 4.4;
69		177
70	@EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat	178	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	179	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
72	aio_fsync aio_fdatasync aio_readahead);	180	aio_scandir aio_symlink aio_readlink aio_realpath aio_fcntl aio_ioctl
73	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel	181	aio_sync aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range
74	max_outstanding nreqs);	182	aio_pathsync aio_readahead aio_fiemap aio_allocate
		183	aio_rename aio_rename2 aio_link aio_move aio_copy aio_group
		184	aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
		185	aio_chmod aio_utime aio_truncate
		186	aio_msync aio_mtouch aio_mlock aio_mlockall
		187	aio_statvfs
		188	aio_slurp
		189	aio_wd);
		190
		191	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		192	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		193	min_parallel max_parallel max_idle idle_timeout
		194	nreqs nready npending nthreads
		195	max_poll_time max_poll_reqs
		196	sendfile fadvise madvise
		197	mmap munmap munlock munlockall);
		198
		199	push @AIO_REQ, qw(aio_busy); # not exported
		200
		201	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
75		202
76	require XSLoader;	203	require XSLoader;
77	XSLoader::load IO::AIO, $VERSION;	204	XSLoader::load ("IO::AIO", $VERSION);
78	}	205	}
79		206
80	=head1 FUNCTIONS	207	=head1 FUNCTIONS
81		208
82	=head2 AIO FUNCTIONS	209	=head2 QUICK OVERVIEW
		210
		211	This section simply lists the prototypes most of the functions for
		212	quick reference. See the following sections for function-by-function
		213	documentation.
		214
		215	aio_wd $pathname, $callback->($wd)
		216	aio_open $pathname, $flags, $mode, $callback->($fh)
		217	aio_close $fh, $callback->($status)
		218	aio_seek $fh,$offset,$whence, $callback->($offs)
		219	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		220	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		221	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		222	aio_readahead $fh,$offset,$length, $callback->($retval)
		223	aio_stat $fh_or_path, $callback->($status)
		224	aio_lstat $fh, $callback->($status)
		225	aio_statvfs $fh_or_path, $callback->($statvfs)
		226	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		227	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		228	aio_chmod $fh_or_path, $mode, $callback->($status)
		229	aio_truncate $fh_or_path, $offset, $callback->($status)
		230	aio_allocate $fh, $mode, $offset, $len, $callback->($status)
		231	aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		232	aio_unlink $pathname, $callback->($status)
		233	aio_mknod $pathname, $mode, $dev, $callback->($status)
		234	aio_link $srcpath, $dstpath, $callback->($status)
		235	aio_symlink $srcpath, $dstpath, $callback->($status)
		236	aio_readlink $pathname, $callback->($link)
		237	aio_realpath $pathname, $callback->($path)
		238	aio_rename $srcpath, $dstpath, $callback->($status)
		239	aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
		240	aio_mkdir $pathname, $mode, $callback->($status)
		241	aio_rmdir $pathname, $callback->($status)
		242	aio_readdir $pathname, $callback->($entries)
		243	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		244	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		245	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		246	aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		247	aio_load $pathname, $data, $callback->($status)
		248	aio_copy $srcpath, $dstpath, $callback->($status)
		249	aio_move $srcpath, $dstpath, $callback->($status)
		250	aio_rmtree $pathname, $callback->($status)
		251	aio_fcntl $fh, $cmd, $arg, $callback->($status)
		252	aio_ioctl $fh, $request, $buf, $callback->($status)
		253	aio_sync $callback->($status)
		254	aio_syncfs $fh, $callback->($status)
		255	aio_fsync $fh, $callback->($status)
		256	aio_fdatasync $fh, $callback->($status)
		257	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		258	aio_pathsync $pathname, $callback->($status)
		259	aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		260	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		261	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		262	aio_mlockall $flags, $callback->($status)
		263	aio_group $callback->(...)
		264	aio_nop $callback->()
		265
		266	$prev_pri = aioreq_pri [$pri]
		267	aioreq_nice $pri_adjust
		268
		269	IO::AIO::poll_wait
		270	IO::AIO::poll_cb
		271	IO::AIO::poll
		272	IO::AIO::flush
		273	IO::AIO::max_poll_reqs $nreqs
		274	IO::AIO::max_poll_time $seconds
		275	IO::AIO::min_parallel $nthreads
		276	IO::AIO::max_parallel $nthreads
		277	IO::AIO::max_idle $nthreads
		278	IO::AIO::idle_timeout $seconds
		279	IO::AIO::max_outstanding $maxreqs
		280	IO::AIO::nreqs
		281	IO::AIO::nready
		282	IO::AIO::npending
		283	$nfd = IO::AIO::get_fdlimit [EXPERIMENTAL]
		284	IO::AIO::min_fdlimit $nfd [EXPERIMENTAL]
		285
		286	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		287	IO::AIO::fadvise $fh, $offset, $len, $advice
		288	IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
		289	IO::AIO::munmap $scalar
		290	IO::AIO::madvise $scalar, $offset, $length, $advice
		291	IO::AIO::mprotect $scalar, $offset, $length, $protect
		292	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		293	IO::AIO::munlockall
		294
		295	=head2 API NOTES
83		296
84	All the C<aio_*> calls are more or less thin wrappers around the syscall	297	All the C<aio_*> calls are more or less thin wrappers around the syscall
85	with the same name (sans C<aio_>). The arguments are similar or identical,	298	with the same name (sans C<aio_>). The arguments are similar or identical,
86	and they all accept an additional (and optional) C<$callback> argument	299	and they all accept an additional (and optional) C<$callback> argument
87	which must be a code reference. This code reference will get called with	300	which must be a code reference. This code reference will be called after
88	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	301	the syscall has been executed in an asynchronous fashion. The results
89	perl, which usually delivers "false") as it's sole argument when the given	302	of the request will be passed as arguments to the callback (and, if an
90	syscall has been executed asynchronously.	303	error occured, in C<$!>) - for most requests the syscall return code (e.g.
		304	most syscalls return C<-1> on error, unlike perl, which usually delivers
		305	"false").
		306
		307	Some requests (such as C<aio_readdir>) pass the actual results and
		308	communicate failures by passing C<undef>.
91		309
92	All functions expecting a filehandle keep a copy of the filehandle	310	All functions expecting a filehandle keep a copy of the filehandle
93	internally until the request has finished.	311	internally until the request has finished.
94		312
		313	All functions return request objects of type L<IO::AIO::REQ> that allow
		314	further manipulation of those requests while they are in-flight.
		315
95	The pathnames you pass to these routines I<must> be absolute and	316	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	317	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	318	current working directory could have changed. Alternatively, you can
98	changed. Alternatively, you can make sure that you never change the	319	make sure that you never change the current working directory anywhere
99	current working directory.	320	in the program and then use relative paths. You can also take advantage
		321	of IO::AIOs working directory abstraction, that lets you specify paths
		322	relative to some previously-opened "working directory object" - see the
		323	description of the C<IO::AIO::WD> class later in this document.
100		324
101	To encode pathnames to byte form, either make sure you either: a)	325	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	326	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	327	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	328	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)	329	effect in the user environment, d) use Glib::filename_from_unicode on
106	use something else.	330	unicode filenames or e) use something else to ensure your scalar has the
		331	correct contents.
		332
		333	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		334	handles correctly whether it is set or not.
		335
		336	=head2 AIO REQUEST FUNCTIONS
107		337
108	=over 4	338	=over 4
109		339
		340	=item $prev_pri = aioreq_pri [$pri]
		341
		342	Returns the priority value that would be used for the next request and, if
		343	C<$pri> is given, sets the priority for the next aio request.
		344
		345	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		346	and C<4>, respectively. Requests with higher priority will be serviced
		347	first.
		348
		349	The priority will be reset to C<0> after each call to one of the C<aio_*>
		350	functions.
		351
		352	Example: open a file with low priority, then read something from it with
		353	higher priority so the read request is serviced before other low priority
		354	open requests (potentially spamming the cache):
		355
		356	aioreq_pri -3;
		357	aio_open ..., sub {
		358	return unless $_[0];
		359
		360	aioreq_pri -2;
		361	aio_read $_[0], ..., sub {
		362	...
		363	};
		364	};
		365
		366
		367	=item aioreq_nice $pri_adjust
		368
		369	Similar to C<aioreq_pri>, but subtracts the given value from the current
		370	priority, so the effect is cumulative.
		371
		372
110	=item aio_open $pathname, $flags, $mode, $callback	373	=item aio_open $pathname, $flags, $mode, $callback->($fh)
111		374
112	Asynchronously open or create a file and call the callback with a newly	375	Asynchronously open or create a file and call the callback with a newly
113	created filehandle for the file.	376	created filehandle for the file (or C<undef> in case of an error).
114		377
115	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	378	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
116	for an explanation.	379	for an explanation.
117		380
118	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a	381	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
119	list. They are the same as used by C<sysopen>.	382	list. They are the same as used by C<sysopen>.
120		383
121	Likewise, C<$mode> specifies the mode of the newly created file, if it	384	Likewise, C<$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>,	385	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,	386	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).	387	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		388	by the umask in effect then the request is being executed, so better never
		389	change the umask.
125		390
126	Example:	391	Example:
127		392
128	aio_open "/etc/passwd", O_RDONLY, 0, sub {	393	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
129	if ($_[0]) {	394	if ($_[0]) {
130	print "open successful, fh is $_[0]\n";	395	print "open successful, fh is $_[0]\n";
131	...	396	...
132	} else {	397	} else {
133	die "open failed: $!\n";	398	die "open failed: $!\n";
134	}	399	}
135	};	400	};
136		401
		402	In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
		403	C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
		404	following POSIX and non-POSIX constants are available (missing ones on
		405	your system are, as usual, C<0>):
		406
		407	C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
		408	C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
		409	C<O_RSYNC>, C<O_SYNC>, C<O_PATH>, C<O_TMPFILE>, and C<O_TTY_INIT>.
		410
		411
137	=item aio_close $fh, $callback	412	=item aio_close $fh, $callback->($status)
138		413
139	Asynchronously close a file and call the callback with the result	414	Asynchronously close a file and call the callback with the result
140	code. I<WARNING:> although accepted, you should not pass in a perl	415	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		416
145	This is supposed to be a bug in the API, so that might change. It's	417	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
146	therefore best to avoid this function.	418	closing the file descriptor associated with the filehandle itself.
147		419
		420	Therefore, C<aio_close> will not close the filehandle - instead it will
		421	use dup2 to overwrite the file descriptor with the write-end of a pipe
		422	(the pipe fd will be created on demand and will be cached).
		423
		424	Or in other words: the file descriptor will be closed, but it will not be
		425	free for reuse until the perl filehandle is closed.
		426
		427	=cut
		428
		429	=item aio_seek $fh, $offset, $whence, $callback->($offs)
		430
		431	Seeks the filehandle to the new C<$offset>, similarly to perl's
		432	C<sysseek>. The C<$whence> can use the traditional values (C<0> for
		433	C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
		434	C<IO::AIO::SEEK_END>).
		435
		436	The resulting absolute offset will be passed to the callback, or C<-1> in
		437	case of an error.
		438
		439	In theory, the C<$whence> constants could be different than the
		440	corresponding values from L<Fcntl>, but perl guarantees they are the same,
		441	so don't panic.
		442
		443	As a GNU/Linux (and maybe Solaris) extension, also the constants
		444	C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
		445	could be found. No guarantees about suitability for use in C<aio_seek> or
		446	Perl's C<sysseek> can be made though, although I would naively assume they
		447	"just work".
		448
148	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	449	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
149		450
150	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	451	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
151		452
152	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	453	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	454	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> and
154	callback without the actual number of bytes read (or -1 on error, just	455	calls the callback with the actual number of bytes transferred (or -1 on
155	like the syscall).	456	error, just like the syscall).
		457
		458	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		459	offset plus the actual number of bytes read.
		460
		461	If C<$offset> is undefined, then the current file descriptor offset will
		462	be used (and updated), otherwise the file descriptor offset will not be
		463	changed by these calls.
		464
		465	If C<$length> is undefined in C<aio_write>, use the remaining length of
		466	C<$data>.
		467
		468	If C<$dataoffset> is less than zero, it will be counted from the end of
		469	C<$data>.
156		470
157	The C<$data> scalar I<MUST NOT> be modified in any way while the request	471	The C<$data> scalar I<MUST NOT> be modified in any way while the request
158	is outstanding. Modifying it can result in segfaults or WW3 (if the	472	is outstanding. Modifying it can result in segfaults or World War III (if
159	necessary/optional hardware is installed).	473	the necessary/optional hardware is installed).
160		474
161	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	475	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
162	offset C<0> within the scalar:	476	offset C<0> within the scalar:
163		477
164	aio_read $fh, 7, 15, $buffer, 0, sub {	478	aio_read $fh, 7, 15, $buffer, 0, sub {
165	$_[0] > 0 or die "read error: $!";	479	$_[0] > 0 or die "read error: $!";
166	print "read $_[0] bytes: <$buffer>\n";	480	print "read $_[0] bytes: <$buffer>\n";
167	};	481	};
168		482
		483
169	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback	484	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
170		485
171	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts	486	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
172	reading at byte offset C<$in_offset>, and starts writing at the current	487	reading at byte offset C<$in_offset>, and starts writing at the current
173	file offset of C<$out_fh>. Because of that, it is not safe to issue more	488	file offset of C<$out_fh>. Because of that, it is not safe to issue more
174	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each	489	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
175	other.	490	other. The same C<$in_fh> works fine though, as this function does not
		491	move or use the file offset of C<$in_fh>.
176		492
		493	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		494	are written, and there is no way to find out how many more bytes have been
		495	read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
		496	number of bytes written to C<$out_fh>. Only if the result value equals
		497	C<$length> one can assume that C<$length> bytes have been read.
		498
		499	Unlike with other C<aio_> functions, it makes a lot of sense to use
		500	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		501	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		502	the socket I/O will be non-blocking. Note, however, that you can run
		503	into a trap where C<aio_sendfile> reads some data with readahead, then
		504	fails to write all data, and when the socket is ready the next time, the
		505	data in the cache is already lost, forcing C<aio_sendfile> to again hit
		506	the disk. Explicit C<aio_read> + C<aio_write> let's you better control
		507	resource usage.
		508
177	This call tries to make use of a native C<sendfile> syscall to provide	509	This call tries to make use of a native C<sendfile>-like syscall to
178	zero-copy operation. For this to work, C<$out_fh> should refer to a	510	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.	511	a socket, and C<$in_fh> should refer to an mmap'able file.
180		512
181	If the native sendfile call fails or is not implemented, it will be	513	If a native sendfile cannot be found or it fails with C<ENOSYS>,
182	emulated, so you can call C<aio_sendfile> on any type of filehandle	514	C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
		515	C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
183	regardless of the limitations of the operating system.	516	type of filehandle regardless of the limitations of the operating system.
184		517
185	Please note, however, that C<aio_sendfile> can read more bytes from	518	As native sendfile syscalls (as practically any non-POSIX interface hacked
186	C<$in_fh> than are written, and there is no way to find out how many	519	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	520	on many systems, this implementation tries to work around some known bugs
188	provides the number of bytes written to C<$out_fh>. Only if the result	521	in Linux and FreeBSD kernels (probably others, too), but that might fail,
189	value equals C<$length> one can assume that C<$length> bytes have been	522	so you really really should check the return value of C<aio_sendfile> -
190	read.	523	fewer bytes than expected might have been transferred.
191		524
		525
192	=item aio_readahead $fh,$offset,$length, $callback	526	=item aio_readahead $fh,$offset,$length, $callback->($retval)
193		527
194	C<aio_readahead> populates the page cache with data from a file so that	528	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<$offset>	529	subsequent reads from that file will not block on disk I/O. The C<$offset>
196	argument specifies the starting point from which data is to be read and	530	argument specifies the starting point from which data is to be read and
197	C<$length> specifies the number of bytes to be read. I/O is performed in	531	C<$length> specifies the number of bytes to be read. I/O is performed in
198	whole pages, so that offset is effectively rounded down to a page boundary	532	whole pages, so that offset is effectively rounded down to a page boundary
199	and bytes are read up to the next page boundary greater than or equal to	533	and bytes are read up to the next page boundary greater than or equal to
200	(off-set+length). C<aio_readahead> does not read beyond the end of the	534	(off-set+length). C<aio_readahead> does not read beyond the end of the
201	file. The current file offset of the file is left unchanged.	535	file. The current file offset of the file is left unchanged.
202		536
203	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	537	If that syscall doesn't exist (likely if your kernel isn't Linux) it will
204	emulated by simply reading the data, which would have a similar effect.	538	be emulated by simply reading the data, which would have a similar effect.
205		539
		540
206	=item aio_stat $fh_or_path, $callback	541	=item aio_stat $fh_or_path, $callback->($status)
207		542
208	=item aio_lstat $fh, $callback	543	=item aio_lstat $fh, $callback->($status)
209		544
210	Works like perl's C<stat> or C<lstat> in void context. The callback will	545	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 _>	546	be called after the stat and the results will be available using C<stat _>
212	or C<-s _> etc...	547	or C<-s _> etc...
213		548
…		…
215	for an explanation.	550	for an explanation.
216		551
217	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	552	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	553	error when stat'ing a large file, the results will be silently truncated
219	unless perl itself is compiled with large file support.	554	unless perl itself is compiled with large file support.
		555
		556	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		557	following constants and functions (if not implemented, the constants will
		558	be C<0> and the functions will either C<croak> or fall back on traditional
		559	behaviour).
		560
		561	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		562	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		563	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
220		564
221	Example: Print the length of F</etc/passwd>:	565	Example: Print the length of F</etc/passwd>:
222		566
223	aio_stat "/etc/passwd", sub {	567	aio_stat "/etc/passwd", sub {
224	$_[0] and die "stat failed: $!";	568	$_[0] and die "stat failed: $!";
225	print "size is ", -s _, "\n";	569	print "size is ", -s _, "\n";
226	};	570	};
227		571
		572
		573	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		574
		575	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		576	whether a file handle or path was passed.
		577
		578	On success, the callback is passed a hash reference with the following
		579	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		580	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		581	is passed.
		582
		583	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		584	C<ST_NOSUID>.
		585
		586	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		587	their correct value when available, or to C<0> on systems that do
		588	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		589	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		590	C<ST_NODIRATIME> and C<ST_RELATIME>.
		591
		592	Example: stat C</wd> and dump out the data if successful.
		593
		594	aio_statvfs "/wd", sub {
		595	my $f = $_[0]
		596	or die "statvfs: $!";
		597
		598	use Data::Dumper;
		599	say Dumper $f;
		600	};
		601
		602	# result:
		603	{
		604	bsize => 1024,
		605	bfree => 4333064312,
		606	blocks => 10253828096,
		607	files => 2050765568,
		608	flag => 4096,
		609	favail => 2042092649,
		610	bavail => 4333064312,
		611	ffree => 2042092649,
		612	namemax => 255,
		613	frsize => 1024,
		614	fsid => 1810
		615	}
		616
		617	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		618
		619	Works like perl's C<utime> function (including the special case of $atime
		620	and $mtime being undef). Fractional times are supported if the underlying
		621	syscalls support them.
		622
		623	When called with a pathname, uses utimes(2) if available, otherwise
		624	utime(2). If called on a file descriptor, uses futimes(2) if available,
		625	otherwise returns ENOSYS, so this is not portable.
		626
		627	Examples:
		628
		629	# set atime and mtime to current time (basically touch(1)):
		630	aio_utime "path", undef, undef;
		631	# set atime to current time and mtime to beginning of the epoch:
		632	aio_utime "path", time, undef; # undef==0
		633
		634
		635	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		636
		637	Works like perl's C<chown> function, except that C<undef> for either $uid
		638	or $gid is being interpreted as "do not change" (but -1 can also be used).
		639
		640	Examples:
		641
		642	# same as "chown root path" in the shell:
		643	aio_chown "path", 0, -1;
		644	# same as above:
		645	aio_chown "path", 0, undef;
		646
		647
		648	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		649
		650	Works like truncate(2) or ftruncate(2).
		651
		652
		653	=item aio_allocate $fh, $mode, $offset, $len, $callback->($status)
		654
		655	Allocates or frees disk space according to the C<$mode> argument. See the
		656	linux C<fallocate> documentation for details.
		657
		658	C<$mode> is usually C<0> or C<IO::AIO::FALLOC_FL_KEEP_SIZE> to allocate
		659	space, or C<IO::AIO::FALLOC_FL_PUNCH_HOLE | IO::AIO::FALLOC_FL_KEEP_SIZE>,
		660	to deallocate a file range.
		661
		662	IO::AIO also supports C<FALLOC_FL_COLLAPSE_RANGE>, to remove a range
		663	(without leaving a hole), C<FALLOC_FL_ZERO_RANGE>, to zero a range,
		664	C<FALLOC_FL_INSERT_RANGE> to insert a range and C<FALLOC_FL_UNSHARE_RANGE>
		665	to unshare shared blocks (see your L<fallocate(2)> manpage).
		666
		667	The file system block size used by C<fallocate> is presumably the
		668	C<f_bsize> returned by C<statvfs>, but different filesystems and filetypes
		669	can dictate other limitations.
		670
		671	If C<fallocate> isn't available or cannot be emulated (currently no
		672	emulation will be attempted), passes C<-1> and sets C<$!> to C<ENOSYS>.
		673
		674
		675	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		676
		677	Works like perl's C<chmod> function.
		678
		679
228	=item aio_unlink $pathname, $callback	680	=item aio_unlink $pathname, $callback->($status)
229		681
230	Asynchronously unlink (delete) a file and call the callback with the	682	Asynchronously unlink (delete) a file and call the callback with the
231	result code.	683	result code.
232		684
		685
		686	=item aio_mknod $pathname, $mode, $dev, $callback->($status)
		687
		688	[EXPERIMENTAL]
		689
		690	Asynchronously create a device node (or fifo). See mknod(2).
		691
		692	The only (POSIX-) portable way of calling this function is:
		693
		694	aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		695
		696	See C<aio_stat> for info about some potentially helpful extra constants
		697	and functions.
		698
		699	=item aio_link $srcpath, $dstpath, $callback->($status)
		700
		701	Asynchronously create a new link to the existing object at C<$srcpath> at
		702	the path C<$dstpath> and call the callback with the result code.
		703
		704
		705	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		706
		707	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		708	the path C<$dstpath> and call the callback with the result code.
		709
		710
		711	=item aio_readlink $pathname, $callback->($link)
		712
		713	Asynchronously read the symlink specified by C<$path> and pass it to
		714	the callback. If an error occurs, nothing or undef gets passed to the
		715	callback.
		716
		717
		718	=item aio_realpath $pathname, $callback->($path)
		719
		720	Asynchronously make the path absolute and resolve any symlinks in
		721	C<$path>. The resulting path only consists of directories (same as
		722	L<Cwd::realpath>).
		723
		724	This request can be used to get the absolute path of the current working
		725	directory by passing it a path of F<.> (a single dot).
		726
		727
		728	=item aio_rename $srcpath, $dstpath, $callback->($status)
		729
		730	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		731	rename(2) and call the callback with the result code.
		732
		733	On systems that support the AIO::WD working directory abstraction
		734	natively, the case C<[$wd, "."]> as C<$srcpath> is specialcased - instead
		735	of failing, C<rename> is called on the absolute path of C<$wd>.
		736
		737
		738	=item aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
		739
		740	Basically a version of C<aio_rename> with an additional C<$flags>
		741	argument. Calling this with C<$flags=0> is the same as calling
		742	C<aio_rename>.
		743
		744	Non-zero flags are currently only supported on GNU/Linux systems that
		745	support renameat2. Other systems fail with C<ENOSYS> in this case.
		746
		747	The following constants are available (missing ones are, as usual C<0>),
		748	see renameat2(2) for details:
		749
		750	C<IO::AIO::RENAME_NOREPLACE>, C<IO::AIO::RENAME_EXCHANGE>
		751	and C<IO::AIO::RENAME_WHITEOUT>.
		752
		753
		754	=item aio_mkdir $pathname, $mode, $callback->($status)
		755
		756	Asynchronously mkdir (create) a directory and call the callback with
		757	the result code. C<$mode> will be modified by the umask at the time the
		758	request is executed, so do not change your umask.
		759
		760
233	=item aio_rmdir $pathname, $callback	761	=item aio_rmdir $pathname, $callback->($status)
234		762
235	Asynchronously rmdir (delete) a directory and call the callback with the	763	Asynchronously rmdir (delete) a directory and call the callback with the
236	result code.	764	result code.
237		765
		766	On systems that support the AIO::WD working directory abstraction
		767	natively, the case C<[$wd, "."]> is specialcased - instead of failing,
		768	C<rmdir> is called on the absolute path of C<$wd>.
		769
		770
238	=item aio_readdir $pathname $callback	771	=item aio_readdir $pathname, $callback->($entries)
239		772
240	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire	773	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	774	directory (i.e. opendir + readdir + closedir). The entries will not be
242	sorted, and will B<NOT> include the C<.> and C<..> entries.	775	sorted, and will B<NOT> include the C<.> and C<..> entries.
243		776
244	The callback a single argument which is either C<undef> or an array-ref	777	The callback is passed a single argument which is either C<undef> or an
245	with the filenames.	778	array-ref with the filenames.
246		779
		780
		781	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		782
		783	Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
		784	tune behaviour and output format. In case of an error, C<$entries> will be
		785	C<undef>.
		786
		787	The flags are a combination of the following constants, ORed together (the
		788	flags will also be passed to the callback, possibly modified):
		789
		790	=over 4
		791
		792	=item IO::AIO::READDIR_DENTS
		793
		794	Normally the callback gets an arrayref consisting of names only (as
		795	with C<aio_readdir>). If this flag is set, then the callback gets an
		796	arrayref with C<[$name, $type, $inode]> arrayrefs, each describing a
		797	single directory entry in more detail:
		798
		799	C<$name> is the name of the entry.
		800
		801	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		802
		803	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		804	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		805	C<IO::AIO::DT_WHT>.
		806
		807	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need
		808	to know, you have to run stat yourself. Also, for speed/memory reasons,
		809	the C<$type> scalars are read-only: you must not modify them.
		810
		811	C<$inode> is the inode number (which might not be exact on systems with 64
		812	bit inode numbers and 32 bit perls). This field has unspecified content on
		813	systems that do not deliver the inode information.
		814
		815	=item IO::AIO::READDIR_DIRS_FIRST
		816
		817	When this flag is set, then the names will be returned in an order where
		818	likely directories come first, in optimal stat order. This is useful when
		819	you need to quickly find directories, or you want to find all directories
		820	while avoiding to stat() each entry.
		821
		822	If the system returns type information in readdir, then this is used
		823	to find directories directly. Otherwise, likely directories are names
		824	beginning with ".", or otherwise names with no dots, of which names with
		825	short names are tried first.
		826
		827	=item IO::AIO::READDIR_STAT_ORDER
		828
		829	When this flag is set, then the names will be returned in an order
		830	suitable for stat()'ing each one. That is, when you plan to stat() most or
		831	all files in the given directory, then the returned order will likely be
		832	faster.
		833
		834	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified,
		835	then the likely dirs come first, resulting in a less optimal stat order
		836	for stat'ing all entries, but likely a more optimal order for finding
		837	subdirectories.
		838
		839	=item IO::AIO::READDIR_FOUND_UNKNOWN
		840
		841	This flag should not be set when calling C<aio_readdirx>. Instead, it
		842	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		843	C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
		844	C<$type>'s are known, which can be used to speed up some algorithms.
		845
		846	=back
		847
		848
		849	=item aio_slurp $pathname, $offset, $length, $data, $callback->($status)
		850
		851	Opens, reads and closes the given file. The data is put into C<$data>,
		852	which is resized as required.
		853
		854	If C<$offset> is negative, then it is counted from the end of the file.
		855
		856	If C<$length> is zero, then the remaining length of the file is
		857	used. Also, in this case, the same limitations to modifying C<$data> apply
		858	as when IO::AIO::mmap is used, i.e. it must only be modified in-place
		859	with C<substr>. If the size of the file is known, specifying a non-zero
		860	C<$length> results in a performance advantage.
		861
		862	This request is similar to the older C<aio_load> request, but since it is
		863	a single request, it might be more efficient to use.
		864
		865	Example: load F</etc/passwd> into C<$passwd>.
		866
		867	my $passwd;
		868	aio_slurp "/etc/passwd", 0, 0, $passwd, sub {
		869	$_[0] >= 0
		870	or die "/etc/passwd: $!\n";
		871
		872	printf "/etc/passwd is %d bytes long, and contains:\n", length $passwd;
		873	print $passwd;
		874	};
		875	IO::AIO::flush;
		876
		877
		878	=item aio_load $pathname, $data, $callback->($status)
		879
		880	This is a composite request that tries to fully load the given file into
		881	memory. Status is the same as with aio_read.
		882
		883	Using C<aio_slurp> might be more efficient, as it is a single request.
		884
		885	=cut
		886
		887	sub aio_load($$;$) {
		888	my ($path, undef, $cb) = @_;
		889	my $data = \$_[1];
		890
		891	my $pri = aioreq_pri;
		892	my $grp = aio_group $cb;
		893
		894	aioreq_pri $pri;
		895	add $grp aio_open $path, O_RDONLY, 0, sub {
		896	my $fh = shift
		897	or return $grp->result (-1);
		898
		899	aioreq_pri $pri;
		900	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		901	$grp->result ($_[0]);
		902	};
		903	};
		904
		905	$grp
		906	}
		907
		908	=item aio_copy $srcpath, $dstpath, $callback->($status)
		909
		910	Try to copy the I<file> (directories not supported as either source or
		911	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		912	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		913
		914	Existing destination files will be truncated.
		915
		916	This is a composite request that creates the destination file with
		917	mode 0200 and copies the contents of the source file into it using
		918	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		919	uid/gid, in that order.
		920
		921	If an error occurs, the partial destination file will be unlinked, if
		922	possible, except when setting atime, mtime, access mode and uid/gid, where
		923	errors are being ignored.
		924
		925	=cut
		926
		927	sub aio_copy($$;$) {
		928	my ($src, $dst, $cb) = @_;
		929
		930	my $pri = aioreq_pri;
		931	my $grp = aio_group $cb;
		932
		933	aioreq_pri $pri;
		934	add $grp aio_open $src, O_RDONLY, 0, sub {
		935	if (my $src_fh = $_[0]) {
		936	my @stat = stat $src_fh; # hmm, might block over nfs?
		937
		938	aioreq_pri $pri;
		939	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		940	if (my $dst_fh = $_[0]) {
		941	aioreq_pri $pri;
		942	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		943	if ($_[0] == $stat[7]) {
		944	$grp->result (0);
		945	close $src_fh;
		946
		947	my $ch = sub {
		948	aioreq_pri $pri;
		949	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		950	aioreq_pri $pri;
		951	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		952	aioreq_pri $pri;
		953	add $grp aio_close $dst_fh;
		954	}
		955	};
		956	};
		957
		958	aioreq_pri $pri;
		959	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		960	if ($_[0] < 0 && $! == ENOSYS) {
		961	aioreq_pri $pri;
		962	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		963	} else {
		964	$ch->();
		965	}
		966	};
		967	} else {
		968	$grp->result (-1);
		969	close $src_fh;
		970	close $dst_fh;
		971
		972	aioreq $pri;
		973	add $grp aio_unlink $dst;
		974	}
		975	};
		976	} else {
		977	$grp->result (-1);
		978	}
		979	},
		980
		981	} else {
		982	$grp->result (-1);
		983	}
		984	};
		985
		986	$grp
		987	}
		988
		989	=item aio_move $srcpath, $dstpath, $callback->($status)
		990
		991	Try to move the I<file> (directories not supported as either source or
		992	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		993	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		994
		995	This is a composite request that tries to rename(2) the file first; if
		996	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		997	that is successful, unlinks the C<$srcpath>.
		998
		999	=cut
		1000
		1001	sub aio_move($$;$) {
		1002	my ($src, $dst, $cb) = @_;
		1003
		1004	my $pri = aioreq_pri;
		1005	my $grp = aio_group $cb;
		1006
		1007	aioreq_pri $pri;
		1008	add $grp aio_rename $src, $dst, sub {
		1009	if ($_[0] && $! == EXDEV) {
		1010	aioreq_pri $pri;
		1011	add $grp aio_copy $src, $dst, sub {
		1012	$grp->result ($_[0]);
		1013
		1014	unless ($_[0]) {
		1015	aioreq_pri $pri;
		1016	add $grp aio_unlink $src;
		1017	}
		1018	};
		1019	} else {
		1020	$grp->result ($_[0]);
		1021	}
		1022	};
		1023
		1024	$grp
		1025	}
		1026
		1027	=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		1028
		1029	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		1030	efficiently separate the entries of directory C<$path> into two sets of
		1031	names, directories you can recurse into (directories), and ones you cannot
		1032	recurse into (everything else, including symlinks to directories).
		1033
		1034	C<aio_scandir> is a composite request that generates many sub requests.
		1035	C<$maxreq> specifies the maximum number of outstanding aio requests that
		1036	this function generates. If it is C<< <= 0 >>, then a suitable default
		1037	will be chosen (currently 4).
		1038
		1039	On error, the callback is called without arguments, otherwise it receives
		1040	two array-refs with path-relative entry names.
		1041
		1042	Example:
		1043
		1044	aio_scandir $dir, 0, sub {
		1045	my ($dirs, $nondirs) = @_;
		1046	print "real directories: @$dirs\n";
		1047	print "everything else: @$nondirs\n";
		1048	};
		1049
		1050	Implementation notes.
		1051
		1052	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		1053
		1054	If readdir returns file type information, then this is used directly to
		1055	find directories.
		1056
		1057	Otherwise, after reading the directory, the modification time, size etc.
		1058	of the directory before and after the readdir is checked, and if they
		1059	match (and isn't the current time), the link count will be used to decide
		1060	how many entries are directories (if >= 2). Otherwise, no knowledge of the
		1061	number of subdirectories will be assumed.
		1062
		1063	Then entries will be sorted into likely directories a non-initial dot
		1064	currently) and likely non-directories (see C<aio_readdirx>). Then every
		1065	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
		1066	in order of their inode numbers. If that succeeds, it assumes that the
		1067	entry is a directory or a symlink to directory (which will be checked
		1068	separately). This is often faster than stat'ing the entry itself because
		1069	filesystems might detect the type of the entry without reading the inode
		1070	data (e.g. ext2fs filetype feature), even on systems that cannot return
		1071	the filetype information on readdir.
		1072
		1073	If the known number of directories (link count - 2) has been reached, the
		1074	rest of the entries is assumed to be non-directories.
		1075
		1076	This only works with certainty on POSIX (= UNIX) filesystems, which
		1077	fortunately are the vast majority of filesystems around.
		1078
		1079	It will also likely work on non-POSIX filesystems with reduced efficiency
		1080	as those tend to return 0 or 1 as link counts, which disables the
		1081	directory counting heuristic.
		1082
		1083	=cut
		1084
		1085	sub aio_scandir($$;$) {
		1086	my ($path, $maxreq, $cb) = @_;
		1087
		1088	my $pri = aioreq_pri;
		1089
		1090	my $grp = aio_group $cb;
		1091
		1092	$maxreq = 4 if $maxreq <= 0;
		1093
		1094	# get a wd object
		1095	aioreq_pri $pri;
		1096	add $grp aio_wd $path, sub {
		1097	$_[0]
		1098	or return $grp->result ();
		1099
		1100	my $wd = [shift, "."];
		1101
		1102	# stat once
		1103	aioreq_pri $pri;
		1104	add $grp aio_stat $wd, sub {
		1105	return $grp->result () if $_[0];
		1106	my $now = time;
		1107	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		1108
		1109	# read the directory entries
		1110	aioreq_pri $pri;
		1111	add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
		1112	my $entries = shift
		1113	or return $grp->result ();
		1114
		1115	# stat the dir another time
		1116	aioreq_pri $pri;
		1117	add $grp aio_stat $wd, sub {
		1118	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		1119
		1120	my $ndirs;
		1121
		1122	# take the slow route if anything looks fishy
		1123	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		1124	$ndirs = -1;
		1125	} else {
		1126	# if nlink == 2, we are finished
		1127	# for non-posix-fs's, we rely on nlink < 2
		1128	$ndirs = (stat _)[3] - 2
		1129	or return $grp->result ([], $entries);
		1130	}
		1131
		1132	my (@dirs, @nondirs);
		1133
		1134	my $statgrp = add $grp aio_group sub {
		1135	$grp->result (\@dirs, \@nondirs);
		1136	};
		1137
		1138	limit $statgrp $maxreq;
		1139	feed $statgrp sub {
		1140	return unless @$entries;
		1141	my $entry = shift @$entries;
		1142
		1143	aioreq_pri $pri;
		1144	$wd->[1] = "$entry/.";
		1145	add $statgrp aio_stat $wd, sub {
		1146	if ($_[0] < 0) {
		1147	push @nondirs, $entry;
		1148	} else {
		1149	# need to check for real directory
		1150	aioreq_pri $pri;
		1151	$wd->[1] = $entry;
		1152	add $statgrp aio_lstat $wd, sub {
		1153	if (-d _) {
		1154	push @dirs, $entry;
		1155
		1156	unless (--$ndirs) {
		1157	push @nondirs, @$entries;
		1158	feed $statgrp;
		1159	}
		1160	} else {
		1161	push @nondirs, $entry;
		1162	}
		1163	}
		1164	}
		1165	};
		1166	};
		1167	};
		1168	};
		1169	};
		1170	};
		1171
		1172	$grp
		1173	}
		1174
		1175	=item aio_rmtree $pathname, $callback->($status)
		1176
		1177	Delete a directory tree starting (and including) C<$path>, return the
		1178	status of the final C<rmdir> only. This is a composite request that
		1179	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1180	everything else.
		1181
		1182	=cut
		1183
		1184	sub aio_rmtree;
		1185	sub aio_rmtree($;$) {
		1186	my ($path, $cb) = @_;
		1187
		1188	my $pri = aioreq_pri;
		1189	my $grp = aio_group $cb;
		1190
		1191	aioreq_pri $pri;
		1192	add $grp aio_scandir $path, 0, sub {
		1193	my ($dirs, $nondirs) = @_;
		1194
		1195	my $dirgrp = aio_group sub {
		1196	add $grp aio_rmdir $path, sub {
		1197	$grp->result ($_[0]);
		1198	};
		1199	};
		1200
		1201	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1202	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1203
		1204	add $grp $dirgrp;
		1205	};
		1206
		1207	$grp
		1208	}
		1209
		1210	=item aio_fcntl $fh, $cmd, $arg, $callback->($status)
		1211
		1212	=item aio_ioctl $fh, $request, $buf, $callback->($status)
		1213
		1214	These work just like the C<fcntl> and C<ioctl> built-in functions, except
		1215	they execute asynchronously and pass the return value to the callback.
		1216
		1217	Both calls can be used for a lot of things, some of which make more sense
		1218	to run asynchronously in their own thread, while some others make less
		1219	sense. For example, calls that block waiting for external events, such
		1220	as locking, will also lock down an I/O thread while it is waiting, which
		1221	can deadlock the whole I/O system. At the same time, there might be no
		1222	alternative to using a thread to wait.
		1223
		1224	So in general, you should only use these calls for things that do
		1225	(filesystem) I/O, not for things that wait for other events (network,
		1226	other processes), although if you are careful and know what you are doing,
		1227	you still can.
		1228
		1229	The following constants are available (missing ones are, as usual C<0>):
		1230
		1231	C<F_DUPFD_CLOEXEC>,
		1232
		1233	C<F_OFD_GETLK>, C<F_OFD_SETLK>, C<F_OFD_GETLKW>,
		1234
		1235	C<FIFREEZE>, C<FITHAW>, C<FITRIM>, C<FICLONE>, C<FICLONERANGE>, C<FIDEDUPERANGE>.
		1236
		1237	C<FS_IOC_GETFLAGS>, C<FS_IOC_SETFLAGS>, C<FS_IOC_GETVERSION>, C<FS_IOC_SETVERSION>,
		1238	C<FS_IOC_FIEMAP>.
		1239
		1240	C<FS_IOC_FSGETXATTR>, C<FS_IOC_FSSETXATTR>, C<FS_IOC_SET_ENCRYPTION_POLICY>,
		1241	C<FS_IOC_GET_ENCRYPTION_PWSALT>, C<FS_IOC_GET_ENCRYPTION_POLICY>, C<FS_KEY_DESCRIPTOR_SIZE>.
		1242
		1243	C<FS_SECRM_FL>, C<FS_UNRM_FL>, C<FS_COMPR_FL>, C<FS_SYNC_FL>, C<FS_IMMUTABLE_FL>,
		1244	C<FS_APPEND_FL>, C<FS_NODUMP_FL>, C<FS_NOATIME_FL>, C<FS_DIRTY_FL>,
		1245	C<FS_COMPRBLK_FL>, C<FS_NOCOMP_FL>, C<FS_ENCRYPT_FL>, C<FS_BTREE_FL>,
		1246	C<FS_INDEX_FL>, C<FS_JOURNAL_DATA_FL>, C<FS_NOTAIL_FL>, C<FS_DIRSYNC_FL>, C<FS_TOPDIR_FL>,
		1247	C<FS_FL_USER_MODIFIABLE>.
		1248
		1249	C<FS_XFLAG_REALTIME>, C<FS_XFLAG_PREALLOC>, C<FS_XFLAG_IMMUTABLE>, C<FS_XFLAG_APPEND>,
		1250	C<FS_XFLAG_SYNC>, C<FS_XFLAG_NOATIME>, C<FS_XFLAG_NODUMP>, C<FS_XFLAG_RTINHERIT>,
		1251	C<FS_XFLAG_PROJINHERIT>, C<FS_XFLAG_NOSYMLINKS>, C<FS_XFLAG_EXTSIZE>, C<FS_XFLAG_EXTSZINHERIT>,
		1252	C<FS_XFLAG_NODEFRAG>, C<FS_XFLAG_FILESTREAM>, C<FS_XFLAG_DAX>, C<FS_XFLAG_HASATTR>,
		1253
		1254	=item aio_sync $callback->($status)
		1255
		1256	Asynchronously call sync and call the callback when finished.
		1257
247	=item aio_fsync $fh, $callback	1258	=item aio_fsync $fh, $callback->($status)
248		1259
249	Asynchronously call fsync on the given filehandle and call the callback	1260	Asynchronously call fsync on the given filehandle and call the callback
250	with the fsync result code.	1261	with the fsync result code.
251		1262
252	=item aio_fdatasync $fh, $callback	1263	=item aio_fdatasync $fh, $callback->($status)
253		1264
254	Asynchronously call fdatasync on the given filehandle and call the	1265	Asynchronously call fdatasync on the given filehandle and call the
255	callback with the fdatasync result code.	1266	callback with the fdatasync result code.
256		1267
257	If this call isn't available because your OS lacks it or it couldn't be	1268	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.	1269	detected, it will be emulated by calling C<fsync> instead.
259		1270
		1271	=item aio_syncfs $fh, $callback->($status)
		1272
		1273	Asynchronously call the syncfs syscall to sync the filesystem associated
		1274	to the given filehandle and call the callback with the syncfs result
		1275	code. If syncfs is not available, calls sync(), but returns C<-1> and sets
		1276	errno to C<ENOSYS> nevertheless.
		1277
		1278	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1279
		1280	Sync the data portion of the file specified by C<$offset> and C<$length>
		1281	to disk (but NOT the metadata), by calling the Linux-specific
		1282	sync_file_range call. If sync_file_range is not available or it returns
		1283	ENOSYS, then fdatasync or fsync is being substituted.
		1284
		1285	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1286	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1287	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1288	manpage for details.
		1289
		1290	=item aio_pathsync $pathname, $callback->($status)
		1291
		1292	This request tries to open, fsync and close the given path. This is a
		1293	composite request intended to sync directories after directory operations
		1294	(E.g. rename). This might not work on all operating systems or have any
		1295	specific effect, but usually it makes sure that directory changes get
		1296	written to disc. It works for anything that can be opened for read-only,
		1297	not just directories.
		1298
		1299	Future versions of this function might fall back to other methods when
		1300	C<fsync> on the directory fails (such as calling C<sync>).
		1301
		1302	Passes C<0> when everything went ok, and C<-1> on error.
		1303
		1304	=cut
		1305
		1306	sub aio_pathsync($;$) {
		1307	my ($path, $cb) = @_;
		1308
		1309	my $pri = aioreq_pri;
		1310	my $grp = aio_group $cb;
		1311
		1312	aioreq_pri $pri;
		1313	add $grp aio_open $path, O_RDONLY, 0, sub {
		1314	my ($fh) = @_;
		1315	if ($fh) {
		1316	aioreq_pri $pri;
		1317	add $grp aio_fsync $fh, sub {
		1318	$grp->result ($_[0]);
		1319
		1320	aioreq_pri $pri;
		1321	add $grp aio_close $fh;
		1322	};
		1323	} else {
		1324	$grp->result (-1);
		1325	}
		1326	};
		1327
		1328	$grp
		1329	}
		1330
		1331	=item aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		1332
		1333	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1334	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1335	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1336	scalar must only be modified in-place while an aio operation is pending on
		1337	it).
		1338
		1339	It calls the C<msync> function of your OS, if available, with the memory
		1340	area starting at C<$offset> in the string and ending C<$length> bytes
		1341	later. If C<$length> is negative, counts from the end, and if C<$length>
		1342	is C<undef>, then it goes till the end of the string. The flags can be
		1343	either C<IO::AIO::MS_ASYNC> or C<IO::AIO::MS_SYNC>, plus an optional
		1344	C<IO::AIO::MS_INVALIDATE>.
		1345
		1346	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1347
		1348	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1349	scalars.
		1350
		1351	It touches (reads or writes) all memory pages in the specified
		1352	range inside the scalar. All caveats and parameters are the same
		1353	as for C<aio_msync>, above, except for flags, which must be either
		1354	C<0> (which reads all pages and ensures they are instantiated) or
		1355	C<IO::AIO::MT_MODIFY>, which modifies the memory pages (by reading and
		1356	writing an octet from it, which dirties the page).
		1357
		1358	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1359
		1360	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1361	scalars.
		1362
		1363	It reads in all the pages of the underlying storage into memory (if any)
		1364	and locks them, so they are not getting swapped/paged out or removed.
		1365
		1366	If C<$length> is undefined, then the scalar will be locked till the end.
		1367
		1368	On systems that do not implement C<mlock>, this function returns C<-1>
		1369	and sets errno to C<ENOSYS>.
		1370
		1371	Note that the corresponding C<munlock> is synchronous and is
		1372	documented under L<MISCELLANEOUS FUNCTIONS>.
		1373
		1374	Example: open a file, mmap and mlock it - both will be undone when
		1375	C<$data> gets destroyed.
		1376
		1377	open my $fh, "<", $path or die "$path: $!";
		1378	my $data;
		1379	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1380	aio_mlock $data; # mlock in background
		1381
		1382	=item aio_mlockall $flags, $callback->($status)
		1383
		1384	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1385	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1386
		1387	On systems that do not implement C<mlockall>, this function returns C<-1>
		1388	and sets errno to C<ENOSYS>.
		1389
		1390	Note that the corresponding C<munlockall> is synchronous and is
		1391	documented under L<MISCELLANEOUS FUNCTIONS>.
		1392
		1393	Example: asynchronously lock all current and future pages into memory.
		1394
		1395	aio_mlockall IO::AIO::MCL_FUTURE;
		1396
		1397	=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		1398
		1399	Queries the extents of the given file (by calling the Linux C<FIEMAP>
		1400	ioctl, see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If
		1401	the ioctl is not available on your OS, then this request will fail with
		1402	C<ENOSYS>.
		1403
		1404	C<$start> is the starting offset to query extents for, C<$length> is the
		1405	size of the range to query - if it is C<undef>, then the whole file will
		1406	be queried.
		1407
		1408	C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
		1409	C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
		1410	exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
		1411	the data portion.
		1412
		1413	C<$count> is the maximum number of extent records to return. If it is
		1414	C<undef>, then IO::AIO queries all extents of the range. As a very special
		1415	case, if it is C<0>, then the callback receives the number of extents
		1416	instead of the extents themselves (which is unreliable, see below).
		1417
		1418	If an error occurs, the callback receives no arguments. The special
		1419	C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
		1420
		1421	Otherwise, the callback receives an array reference with extent
		1422	structures. Each extent structure is an array reference itself, with the
		1423	following members:
		1424
		1425	[$logical, $physical, $length, $flags]
		1426
		1427	Flags is any combination of the following flag values (typically either C<0>
		1428	or C<IO::AIO::FIEMAP_EXTENT_LAST> (1)):
		1429
		1430	C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
		1431	C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
		1432	C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
		1433	C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
		1434	C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
		1435	C<IO::AIO::FIEMAP_EXTENT_SHARED>.
		1436
		1437	At the time of this writing (Linux 3.2), this request is unreliable unless
		1438	C<$count> is C<undef>, as the kernel has all sorts of bugs preventing
		1439	it to return all extents of a range for files with a large number of
		1440	extents. The code (only) works around all these issues if C<$count> is
		1441	C<undef>.
		1442
		1443	=item aio_group $callback->(...)
		1444
		1445	This is a very special aio request: Instead of doing something, it is a
		1446	container for other aio requests, which is useful if you want to bundle
		1447	many requests into a single, composite, request with a definite callback
		1448	and the ability to cancel the whole request with its subrequests.
		1449
		1450	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		1451	for more info.
		1452
		1453	Example:
		1454
		1455	my $grp = aio_group sub {
		1456	print "all stats done\n";
		1457	};
		1458
		1459	add $grp
		1460	(aio_stat ...),
		1461	(aio_stat ...),
		1462	...;
		1463
		1464	=item aio_nop $callback->()
		1465
		1466	This is a special request - it does nothing in itself and is only used for
		1467	side effects, such as when you want to add a dummy request to a group so
		1468	that finishing the requests in the group depends on executing the given
		1469	code.
		1470
		1471	While this request does nothing, it still goes through the execution
		1472	phase and still requires a worker thread. Thus, the callback will not
		1473	be executed immediately but only after other requests in the queue have
		1474	entered their execution phase. This can be used to measure request
		1475	latency.
		1476
		1477	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		1478
		1479	Mainly used for debugging and benchmarking, this aio request puts one of
		1480	the request workers to sleep for the given time.
		1481
		1482	While it is theoretically handy to have simple I/O scheduling requests
		1483	like sleep and file handle readable/writable, the overhead this creates is
		1484	immense (it blocks a thread for a long time) so do not use this function
		1485	except to put your application under artificial I/O pressure.
		1486
260	=back	1487	=back
261		1488
		1489
		1490	=head2 IO::AIO::WD - multiple working directories
		1491
		1492	Your process only has one current working directory, which is used by all
		1493	threads. This makes it hard to use relative paths (some other component
		1494	could call C<chdir> at any time, and it is hard to control when the path
		1495	will be used by IO::AIO).
		1496
		1497	One solution for this is to always use absolute paths. This usually works,
		1498	but can be quite slow (the kernel has to walk the whole path on every
		1499	access), and can also be a hassle to implement.
		1500
		1501	Newer POSIX systems have a number of functions (openat, fdopendir,
		1502	futimensat and so on) that make it possible to specify working directories
		1503	per operation.
		1504
		1505	For portability, and because the clowns who "designed", or shall I write,
		1506	perpetrated this new interface were obviously half-drunk, this abstraction
		1507	cannot be perfect, though.
		1508
		1509	IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
		1510	object. This object stores the canonicalised, absolute version of the
		1511	path, and on systems that allow it, also a directory file descriptor.
		1512
		1513	Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
		1514	or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
		1515	object and a pathname instead (or the IO::AIO::WD object alone, which
		1516	gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
		1517	IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
		1518	to that IO::AIO::WD object.
		1519
		1520	For example, to get a wd object for F</etc> and then stat F<passwd>
		1521	inside, you would write:
		1522
		1523	aio_wd "/etc", sub {
		1524	my $etcdir = shift;
		1525
		1526	# although $etcdir can be undef on error, there is generally no reason
		1527	# to check for errors here, as aio_stat will fail with ENOENT
		1528	# when $etcdir is undef.
		1529
		1530	aio_stat [$etcdir, "passwd"], sub {
		1531	# yay
		1532	};
		1533	};
		1534
		1535	The fact that C<aio_wd> is a request and not a normal function shows that
		1536	creating an IO::AIO::WD object is itself a potentially blocking operation,
		1537	which is why it is done asynchronously.
		1538
		1539	To stat the directory obtained with C<aio_wd> above, one could write
		1540	either of the following three request calls:
		1541
		1542	aio_lstat "/etc" , sub { ... # pathname as normal string
		1543	aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
		1544	aio_lstat $wd , sub { ... # shorthand for the previous
		1545
		1546	As with normal pathnames, IO::AIO keeps a copy of the working directory
		1547	object and the pathname string, so you could write the following without
		1548	causing any issues due to C<$path> getting reused:
		1549
		1550	my $path = [$wd, undef];
		1551
		1552	for my $name (qw(abc def ghi)) {
		1553	$path->[1] = $name;
		1554	aio_stat $path, sub {
		1555	# ...
		1556	};
		1557	}
		1558
		1559	There are some caveats: when directories get renamed (or deleted), the
		1560	pathname string doesn't change, so will point to the new directory (or
		1561	nowhere at all), while the directory fd, if available on the system,
		1562	will still point to the original directory. Most functions accepting a
		1563	pathname will use the directory fd on newer systems, and the string on
		1564	older systems. Some functions (such as C<aio_realpath>) will always rely on
		1565	the string form of the pathname.
		1566
		1567	So this functionality is mainly useful to get some protection against
		1568	C<chdir>, to easily get an absolute path out of a relative path for future
		1569	reference, and to speed up doing many operations in the same directory
		1570	(e.g. when stat'ing all files in a directory).
		1571
		1572	The following functions implement this working directory abstraction:
		1573
		1574	=over 4
		1575
		1576	=item aio_wd $pathname, $callback->($wd)
		1577
		1578	Asynchonously canonicalise the given pathname and convert it to an
		1579	IO::AIO::WD object representing it. If possible and supported on the
		1580	system, also open a directory fd to speed up pathname resolution relative
		1581	to this working directory.
		1582
		1583	If something goes wrong, then C<undef> is passwd to the callback instead
		1584	of a working directory object and C<$!> is set appropriately. Since
		1585	passing C<undef> as working directory component of a pathname fails the
		1586	request with C<ENOENT>, there is often no need for error checking in the
		1587	C<aio_wd> callback, as future requests using the value will fail in the
		1588	expected way.
		1589
		1590	=item IO::AIO::CWD
		1591
		1592	This is a compiletime constant (object) that represents the process
		1593	current working directory.
		1594
		1595	Specifying this object as working directory object for a pathname is as if
		1596	the pathname would be specified directly, without a directory object. For
		1597	example, these calls are functionally identical:
		1598
		1599	aio_stat "somefile", sub { ... };
		1600	aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
		1601
		1602	=back
		1603
		1604	To recover the path associated with an IO::AIO::WD object, you can use
		1605	C<aio_realpath>:
		1606
		1607	aio_realpath $wd, sub {
		1608	warn "path is $_[0]\n";
		1609	};
		1610
		1611	Currently, C<aio_statvfs> always, and C<aio_rename> and C<aio_rmdir>
		1612	sometimes, fall back to using an absolue path.
		1613
		1614	=head2 IO::AIO::REQ CLASS
		1615
		1616	All non-aggregate C<aio_*> functions return an object of this class when
		1617	called in non-void context.
		1618
		1619	=over 4
		1620
		1621	=item cancel $req
		1622
		1623	Cancels the request, if possible. Has the effect of skipping execution
		1624	when entering the B<execute> state and skipping calling the callback when
		1625	entering the the B<result> state, but will leave the request otherwise
		1626	untouched (with the exception of readdir). That means that requests that
		1627	currently execute will not be stopped and resources held by the request
		1628	will not be freed prematurely.
		1629
		1630	=item cb $req $callback->(...)
		1631
		1632	Replace (or simply set) the callback registered to the request.
		1633
		1634	=back
		1635
		1636	=head2 IO::AIO::GRP CLASS
		1637
		1638	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		1639	objects of this class, too.
		1640
		1641	A IO::AIO::GRP object is a special request that can contain multiple other
		1642	aio requests.
		1643
		1644	You create one by calling the C<aio_group> constructing function with a
		1645	callback that will be called when all contained requests have entered the
		1646	C<done> state:
		1647
		1648	my $grp = aio_group sub {
		1649	print "all requests are done\n";
		1650	};
		1651
		1652	You add requests by calling the C<add> method with one or more
		1653	C<IO::AIO::REQ> objects:
		1654
		1655	$grp->add (aio_unlink "...");
		1656
		1657	add $grp aio_stat "...", sub {
		1658	$_[0] or return $grp->result ("error");
		1659
		1660	# add another request dynamically, if first succeeded
		1661	add $grp aio_open "...", sub {
		1662	$grp->result ("ok");
		1663	};
		1664	};
		1665
		1666	This makes it very easy to create composite requests (see the source of
		1667	C<aio_move> for an application) that work and feel like simple requests.
		1668
		1669	=over 4
		1670
		1671	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		1672	C<IO::AIO::poll_cb>, just like any other request.
		1673
		1674	=item * They can be canceled like any other request. Canceling will cancel not
		1675	only the request itself, but also all requests it contains.
		1676
		1677	=item * They can also can also be added to other IO::AIO::GRP objects.
		1678
		1679	=item * You must not add requests to a group from within the group callback (or
		1680	any later time).
		1681
		1682	=back
		1683
		1684	Their lifetime, simplified, looks like this: when they are empty, they
		1685	will finish very quickly. If they contain only requests that are in the
		1686	C<done> state, they will also finish. Otherwise they will continue to
		1687	exist.
		1688
		1689	That means after creating a group you have some time to add requests
		1690	(precisely before the callback has been invoked, which is only done within
		1691	the C<poll_cb>). And in the callbacks of those requests, you can add
		1692	further requests to the group. And only when all those requests have
		1693	finished will the the group itself finish.
		1694
		1695	=over 4
		1696
		1697	=item add $grp ...
		1698
		1699	=item $grp->add (...)
		1700
		1701	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1702	be added, including other groups, as long as you do not create circular
		1703	dependencies.
		1704
		1705	Returns all its arguments.
		1706
		1707	=item $grp->cancel_subs
		1708
		1709	Cancel all subrequests and clears any feeder, but not the group request
		1710	itself. Useful when you queued a lot of events but got a result early.
		1711
		1712	The group request will finish normally (you cannot add requests to the
		1713	group).
		1714
		1715	=item $grp->result (...)
		1716
		1717	Set the result value(s) that will be passed to the group callback when all
		1718	subrequests have finished and set the groups errno to the current value
		1719	of errno (just like calling C<errno> without an error number). By default,
		1720	no argument will be passed and errno is zero.
		1721
		1722	=item $grp->errno ([$errno])
		1723
		1724	Sets the group errno value to C<$errno>, or the current value of errno
		1725	when the argument is missing.
		1726
		1727	Every aio request has an associated errno value that is restored when
		1728	the callback is invoked. This method lets you change this value from its
		1729	default (0).
		1730
		1731	Calling C<result> will also set errno, so make sure you either set C<$!>
		1732	before the call to C<result>, or call c<errno> after it.
		1733
		1734	=item feed $grp $callback->($grp)
		1735
		1736	Sets a feeder/generator on this group: every group can have an attached
		1737	generator that generates requests if idle. The idea behind this is that,
		1738	although you could just queue as many requests as you want in a group,
		1739	this might starve other requests for a potentially long time. For example,
		1740	C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
		1741	requests, delaying any later requests for a long time.
		1742
		1743	To avoid this, and allow incremental generation of requests, you can
		1744	instead a group and set a feeder on it that generates those requests. The
		1745	feed callback will be called whenever there are few enough (see C<limit>,
		1746	below) requests active in the group itself and is expected to queue more
		1747	requests.
		1748
		1749	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1750	not impose any limits).
		1751
		1752	If the feed does not queue more requests when called, it will be
		1753	automatically removed from the group.
		1754
		1755	If the feed limit is C<0> when this method is called, it will be set to
		1756	C<2> automatically.
		1757
		1758	Example:
		1759
		1760	# stat all files in @files, but only ever use four aio requests concurrently:
		1761
		1762	my $grp = aio_group sub { print "finished\n" };
		1763	limit $grp 4;
		1764	feed $grp sub {
		1765	my $file = pop @files
		1766	or return;
		1767
		1768	add $grp aio_stat $file, sub { ... };
		1769	};
		1770
		1771	=item limit $grp $num
		1772
		1773	Sets the feeder limit for the group: The feeder will be called whenever
		1774	the group contains less than this many requests.
		1775
		1776	Setting the limit to C<0> will pause the feeding process.
		1777
		1778	The default value for the limit is C<0>, but note that setting a feeder
		1779	automatically bumps it up to C<2>.
		1780
		1781	=back
		1782
262	=head2 SUPPORT FUNCTIONS	1783	=head2 SUPPORT FUNCTIONS
263		1784
		1785	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1786
264	=over 4	1787	=over 4
265		1788
266	=item $fileno = IO::AIO::poll_fileno	1789	=item $fileno = IO::AIO::poll_fileno
267		1790
268	Return the I<request result pipe file descriptor>. This filehandle must be	1791	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	1792	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	1793	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
271	to call C<poll_cb> to check the results.	1794	you have to call C<poll_cb> to check the results.
272		1795
273	See C<poll_cb> for an example.	1796	See C<poll_cb> for an example.
274		1797
275	=item IO::AIO::poll_cb	1798	=item IO::AIO::poll_cb
276		1799
277	Process all outstanding events on the result pipe. You have to call this	1800	Process some requests that have reached the result phase (i.e. they have
278	regularly. Returns the number of events processed. Returns immediately	1801	been executed but the results are not yet reported). You have to call
279	when no events are outstanding.	1802	this "regularly" to finish outstanding requests.
		1803
		1804	Returns C<0> if all events could be processed (or there were no
		1805	events to process), or C<-1> if it returned earlier for whatever
		1806	reason. Returns immediately when no events are outstanding. The amount
		1807	of events processed depends on the settings of C<IO::AIO::max_poll_req>,
		1808	C<IO::AIO::max_poll_time> and C<IO::AIO::max_outstanding>.
		1809
		1810	If not all requests were processed for whatever reason, the poll file
		1811	descriptor will still be ready when C<poll_cb> returns, so normally you
		1812	don't have to do anything special to have it called later.
		1813
		1814	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1815	ready, it can be beneficial to call this function from loops which submit
		1816	a lot of requests, to make sure the results get processed when they become
		1817	available and not just when the loop is finished and the event loop takes
		1818	over again. This function returns very fast when there are no outstanding
		1819	requests.
280		1820
281	Example: Install an Event watcher that automatically calls	1821	Example: Install an Event watcher that automatically calls
282	IO::AIO::poll_cb with high priority:	1822	IO::AIO::poll_cb with high priority (more examples can be found in the
		1823	SYNOPSIS section, at the top of this document):
283		1824
284	Event->io (fd => IO::AIO::poll_fileno,	1825	Event->io (fd => IO::AIO::poll_fileno,
285	poll => 'r', async => 1,	1826	poll => 'r', async => 1,
286	cb => \&IO::AIO::poll_cb);	1827	cb => \&IO::AIO::poll_cb);
287		1828
288	=item IO::AIO::poll_wait	1829	=item IO::AIO::poll_wait
289		1830
290	Wait till the result filehandle becomes ready for reading (simply does a	1831	Wait until either at least one request is in the result phase or no
291	C<select> on the filehandle. This is useful if you want to synchronously wait	1832	requests are outstanding anymore.
292	for some requests to finish).	1833
		1834	This is useful if you want to synchronously wait for some requests to
		1835	become ready, without actually handling them.
293		1836
294	See C<nreqs> for an example.	1837	See C<nreqs> for an example.
295		1838
		1839	=item IO::AIO::poll
		1840
		1841	Waits until some requests have been handled.
		1842
		1843	Returns the number of requests processed, but is otherwise strictly
		1844	equivalent to:
		1845
		1846	IO::AIO::poll_wait, IO::AIO::poll_cb
		1847
296	=item IO::AIO::nreqs	1848	=item IO::AIO::flush
297		1849
298	Returns the number of requests currently outstanding (i.e. for which their	1850	Wait till all outstanding AIO requests have been handled.
299	callback has not been invoked yet).
300		1851
301	Example: wait till there are no outstanding requests anymore:	1852	Strictly equivalent to:
302		1853
303	IO::AIO::poll_wait, IO::AIO::poll_cb	1854	IO::AIO::poll_wait, IO::AIO::poll_cb
304	while IO::AIO::nreqs;	1855	while IO::AIO::nreqs;
305		1856
306	=item IO::AIO::flush	1857	=item IO::AIO::max_poll_reqs $nreqs
307		1858
308	Wait till all outstanding AIO requests have been handled.	1859	=item IO::AIO::max_poll_time $seconds
309		1860
310	Strictly equivalent to:	1861	These set the maximum number of requests (default C<0>, meaning infinity)
		1862	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1863	the maximum amount of time (default C<0>, meaning infinity) spent in
		1864	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1865	of time C<poll_cb> is allowed to use).
311		1866
312	IO::AIO::poll_wait, IO::AIO::poll_cb	1867	Setting C<max_poll_time> to a non-zero value creates an overhead of one
313	while IO::AIO::nreqs;	1868	syscall per request processed, which is not normally a problem unless your
		1869	callbacks are really really fast or your OS is really really slow (I am
		1870	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
314		1871
315	=item IO::AIO::poll	1872	Setting these is useful if you want to ensure some level of
		1873	interactiveness when perl is not fast enough to process all requests in
		1874	time.
316		1875
317	Waits until some requests have been handled.	1876	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
318		1877
319	Strictly equivalent to:	1878	Example: Install an Event watcher that automatically calls
		1879	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1880	program get the CPU sometimes even under high AIO load.
320		1881
321	IO::AIO::poll_wait, IO::AIO::poll_cb	1882	# try not to spend much more than 0.1s in poll_cb
322	if IO::AIO::nreqs;	1883	IO::AIO::max_poll_time 0.1;
		1884
		1885	# use a low priority so other tasks have priority
		1886	Event->io (fd => IO::AIO::poll_fileno,
		1887	poll => 'r', nice => 1,
		1888	cb => &IO::AIO::poll_cb);
		1889
		1890	=back
		1891
		1892	=head3 CONTROLLING THE NUMBER OF THREADS
		1893
		1894	=over
323		1895
324	=item IO::AIO::min_parallel $nthreads	1896	=item IO::AIO::min_parallel $nthreads
325		1897
326	Set the minimum number of AIO threads to C<$nthreads>. The current default	1898	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	1899	default is C<8>, which means eight asynchronous operations can execute
328	(the number of outstanding operations, however, is unlimited).	1900	concurrently at any one time (the number of outstanding requests,
		1901	however, is unlimited).
329		1902
330	IO::AIO starts threads only on demand, when an AIO request is queued and	1903	IO::AIO starts threads only on demand, when an AIO request is queued and
331	no free thread exists.	1904	no free thread exists. Please note that queueing up a hundred requests can
		1905	create demand for a hundred threads, even if it turns out that everything
		1906	is in the cache and could have been processed faster by a single thread.
332		1907
333	It is recommended to keep the number of threads low, as some Linux	1908	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	1909	Linux kernel versions will scale negatively with the number of threads
335	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1910	(higher parallelity => MUCH higher latency). With current Linux 2.6
336	threads should be fine.	1911	versions, 4-32 threads should be fine.
337		1912
338	Under most circumstances you don't need to call this function, as the	1913	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.	1914	module selects a default that is suitable for low to moderate load.
340		1915
341	=item IO::AIO::max_parallel $nthreads	1916	=item IO::AIO::max_parallel $nthreads
…		…
350	This module automatically runs C<max_parallel 0> at program end, to ensure	1925	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.	1926	that all threads are killed and that there are no outstanding requests.
352		1927
353	Under normal circumstances you don't need to call this function.	1928	Under normal circumstances you don't need to call this function.
354		1929
		1930	=item IO::AIO::max_idle $nthreads
		1931
		1932	Limit the number of threads (default: 4) that are allowed to idle
		1933	(i.e., threads that did not get a request to process within the idle
		1934	timeout (default: 10 seconds). That means if a thread becomes idle while
		1935	C<$nthreads> other threads are also idle, it will free its resources and
		1936	exit.
		1937
		1938	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1939	to allow for extremely high load situations, but want to free resources
		1940	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1941
		1942	The default is probably ok in most situations, especially if thread
		1943	creation is fast. If thread creation is very slow on your system you might
		1944	want to use larger values.
		1945
		1946	=item IO::AIO::idle_timeout $seconds
		1947
		1948	Sets the minimum idle timeout (default 10) after which worker threads are
		1949	allowed to exit. SEe C<IO::AIO::max_idle>.
		1950
355	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1951	=item IO::AIO::max_outstanding $maxreqs
356		1952
357	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1953	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	1954	you do queue up more than this number of requests, the next call to
359	some requests have been handled.	1955	C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
		1956	C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
		1957	longer exceeded.
360		1958
361	The default is very large, so normally there is no practical limit. If you	1959	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	1960	used to make poll functions block if the limit is exceeded.
363	this to a relatively low number, such as C<100>.
364		1961
365	Under normal circumstances you don't need to call this function.	1962	This is a very bad function to use in interactive programs because it
		1963	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1964	use an C<aio_group> together with a feed callback.
		1965
		1966	Its main use is in scripts without an event loop - when you want to stat
		1967	a lot of files, you can write something like this:
		1968
		1969	IO::AIO::max_outstanding 32;
		1970
		1971	for my $path (...) {
		1972	aio_stat $path , ...;
		1973	IO::AIO::poll_cb;
		1974	}
		1975
		1976	IO::AIO::flush;
		1977
		1978	The call to C<poll_cb> inside the loop will normally return instantly, but
		1979	as soon as more thna C<32> reqeusts are in-flight, it will block until
		1980	some requests have been handled. This keeps the loop from pushing a large
		1981	number of C<aio_stat> requests onto the queue.
		1982
		1983	The default value for C<max_outstanding> is very large, so there is no
		1984	practical limit on the number of outstanding requests.
366		1985
367	=back	1986	=back
368		1987
		1988	=head3 STATISTICAL INFORMATION
		1989
		1990	=over
		1991
		1992	=item IO::AIO::nreqs
		1993
		1994	Returns the number of requests currently in the ready, execute or pending
		1995	states (i.e. for which their callback has not been invoked yet).
		1996
		1997	Example: wait till there are no outstanding requests anymore:
		1998
		1999	IO::AIO::poll_wait, IO::AIO::poll_cb
		2000	while IO::AIO::nreqs;
		2001
		2002	=item IO::AIO::nready
		2003
		2004	Returns the number of requests currently in the ready state (not yet
		2005	executed).
		2006
		2007	=item IO::AIO::npending
		2008
		2009	Returns the number of requests currently in the pending state (executed,
		2010	but not yet processed by poll_cb).
		2011
		2012	=back
		2013
		2014	=head3 MISCELLANEOUS FUNCTIONS
		2015
		2016	IO::AIO implements some functions that are useful when you want to use
		2017	some "Advanced I/O" function not available to in Perl, without going the
		2018	"Asynchronous I/O" route. Many of these have an asynchronous C<aio_*>
		2019	counterpart.
		2020
		2021	=over 4
		2022
		2023	=item $numfd = IO::AIO::get_fdlimit
		2024
		2025	This function is I<EXPERIMENTAL> and subject to change.
		2026
		2027	Tries to find the current file descriptor limit and returns it, or
		2028	C<undef> and sets C<$!> in case of an error. The limit is one larger than
		2029	the highest valid file descriptor number.
		2030
		2031	=item IO::AIO::min_fdlimit [$numfd]
		2032
		2033	This function is I<EXPERIMENTAL> and subject to change.
		2034
		2035	Try to increase the current file descriptor limit(s) to at least C<$numfd>
		2036	by changing the soft or hard file descriptor resource limit. If C<$numfd>
		2037	is missing, it will try to set a very high limit, although this is not
		2038	recommended when you know the actual minimum that you require.
		2039
		2040	If the limit cannot be raised enough, the function makes a best-effort
		2041	attempt to increase the limit as much as possible, using various
		2042	tricks, while still failing. You can query the resulting limit using
		2043	C<IO::AIO::get_fdlimit>.
		2044
		2045	If an error occurs, returns C<undef> and sets C<$!>, otherwise returns
		2046	true.
		2047
		2048	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		2049
		2050	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		2051	but is blocking (this makes most sense if you know the input data is
		2052	likely cached already and the output filehandle is set to non-blocking
		2053	operations).
		2054
		2055	Returns the number of bytes copied, or C<-1> on error.
		2056
		2057	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		2058
		2059	Simply calls the C<posix_fadvise> function (see its
		2060	manpage for details). The following advice constants are
		2061	available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		2062	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		2063	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		2064
		2065	On systems that do not implement C<posix_fadvise>, this function returns
		2066	ENOSYS, otherwise the return value of C<posix_fadvise>.
		2067
		2068	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		2069
		2070	Simply calls the C<posix_madvise> function (see its
		2071	manpage for details). The following advice constants are
		2072	available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		2073	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>,
		2074	C<IO::AIO::MADV_DONTNEED>.
		2075
		2076	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2077	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2078	will be reduced to fit into the C<$scalar>.
		2079
		2080	On systems that do not implement C<posix_madvise>, this function returns
		2081	ENOSYS, otherwise the return value of C<posix_madvise>.
		2082
		2083	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		2084
		2085	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		2086	$scalar (see its manpage for details). The following protect
		2087	constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		2088	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		2089
		2090	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2091	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2092	will be reduced to fit into the C<$scalar>.
		2093
		2094	On systems that do not implement C<mprotect>, this function returns
		2095	ENOSYS, otherwise the return value of C<mprotect>.
		2096
		2097	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		2098
		2099	Memory-maps a file (or anonymous memory range) and attaches it to the
		2100	given C<$scalar>, which will act like a string scalar. Returns true on
		2101	success, and false otherwise.
		2102
		2103	The scalar must exist, but its contents do not matter - this means you
		2104	cannot use a nonexistant array or hash element. When in doubt, C<undef>
		2105	the scalar first.
		2106
		2107	The only operations allowed on the mmapped scalar are C<substr>/C<vec>,
		2108	which don't change the string length, and most read-only operations such
		2109	as copying it or searching it with regexes and so on.
		2110
		2111	Anything else is unsafe and will, at best, result in memory leaks.
		2112
		2113	The memory map associated with the C<$scalar> is automatically removed
		2114	when the C<$scalar> is undef'd or destroyed, or when the C<IO::AIO::mmap>
		2115	or C<IO::AIO::munmap> functions are called on it.
		2116
		2117	This calls the C<mmap>(2) function internally. See your system's manual
		2118	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		2119
		2120	The C<$length> must be larger than zero and smaller than the actual
		2121	filesize.
		2122
		2123	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		2124	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		2125
		2126	C<$flags> can be a combination of
		2127	C<IO::AIO::MAP_SHARED> or
		2128	C<IO::AIO::MAP_PRIVATE>,
		2129	or a number of system-specific flags (when not available, the are C<0>):
		2130	C<IO::AIO::MAP_ANONYMOUS> (which is set to C<MAP_ANON> if your system only provides this constant),
		2131	C<IO::AIO::MAP_LOCKED>,
		2132	C<IO::AIO::MAP_NORESERVE>,
		2133	C<IO::AIO::MAP_POPULATE>,
		2134	C<IO::AIO::MAP_NONBLOCK>,
		2135	C<IO::AIO::MAP_FIXED>,
		2136	C<IO::AIO::MAP_GROWSDOWN>,
		2137	C<IO::AIO::MAP_32BIT>,
		2138	C<IO::AIO::MAP_HUGETLB> or
		2139	C<IO::AIO::MAP_STACK>.
		2140
		2141	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		2142
		2143	C<$offset> is the offset from the start of the file - it generally must be
		2144	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		2145
		2146	Example:
		2147
		2148	use Digest::MD5;
		2149	use IO::AIO;
		2150
		2151	open my $fh, "<verybigfile"
		2152	or die "$!";
		2153
		2154	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		2155	or die "verybigfile: $!";
		2156
		2157	my $fast_md5 = md5 $data;
		2158
		2159	=item IO::AIO::munmap $scalar
		2160
		2161	Removes a previous mmap and undefines the C<$scalar>.
		2162
		2163	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		2164
		2165	Calls the C<munlock> function, undoing the effects of a previous
		2166	C<aio_mlock> call (see its description for details).
		2167
		2168	=item IO::AIO::munlockall
		2169
		2170	Calls the C<munlockall> function.
		2171
		2172	On systems that do not implement C<munlockall>, this function returns
		2173	ENOSYS, otherwise the return value of C<munlockall>.
		2174
		2175	=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
		2176
		2177	Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
		2178	C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
		2179	should be the file offset.
		2180
		2181	C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
		2182	silently corrupt the data in this case.
		2183
		2184	The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
		2185	C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
		2186	C<IO::AIO::SPLICE_F_GIFT>.
		2187
		2188	See the C<splice(2)> manpage for details.
		2189
		2190	=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
		2191
		2192	Calls the GNU/Linux C<tee(2)> syscall, see its manpage and the
		2193	description for C<IO::AIO::splice> above for details.
		2194
		2195	=item $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
		2196
		2197	Attempts to query or change the pipe buffer size. Obviously works only
		2198	on pipes, and currently works only on GNU/Linux systems, and fails with
		2199	C<-1>/C<ENOSYS> everywhere else. If anybody knows how to influence pipe buffer
		2200	size on other systems, drop me a note.
		2201
		2202	=item ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
		2203
		2204	This is a direct interface to the Linux L<pipe2(2)> system call. If
		2205	C<$flags> is missing or C<0>, then this should be the same as a call to
		2206	perl's built-in C<pipe> function and create a new pipe, and works on
		2207	systems that lack the pipe2 syscall. On win32, this case invokes C<_pipe
		2208	(..., 4096, O_BINARY)>.
		2209
		2210	If C<$flags> is non-zero, it tries to invoke the pipe2 system call with
		2211	the given flags (Linux 2.6.27, glibc 2.9).
		2212
		2213	On success, the read and write file handles are returned.
		2214
		2215	On error, nothing will be returned. If the pipe2 syscall is missing and
		2216	C<$flags> is non-zero, fails with C<ENOSYS>.
		2217
		2218	Please refer to L<pipe2(2)> for more info on the C<$flags>, but at the
		2219	time of this writing, C<IO::AIO::O_CLOEXEC>, C<IO::AIO::O_NONBLOCK> and
		2220	C<IO::AIO::O_DIRECT> (Linux 3.4, for packet-based pipes) were supported.
		2221
		2222	Example: create a pipe race-free w.r.t. threads and fork:
		2223
		2224	my ($rfh, $wfh) = IO::AIO::pipe2 IO::AIO::O_CLOEXEC
		2225	or die "pipe2: $!\n";
		2226
		2227	=item $fh = IO::AIO::eventfd [$initval, [$flags]]
		2228
		2229	This is a direct interface to the Linux L<eventfd(2)> system call. The
		2230	(unhelpful) defaults for C<$initval> and C<$flags> are C<0> for both.
		2231
		2232	On success, the new eventfd filehandle is returned, otherwise returns
		2233	C<undef>. If the eventfd syscall is missing, fails with C<ENOSYS>.
		2234
		2235	Please refer to L<eventfd(2)> for more info on this call.
		2236
		2237	The following symbol flag values are available: C<IO::AIO::EFD_CLOEXEC>,
		2238	C<IO::AIO::EFD_NONBLOCK> and C<IO::AIO::EFD_SEMAPHORE> (Linux 2.6.30).
		2239
		2240	Example: create a new eventfd filehandle:
		2241
		2242	$fh = IO::AIO::eventfd 0, IO::AIO::O_CLOEXEC
		2243	or die "eventfd: $!\n";
		2244
		2245	=item $fh = IO::AIO::timerfd_create $clockid[, $flags]
		2246
		2247	This is a direct interface to the Linux L<timerfd_create(2)> system call. The
		2248	(unhelpful) default for C<$flags> is C<0>.
		2249
		2250	On success, the new timerfd filehandle is returned, otherwise returns
		2251	C<undef>. If the eventfd syscall is missing, fails with C<ENOSYS>.
		2252
		2253	Please refer to L<timerfd_create(2)> for more info on this call.
		2254
		2255	The following C<$clockid> values are
		2256	available: C<IO::AIO::CLOCK_REALTIME>, C<IO::AIO::CLOCK_MONOTONIC>
		2257	C<IO::AIO::CLOCK_CLOCK_BOOTTIME> (Linux 3.15)
		2258	C<IO::AIO::CLOCK_CLOCK_REALTIME_ALARM> (Linux 3.11) and
		2259	C<IO::AIO::CLOCK_CLOCK_BOOTTIME_ALARM> (Linux 3.11).
		2260
		2261	The following C<$flags> values are available (Linux
		2262	2.6.27): C<IO::AIO::TFD_NONBLOCK> and C<IO::AIO::TFD_CLOEXEC>.
		2263
		2264	Example: create a new timerfd and set it to one-second repeated alarms,
		2265	then wait for two alarms:
		2266
		2267	my $fh = IO::AIO::timerfd_create IO::AIO::CLOCK_BOOTTIME, IO::AIO::TFD_CLOEXEC
		2268	or die "timerfd_create: $!\n";
		2269
		2270	defined IO::AIO::timerfd_settime $fh, 0, 1, 1
		2271	or die "timerfd_settime: $!\n";
		2272
		2273	for (1..2) {
		2274	8 == sysread $fh, my $buf, 8
		2275	or die "timerfd read failure\n";
		2276
		2277	printf "number of expirations (likely 1): %d\n",
		2278	unpack "Q", $buf;
		2279	}
		2280
		2281	=item ($cur_interval, $cur_value) = IO::AIO::timerfd_settime $fh, $flags, $new_interval, $nbw_value
		2282
		2283	This is a direct interface to the Linux L<timerfd_settime(2)> system
		2284	call. Please refer to its manpage for more info on this call.
		2285
		2286	The new itimerspec is specified using two (possibly fractional) second
		2287	values, C<$new_interval> and C<$new_value>).
		2288
		2289	On success, the current interval and value are returned (as per
		2290	C<timerfd_gettime>). On failure, the empty list is returned.
		2291
		2292	The following C<$flags> values are
		2293	available: C<IO::AIO::TFD_TIMER_ABSTIME> and
		2294	C<IO::AIO::TFD_TIMER_CANCEL_ON_SET>.
		2295
		2296	See C<IO::AIO::timerfd_create> for a full example.
		2297
		2298	=item ($cur_interval, $cur_value) = IO::AIO::timerfd_gettime $fh
		2299
		2300	This is a direct interface to the Linux L<timerfd_gettime(2)> system
		2301	call. Please refer to its manpage for more info on this call.
		2302
		2303	On success, returns the current values of interval and value for the given
		2304	timerfd (as potentially fractional second values). On failure, the empty
		2305	list is returned.
		2306
		2307	=back
		2308
369	=cut	2309	=cut
370		2310
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;	2311	min_parallel 8;
388		2312
389	END {	2313	END { flush }
390	max_parallel 0;
391	}
392		2314
393	1;	2315	1;
394		2316
		2317	=head1 EVENT LOOP INTEGRATION
		2318
		2319	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		2320	automatically into many event loops:
		2321
		2322	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		2323	use AnyEvent::AIO;
		2324
		2325	You can also integrate IO::AIO manually into many event loops, here are
		2326	some examples of how to do this:
		2327
		2328	# EV integration
		2329	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		2330
		2331	# Event integration
		2332	Event->io (fd => IO::AIO::poll_fileno,
		2333	poll => 'r',
		2334	cb => \&IO::AIO::poll_cb);
		2335
		2336	# Glib/Gtk2 integration
		2337	add_watch Glib::IO IO::AIO::poll_fileno,
		2338	in => sub { IO::AIO::poll_cb; 1 };
		2339
		2340	# Tk integration
		2341	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		2342	readable => \&IO::AIO::poll_cb);
		2343
		2344	# Danga::Socket integration
		2345	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		2346	\&IO::AIO::poll_cb);
		2347
395	=head2 FORK BEHAVIOUR	2348	=head2 FORK BEHAVIOUR
396		2349
397	Before the fork, IO::AIO enters a quiescent state where no requests	2350	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	2351	considerably. Specifically, only async-safe functions can be called after
399	the fork the parent simply leaves the quiescent state and continues	2352	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	2353	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	2354	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	2355	reasons) perl itself often is linked against pthreads, so this limitation
403	parent process has been reached again.	2356	applies to quite a lot of perls.
		2357
		2358	This module no longer tries to fight your OS, or POSIX. That means IO::AIO
		2359	only works in the process that loaded it. Forking is fully supported, but
		2360	using IO::AIO in the child is not.
		2361
		2362	You might get around by not I<using> IO::AIO before (or after)
		2363	forking. You could also try to call the L<IO::AIO::reinit> function in the
		2364	child:
		2365
		2366	=over 4
		2367
		2368	=item IO::AIO::reinit
		2369
		2370	Abandons all current requests and I/O threads and simply reinitialises all
		2371	data structures. This is not an operation supported by any standards, but
		2372	happens to work on GNU/Linux and some newer BSD systems.
		2373
		2374	The only reasonable use for this function is to call it after forking, if
		2375	C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
		2376	the process will result in undefined behaviour. Calling it at any time
		2377	will also result in any undefined (by POSIX) behaviour.
		2378
		2379	=back
		2380
		2381	=head2 LINUX-SPECIFIC CALLS
		2382
		2383	When a call is documented as "linux-specific" then this means it
		2384	originated on GNU/Linux. C<IO::AIO> will usually try to autodetect the
		2385	availability and compatibility of such calls regardless of the platform
		2386	it is compiled on, so platforms such as FreeBSD which often implement
		2387	these calls will work. When in doubt, call them and see if they fail wth
		2388	C<ENOSYS>.
		2389
		2390	=head2 MEMORY USAGE
		2391
		2392	Per-request usage:
		2393
		2394	Each aio request uses - depending on your architecture - around 100-200
		2395	bytes of memory. In addition, stat requests need a stat buffer (possibly
		2396	a few hundred bytes), readdir requires a result buffer and so on. Perl
		2397	scalars and other data passed into aio requests will also be locked and
		2398	will consume memory till the request has entered the done state.
		2399
		2400	This is not awfully much, so queuing lots of requests is not usually a
		2401	problem.
		2402
		2403	Per-thread usage:
		2404
		2405	In the execution phase, some aio requests require more memory for
		2406	temporary buffers, and each thread requires a stack and other data
		2407	structures (usually around 16k-128k, depending on the OS).
		2408
		2409	=head1 KNOWN BUGS
		2410
		2411	Known bugs will be fixed in the next release :)
		2412
		2413	=head1 KNOWN ISSUES
		2414
		2415	Calls that try to "import" foreign memory areas (such as C<IO::AIO::mmap>
		2416	or C<IO::AIO::aio_slurp>) do not work with generic lvalues, such as
		2417	non-created hash slots or other scalars I didn't think of. It's best to
		2418	avoid such and either use scalar variables or making sure that the scalar
		2419	exists (e.g. by storing C<undef>) and isn't "funny" (e.g. tied).
		2420
		2421	I am not sure anything can be done about this, so this is considered a
		2422	known issue, rather than a bug.
404		2423
405	=head1 SEE ALSO	2424	=head1 SEE ALSO
406		2425
407	L<Coro>, L<Linux::AIO>.	2426	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		2427	more natural syntax.
408		2428
409	=head1 AUTHOR	2429	=head1 AUTHOR
410		2430
411	Marc Lehmann <schmorp@schmorp.de>	2431	Marc Lehmann <schmorp@schmorp.de>
412	http://home.schmorp.de/	2432	http://home.schmorp.de/

Diff Legend

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