ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.14 by root, Mon Jul 11 02:53:59 2005 UTC vs.
Revision 1.273 by root, Fri Jun 23 22:33:06 2017 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	\&IO::AIO::poll_cb;	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.	68	it is currently not reentrant in any way, so use appropriate locking
		69	yourself, always call C<poll_cb> from within the same thread, or never
		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
54		164
55	=cut	165	=cut
56		166
57	package IO::AIO;	167	package IO::AIO;
58		168
		169	use Carp ();
		170
		171	use common::sense;
		172
59	use base 'Exporter';	173	use base 'Exporter';
60		174
61	use Fcntl ();
62
63	BEGIN {	175	BEGIN {
64	$VERSION = 0.3;	176	our $VERSION = 4.35;
65		177
66	@EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink	178	our @AIO_REQ = qw(aio_sendfile aio_seek aio_read aio_write aio_open aio_close
67	aio_fsync aio_fdatasync aio_readahead);	179	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
68	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);	180	aio_scandir aio_symlink aio_readlink aio_realpath aio_fcntl aio_ioctl
		181	aio_sync aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range
		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_wd);
		189
		190	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		191	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		192	min_parallel max_parallel max_idle idle_timeout
		193	nreqs nready npending nthreads
		194	max_poll_time max_poll_reqs
		195	sendfile fadvise madvise
		196	mmap munmap munlock munlockall);
		197
		198	push @AIO_REQ, qw(aio_busy); # not exported
		199
		200	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
69		201
70	require XSLoader;	202	require XSLoader;
71	XSLoader::load IO::AIO, $VERSION;	203	XSLoader::load ("IO::AIO", $VERSION);
72	}	204	}
73		205
74	=head1 FUNCTIONS	206	=head1 FUNCTIONS
75		207
76	=head2 AIO FUNCTIONS	208	=head2 QUICK OVERVIEW
		209
		210	This section simply lists the prototypes most of the functions for
		211	quick reference. See the following sections for function-by-function
		212	documentation.
		213
		214	aio_wd $pathname, $callback->($wd)
		215	aio_open $pathname, $flags, $mode, $callback->($fh)
		216	aio_close $fh, $callback->($status)
		217	aio_seek $fh,$offset,$whence, $callback->($offs)
		218	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		219	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		220	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		221	aio_readahead $fh,$offset,$length, $callback->($retval)
		222	aio_stat $fh_or_path, $callback->($status)
		223	aio_lstat $fh, $callback->($status)
		224	aio_statvfs $fh_or_path, $callback->($statvfs)
		225	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		226	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		227	aio_chmod $fh_or_path, $mode, $callback->($status)
		228	aio_truncate $fh_or_path, $offset, $callback->($status)
		229	aio_allocate $fh, $mode, $offset, $len, $callback->($status)
		230	aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		231	aio_unlink $pathname, $callback->($status)
		232	aio_mknod $pathname, $mode, $dev, $callback->($status)
		233	aio_link $srcpath, $dstpath, $callback->($status)
		234	aio_symlink $srcpath, $dstpath, $callback->($status)
		235	aio_readlink $pathname, $callback->($link)
		236	aio_realpath $pathname, $callback->($path)
		237	aio_rename $srcpath, $dstpath, $callback->($status)
		238	aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
		239	aio_mkdir $pathname, $mode, $callback->($status)
		240	aio_rmdir $pathname, $callback->($status)
		241	aio_readdir $pathname, $callback->($entries)
		242	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		243	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		244	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		245	aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		246	aio_load $pathname, $data, $callback->($status)
		247	aio_copy $srcpath, $dstpath, $callback->($status)
		248	aio_move $srcpath, $dstpath, $callback->($status)
		249	aio_rmtree $pathname, $callback->($status)
		250	aio_fcntl $fh, $cmd, $arg, $callback->($status)
		251	aio_ioctl $fh, $request, $buf, $callback->($status)
		252	aio_sync $callback->($status)
		253	aio_syncfs $fh, $callback->($status)
		254	aio_fsync $fh, $callback->($status)
		255	aio_fdatasync $fh, $callback->($status)
		256	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		257	aio_pathsync $pathname, $callback->($status)
		258	aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		259	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		260	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		261	aio_mlockall $flags, $callback->($status)
		262	aio_group $callback->(...)
		263	aio_nop $callback->()
		264
		265	$prev_pri = aioreq_pri [$pri]
		266	aioreq_nice $pri_adjust
		267
		268	IO::AIO::poll_wait
		269	IO::AIO::poll_cb
		270	IO::AIO::poll
		271	IO::AIO::flush
		272	IO::AIO::max_poll_reqs $nreqs
		273	IO::AIO::max_poll_time $seconds
		274	IO::AIO::min_parallel $nthreads
		275	IO::AIO::max_parallel $nthreads
		276	IO::AIO::max_idle $nthreads
		277	IO::AIO::idle_timeout $seconds
		278	IO::AIO::max_outstanding $maxreqs
		279	IO::AIO::nreqs
		280	IO::AIO::nready
		281	IO::AIO::npending
		282
		283	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		284	IO::AIO::fadvise $fh, $offset, $len, $advice
		285	IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
		286	IO::AIO::munmap $scalar
		287	IO::AIO::madvise $scalar, $offset, $length, $advice
		288	IO::AIO::mprotect $scalar, $offset, $length, $protect
		289	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		290	IO::AIO::munlockall
		291
		292	=head2 API NOTES
77		293
78	All the C<aio_*> calls are more or less thin wrappers around the syscall	294	All the C<aio_*> calls are more or less thin wrappers around the syscall
79	with the same name (sans C<aio_>). The arguments are similar or identical,	295	with the same name (sans C<aio_>). The arguments are similar or identical,
80	and they all accept an additional (and optional) C<$callback> argument	296	and they all accept an additional (and optional) C<$callback> argument
81	which must be a code reference. This code reference will get called with	297	which must be a code reference. This code reference will be called after
82	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	298	the syscall has been executed in an asynchronous fashion. The results
83	perl, which usually delivers "false") as it's sole argument when the given	299	of the request will be passed as arguments to the callback (and, if an
84	syscall has been executed asynchronously.	300	error occured, in C<$!>) - for most requests the syscall return code (e.g.
		301	most syscalls return C<-1> on error, unlike perl, which usually delivers
		302	"false").
85		303
86	All functions that expect a filehandle will also accept a file descriptor.	304	Some requests (such as C<aio_readdir>) pass the actual results and
		305	communicate failures by passing C<undef>.
87		306
		307	All functions expecting a filehandle keep a copy of the filehandle
		308	internally until the request has finished.
		309
		310	All functions return request objects of type L<IO::AIO::REQ> that allow
		311	further manipulation of those requests while they are in-flight.
		312
88	The filenames you pass to these routines I<must> be absolute. The reason	313	The pathnames you pass to these routines I<should> be absolute. The
89	is that at the time the request is being executed, the current working	314	reason for this is that at the time the request is being executed, the
90	directory could have changed. Alternatively, you can make sure that you	315	current working directory could have changed. Alternatively, you can
91	never change the current working directory.	316	make sure that you never change the current working directory anywhere
		317	in the program and then use relative paths. You can also take advantage
		318	of IO::AIOs working directory abstraction, that lets you specify paths
		319	relative to some previously-opened "working directory object" - see the
		320	description of the C<IO::AIO::WD> class later in this document.
		321
		322	To encode pathnames as octets, either make sure you either: a) always pass
		323	in filenames you got from outside (command line, readdir etc.) without
		324	tinkering, b) are in your native filesystem encoding, c) use the Encode
		325	module and encode your pathnames to the locale (or other) encoding in
		326	effect in the user environment, d) use Glib::filename_from_unicode on
		327	unicode filenames or e) use something else to ensure your scalar has the
		328	correct contents.
		329
		330	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		331	handles correctly whether it is set or not.
		332
		333	=head2 AIO REQUEST FUNCTIONS
92		334
93	=over 4	335	=over 4
94		336
		337	=item $prev_pri = aioreq_pri [$pri]
		338
		339	Returns the priority value that would be used for the next request and, if
		340	C<$pri> is given, sets the priority for the next aio request.
		341
		342	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		343	and C<4>, respectively. Requests with higher priority will be serviced
		344	first.
		345
		346	The priority will be reset to C<0> after each call to one of the C<aio_*>
		347	functions.
		348
		349	Example: open a file with low priority, then read something from it with
		350	higher priority so the read request is serviced before other low priority
		351	open requests (potentially spamming the cache):
		352
		353	aioreq_pri -3;
		354	aio_open ..., sub {
		355	return unless $_[0];
		356
		357	aioreq_pri -2;
		358	aio_read $_[0], ..., sub {
		359	...
		360	};
		361	};
		362
		363
		364	=item aioreq_nice $pri_adjust
		365
		366	Similar to C<aioreq_pri>, but subtracts the given value from the current
		367	priority, so the effect is cumulative.
		368
		369
95	=item aio_open $pathname, $flags, $mode, $callback	370	=item aio_open $pathname, $flags, $mode, $callback->($fh)
96		371
97	Asynchronously open or create a file and call the callback with a newly	372	Asynchronously open or create a file and call the callback with a newly
98	created filehandle for the file.	373	created filehandle for the file (or C<undef> in case of an error).
99		374
100	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	375	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
101	for an explanation.	376	for an explanation.
102		377
103	The C<$mode> argument is a bitmask. See the C<Fcntl> module for a	378	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
104	list. They are the same as used in C<sysopen>.	379	list. They are the same as used by C<sysopen>.
		380
		381	Likewise, C<$mode> specifies the mode of the newly created file, if it
		382	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
		383	except that it is mandatory (i.e. use C<0> if you don't create new files,
		384	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		385	by the umask in effect then the request is being executed, so better never
		386	change the umask.
105		387
106	Example:	388	Example:
107		389
108	aio_open "/etc/passwd", O_RDONLY, 0, sub {	390	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
109	if ($_[0]) {	391	if ($_[0]) {
110	print "open successful, fh is $_[0]\n";	392	print "open successful, fh is $_[0]\n";
111	...	393	...
112	} else {	394	} else {
113	die "open failed: $!\n";	395	die "open failed: $!\n";
114	}	396	}
115	};	397	};
116		398
		399	In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
		400	C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
		401	following POSIX and non-POSIX constants are available (missing ones on
		402	your system are, as usual, C<0>):
		403
		404	C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
		405	C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
		406	C<O_RSYNC>, C<O_SYNC>, C<O_PATH>, C<O_TMPFILE>, and C<O_TTY_INIT>.
		407
		408
117	=item aio_close $fh, $callback	409	=item aio_close $fh, $callback->($status)
118		410
119	Asynchronously close a file and call the callback with the result	411	Asynchronously close a file and call the callback with the result
120	code. I<WARNING:> although accepted, you should not pass in a perl	412	code.
121	filehandle here, as perl will likely close the file descriptor itself when
122	the filehandle is destroyed. Normally, you can safely call perls C<close>
123	or just let filehandles go out of scope.
124		413
		414	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
		415	closing the file descriptor associated with the filehandle itself.
		416
		417	Therefore, C<aio_close> will not close the filehandle - instead it will
		418	use dup2 to overwrite the file descriptor with the write-end of a pipe
		419	(the pipe fd will be created on demand and will be cached).
		420
		421	Or in other words: the file descriptor will be closed, but it will not be
		422	free for reuse until the perl filehandle is closed.
		423
		424	=cut
		425
		426	=item aio_seek $fh, $offset, $whence, $callback->($offs)
		427
		428	Seeks the filehandle to the new C<$offset>, similarly to perl's
		429	C<sysseek>. The C<$whence> can use the traditional values (C<0> for
		430	C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
		431	C<IO::AIO::SEEK_END>).
		432
		433	The resulting absolute offset will be passed to the callback, or C<-1> in
		434	case of an error.
		435
		436	In theory, the C<$whence> constants could be different than the
		437	corresponding values from L<Fcntl>, but perl guarantees they are the same,
		438	so don't panic.
		439
		440	As a GNU/Linux (and maybe Solaris) extension, also the constants
		441	C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
		442	could be found. No guarantees about suitability for use in C<aio_seek> or
		443	Perl's C<sysseek> can be made though, although I would naively assume they
		444	"just work".
		445
125	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	446	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
126		447
127	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	448	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
128		449
129	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	450	Reads or writes C<$length> bytes from or to the specified C<$fh> and
130	into the scalar given by C<data> and offset C<dataoffset> and calls the	451	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> and
131	callback without the actual number of bytes read (or -1 on error, just	452	calls the callback with the actual number of bytes transferred (or -1 on
132	like the syscall).	453	error, just like the syscall).
133		454
		455	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		456	offset plus the actual number of bytes read.
		457
		458	If C<$offset> is undefined, then the current file descriptor offset will
		459	be used (and updated), otherwise the file descriptor offset will not be
		460	changed by these calls.
		461
		462	If C<$length> is undefined in C<aio_write>, use the remaining length of
		463	C<$data>.
		464
		465	If C<$dataoffset> is less than zero, it will be counted from the end of
		466	C<$data>.
		467
		468	The C<$data> scalar I<MUST NOT> be modified in any way while the request
		469	is outstanding. Modifying it can result in segfaults or World War III (if
		470	the necessary/optional hardware is installed).
		471
134	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at	472	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
135	offset C<0> within the scalar:	473	offset C<0> within the scalar:
136		474
137	aio_read $fh, 7, 15, $buffer, 0, sub {	475	aio_read $fh, 7, 15, $buffer, 0, sub {
138	$_[0] > 0 or die "read error: $!";	476	$_[0] > 0 or die "read error: $!";
139	print "read $_[0] bytes: <$buffer>\n";	477	print "read $_[0] bytes: <$buffer>\n";
140	};	478	};
141		479
		480
		481	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		482
		483	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
		484	reading at byte offset C<$in_offset>, and starts writing at the current
		485	file offset of C<$out_fh>. Because of that, it is not safe to issue more
		486	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
		487	other. The same C<$in_fh> works fine though, as this function does not
		488	move or use the file offset of C<$in_fh>.
		489
		490	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		491	are written, and there is no way to find out how many more bytes have been
		492	read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
		493	number of bytes written to C<$out_fh>. Only if the result value equals
		494	C<$length> one can assume that C<$length> bytes have been read.
		495
		496	Unlike with other C<aio_> functions, it makes a lot of sense to use
		497	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		498	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		499	the socket I/O will be non-blocking. Note, however, that you can run
		500	into a trap where C<aio_sendfile> reads some data with readahead, then
		501	fails to write all data, and when the socket is ready the next time, the
		502	data in the cache is already lost, forcing C<aio_sendfile> to again hit
		503	the disk. Explicit C<aio_read> + C<aio_write> let's you better control
		504	resource usage.
		505
		506	This call tries to make use of a native C<sendfile>-like syscall to
		507	provide zero-copy operation. For this to work, C<$out_fh> should refer to
		508	a socket, and C<$in_fh> should refer to an mmap'able file.
		509
		510	If a native sendfile cannot be found or it fails with C<ENOSYS>,
		511	C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
		512	C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
		513	type of filehandle regardless of the limitations of the operating system.
		514
		515	As native sendfile syscalls (as practically any non-POSIX interface hacked
		516	together in a hurry to improve benchmark numbers) tend to be rather buggy
		517	on many systems, this implementation tries to work around some known bugs
		518	in Linux and FreeBSD kernels (probably others, too), but that might fail,
		519	so you really really should check the return value of C<aio_sendfile> -
		520	fewer bytes than expected might have been transferred.
		521
		522
142	=item aio_readahead $fh,$offset,$length, $callback	523	=item aio_readahead $fh,$offset,$length, $callback->($retval)
143		524
144	Asynchronously reads the specified byte range into the page cache, using
145	the C<readahead> syscall. If that syscall doesn't exist the status will be
146	C<-1> and C<$!> is set to ENOSYS.
147
148	readahead() populates the page cache with data from a file so that	525	C<aio_readahead> populates the page cache with data from a file so that
149	subsequent reads from that file will not block on disk I/O. The C<$offset>	526	subsequent reads from that file will not block on disk I/O. The C<$offset>
150	argument specifies the starting point from which data is to be read and	527	argument specifies the starting point from which data is to be read and
151	C<$length> specifies the number of bytes to be read. I/O is performed in	528	C<$length> specifies the number of bytes to be read. I/O is performed in
152	whole pages, so that offset is effectively rounded down to a page boundary	529	whole pages, so that offset is effectively rounded down to a page boundary
153	and bytes are read up to the next page boundary greater than or equal to	530	and bytes are read up to the next page boundary greater than or equal to
154	(off-set+length). aio_readahead() does not read beyond the end of the	531	(off-set+length). C<aio_readahead> does not read beyond the end of the
155	file. The current file offset of the file is left unchanged.	532	file. The current file offset of the file is left unchanged.
156		533
		534	If that syscall doesn't exist (likely if your kernel isn't Linux) it will
		535	be emulated by simply reading the data, which would have a similar effect.
		536
		537
157	=item aio_stat $fh_or_path, $callback	538	=item aio_stat $fh_or_path, $callback->($status)
158		539
159	=item aio_lstat $fh, $callback	540	=item aio_lstat $fh, $callback->($status)
160		541
161	Works like perl's C<stat> or C<lstat> in void context. The callback will	542	Works like perl's C<stat> or C<lstat> in void context. The callback will
162	be called after the stat and the results will be available using C<stat _>	543	be called after the stat and the results will be available using C<stat _>
163	or C<-s _> etc...	544	or C<-s _> etc...
164		545
…		…
166	for an explanation.	547	for an explanation.
167		548
168	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	549	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
169	error when stat'ing a large file, the results will be silently truncated	550	error when stat'ing a large file, the results will be silently truncated
170	unless perl itself is compiled with large file support.	551	unless perl itself is compiled with large file support.
		552
		553	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		554	following constants and functions (if not implemented, the constants will
		555	be C<0> and the functions will either C<croak> or fall back on traditional
		556	behaviour).
		557
		558	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		559	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		560	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
171		561
172	Example: Print the length of F</etc/passwd>:	562	Example: Print the length of F</etc/passwd>:
173		563
174	aio_stat "/etc/passwd", sub {	564	aio_stat "/etc/passwd", sub {
175	$_[0] and die "stat failed: $!";	565	$_[0] and die "stat failed: $!";
176	print "size is ", -s _, "\n";	566	print "size is ", -s _, "\n";
177	};	567	};
178		568
		569
		570	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		571
		572	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		573	whether a file handle or path was passed.
		574
		575	On success, the callback is passed a hash reference with the following
		576	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		577	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		578	is passed.
		579
		580	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		581	C<ST_NOSUID>.
		582
		583	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		584	their correct value when available, or to C<0> on systems that do
		585	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		586	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		587	C<ST_NODIRATIME> and C<ST_RELATIME>.
		588
		589	Example: stat C</wd> and dump out the data if successful.
		590
		591	aio_statvfs "/wd", sub {
		592	my $f = $_[0]
		593	or die "statvfs: $!";
		594
		595	use Data::Dumper;
		596	say Dumper $f;
		597	};
		598
		599	# result:
		600	{
		601	bsize => 1024,
		602	bfree => 4333064312,
		603	blocks => 10253828096,
		604	files => 2050765568,
		605	flag => 4096,
		606	favail => 2042092649,
		607	bavail => 4333064312,
		608	ffree => 2042092649,
		609	namemax => 255,
		610	frsize => 1024,
		611	fsid => 1810
		612	}
		613
		614	Here is a (likely partial - send me updates!) list of fsid values used by
		615	Linux - it is safe to hardcode these when C<$^O> is C<linux>:
		616
		617	0x0000adf5 adfs
		618	0x0000adff affs
		619	0x5346414f afs
		620	0x09041934 anon-inode filesystem
		621	0x00000187 autofs
		622	0x42465331 befs
		623	0x1badface bfs
		624	0x42494e4d binfmt_misc
		625	0x9123683e btrfs
		626	0x0027e0eb cgroupfs
		627	0xff534d42 cifs
		628	0x73757245 coda
		629	0x012ff7b7 coh
		630	0x28cd3d45 cramfs
		631	0x453dcd28 cramfs-wend (wrong endianness)
		632	0x64626720 debugfs
		633	0x00001373 devfs
		634	0x00001cd1 devpts
		635	0x0000f15f ecryptfs
		636	0x00414a53 efs
		637	0x0000137d ext
		638	0x0000ef53 ext2/ext3/ext4
		639	0x0000ef51 ext2
		640	0xf2f52010 f2fs
		641	0x00004006 fat
		642	0x65735546 fuseblk
		643	0x65735543 fusectl
		644	0x0bad1dea futexfs
		645	0x01161970 gfs2
		646	0x47504653 gpfs
		647	0x00004244 hfs
		648	0xf995e849 hpfs
		649	0x00c0ffee hostfs
		650	0x958458f6 hugetlbfs
		651	0x2bad1dea inotifyfs
		652	0x00009660 isofs
		653	0x000072b6 jffs2
		654	0x3153464a jfs
		655	0x6b414653 k-afs
		656	0x0bd00bd0 lustre
		657	0x0000137f minix
		658	0x0000138f minix 30 char names
		659	0x00002468 minix v2
		660	0x00002478 minix v2 30 char names
		661	0x00004d5a minix v3
		662	0x19800202 mqueue
		663	0x00004d44 msdos
		664	0x0000564c novell
		665	0x00006969 nfs
		666	0x6e667364 nfsd
		667	0x00003434 nilfs
		668	0x5346544e ntfs
		669	0x00009fa1 openprom
		670	0x7461636F ocfs2
		671	0x00009fa0 proc
		672	0x6165676c pstorefs
		673	0x0000002f qnx4
		674	0x68191122 qnx6
		675	0x858458f6 ramfs
		676	0x52654973 reiserfs
		677	0x00007275 romfs
		678	0x67596969 rpc_pipefs
		679	0x73636673 securityfs
		680	0xf97cff8c selinux
		681	0x0000517b smb
		682	0x534f434b sockfs
		683	0x73717368 squashfs
		684	0x62656572 sysfs
		685	0x012ff7b6 sysv2
		686	0x012ff7b5 sysv4
		687	0x01021994 tmpfs
		688	0x15013346 udf
		689	0x00011954 ufs
		690	0x54190100 ufs byteswapped
		691	0x00009fa2 usbdevfs
		692	0x01021997 v9fs
		693	0xa501fcf5 vxfs
		694	0xabba1974 xenfs
		695	0x012ff7b4 xenix
		696	0x58465342 xfs
		697	0x012fd16d xia
		698
		699	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		700
		701	Works like perl's C<utime> function (including the special case of $atime
		702	and $mtime being undef). Fractional times are supported if the underlying
		703	syscalls support them.
		704
		705	When called with a pathname, uses utimes(2) if available, otherwise
		706	utime(2). If called on a file descriptor, uses futimes(2) if available,
		707	otherwise returns ENOSYS, so this is not portable.
		708
		709	Examples:
		710
		711	# set atime and mtime to current time (basically touch(1)):
		712	aio_utime "path", undef, undef;
		713	# set atime to current time and mtime to beginning of the epoch:
		714	aio_utime "path", time, undef; # undef==0
		715
		716
		717	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		718
		719	Works like perl's C<chown> function, except that C<undef> for either $uid
		720	or $gid is being interpreted as "do not change" (but -1 can also be used).
		721
		722	Examples:
		723
		724	# same as "chown root path" in the shell:
		725	aio_chown "path", 0, -1;
		726	# same as above:
		727	aio_chown "path", 0, undef;
		728
		729
		730	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		731
		732	Works like truncate(2) or ftruncate(2).
		733
		734
		735	=item aio_allocate $fh, $mode, $offset, $len, $callback->($status)
		736
		737	Allocates or frees disk space according to the C<$mode> argument. See the
		738	linux C<fallocate> documentation for details.
		739
		740	C<$mode> is usually C<0> or C<IO::AIO::FALLOC_FL_KEEP_SIZE> to allocate
		741	space, or C<IO::AIO::FALLOC_FL_PUNCH_HOLE | IO::AIO::FALLOC_FL_KEEP_SIZE>,
		742	to deallocate a file range.
		743
		744	IO::AIO also supports C<FALLOC_FL_COLLAPSE_RANGE>, to remove a range
		745	(without leaving a hole), C<FALLOC_FL_ZERO_RANGE>, to zero a range,
		746	C<FALLOC_FL_INSERT_RANGE> to insert a range and C<FALLOC_FL_UNSHARE_RANGE>
		747	to unshare shared blocks (see your L<fallocate(2)> manpage).
		748
		749	The file system block size used by C<fallocate> is presumably the
		750	C<f_bsize> returned by C<statvfs>, but different filesystems and filetypes
		751	can dictate other limitations.
		752
		753	If C<fallocate> isn't available or cannot be emulated (currently no
		754	emulation will be attempted), passes C<-1> and sets C<$!> to C<ENOSYS>.
		755
		756
		757	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		758
		759	Works like perl's C<chmod> function.
		760
		761
179	=item aio_unlink $pathname, $callback	762	=item aio_unlink $pathname, $callback->($status)
180		763
181	Asynchronously unlink (delete) a file and call the callback with the	764	Asynchronously unlink (delete) a file and call the callback with the
182	result code.	765	result code.
183		766
		767
		768	=item aio_mknod $pathname, $mode, $dev, $callback->($status)
		769
		770	[EXPERIMENTAL]
		771
		772	Asynchronously create a device node (or fifo). See mknod(2).
		773
		774	The only (POSIX-) portable way of calling this function is:
		775
		776	aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		777
		778	See C<aio_stat> for info about some potentially helpful extra constants
		779	and functions.
		780
		781	=item aio_link $srcpath, $dstpath, $callback->($status)
		782
		783	Asynchronously create a new link to the existing object at C<$srcpath> at
		784	the path C<$dstpath> and call the callback with the result code.
		785
		786
		787	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		788
		789	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		790	the path C<$dstpath> and call the callback with the result code.
		791
		792
		793	=item aio_readlink $pathname, $callback->($link)
		794
		795	Asynchronously read the symlink specified by C<$path> and pass it to
		796	the callback. If an error occurs, nothing or undef gets passed to the
		797	callback.
		798
		799
		800	=item aio_realpath $pathname, $callback->($path)
		801
		802	Asynchronously make the path absolute and resolve any symlinks in
		803	C<$path>. The resulting path only consists of directories (same as
		804	L<Cwd::realpath>).
		805
		806	This request can be used to get the absolute path of the current working
		807	directory by passing it a path of F<.> (a single dot).
		808
		809
		810	=item aio_rename $srcpath, $dstpath, $callback->($status)
		811
		812	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		813	rename(2) and call the callback with the result code.
		814
		815	On systems that support the AIO::WD working directory abstraction
		816	natively, the case C<[$wd, "."]> as C<$srcpath> is specialcased - instead
		817	of failing, C<rename> is called on the absolute path of C<$wd>.
		818
		819
		820	=item aio_rename2 $srcpath, $dstpath, $flags, $callback->($status)
		821
		822	Basically a version of C<aio_rename> with an additional C<$flags>
		823	argument. Calling this with C<$flags=0> is the same as calling
		824	C<aio_rename>.
		825
		826	Non-zero flags are currently only supported on GNU/Linux systems that
		827	support renameat2. Other systems fail with C<ENOSYS> in this case.
		828
		829	The following constants are available (missing ones are, as usual C<0>),
		830	see renameat2(2) for details:
		831
		832	C<IO::AIO::RENAME_NOREPLACE>, C<IO::AIO::RENAME_EXCHANGE>
		833	and C<IO::AIO::RENAME_WHITEOUT>.
		834
		835
		836	=item aio_mkdir $pathname, $mode, $callback->($status)
		837
		838	Asynchronously mkdir (create) a directory and call the callback with
		839	the result code. C<$mode> will be modified by the umask at the time the
		840	request is executed, so do not change your umask.
		841
		842
		843	=item aio_rmdir $pathname, $callback->($status)
		844
		845	Asynchronously rmdir (delete) a directory and call the callback with the
		846	result code.
		847
		848	On systems that support the AIO::WD working directory abstraction
		849	natively, the case C<[$wd, "."]> is specialcased - instead of failing,
		850	C<rmdir> is called on the absolute path of C<$wd>.
		851
		852
		853	=item aio_readdir $pathname, $callback->($entries)
		854
		855	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
		856	directory (i.e. opendir + readdir + closedir). The entries will not be
		857	sorted, and will B<NOT> include the C<.> and C<..> entries.
		858
		859	The callback is passed a single argument which is either C<undef> or an
		860	array-ref with the filenames.
		861
		862
		863	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		864
		865	Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
		866	tune behaviour and output format. In case of an error, C<$entries> will be
		867	C<undef>.
		868
		869	The flags are a combination of the following constants, ORed together (the
		870	flags will also be passed to the callback, possibly modified):
		871
		872	=over 4
		873
		874	=item IO::AIO::READDIR_DENTS
		875
		876	When this flag is off, then the callback gets an arrayref consisting of
		877	names only (as with C<aio_readdir>), otherwise it gets an arrayref with
		878	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		879	entry in more detail.
		880
		881	C<$name> is the name of the entry.
		882
		883	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		884
		885	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		886	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		887	C<IO::AIO::DT_WHT>.
		888
		889	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		890	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		891	scalars are read-only: you can not modify them.
		892
		893	C<$inode> is the inode number (which might not be exact on systems with 64
		894	bit inode numbers and 32 bit perls). This field has unspecified content on
		895	systems that do not deliver the inode information.
		896
		897	=item IO::AIO::READDIR_DIRS_FIRST
		898
		899	When this flag is set, then the names will be returned in an order where
		900	likely directories come first, in optimal stat order. This is useful when
		901	you need to quickly find directories, or you want to find all directories
		902	while avoiding to stat() each entry.
		903
		904	If the system returns type information in readdir, then this is used
		905	to find directories directly. Otherwise, likely directories are names
		906	beginning with ".", or otherwise names with no dots, of which names with
		907	short names are tried first.
		908
		909	=item IO::AIO::READDIR_STAT_ORDER
		910
		911	When this flag is set, then the names will be returned in an order
		912	suitable for stat()'ing each one. That is, when you plan to stat()
		913	all files in the given directory, then the returned order will likely
		914	be fastest.
		915
		916	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		917	the likely dirs come first, resulting in a less optimal stat order.
		918
		919	=item IO::AIO::READDIR_FOUND_UNKNOWN
		920
		921	This flag should not be set when calling C<aio_readdirx>. Instead, it
		922	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		923	C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
		924	C<$type>'s are known, which can be used to speed up some algorithms.
		925
		926	=back
		927
		928
		929	=item aio_load $pathname, $data, $callback->($status)
		930
		931	This is a composite request that tries to fully load the given file into
		932	memory. Status is the same as with aio_read.
		933
		934	=cut
		935
		936	sub aio_load($$;$) {
		937	my ($path, undef, $cb) = @_;
		938	my $data = \$_[1];
		939
		940	my $pri = aioreq_pri;
		941	my $grp = aio_group $cb;
		942
		943	aioreq_pri $pri;
		944	add $grp aio_open $path, O_RDONLY, 0, sub {
		945	my $fh = shift
		946	or return $grp->result (-1);
		947
		948	aioreq_pri $pri;
		949	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		950	$grp->result ($_[0]);
		951	};
		952	};
		953
		954	$grp
		955	}
		956
		957	=item aio_copy $srcpath, $dstpath, $callback->($status)
		958
		959	Try to copy the I<file> (directories not supported as either source or
		960	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		961	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		962
		963	This is a composite request that creates the destination file with
		964	mode 0200 and copies the contents of the source file into it using
		965	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		966	uid/gid, in that order.
		967
		968	If an error occurs, the partial destination file will be unlinked, if
		969	possible, except when setting atime, mtime, access mode and uid/gid, where
		970	errors are being ignored.
		971
		972	=cut
		973
		974	sub aio_copy($$;$) {
		975	my ($src, $dst, $cb) = @_;
		976
		977	my $pri = aioreq_pri;
		978	my $grp = aio_group $cb;
		979
		980	aioreq_pri $pri;
		981	add $grp aio_open $src, O_RDONLY, 0, sub {
		982	if (my $src_fh = $_[0]) {
		983	my @stat = stat $src_fh; # hmm, might block over nfs?
		984
		985	aioreq_pri $pri;
		986	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		987	if (my $dst_fh = $_[0]) {
		988	aioreq_pri $pri;
		989	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		990	if ($_[0] == $stat[7]) {
		991	$grp->result (0);
		992	close $src_fh;
		993
		994	my $ch = sub {
		995	aioreq_pri $pri;
		996	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		997	aioreq_pri $pri;
		998	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		999	aioreq_pri $pri;
		1000	add $grp aio_close $dst_fh;
		1001	}
		1002	};
		1003	};
		1004
		1005	aioreq_pri $pri;
		1006	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		1007	if ($_[0] < 0 && $! == ENOSYS) {
		1008	aioreq_pri $pri;
		1009	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		1010	} else {
		1011	$ch->();
		1012	}
		1013	};
		1014	} else {
		1015	$grp->result (-1);
		1016	close $src_fh;
		1017	close $dst_fh;
		1018
		1019	aioreq $pri;
		1020	add $grp aio_unlink $dst;
		1021	}
		1022	};
		1023	} else {
		1024	$grp->result (-1);
		1025	}
		1026	},
		1027
		1028	} else {
		1029	$grp->result (-1);
		1030	}
		1031	};
		1032
		1033	$grp
		1034	}
		1035
		1036	=item aio_move $srcpath, $dstpath, $callback->($status)
		1037
		1038	Try to move the I<file> (directories not supported as either source or
		1039	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		1040	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		1041
		1042	This is a composite request that tries to rename(2) the file first; if
		1043	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		1044	that is successful, unlinks the C<$srcpath>.
		1045
		1046	=cut
		1047
		1048	sub aio_move($$;$) {
		1049	my ($src, $dst, $cb) = @_;
		1050
		1051	my $pri = aioreq_pri;
		1052	my $grp = aio_group $cb;
		1053
		1054	aioreq_pri $pri;
		1055	add $grp aio_rename $src, $dst, sub {
		1056	if ($_[0] && $! == EXDEV) {
		1057	aioreq_pri $pri;
		1058	add $grp aio_copy $src, $dst, sub {
		1059	$grp->result ($_[0]);
		1060
		1061	unless ($_[0]) {
		1062	aioreq_pri $pri;
		1063	add $grp aio_unlink $src;
		1064	}
		1065	};
		1066	} else {
		1067	$grp->result ($_[0]);
		1068	}
		1069	};
		1070
		1071	$grp
		1072	}
		1073
		1074	=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		1075
		1076	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		1077	efficiently separate the entries of directory C<$path> into two sets of
		1078	names, directories you can recurse into (directories), and ones you cannot
		1079	recurse into (everything else, including symlinks to directories).
		1080
		1081	C<aio_scandir> is a composite request that creates of many sub requests_
		1082	C<$maxreq> specifies the maximum number of outstanding aio requests that
		1083	this function generates. If it is C<< <= 0 >>, then a suitable default
		1084	will be chosen (currently 4).
		1085
		1086	On error, the callback is called without arguments, otherwise it receives
		1087	two array-refs with path-relative entry names.
		1088
		1089	Example:
		1090
		1091	aio_scandir $dir, 0, sub {
		1092	my ($dirs, $nondirs) = @_;
		1093	print "real directories: @$dirs\n";
		1094	print "everything else: @$nondirs\n";
		1095	};
		1096
		1097	Implementation notes.
		1098
		1099	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		1100
		1101	If readdir returns file type information, then this is used directly to
		1102	find directories.
		1103
		1104	Otherwise, after reading the directory, the modification time, size etc.
		1105	of the directory before and after the readdir is checked, and if they
		1106	match (and isn't the current time), the link count will be used to decide
		1107	how many entries are directories (if >= 2). Otherwise, no knowledge of the
		1108	number of subdirectories will be assumed.
		1109
		1110	Then entries will be sorted into likely directories a non-initial dot
		1111	currently) and likely non-directories (see C<aio_readdirx>). Then every
		1112	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
		1113	in order of their inode numbers. If that succeeds, it assumes that the
		1114	entry is a directory or a symlink to directory (which will be checked
		1115	separately). This is often faster than stat'ing the entry itself because
		1116	filesystems might detect the type of the entry without reading the inode
		1117	data (e.g. ext2fs filetype feature), even on systems that cannot return
		1118	the filetype information on readdir.
		1119
		1120	If the known number of directories (link count - 2) has been reached, the
		1121	rest of the entries is assumed to be non-directories.
		1122
		1123	This only works with certainty on POSIX (= UNIX) filesystems, which
		1124	fortunately are the vast majority of filesystems around.
		1125
		1126	It will also likely work on non-POSIX filesystems with reduced efficiency
		1127	as those tend to return 0 or 1 as link counts, which disables the
		1128	directory counting heuristic.
		1129
		1130	=cut
		1131
		1132	sub aio_scandir($$;$) {
		1133	my ($path, $maxreq, $cb) = @_;
		1134
		1135	my $pri = aioreq_pri;
		1136
		1137	my $grp = aio_group $cb;
		1138
		1139	$maxreq = 4 if $maxreq <= 0;
		1140
		1141	# get a wd object
		1142	aioreq_pri $pri;
		1143	add $grp aio_wd $path, sub {
		1144	$_[0]
		1145	or return $grp->result ();
		1146
		1147	my $wd = [shift, "."];
		1148
		1149	# stat once
		1150	aioreq_pri $pri;
		1151	add $grp aio_stat $wd, sub {
		1152	return $grp->result () if $_[0];
		1153	my $now = time;
		1154	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		1155
		1156	# read the directory entries
		1157	aioreq_pri $pri;
		1158	add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
		1159	my $entries = shift
		1160	or return $grp->result ();
		1161
		1162	# stat the dir another time
		1163	aioreq_pri $pri;
		1164	add $grp aio_stat $wd, sub {
		1165	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		1166
		1167	my $ndirs;
		1168
		1169	# take the slow route if anything looks fishy
		1170	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		1171	$ndirs = -1;
		1172	} else {
		1173	# if nlink == 2, we are finished
		1174	# for non-posix-fs's, we rely on nlink < 2
		1175	$ndirs = (stat _)[3] - 2
		1176	or return $grp->result ([], $entries);
		1177	}
		1178
		1179	my (@dirs, @nondirs);
		1180
		1181	my $statgrp = add $grp aio_group sub {
		1182	$grp->result (\@dirs, \@nondirs);
		1183	};
		1184
		1185	limit $statgrp $maxreq;
		1186	feed $statgrp sub {
		1187	return unless @$entries;
		1188	my $entry = shift @$entries;
		1189
		1190	aioreq_pri $pri;
		1191	$wd->[1] = "$entry/.";
		1192	add $statgrp aio_stat $wd, sub {
		1193	if ($_[0] < 0) {
		1194	push @nondirs, $entry;
		1195	} else {
		1196	# need to check for real directory
		1197	aioreq_pri $pri;
		1198	$wd->[1] = $entry;
		1199	add $statgrp aio_lstat $wd, sub {
		1200	if (-d _) {
		1201	push @dirs, $entry;
		1202
		1203	unless (--$ndirs) {
		1204	push @nondirs, @$entries;
		1205	feed $statgrp;
		1206	}
		1207	} else {
		1208	push @nondirs, $entry;
		1209	}
		1210	}
		1211	}
		1212	};
		1213	};
		1214	};
		1215	};
		1216	};
		1217	};
		1218
		1219	$grp
		1220	}
		1221
		1222	=item aio_rmtree $pathname, $callback->($status)
		1223
		1224	Delete a directory tree starting (and including) C<$path>, return the
		1225	status of the final C<rmdir> only. This is a composite request that
		1226	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1227	everything else.
		1228
		1229	=cut
		1230
		1231	sub aio_rmtree;
		1232	sub aio_rmtree($;$) {
		1233	my ($path, $cb) = @_;
		1234
		1235	my $pri = aioreq_pri;
		1236	my $grp = aio_group $cb;
		1237
		1238	aioreq_pri $pri;
		1239	add $grp aio_scandir $path, 0, sub {
		1240	my ($dirs, $nondirs) = @_;
		1241
		1242	my $dirgrp = aio_group sub {
		1243	add $grp aio_rmdir $path, sub {
		1244	$grp->result ($_[0]);
		1245	};
		1246	};
		1247
		1248	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1249	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1250
		1251	add $grp $dirgrp;
		1252	};
		1253
		1254	$grp
		1255	}
		1256
		1257	=item aio_fcntl $fh, $cmd, $arg, $callback->($status)
		1258
		1259	=item aio_ioctl $fh, $request, $buf, $callback->($status)
		1260
		1261	These work just like the C<fcntl> and C<ioctl> built-in functions, except
		1262	they execute asynchronously and pass the return value to the callback.
		1263
		1264	Both calls can be used for a lot of things, some of which make more sense
		1265	to run asynchronously in their own thread, while some others make less
		1266	sense. For example, calls that block waiting for external events, such
		1267	as locking, will also lock down an I/O thread while it is waiting, which
		1268	can deadlock the whole I/O system. At the same time, there might be no
		1269	alternative to using a thread to wait.
		1270
		1271	So in general, you should only use these calls for things that do
		1272	(filesystem) I/O, not for things that wait for other events (network,
		1273	other processes), although if you are careful and know what you are doing,
		1274	you still can.
		1275
		1276	The following constants are available (missing ones are, as usual C<0>):
		1277
		1278	C<F_DUPFD_CLOEXEC>,
		1279
		1280	C<F_OFD_GETLK>, C<F_OFD_SETLK>, C<F_OFD_GETLKW>,
		1281
		1282	C<FIFREEZE>, C<FITHAW>, C<FITRIM>, C<FICLONE>, C<FICLONERANGE>, C<FIDEDUPERANGE>.
		1283
		1284	C<FS_IOC_GETFLAGS>, C<FS_IOC_SETFLAGS>, C<FS_IOC_GETVERSION>, C<FS_IOC_SETVERSION>,
		1285	C<FS_IOC_FIEMAP>.
		1286
		1287	C<FS_IOC_FSGETXATTR>, C<FS_IOC_FSSETXATTR>, C<FS_IOC_SET_ENCRYPTION_POLICY>,
		1288	C<FS_IOC_GET_ENCRYPTION_PWSALT>, C<FS_IOC_GET_ENCRYPTION_POLICY>, C<FS_KEY_DESCRIPTOR_SIZE>.
		1289
		1290	C<FS_SECRM_FL>, C<FS_UNRM_FL>, C<FS_COMPR_FL>, C<FS_SYNC_FL>, C<FS_IMMUTABLE_FL>,
		1291	C<FS_APPEND_FL>, C<FS_NODUMP_FL>, C<FS_NOATIME_FL>, C<FS_DIRTY_FL>,
		1292	C<FS_COMPRBLK_FL>, C<FS_NOCOMP_FL>, C<FS_ENCRYPT_FL>, C<FS_BTREE_FL>,
		1293	C<FS_INDEX_FL>, C<FS_JOURNAL_DATA_FL>, C<FS_NOTAIL_FL>, C<FS_DIRSYNC_FL>, C<FS_TOPDIR_FL>,
		1294	C<FS_FL_USER_MODIFIABLE>.
		1295
		1296	C<FS_XFLAG_REALTIME>, C<FS_XFLAG_PREALLOC>, C<FS_XFLAG_IMMUTABLE>, C<FS_XFLAG_APPEND>,
		1297	C<FS_XFLAG_SYNC>, C<FS_XFLAG_NOATIME>, C<FS_XFLAG_NODUMP>, C<FS_XFLAG_RTINHERIT>,
		1298	C<FS_XFLAG_PROJINHERIT>, C<FS_XFLAG_NOSYMLINKS>, C<FS_XFLAG_EXTSIZE>, C<FS_XFLAG_EXTSZINHERIT>,
		1299	C<FS_XFLAG_NODEFRAG>, C<FS_XFLAG_FILESTREAM>, C<FS_XFLAG_DAX>, C<FS_XFLAG_HASATTR>,
		1300
		1301	=item aio_sync $callback->($status)
		1302
		1303	Asynchronously call sync and call the callback when finished.
		1304
184	=item aio_fsync $fh, $callback	1305	=item aio_fsync $fh, $callback->($status)
185		1306
186	Asynchronously call fsync on the given filehandle and call the callback	1307	Asynchronously call fsync on the given filehandle and call the callback
187	with the fsync result code.	1308	with the fsync result code.
188		1309
189	=item aio_fdatasync $fh, $callback	1310	=item aio_fdatasync $fh, $callback->($status)
190		1311
191	Asynchronously call fdatasync on the given filehandle and call the	1312	Asynchronously call fdatasync on the given filehandle and call the
192	callback with the fdatasync result code.	1313	callback with the fdatasync result code.
193		1314
		1315	If this call isn't available because your OS lacks it or it couldn't be
		1316	detected, it will be emulated by calling C<fsync> instead.
		1317
		1318	=item aio_syncfs $fh, $callback->($status)
		1319
		1320	Asynchronously call the syncfs syscall to sync the filesystem associated
		1321	to the given filehandle and call the callback with the syncfs result
		1322	code. If syncfs is not available, calls sync(), but returns C<-1> and sets
		1323	errno to C<ENOSYS> nevertheless.
		1324
		1325	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1326
		1327	Sync the data portion of the file specified by C<$offset> and C<$length>
		1328	to disk (but NOT the metadata), by calling the Linux-specific
		1329	sync_file_range call. If sync_file_range is not available or it returns
		1330	ENOSYS, then fdatasync or fsync is being substituted.
		1331
		1332	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1333	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1334	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1335	manpage for details.
		1336
		1337	=item aio_pathsync $pathname, $callback->($status)
		1338
		1339	This request tries to open, fsync and close the given path. This is a
		1340	composite request intended to sync directories after directory operations
		1341	(E.g. rename). This might not work on all operating systems or have any
		1342	specific effect, but usually it makes sure that directory changes get
		1343	written to disc. It works for anything that can be opened for read-only,
		1344	not just directories.
		1345
		1346	Future versions of this function might fall back to other methods when
		1347	C<fsync> on the directory fails (such as calling C<sync>).
		1348
		1349	Passes C<0> when everything went ok, and C<-1> on error.
		1350
		1351	=cut
		1352
		1353	sub aio_pathsync($;$) {
		1354	my ($path, $cb) = @_;
		1355
		1356	my $pri = aioreq_pri;
		1357	my $grp = aio_group $cb;
		1358
		1359	aioreq_pri $pri;
		1360	add $grp aio_open $path, O_RDONLY, 0, sub {
		1361	my ($fh) = @_;
		1362	if ($fh) {
		1363	aioreq_pri $pri;
		1364	add $grp aio_fsync $fh, sub {
		1365	$grp->result ($_[0]);
		1366
		1367	aioreq_pri $pri;
		1368	add $grp aio_close $fh;
		1369	};
		1370	} else {
		1371	$grp->result (-1);
		1372	}
		1373	};
		1374
		1375	$grp
		1376	}
		1377
		1378	=item aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		1379
		1380	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1381	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1382	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1383	scalar must only be modified in-place while an aio operation is pending on
		1384	it).
		1385
		1386	It calls the C<msync> function of your OS, if available, with the memory
		1387	area starting at C<$offset> in the string and ending C<$length> bytes
		1388	later. If C<$length> is negative, counts from the end, and if C<$length>
		1389	is C<undef>, then it goes till the end of the string. The flags can be
		1390	either C<IO::AIO::MS_ASYNC> or C<IO::AIO::MS_SYNC>, plus an optional
		1391	C<IO::AIO::MS_INVALIDATE>.
		1392
		1393	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1394
		1395	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1396	scalars.
		1397
		1398	It touches (reads or writes) all memory pages in the specified
		1399	range inside the scalar. All caveats and parameters are the same
		1400	as for C<aio_msync>, above, except for flags, which must be either
		1401	C<0> (which reads all pages and ensures they are instantiated) or
		1402	C<IO::AIO::MT_MODIFY>, which modifies the memory pages (by reading and
		1403	writing an octet from it, which dirties the page).
		1404
		1405	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1406
		1407	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1408	scalars.
		1409
		1410	It reads in all the pages of the underlying storage into memory (if any)
		1411	and locks them, so they are not getting swapped/paged out or removed.
		1412
		1413	If C<$length> is undefined, then the scalar will be locked till the end.
		1414
		1415	On systems that do not implement C<mlock>, this function returns C<-1>
		1416	and sets errno to C<ENOSYS>.
		1417
		1418	Note that the corresponding C<munlock> is synchronous and is
		1419	documented under L<MISCELLANEOUS FUNCTIONS>.
		1420
		1421	Example: open a file, mmap and mlock it - both will be undone when
		1422	C<$data> gets destroyed.
		1423
		1424	open my $fh, "<", $path or die "$path: $!";
		1425	my $data;
		1426	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1427	aio_mlock $data; # mlock in background
		1428
		1429	=item aio_mlockall $flags, $callback->($status)
		1430
		1431	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1432	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1433
		1434	On systems that do not implement C<mlockall>, this function returns C<-1>
		1435	and sets errno to C<ENOSYS>.
		1436
		1437	Note that the corresponding C<munlockall> is synchronous and is
		1438	documented under L<MISCELLANEOUS FUNCTIONS>.
		1439
		1440	Example: asynchronously lock all current and future pages into memory.
		1441
		1442	aio_mlockall IO::AIO::MCL_FUTURE;
		1443
		1444	=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		1445
		1446	Queries the extents of the given file (by calling the Linux C<FIEMAP>
		1447	ioctl, see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If
		1448	the ioctl is not available on your OS, then this request will fail with
		1449	C<ENOSYS>.
		1450
		1451	C<$start> is the starting offset to query extents for, C<$length> is the
		1452	size of the range to query - if it is C<undef>, then the whole file will
		1453	be queried.
		1454
		1455	C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
		1456	C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
		1457	exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
		1458	the data portion.
		1459
		1460	C<$count> is the maximum number of extent records to return. If it is
		1461	C<undef>, then IO::AIO queries all extents of the range. As a very special
		1462	case, if it is C<0>, then the callback receives the number of extents
		1463	instead of the extents themselves (which is unreliable, see below).
		1464
		1465	If an error occurs, the callback receives no arguments. The special
		1466	C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
		1467
		1468	Otherwise, the callback receives an array reference with extent
		1469	structures. Each extent structure is an array reference itself, with the
		1470	following members:
		1471
		1472	[$logical, $physical, $length, $flags]
		1473
		1474	Flags is any combination of the following flag values (typically either C<0>
		1475	or C<IO::AIO::FIEMAP_EXTENT_LAST> (1)):
		1476
		1477	C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
		1478	C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
		1479	C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
		1480	C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
		1481	C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
		1482	C<IO::AIO::FIEMAP_EXTENT_SHARED>.
		1483
		1484	At the time of this writing (Linux 3.2), this requets is unreliable unless
		1485	C<$count> is C<undef>, as the kernel has all sorts of bugs preventing
		1486	it to return all extents of a range for files with large number of
		1487	extents. The code works around all these issues if C<$count> is undef.
		1488
		1489	=item aio_group $callback->(...)
		1490
		1491	This is a very special aio request: Instead of doing something, it is a
		1492	container for other aio requests, which is useful if you want to bundle
		1493	many requests into a single, composite, request with a definite callback
		1494	and the ability to cancel the whole request with its subrequests.
		1495
		1496	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		1497	for more info.
		1498
		1499	Example:
		1500
		1501	my $grp = aio_group sub {
		1502	print "all stats done\n";
		1503	};
		1504
		1505	add $grp
		1506	(aio_stat ...),
		1507	(aio_stat ...),
		1508	...;
		1509
		1510	=item aio_nop $callback->()
		1511
		1512	This is a special request - it does nothing in itself and is only used for
		1513	side effects, such as when you want to add a dummy request to a group so
		1514	that finishing the requests in the group depends on executing the given
		1515	code.
		1516
		1517	While this request does nothing, it still goes through the execution
		1518	phase and still requires a worker thread. Thus, the callback will not
		1519	be executed immediately but only after other requests in the queue have
		1520	entered their execution phase. This can be used to measure request
		1521	latency.
		1522
		1523	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		1524
		1525	Mainly used for debugging and benchmarking, this aio request puts one of
		1526	the request workers to sleep for the given time.
		1527
		1528	While it is theoretically handy to have simple I/O scheduling requests
		1529	like sleep and file handle readable/writable, the overhead this creates is
		1530	immense (it blocks a thread for a long time) so do not use this function
		1531	except to put your application under artificial I/O pressure.
		1532
194	=back	1533	=back
195		1534
		1535
		1536	=head2 IO::AIO::WD - multiple working directories
		1537
		1538	Your process only has one current working directory, which is used by all
		1539	threads. This makes it hard to use relative paths (some other component
		1540	could call C<chdir> at any time, and it is hard to control when the path
		1541	will be used by IO::AIO).
		1542
		1543	One solution for this is to always use absolute paths. This usually works,
		1544	but can be quite slow (the kernel has to walk the whole path on every
		1545	access), and can also be a hassle to implement.
		1546
		1547	Newer POSIX systems have a number of functions (openat, fdopendir,
		1548	futimensat and so on) that make it possible to specify working directories
		1549	per operation.
		1550
		1551	For portability, and because the clowns who "designed", or shall I write,
		1552	perpetrated this new interface were obviously half-drunk, this abstraction
		1553	cannot be perfect, though.
		1554
		1555	IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
		1556	object. This object stores the canonicalised, absolute version of the
		1557	path, and on systems that allow it, also a directory file descriptor.
		1558
		1559	Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
		1560	or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
		1561	object and a pathname instead (or the IO::AIO::WD object alone, which
		1562	gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
		1563	IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
		1564	to that IO::AIO::WD object.
		1565
		1566	For example, to get a wd object for F</etc> and then stat F<passwd>
		1567	inside, you would write:
		1568
		1569	aio_wd "/etc", sub {
		1570	my $etcdir = shift;
		1571
		1572	# although $etcdir can be undef on error, there is generally no reason
		1573	# to check for errors here, as aio_stat will fail with ENOENT
		1574	# when $etcdir is undef.
		1575
		1576	aio_stat [$etcdir, "passwd"], sub {
		1577	# yay
		1578	};
		1579	};
		1580
		1581	The fact that C<aio_wd> is a request and not a normal function shows that
		1582	creating an IO::AIO::WD object is itself a potentially blocking operation,
		1583	which is why it is done asynchronously.
		1584
		1585	To stat the directory obtained with C<aio_wd> above, one could write
		1586	either of the following three request calls:
		1587
		1588	aio_lstat "/etc" , sub { ... # pathname as normal string
		1589	aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
		1590	aio_lstat $wd , sub { ... # shorthand for the previous
		1591
		1592	As with normal pathnames, IO::AIO keeps a copy of the working directory
		1593	object and the pathname string, so you could write the following without
		1594	causing any issues due to C<$path> getting reused:
		1595
		1596	my $path = [$wd, undef];
		1597
		1598	for my $name (qw(abc def ghi)) {
		1599	$path->[1] = $name;
		1600	aio_stat $path, sub {
		1601	# ...
		1602	};
		1603	}
		1604
		1605	There are some caveats: when directories get renamed (or deleted), the
		1606	pathname string doesn't change, so will point to the new directory (or
		1607	nowhere at all), while the directory fd, if available on the system,
		1608	will still point to the original directory. Most functions accepting a
		1609	pathname will use the directory fd on newer systems, and the string on
		1610	older systems. Some functions (such as realpath) will always rely on the
		1611	string form of the pathname.
		1612
		1613	So this functionality is mainly useful to get some protection against
		1614	C<chdir>, to easily get an absolute path out of a relative path for future
		1615	reference, and to speed up doing many operations in the same directory
		1616	(e.g. when stat'ing all files in a directory).
		1617
		1618	The following functions implement this working directory abstraction:
		1619
		1620	=over 4
		1621
		1622	=item aio_wd $pathname, $callback->($wd)
		1623
		1624	Asynchonously canonicalise the given pathname and convert it to an
		1625	IO::AIO::WD object representing it. If possible and supported on the
		1626	system, also open a directory fd to speed up pathname resolution relative
		1627	to this working directory.
		1628
		1629	If something goes wrong, then C<undef> is passwd to the callback instead
		1630	of a working directory object and C<$!> is set appropriately. Since
		1631	passing C<undef> as working directory component of a pathname fails the
		1632	request with C<ENOENT>, there is often no need for error checking in the
		1633	C<aio_wd> callback, as future requests using the value will fail in the
		1634	expected way.
		1635
		1636	=item IO::AIO::CWD
		1637
		1638	This is a compiletime constant (object) that represents the process
		1639	current working directory.
		1640
		1641	Specifying this object as working directory object for a pathname is as if
		1642	the pathname would be specified directly, without a directory object. For
		1643	example, these calls are functionally identical:
		1644
		1645	aio_stat "somefile", sub { ... };
		1646	aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
		1647
		1648	=back
		1649
		1650	To recover the path associated with an IO::AIO::WD object, you can use
		1651	C<aio_realpath>:
		1652
		1653	aio_realpath $wd, sub {
		1654	warn "path is $_[0]\n";
		1655	};
		1656
		1657	Currently, C<aio_statvfs> always, and C<aio_rename> and C<aio_rmdir>
		1658	sometimes, fall back to using an absolue path.
		1659
		1660	=head2 IO::AIO::REQ CLASS
		1661
		1662	All non-aggregate C<aio_*> functions return an object of this class when
		1663	called in non-void context.
		1664
		1665	=over 4
		1666
		1667	=item cancel $req
		1668
		1669	Cancels the request, if possible. Has the effect of skipping execution
		1670	when entering the B<execute> state and skipping calling the callback when
		1671	entering the the B<result> state, but will leave the request otherwise
		1672	untouched (with the exception of readdir). That means that requests that
		1673	currently execute will not be stopped and resources held by the request
		1674	will not be freed prematurely.
		1675
		1676	=item cb $req $callback->(...)
		1677
		1678	Replace (or simply set) the callback registered to the request.
		1679
		1680	=back
		1681
		1682	=head2 IO::AIO::GRP CLASS
		1683
		1684	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		1685	objects of this class, too.
		1686
		1687	A IO::AIO::GRP object is a special request that can contain multiple other
		1688	aio requests.
		1689
		1690	You create one by calling the C<aio_group> constructing function with a
		1691	callback that will be called when all contained requests have entered the
		1692	C<done> state:
		1693
		1694	my $grp = aio_group sub {
		1695	print "all requests are done\n";
		1696	};
		1697
		1698	You add requests by calling the C<add> method with one or more
		1699	C<IO::AIO::REQ> objects:
		1700
		1701	$grp->add (aio_unlink "...");
		1702
		1703	add $grp aio_stat "...", sub {
		1704	$_[0] or return $grp->result ("error");
		1705
		1706	# add another request dynamically, if first succeeded
		1707	add $grp aio_open "...", sub {
		1708	$grp->result ("ok");
		1709	};
		1710	};
		1711
		1712	This makes it very easy to create composite requests (see the source of
		1713	C<aio_move> for an application) that work and feel like simple requests.
		1714
		1715	=over 4
		1716
		1717	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		1718	C<IO::AIO::poll_cb>, just like any other request.
		1719
		1720	=item * They can be canceled like any other request. Canceling will cancel not
		1721	only the request itself, but also all requests it contains.
		1722
		1723	=item * They can also can also be added to other IO::AIO::GRP objects.
		1724
		1725	=item * You must not add requests to a group from within the group callback (or
		1726	any later time).
		1727
		1728	=back
		1729
		1730	Their lifetime, simplified, looks like this: when they are empty, they
		1731	will finish very quickly. If they contain only requests that are in the
		1732	C<done> state, they will also finish. Otherwise they will continue to
		1733	exist.
		1734
		1735	That means after creating a group you have some time to add requests
		1736	(precisely before the callback has been invoked, which is only done within
		1737	the C<poll_cb>). And in the callbacks of those requests, you can add
		1738	further requests to the group. And only when all those requests have
		1739	finished will the the group itself finish.
		1740
		1741	=over 4
		1742
		1743	=item add $grp ...
		1744
		1745	=item $grp->add (...)
		1746
		1747	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1748	be added, including other groups, as long as you do not create circular
		1749	dependencies.
		1750
		1751	Returns all its arguments.
		1752
		1753	=item $grp->cancel_subs
		1754
		1755	Cancel all subrequests and clears any feeder, but not the group request
		1756	itself. Useful when you queued a lot of events but got a result early.
		1757
		1758	The group request will finish normally (you cannot add requests to the
		1759	group).
		1760
		1761	=item $grp->result (...)
		1762
		1763	Set the result value(s) that will be passed to the group callback when all
		1764	subrequests have finished and set the groups errno to the current value
		1765	of errno (just like calling C<errno> without an error number). By default,
		1766	no argument will be passed and errno is zero.
		1767
		1768	=item $grp->errno ([$errno])
		1769
		1770	Sets the group errno value to C<$errno>, or the current value of errno
		1771	when the argument is missing.
		1772
		1773	Every aio request has an associated errno value that is restored when
		1774	the callback is invoked. This method lets you change this value from its
		1775	default (0).
		1776
		1777	Calling C<result> will also set errno, so make sure you either set C<$!>
		1778	before the call to C<result>, or call c<errno> after it.
		1779
		1780	=item feed $grp $callback->($grp)
		1781
		1782	Sets a feeder/generator on this group: every group can have an attached
		1783	generator that generates requests if idle. The idea behind this is that,
		1784	although you could just queue as many requests as you want in a group,
		1785	this might starve other requests for a potentially long time. For example,
		1786	C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
		1787	requests, delaying any later requests for a long time.
		1788
		1789	To avoid this, and allow incremental generation of requests, you can
		1790	instead a group and set a feeder on it that generates those requests. The
		1791	feed callback will be called whenever there are few enough (see C<limit>,
		1792	below) requests active in the group itself and is expected to queue more
		1793	requests.
		1794
		1795	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1796	not impose any limits).
		1797
		1798	If the feed does not queue more requests when called, it will be
		1799	automatically removed from the group.
		1800
		1801	If the feed limit is C<0> when this method is called, it will be set to
		1802	C<2> automatically.
		1803
		1804	Example:
		1805
		1806	# stat all files in @files, but only ever use four aio requests concurrently:
		1807
		1808	my $grp = aio_group sub { print "finished\n" };
		1809	limit $grp 4;
		1810	feed $grp sub {
		1811	my $file = pop @files
		1812	or return;
		1813
		1814	add $grp aio_stat $file, sub { ... };
		1815	};
		1816
		1817	=item limit $grp $num
		1818
		1819	Sets the feeder limit for the group: The feeder will be called whenever
		1820	the group contains less than this many requests.
		1821
		1822	Setting the limit to C<0> will pause the feeding process.
		1823
		1824	The default value for the limit is C<0>, but note that setting a feeder
		1825	automatically bumps it up to C<2>.
		1826
		1827	=back
		1828
196	=head2 SUPPORT FUNCTIONS	1829	=head2 SUPPORT FUNCTIONS
197		1830
		1831	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1832
198	=over 4	1833	=over 4
199		1834
200	=item $fileno = IO::AIO::poll_fileno	1835	=item $fileno = IO::AIO::poll_fileno
201		1836
202	Return the I<request result pipe filehandle>. This filehandle must be	1837	Return the I<request result pipe file descriptor>. This filehandle must be
203	polled for reading by some mechanism outside this module (e.g. Event	1838	polled for reading by some mechanism outside this module (e.g. EV, Glib,
204	or select, see below). If the pipe becomes readable you have to call	1839	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
205	C<poll_cb> to check the results.	1840	you have to call C<poll_cb> to check the results.
206		1841
207	See C<poll_cb> for an example.	1842	See C<poll_cb> for an example.
208		1843
209	=item IO::AIO::poll_cb	1844	=item IO::AIO::poll_cb
210		1845
211	Process all outstanding events on the result pipe. You have to call this	1846	Process some requests that have reached the result phase (i.e. they have
212	regularly. Returns the number of events processed. Returns immediately	1847	been executed but the results are not yet reported). You have to call
213	when no events are outstanding.	1848	this "regularly" to finish outstanding requests.
214		1849
215	You can use Event to multiplex, e.g.:	1850	Returns C<0> if all events could be processed (or there were no
		1851	events to process), or C<-1> if it returned earlier for whatever
		1852	reason. Returns immediately when no events are outstanding. The amount
		1853	of events processed depends on the settings of C<IO::AIO::max_poll_req>,
		1854	C<IO::AIO::max_poll_time> and C<IO::AIO::max_outstanding>.
		1855
		1856	If not all requests were processed for whatever reason, the poll file
		1857	descriptor will still be ready when C<poll_cb> returns, so normally you
		1858	don't have to do anything special to have it called later.
		1859
		1860	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1861	ready, it can be beneficial to call this function from loops which submit
		1862	a lot of requests, to make sure the results get processed when they become
		1863	available and not just when the loop is finished and the event loop takes
		1864	over again. This function returns very fast when there are no outstanding
		1865	requests.
		1866
		1867	Example: Install an Event watcher that automatically calls
		1868	IO::AIO::poll_cb with high priority (more examples can be found in the
		1869	SYNOPSIS section, at the top of this document):
216		1870
217	Event->io (fd => IO::AIO::poll_fileno,	1871	Event->io (fd => IO::AIO::poll_fileno,
218	poll => 'r', async => 1,	1872	poll => 'r', async => 1,
219	cb => \&IO::AIO::poll_cb);	1873	cb => \&IO::AIO::poll_cb);
220		1874
221	=item IO::AIO::poll_wait	1875	=item IO::AIO::poll_wait
222		1876
223	Wait till the result filehandle becomes ready for reading (simply does a	1877	Wait until either at least one request is in the result phase or no
224	select on the filehandle. This is useful if you want to synchronously wait	1878	requests are outstanding anymore.
225	for some requests to finish).	1879
		1880	This is useful if you want to synchronously wait for some requests to
		1881	become ready, without actually handling them.
226		1882
227	See C<nreqs> for an example.	1883	See C<nreqs> for an example.
228		1884
		1885	=item IO::AIO::poll
		1886
		1887	Waits until some requests have been handled.
		1888
		1889	Returns the number of requests processed, but is otherwise strictly
		1890	equivalent to:
		1891
		1892	IO::AIO::poll_wait, IO::AIO::poll_cb
		1893
229	=item IO::AIO::nreqs	1894	=item IO::AIO::flush
230		1895
231	Returns the number of requests currently outstanding.	1896	Wait till all outstanding AIO requests have been handled.
232		1897
233	Example: wait till there are no outstanding requests anymore:	1898	Strictly equivalent to:
234		1899
235	IO::AIO::poll_wait, IO::AIO::poll_cb	1900	IO::AIO::poll_wait, IO::AIO::poll_cb
236	while IO::AIO::nreqs;	1901	while IO::AIO::nreqs;
237		1902
		1903	=item IO::AIO::max_poll_reqs $nreqs
		1904
		1905	=item IO::AIO::max_poll_time $seconds
		1906
		1907	These set the maximum number of requests (default C<0>, meaning infinity)
		1908	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1909	the maximum amount of time (default C<0>, meaning infinity) spent in
		1910	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1911	of time C<poll_cb> is allowed to use).
		1912
		1913	Setting C<max_poll_time> to a non-zero value creates an overhead of one
		1914	syscall per request processed, which is not normally a problem unless your
		1915	callbacks are really really fast or your OS is really really slow (I am
		1916	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
		1917
		1918	Setting these is useful if you want to ensure some level of
		1919	interactiveness when perl is not fast enough to process all requests in
		1920	time.
		1921
		1922	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
		1923
		1924	Example: Install an Event watcher that automatically calls
		1925	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1926	program get the CPU sometimes even under high AIO load.
		1927
		1928	# try not to spend much more than 0.1s in poll_cb
		1929	IO::AIO::max_poll_time 0.1;
		1930
		1931	# use a low priority so other tasks have priority
		1932	Event->io (fd => IO::AIO::poll_fileno,
		1933	poll => 'r', nice => 1,
		1934	cb => &IO::AIO::poll_cb);
		1935
		1936	=back
		1937
		1938	=head3 CONTROLLING THE NUMBER OF THREADS
		1939
		1940	=over
		1941
		1942	=item IO::AIO::min_parallel $nthreads
		1943
		1944	Set the minimum number of AIO threads to C<$nthreads>. The current
		1945	default is C<8>, which means eight asynchronous operations can execute
		1946	concurrently at any one time (the number of outstanding requests,
		1947	however, is unlimited).
		1948
		1949	IO::AIO starts threads only on demand, when an AIO request is queued and
		1950	no free thread exists. Please note that queueing up a hundred requests can
		1951	create demand for a hundred threads, even if it turns out that everything
		1952	is in the cache and could have been processed faster by a single thread.
		1953
		1954	It is recommended to keep the number of threads relatively low, as some
		1955	Linux kernel versions will scale negatively with the number of threads
		1956	(higher parallelity => MUCH higher latency). With current Linux 2.6
		1957	versions, 4-32 threads should be fine.
		1958
		1959	Under most circumstances you don't need to call this function, as the
		1960	module selects a default that is suitable for low to moderate load.
		1961
		1962	=item IO::AIO::max_parallel $nthreads
		1963
		1964	Sets the maximum number of AIO threads to C<$nthreads>. If more than the
		1965	specified number of threads are currently running, this function kills
		1966	them. This function blocks until the limit is reached.
		1967
		1968	While C<$nthreads> are zero, aio requests get queued but not executed
		1969	until the number of threads has been increased again.
		1970
		1971	This module automatically runs C<max_parallel 0> at program end, to ensure
		1972	that all threads are killed and that there are no outstanding requests.
		1973
		1974	Under normal circumstances you don't need to call this function.
		1975
		1976	=item IO::AIO::max_idle $nthreads
		1977
		1978	Limit the number of threads (default: 4) that are allowed to idle
		1979	(i.e., threads that did not get a request to process within the idle
		1980	timeout (default: 10 seconds). That means if a thread becomes idle while
		1981	C<$nthreads> other threads are also idle, it will free its resources and
		1982	exit.
		1983
		1984	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1985	to allow for extremely high load situations, but want to free resources
		1986	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1987
		1988	The default is probably ok in most situations, especially if thread
		1989	creation is fast. If thread creation is very slow on your system you might
		1990	want to use larger values.
		1991
		1992	=item IO::AIO::idle_timeout $seconds
		1993
		1994	Sets the minimum idle timeout (default 10) after which worker threads are
		1995	allowed to exit. SEe C<IO::AIO::max_idle>.
		1996
		1997	=item IO::AIO::max_outstanding $maxreqs
		1998
		1999	Sets the maximum number of outstanding requests to C<$nreqs>. If
		2000	you do queue up more than this number of requests, the next call to
		2001	C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
		2002	C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
		2003	longer exceeded.
		2004
		2005	In other words, this setting does not enforce a queue limit, but can be
		2006	used to make poll functions block if the limit is exceeded.
		2007
		2008	This is a very bad function to use in interactive programs because it
		2009	blocks, and a bad way to reduce concurrency because it is inexact: Better
		2010	use an C<aio_group> together with a feed callback.
		2011
		2012	Its main use is in scripts without an event loop - when you want to stat
		2013	a lot of files, you can write somehting like this:
		2014
		2015	IO::AIO::max_outstanding 32;
		2016
		2017	for my $path (...) {
		2018	aio_stat $path , ...;
		2019	IO::AIO::poll_cb;
		2020	}
		2021
		2022	IO::AIO::flush;
		2023
		2024	The call to C<poll_cb> inside the loop will normally return instantly, but
		2025	as soon as more thna C<32> reqeusts are in-flight, it will block until
		2026	some requests have been handled. This keeps the loop from pushing a large
		2027	number of C<aio_stat> requests onto the queue.
		2028
		2029	The default value for C<max_outstanding> is very large, so there is no
		2030	practical limit on the number of outstanding requests.
		2031
		2032	=back
		2033
		2034	=head3 STATISTICAL INFORMATION
		2035
		2036	=over
		2037
238	=item IO::AIO::flush	2038	=item IO::AIO::nreqs
239		2039
240	Wait till all outstanding AIO requests have been handled.	2040	Returns the number of requests currently in the ready, execute or pending
		2041	states (i.e. for which their callback has not been invoked yet).
241		2042
242	Strictly equivalent to:	2043	Example: wait till there are no outstanding requests anymore:
243		2044
244	IO::AIO::poll_wait, IO::AIO::poll_cb	2045	IO::AIO::poll_wait, IO::AIO::poll_cb
245	while IO::AIO::nreqs;	2046	while IO::AIO::nreqs;
246		2047
		2048	=item IO::AIO::nready
		2049
		2050	Returns the number of requests currently in the ready state (not yet
		2051	executed).
		2052
247	=item IO::AIO::poll	2053	=item IO::AIO::npending
248		2054
249	Waits until some requests have been handled.	2055	Returns the number of requests currently in the pending state (executed,
250		2056	but not yet processed by poll_cb).
251	Strictly equivalent to:
252
253	IO::AIO::poll_wait, IO::AIO::poll_cb
254	if IO::AIO::nreqs;
255
256	=item IO::AIO::min_parallel $nthreads
257
258	Set the minimum number of AIO threads to C<$nthreads>. The default is
259	C<1>, which means a single asynchronous operation can be done at one time
260	(the number of outstanding operations, however, is unlimited).
261
262	It is recommended to keep the number of threads low, as some Linux
263	kernel versions will scale negatively with the number of threads (higher
264	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32
265	threads should be fine.
266
267	Under normal circumstances you don't need to call this function, as this
268	module automatically starts some threads (the exact number might change,
269	and is currently 4).
270
271	=item IO::AIO::max_parallel $nthreads
272
273	Sets the maximum number of AIO threads to C<$nthreads>. If more than
274	the specified number of threads are currently running, kill them. This
275	function blocks until the limit is reached.
276
277	This module automatically runs C<max_parallel 0> at program end, to ensure
278	that all threads are killed and that there are no outstanding requests.
279
280	Under normal circumstances you don't need to call this function.
281
282	=item $oldnreqs = IO::AIO::max_outstanding $nreqs
283
284	Sets the maximum number of outstanding requests to C<$nreqs>. If you
285	try to queue up more than this number of requests, the caller will block until
286	some requests have been handled.
287
288	The default is very large, so normally there is no practical limit. If you
289	queue up many requests in a loop it it often improves speed if you set
290	this to a relatively low number, such as C<100>.
291
292	Under normal circumstances you don't need to call this function.
293		2057
294	=back	2058	=back
295		2059
		2060	=head3 MISCELLANEOUS FUNCTIONS
		2061
		2062	IO::AIO implements some functions that are useful when you want to use
		2063	some "Advanced I/O" function not available to in Perl, without going the
		2064	"Asynchronous I/O" route. Many of these have an asynchronous C<aio_*>
		2065	counterpart.
		2066
		2067	=over 4
		2068
		2069	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		2070
		2071	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		2072	but is blocking (this makes most sense if you know the input data is
		2073	likely cached already and the output filehandle is set to non-blocking
		2074	operations).
		2075
		2076	Returns the number of bytes copied, or C<-1> on error.
		2077
		2078	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		2079
		2080	Simply calls the C<posix_fadvise> function (see its
		2081	manpage for details). The following advice constants are
		2082	available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		2083	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		2084	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		2085
		2086	On systems that do not implement C<posix_fadvise>, this function returns
		2087	ENOSYS, otherwise the return value of C<posix_fadvise>.
		2088
		2089	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		2090
		2091	Simply calls the C<posix_madvise> function (see its
		2092	manpage for details). The following advice constants are
		2093	available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		2094	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>,
		2095	C<IO::AIO::MADV_DONTNEED>.
		2096
		2097	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2098	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2099	will be reduced to fit into the C<$scalar>.
		2100
		2101	On systems that do not implement C<posix_madvise>, this function returns
		2102	ENOSYS, otherwise the return value of C<posix_madvise>.
		2103
		2104	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		2105
		2106	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		2107	$scalar (see its manpage for details). The following protect
		2108	constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		2109	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		2110
		2111	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2112	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2113	will be reduced to fit into the C<$scalar>.
		2114
		2115	On systems that do not implement C<mprotect>, this function returns
		2116	ENOSYS, otherwise the return value of C<mprotect>.
		2117
		2118	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		2119
		2120	Memory-maps a file (or anonymous memory range) and attaches it to the
		2121	given C<$scalar>, which will act like a string scalar. Returns true on
		2122	success, and false otherwise.
		2123
		2124	The scalar must exist, but its contents do not matter - this means you
		2125	cannot use a nonexistant array or hash element. When in doubt, C<undef>
		2126	the scalar first.
		2127
		2128	The only operations allowed on the mmapped scalar are C<substr>/C<vec>,
		2129	which don't change the string length, and most read-only operations such
		2130	as copying it or searching it with regexes and so on.
		2131
		2132	Anything else is unsafe and will, at best, result in memory leaks.
		2133
		2134	The memory map associated with the C<$scalar> is automatically removed
		2135	when the C<$scalar> is undef'd or destroyed, or when the C<IO::AIO::mmap>
		2136	or C<IO::AIO::munmap> functions are called on it.
		2137
		2138	This calls the C<mmap>(2) function internally. See your system's manual
		2139	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		2140
		2141	The C<$length> must be larger than zero and smaller than the actual
		2142	filesize.
		2143
		2144	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		2145	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		2146
		2147	C<$flags> can be a combination of
		2148	C<IO::AIO::MAP_SHARED> or
		2149	C<IO::AIO::MAP_PRIVATE>,
		2150	or a number of system-specific flags (when not available, the are C<0>):
		2151	C<IO::AIO::MAP_ANONYMOUS> (which is set to C<MAP_ANON> if your system only provides this constant),
		2152	C<IO::AIO::MAP_LOCKED>,
		2153	C<IO::AIO::MAP_NORESERVE>,
		2154	C<IO::AIO::MAP_POPULATE>,
		2155	C<IO::AIO::MAP_NONBLOCK>,
		2156	C<IO::AIO::MAP_FIXED>,
		2157	C<IO::AIO::MAP_GROWSDOWN>,
		2158	C<IO::AIO::MAP_32BIT>,
		2159	C<IO::AIO::MAP_HUGETLB> or
		2160	C<IO::AIO::MAP_STACK>.
		2161
		2162	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		2163
		2164	C<$offset> is the offset from the start of the file - it generally must be
		2165	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		2166
		2167	Example:
		2168
		2169	use Digest::MD5;
		2170	use IO::AIO;
		2171
		2172	open my $fh, "<verybigfile"
		2173	or die "$!";
		2174
		2175	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		2176	or die "verybigfile: $!";
		2177
		2178	my $fast_md5 = md5 $data;
		2179
		2180	=item IO::AIO::munmap $scalar
		2181
		2182	Removes a previous mmap and undefines the C<$scalar>.
		2183
		2184	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		2185
		2186	Calls the C<munlock> function, undoing the effects of a previous
		2187	C<aio_mlock> call (see its description for details).
		2188
		2189	=item IO::AIO::munlockall
		2190
		2191	Calls the C<munlockall> function.
		2192
		2193	On systems that do not implement C<munlockall>, this function returns
		2194	ENOSYS, otherwise the return value of C<munlockall>.
		2195
		2196	=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
		2197
		2198	Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
		2199	C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
		2200	should be the file offset.
		2201
		2202	C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
		2203	silently corrupt the data in this case.
		2204
		2205	The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
		2206	C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
		2207	C<IO::AIO::SPLICE_F_GIFT>.
		2208
		2209	See the C<splice(2)> manpage for details.
		2210
		2211	=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
		2212
		2213	Calls the GNU/Linux C<tee(2)> syscall, see its manpage and the
		2214	description for C<IO::AIO::splice> above for details.
		2215
		2216	=item $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
		2217
		2218	Attempts to query or change the pipe buffer size. Obviously works only
		2219	on pipes, and currently works only on GNU/Linux systems, and fails with
		2220	C<-1>/C<ENOSYS> everywhere else. If anybody knows how to influence pipe buffer
		2221	size on other systems, drop me a note.
		2222
		2223	=item ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
		2224
		2225	This is a direct interface to the Linux L<pipe2(2)> system call. If
		2226	C<$flags> is missing or C<0>, then this should be the same as a call to
		2227	perl's built-in C<pipe> function and create a new pipe, and works on
		2228	systems that lack the pipe2 syscall. On win32, this case invokes C<_pipe
		2229	(..., 4096, O_BINARY)>.
		2230
		2231	If C<$flags> is non-zero, it tries to invoke the pipe2 system call with
		2232	the given flags (Linux 2.6.27, glibc 2.9).
		2233
		2234	On success, the read and write file handles are returned.
		2235
		2236	On error, nothing will be returned. If the pipe2 syscall is missing and
		2237	C<$flags> is non-zero, fails with C<ENOSYS>.
		2238
		2239	Please refer to L<pipe2(2)> for more info on the C<$flags>, but at the
		2240	time of this writing, C<IO::AIO::O_CLOEXEC>, C<IO::AIO::O_NONBLOCK> and
		2241	C<IO::AIO::O_DIRECT> (Linux 3.4, for packet-based pipes) were supported.
		2242
		2243	=back
		2244
296	=cut	2245	=cut
297		2246
298	# support function to convert a fd into a perl filehandle
299	sub _fd2fh {
300	return undef if $_[0] < 0;
301
302	# try to be perl5.6-compatible
303	local *AIO_FH;
304	open AIO_FH, "+<&=$_[0]"
305	or return undef;
306
307	*AIO_FH
308	}
309
310	min_parallel 4;	2247	min_parallel 8;
311		2248
312	END {	2249	END { flush }
313	max_parallel 0;
314	}
315		2250
316	1;	2251	1;
317		2252
		2253	=head1 EVENT LOOP INTEGRATION
		2254
		2255	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		2256	automatically into many event loops:
		2257
		2258	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		2259	use AnyEvent::AIO;
		2260
		2261	You can also integrate IO::AIO manually into many event loops, here are
		2262	some examples of how to do this:
		2263
		2264	# EV integration
		2265	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		2266
		2267	# Event integration
		2268	Event->io (fd => IO::AIO::poll_fileno,
		2269	poll => 'r',
		2270	cb => \&IO::AIO::poll_cb);
		2271
		2272	# Glib/Gtk2 integration
		2273	add_watch Glib::IO IO::AIO::poll_fileno,
		2274	in => sub { IO::AIO::poll_cb; 1 };
		2275
		2276	# Tk integration
		2277	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		2278	readable => \&IO::AIO::poll_cb);
		2279
		2280	# Danga::Socket integration
		2281	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		2282	\&IO::AIO::poll_cb);
		2283
		2284	=head2 FORK BEHAVIOUR
		2285
		2286	Usage of pthreads in a program changes the semantics of fork
		2287	considerably. Specifically, only async-safe functions can be called after
		2288	fork. Perl doesn't know about this, so in general, you cannot call fork
		2289	with defined behaviour in perl if pthreads are involved. IO::AIO uses
		2290	pthreads, so this applies, but many other extensions and (for inexplicable
		2291	reasons) perl itself often is linked against pthreads, so this limitation
		2292	applies to quite a lot of perls.
		2293
		2294	This module no longer tries to fight your OS, or POSIX. That means IO::AIO
		2295	only works in the process that loaded it. Forking is fully supported, but
		2296	using IO::AIO in the child is not.
		2297
		2298	You might get around by not I<using> IO::AIO before (or after)
		2299	forking. You could also try to call the L<IO::AIO::reinit> function in the
		2300	child:
		2301
		2302	=over 4
		2303
		2304	=item IO::AIO::reinit
		2305
		2306	Abandons all current requests and I/O threads and simply reinitialises all
		2307	data structures. This is not an operation supported by any standards, but
		2308	happens to work on GNU/Linux and some newer BSD systems.
		2309
		2310	The only reasonable use for this function is to call it after forking, if
		2311	C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
		2312	the process will result in undefined behaviour. Calling it at any time
		2313	will also result in any undefined (by POSIX) behaviour.
		2314
		2315	=back
		2316
		2317	=head2 MEMORY USAGE
		2318
		2319	Per-request usage:
		2320
		2321	Each aio request uses - depending on your architecture - around 100-200
		2322	bytes of memory. In addition, stat requests need a stat buffer (possibly
		2323	a few hundred bytes), readdir requires a result buffer and so on. Perl
		2324	scalars and other data passed into aio requests will also be locked and
		2325	will consume memory till the request has entered the done state.
		2326
		2327	This is not awfully much, so queuing lots of requests is not usually a
		2328	problem.
		2329
		2330	Per-thread usage:
		2331
		2332	In the execution phase, some aio requests require more memory for
		2333	temporary buffers, and each thread requires a stack and other data
		2334	structures (usually around 16k-128k, depending on the OS).
		2335
		2336	=head1 KNOWN BUGS
		2337
		2338	Known bugs will be fixed in the next release.
		2339
318	=head1 SEE ALSO	2340	=head1 SEE ALSO
319		2341
320	L<Coro>, L<Linux::AIO>.	2342	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		2343	more natural syntax.
321		2344
322	=head1 AUTHOR	2345	=head1 AUTHOR
323		2346
324	Marc Lehmann <schmorp@schmorp.de>	2347	Marc Lehmann <schmorp@schmorp.de>
325	http://home.schmorp.de/	2348	http://home.schmorp.de/

Diff Legend

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