ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.6 by root, Sun Jul 10 22:19:48 2005 UTC vs.
Revision 1.269 by root, Tue Jun 6 04:29:35 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', async => 1,
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		30
33	=head1 DESCRIPTION	31	=head1 DESCRIPTION
34		32
35	This module implements asynchronous I/O using whatever means your	33	This module implements asynchronous I/O using whatever means your
36	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>).
37		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
38	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
39	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
40	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
41	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
42	aio functions available on many operating systems. However, they are often	57	functions available on many operating systems. However, they are often
43	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
44	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
45	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.
46		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
47	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,
48	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
49		164
50	=cut	165	=cut
51		166
52	package IO::AIO;	167	package IO::AIO;
53		168
		169	use Carp ();
		170
		171	use common::sense;
		172
54	use base 'Exporter';	173	use base 'Exporter';
55		174
56	use Fcntl ();
57
58	BEGIN {	175	BEGIN {
59	$VERSION = 0.2;	176	our $VERSION = 4.35;
60		177
61	@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
62	aio_fsync aio_fdatasync aio_readahead);	179	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
63	@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_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';
64		201
65	require XSLoader;	202	require XSLoader;
66	XSLoader::load IO::AIO, $VERSION;	203	XSLoader::load ("IO::AIO", $VERSION);
67	}	204	}
68		205
69	=head1 FUNCTIONS	206	=head1 FUNCTIONS
70		207
71	=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_mkdir $pathname, $mode, $callback->($status)
		239	aio_rmdir $pathname, $callback->($status)
		240	aio_readdir $pathname, $callback->($entries)
		241	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		242	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		243	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		244	aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		245	aio_load $pathname, $data, $callback->($status)
		246	aio_copy $srcpath, $dstpath, $callback->($status)
		247	aio_move $srcpath, $dstpath, $callback->($status)
		248	aio_rmtree $pathname, $callback->($status)
		249	aio_fcntl $fh, $cmd, $arg, $callback->($status)
		250	aio_ioctl $fh, $request, $buf, $callback->($status)
		251	aio_sync $callback->($status)
		252	aio_syncfs $fh, $callback->($status)
		253	aio_fsync $fh, $callback->($status)
		254	aio_fdatasync $fh, $callback->($status)
		255	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		256	aio_pathsync $pathname, $callback->($status)
		257	aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		258	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		259	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		260	aio_mlockall $flags, $callback->($status)
		261	aio_group $callback->(...)
		262	aio_nop $callback->()
		263
		264	$prev_pri = aioreq_pri [$pri]
		265	aioreq_nice $pri_adjust
		266
		267	IO::AIO::poll_wait
		268	IO::AIO::poll_cb
		269	IO::AIO::poll
		270	IO::AIO::flush
		271	IO::AIO::max_poll_reqs $nreqs
		272	IO::AIO::max_poll_time $seconds
		273	IO::AIO::min_parallel $nthreads
		274	IO::AIO::max_parallel $nthreads
		275	IO::AIO::max_idle $nthreads
		276	IO::AIO::idle_timeout $seconds
		277	IO::AIO::max_outstanding $maxreqs
		278	IO::AIO::nreqs
		279	IO::AIO::nready
		280	IO::AIO::npending
		281
		282	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		283	IO::AIO::fadvise $fh, $offset, $len, $advice
		284	IO::AIO::mmap $scalar, $length, $prot, $flags[, $fh[, $offset]]
		285	IO::AIO::munmap $scalar
		286	IO::AIO::madvise $scalar, $offset, $length, $advice
		287	IO::AIO::mprotect $scalar, $offset, $length, $protect
		288	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		289	IO::AIO::munlockall
		290
		291	=head2 API NOTES
72		292
73	All the C<aio_*> calls are more or less thin wrappers around the syscall	293	All the C<aio_*> calls are more or less thin wrappers around the syscall
74	with the same name (sans C<aio_>). The arguments are similar or identical,	294	with the same name (sans C<aio_>). The arguments are similar or identical,
75	and they all accept an additional C<$callback> argument which must be	295	and they all accept an additional (and optional) C<$callback> argument
76	a code reference. This code reference will get called with the syscall	296	which must be a code reference. This code reference will be called after
		297	the syscall has been executed in an asynchronous fashion. The results
		298	of the request will be passed as arguments to the callback (and, if an
		299	error occured, in C<$!>) - for most requests the syscall return code (e.g.
77	return code (e.g. most syscalls return C<-1> on error, unlike perl, which	300	most syscalls return C<-1> on error, unlike perl, which usually delivers
78	usually delivers "false") as it's sole argument when the given syscall has	301	"false").
79	been executed asynchronously.
80		302
81	All functions that expect a filehandle will also accept a file descriptor.	303	Some requests (such as C<aio_readdir>) pass the actual results and
		304	communicate failures by passing C<undef>.
82		305
		306	All functions expecting a filehandle keep a copy of the filehandle
		307	internally until the request has finished.
		308
		309	All functions return request objects of type L<IO::AIO::REQ> that allow
		310	further manipulation of those requests while they are in-flight.
		311
83	The filenames you pass to these routines I<must> be absolute. The reason	312	The pathnames you pass to these routines I<should> be absolute. The
84	is that at the time the request is being executed, the current working	313	reason for this is that at the time the request is being executed, the
85	directory could have changed. Alternatively, you can make sure that you	314	current working directory could have changed. Alternatively, you can
86	never change the current working directory.	315	make sure that you never change the current working directory anywhere
		316	in the program and then use relative paths. You can also take advantage
		317	of IO::AIOs working directory abstraction, that lets you specify paths
		318	relative to some previously-opened "working directory object" - see the
		319	description of the C<IO::AIO::WD> class later in this document.
		320
		321	To encode pathnames as octets, either make sure you either: a) always pass
		322	in filenames you got from outside (command line, readdir etc.) without
		323	tinkering, b) are in your native filesystem encoding, c) use the Encode
		324	module and encode your pathnames to the locale (or other) encoding in
		325	effect in the user environment, d) use Glib::filename_from_unicode on
		326	unicode filenames or e) use something else to ensure your scalar has the
		327	correct contents.
		328
		329	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		330	handles correctly whether it is set or not.
		331
		332	=head2 AIO REQUEST FUNCTIONS
87		333
88	=over 4	334	=over 4
89		335
		336	=item $prev_pri = aioreq_pri [$pri]
		337
		338	Returns the priority value that would be used for the next request and, if
		339	C<$pri> is given, sets the priority for the next aio request.
		340
		341	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		342	and C<4>, respectively. Requests with higher priority will be serviced
		343	first.
		344
		345	The priority will be reset to C<0> after each call to one of the C<aio_*>
		346	functions.
		347
		348	Example: open a file with low priority, then read something from it with
		349	higher priority so the read request is serviced before other low priority
		350	open requests (potentially spamming the cache):
		351
		352	aioreq_pri -3;
		353	aio_open ..., sub {
		354	return unless $_[0];
		355
		356	aioreq_pri -2;
		357	aio_read $_[0], ..., sub {
		358	...
		359	};
		360	};
		361
		362
		363	=item aioreq_nice $pri_adjust
		364
		365	Similar to C<aioreq_pri>, but subtracts the given value from the current
		366	priority, so the effect is cumulative.
		367
		368
90	=item aio_open $pathname, $flags, $mode, $callback	369	=item aio_open $pathname, $flags, $mode, $callback->($fh)
91		370
92	Asynchronously open or create a file and call the callback with a newly	371	Asynchronously open or create a file and call the callback with a newly
93	created filehandle for the file.	372	created filehandle for the file (or C<undef> in case of an error).
94		373
95	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	374	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
96	for an explanation.	375	for an explanation.
97		376
98	The C<$mode> argument is a bitmask. See the C<Fcntl> module for a	377	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
99	list. They are the same as used in C<sysopen>.	378	list. They are the same as used by C<sysopen>.
		379
		380	Likewise, C<$mode> specifies the mode of the newly created file, if it
		381	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
		382	except that it is mandatory (i.e. use C<0> if you don't create new files,
		383	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		384	by the umask in effect then the request is being executed, so better never
		385	change the umask.
100		386
101	Example:	387	Example:
102		388
103	aio_open "/etc/passwd", O_RDONLY, 0, sub {	389	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
104	if ($_[0]) {	390	if ($_[0]) {
105	print "open successful, fh is $_[0]\n";	391	print "open successful, fh is $_[0]\n";
106	...	392	...
107	} else {	393	} else {
108	die "open failed: $!\n";	394	die "open failed: $!\n";
109	}	395	}
110	};	396	};
111		397
		398	In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
		399	C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
		400	following POSIX and non-POSIX constants are available (missing ones on
		401	your system are, as usual, C<0>):
		402
		403	C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
		404	C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
		405	C<O_RSYNC>, C<O_SYNC>, C<O_PATH>, C<O_TMPFILE>, and C<O_TTY_INIT>.
		406
		407
112	=item aio_close $fh, $callback	408	=item aio_close $fh, $callback->($status)
113		409
114	Asynchronously close a file and call the callback with the result	410	Asynchronously close a file and call the callback with the result
115	code. I<WARNING:> although accepted, you should not pass in a perl	411	code.
116	filehandle here, as perl will likely close the file descriptor itself when
117	the filehandle is destroyed. Normally, you can safely call perls C<close>
118	or just let filehandles go out of scope.
119		412
		413	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
		414	closing the file descriptor associated with the filehandle itself.
		415
		416	Therefore, C<aio_close> will not close the filehandle - instead it will
		417	use dup2 to overwrite the file descriptor with the write-end of a pipe
		418	(the pipe fd will be created on demand and will be cached).
		419
		420	Or in other words: the file descriptor will be closed, but it will not be
		421	free for reuse until the perl filehandle is closed.
		422
		423	=cut
		424
		425	=item aio_seek $fh, $offset, $whence, $callback->($offs)
		426
		427	Seeks the filehandle to the new C<$offset>, similarly to perl's
		428	C<sysseek>. The C<$whence> can use the traditional values (C<0> for
		429	C<IO::AIO::SEEK_SET>, C<1> for C<IO::AIO::SEEK_CUR> or C<2> for
		430	C<IO::AIO::SEEK_END>).
		431
		432	The resulting absolute offset will be passed to the callback, or C<-1> in
		433	case of an error.
		434
		435	In theory, the C<$whence> constants could be different than the
		436	corresponding values from L<Fcntl>, but perl guarantees they are the same,
		437	so don't panic.
		438
		439	As a GNU/Linux (and maybe Solaris) extension, also the constants
		440	C<IO::AIO::SEEK_DATA> and C<IO::AIO::SEEK_HOLE> are available, if they
		441	could be found. No guarantees about suitability for use in C<aio_seek> or
		442	Perl's C<sysseek> can be made though, although I would naively assume they
		443	"just work".
		444
120	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	445	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
121		446
122	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	447	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
123		448
124	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	449	Reads or writes C<$length> bytes from or to the specified C<$fh> and
125	into the scalar given by C<data> and offset C<dataoffset> and calls the	450	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset> and
126	callback without the actual number of bytes read (or -1 on error, just	451	calls the callback with the actual number of bytes transferred (or -1 on
127	like the syscall).	452	error, just like the syscall).
128		453
		454	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		455	offset plus the actual number of bytes read.
		456
		457	If C<$offset> is undefined, then the current file descriptor offset will
		458	be used (and updated), otherwise the file descriptor offset will not be
		459	changed by these calls.
		460
		461	If C<$length> is undefined in C<aio_write>, use the remaining length of
		462	C<$data>.
		463
		464	If C<$dataoffset> is less than zero, it will be counted from the end of
		465	C<$data>.
		466
		467	The C<$data> scalar I<MUST NOT> be modified in any way while the request
		468	is outstanding. Modifying it can result in segfaults or World War III (if
		469	the necessary/optional hardware is installed).
		470
129	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at	471	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
130	offset C<0> within the scalar:	472	offset C<0> within the scalar:
131		473
132	aio_read $fh, 7, 15, $buffer, 0, sub {	474	aio_read $fh, 7, 15, $buffer, 0, sub {
133	$_[0] >= 0 or die "read error: $!";	475	$_[0] > 0 or die "read error: $!";
134	print "read <$buffer>\n";	476	print "read $_[0] bytes: <$buffer>\n";
135	};	477	};
136		478
		479
		480	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		481
		482	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
		483	reading at byte offset C<$in_offset>, and starts writing at the current
		484	file offset of C<$out_fh>. Because of that, it is not safe to issue more
		485	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
		486	other. The same C<$in_fh> works fine though, as this function does not
		487	move or use the file offset of C<$in_fh>.
		488
		489	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		490	are written, and there is no way to find out how many more bytes have been
		491	read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
		492	number of bytes written to C<$out_fh>. Only if the result value equals
		493	C<$length> one can assume that C<$length> bytes have been read.
		494
		495	Unlike with other C<aio_> functions, it makes a lot of sense to use
		496	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		497	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		498	the socket I/O will be non-blocking. Note, however, that you can run
		499	into a trap where C<aio_sendfile> reads some data with readahead, then
		500	fails to write all data, and when the socket is ready the next time, the
		501	data in the cache is already lost, forcing C<aio_sendfile> to again hit
		502	the disk. Explicit C<aio_read> + C<aio_write> let's you better control
		503	resource usage.
		504
		505	This call tries to make use of a native C<sendfile>-like syscall to
		506	provide zero-copy operation. For this to work, C<$out_fh> should refer to
		507	a socket, and C<$in_fh> should refer to an mmap'able file.
		508
		509	If a native sendfile cannot be found or it fails with C<ENOSYS>,
		510	C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
		511	C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
		512	type of filehandle regardless of the limitations of the operating system.
		513
		514	As native sendfile syscalls (as practically any non-POSIX interface hacked
		515	together in a hurry to improve benchmark numbers) tend to be rather buggy
		516	on many systems, this implementation tries to work around some known bugs
		517	in Linux and FreeBSD kernels (probably others, too), but that might fail,
		518	so you really really should check the return value of C<aio_sendfile> -
		519	fewer bytes than expected might have been transferred.
		520
		521
137	=item aio_readahead $fh,$offset,$length, $callback	522	=item aio_readahead $fh,$offset,$length, $callback->($retval)
138		523
139	Asynchronously reads the specified byte range into the page cache, using
140	the C<readahead> syscall. If that syscall doesn't exist the status will be
141	C<-1> and C<$!> is set to ENOSYS.
142
143	readahead() populates the page cache with data from a file so that	524	C<aio_readahead> populates the page cache with data from a file so that
144	subsequent reads from that file will not block on disk I/O. The C<$offset>	525	subsequent reads from that file will not block on disk I/O. The C<$offset>
145	argument specifies the starting point from which data is to be read and	526	argument specifies the starting point from which data is to be read and
146	C<$length> specifies the number of bytes to be read. I/O is performed in	527	C<$length> specifies the number of bytes to be read. I/O is performed in
147	whole pages, so that offset is effectively rounded down to a page boundary	528	whole pages, so that offset is effectively rounded down to a page boundary
148	and bytes are read up to the next page boundary greater than or equal to	529	and bytes are read up to the next page boundary greater than or equal to
149	(off-set+length). aio_readahead() does not read beyond the end of the	530	(off-set+length). C<aio_readahead> does not read beyond the end of the
150	file. The current file offset of the file is left unchanged.	531	file. The current file offset of the file is left unchanged.
151		532
		533	If that syscall doesn't exist (likely if your kernel isn't Linux) it will
		534	be emulated by simply reading the data, which would have a similar effect.
		535
		536
152	=item aio_stat $fh_or_path, $callback	537	=item aio_stat $fh_or_path, $callback->($status)
153		538
154	=item aio_lstat $fh, $callback	539	=item aio_lstat $fh, $callback->($status)
155		540
156	Works like perl's C<stat> or C<lstat> in void context. The callback will	541	Works like perl's C<stat> or C<lstat> in void context. The callback will
157	be called after the stat and the results will be available using C<stat _>	542	be called after the stat and the results will be available using C<stat _>
158	or C<-s _> etc...	543	or C<-s _> etc...
159		544
…		…
161	for an explanation.	546	for an explanation.
162		547
163	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	548	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
164	error when stat'ing a large file, the results will be silently truncated	549	error when stat'ing a large file, the results will be silently truncated
165	unless perl itself is compiled with large file support.	550	unless perl itself is compiled with large file support.
		551
		552	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		553	following constants and functions (if not implemented, the constants will
		554	be C<0> and the functions will either C<croak> or fall back on traditional
		555	behaviour).
		556
		557	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		558	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		559	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
166		560
167	Example: Print the length of F</etc/passwd>:	561	Example: Print the length of F</etc/passwd>:
168		562
169	aio_stat "/etc/passwd", sub {	563	aio_stat "/etc/passwd", sub {
170	$_[0] and die "stat failed: $!";	564	$_[0] and die "stat failed: $!";
171	print "size is ", -s _, "\n";	565	print "size is ", -s _, "\n";
172	};	566	};
173		567
		568
		569	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		570
		571	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		572	whether a file handle or path was passed.
		573
		574	On success, the callback is passed a hash reference with the following
		575	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		576	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		577	is passed.
		578
		579	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		580	C<ST_NOSUID>.
		581
		582	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		583	their correct value when available, or to C<0> on systems that do
		584	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		585	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		586	C<ST_NODIRATIME> and C<ST_RELATIME>.
		587
		588	Example: stat C</wd> and dump out the data if successful.
		589
		590	aio_statvfs "/wd", sub {
		591	my $f = $_[0]
		592	or die "statvfs: $!";
		593
		594	use Data::Dumper;
		595	say Dumper $f;
		596	};
		597
		598	# result:
		599	{
		600	bsize => 1024,
		601	bfree => 4333064312,
		602	blocks => 10253828096,
		603	files => 2050765568,
		604	flag => 4096,
		605	favail => 2042092649,
		606	bavail => 4333064312,
		607	ffree => 2042092649,
		608	namemax => 255,
		609	frsize => 1024,
		610	fsid => 1810
		611	}
		612
		613	Here is a (likely partial - send me updates!) list of fsid values used by
		614	Linux - it is safe to hardcode these when C<$^O> is C<linux>:
		615
		616	0x0000adf5 adfs
		617	0x0000adff affs
		618	0x5346414f afs
		619	0x09041934 anon-inode filesystem
		620	0x00000187 autofs
		621	0x42465331 befs
		622	0x1badface bfs
		623	0x42494e4d binfmt_misc
		624	0x9123683e btrfs
		625	0x0027e0eb cgroupfs
		626	0xff534d42 cifs
		627	0x73757245 coda
		628	0x012ff7b7 coh
		629	0x28cd3d45 cramfs
		630	0x453dcd28 cramfs-wend (wrong endianness)
		631	0x64626720 debugfs
		632	0x00001373 devfs
		633	0x00001cd1 devpts
		634	0x0000f15f ecryptfs
		635	0x00414a53 efs
		636	0x0000137d ext
		637	0x0000ef53 ext2/ext3/ext4
		638	0x0000ef51 ext2
		639	0xf2f52010 f2fs
		640	0x00004006 fat
		641	0x65735546 fuseblk
		642	0x65735543 fusectl
		643	0x0bad1dea futexfs
		644	0x01161970 gfs2
		645	0x47504653 gpfs
		646	0x00004244 hfs
		647	0xf995e849 hpfs
		648	0x00c0ffee hostfs
		649	0x958458f6 hugetlbfs
		650	0x2bad1dea inotifyfs
		651	0x00009660 isofs
		652	0x000072b6 jffs2
		653	0x3153464a jfs
		654	0x6b414653 k-afs
		655	0x0bd00bd0 lustre
		656	0x0000137f minix
		657	0x0000138f minix 30 char names
		658	0x00002468 minix v2
		659	0x00002478 minix v2 30 char names
		660	0x00004d5a minix v3
		661	0x19800202 mqueue
		662	0x00004d44 msdos
		663	0x0000564c novell
		664	0x00006969 nfs
		665	0x6e667364 nfsd
		666	0x00003434 nilfs
		667	0x5346544e ntfs
		668	0x00009fa1 openprom
		669	0x7461636F ocfs2
		670	0x00009fa0 proc
		671	0x6165676c pstorefs
		672	0x0000002f qnx4
		673	0x68191122 qnx6
		674	0x858458f6 ramfs
		675	0x52654973 reiserfs
		676	0x00007275 romfs
		677	0x67596969 rpc_pipefs
		678	0x73636673 securityfs
		679	0xf97cff8c selinux
		680	0x0000517b smb
		681	0x534f434b sockfs
		682	0x73717368 squashfs
		683	0x62656572 sysfs
		684	0x012ff7b6 sysv2
		685	0x012ff7b5 sysv4
		686	0x01021994 tmpfs
		687	0x15013346 udf
		688	0x00011954 ufs
		689	0x54190100 ufs byteswapped
		690	0x00009fa2 usbdevfs
		691	0x01021997 v9fs
		692	0xa501fcf5 vxfs
		693	0xabba1974 xenfs
		694	0x012ff7b4 xenix
		695	0x58465342 xfs
		696	0x012fd16d xia
		697
		698	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		699
		700	Works like perl's C<utime> function (including the special case of $atime
		701	and $mtime being undef). Fractional times are supported if the underlying
		702	syscalls support them.
		703
		704	When called with a pathname, uses utimes(2) if available, otherwise
		705	utime(2). If called on a file descriptor, uses futimes(2) if available,
		706	otherwise returns ENOSYS, so this is not portable.
		707
		708	Examples:
		709
		710	# set atime and mtime to current time (basically touch(1)):
		711	aio_utime "path", undef, undef;
		712	# set atime to current time and mtime to beginning of the epoch:
		713	aio_utime "path", time, undef; # undef==0
		714
		715
		716	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		717
		718	Works like perl's C<chown> function, except that C<undef> for either $uid
		719	or $gid is being interpreted as "do not change" (but -1 can also be used).
		720
		721	Examples:
		722
		723	# same as "chown root path" in the shell:
		724	aio_chown "path", 0, -1;
		725	# same as above:
		726	aio_chown "path", 0, undef;
		727
		728
		729	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		730
		731	Works like truncate(2) or ftruncate(2).
		732
		733
		734	=item aio_allocate $fh, $mode, $offset, $len, $callback->($status)
		735
		736	Allocates or frees disk space according to the C<$mode> argument. See the
		737	linux C<fallocate> documentation for details.
		738
		739	C<$mode> is usually C<0> or C<IO::AIO::FALLOC_FL_KEEP_SIZE> to allocate
		740	space, or C<IO::AIO::FALLOC_FL_PUNCH_HOLE | IO::AIO::FALLOC_FL_KEEP_SIZE>,
		741	to deallocate a file range.
		742
		743	IO::AIO also supports C<FALLOC_FL_COLLAPSE_RANGE>, to remove a range
		744	(without leaving a hole) and C<FALLOC_FL_ZERO_RANGE>, to zero a range (see
		745	your L<fallocate(2)> manpage).
		746
		747	The file system block size used by C<fallocate> is presumably the
		748	C<f_bsize> returned by C<statvfs>.
		749
		750	If C<fallocate> isn't available or cannot be emulated (currently no
		751	emulation will be attempted), passes C<-1> and sets C<$!> to C<ENOSYS>.
		752
		753
		754	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		755
		756	Works like perl's C<chmod> function.
		757
		758
174	=item aio_unlink $pathname, $callback	759	=item aio_unlink $pathname, $callback->($status)
175		760
176	Asynchronously unlink (delete) a file and call the callback with the	761	Asynchronously unlink (delete) a file and call the callback with the
177	result code.	762	result code.
178		763
		764
		765	=item aio_mknod $pathname, $mode, $dev, $callback->($status)
		766
		767	[EXPERIMENTAL]
		768
		769	Asynchronously create a device node (or fifo). See mknod(2).
		770
		771	The only (POSIX-) portable way of calling this function is:
		772
		773	aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		774
		775	See C<aio_stat> for info about some potentially helpful extra constants
		776	and functions.
		777
		778	=item aio_link $srcpath, $dstpath, $callback->($status)
		779
		780	Asynchronously create a new link to the existing object at C<$srcpath> at
		781	the path C<$dstpath> and call the callback with the result code.
		782
		783
		784	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		785
		786	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		787	the path C<$dstpath> and call the callback with the result code.
		788
		789
		790	=item aio_readlink $pathname, $callback->($link)
		791
		792	Asynchronously read the symlink specified by C<$path> and pass it to
		793	the callback. If an error occurs, nothing or undef gets passed to the
		794	callback.
		795
		796
		797	=item aio_realpath $pathname, $callback->($path)
		798
		799	Asynchronously make the path absolute and resolve any symlinks in
		800	C<$path>. The resulting path only consists of directories (same as
		801	L<Cwd::realpath>).
		802
		803	This request can be used to get the absolute path of the current working
		804	directory by passing it a path of F<.> (a single dot).
		805
		806
		807	=item aio_rename $srcpath, $dstpath, $callback->($status)
		808
		809	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		810	rename(2) and call the callback with the result code.
		811
		812	On systems that support the AIO::WD working directory abstraction
		813	natively, the case C<[$wd, "."]> as C<$srcpath> is specialcased - instead
		814	of failing, C<rename> is called on the absolute path of C<$wd>.
		815
		816
		817	=item aio_mkdir $pathname, $mode, $callback->($status)
		818
		819	Asynchronously mkdir (create) a directory and call the callback with
		820	the result code. C<$mode> will be modified by the umask at the time the
		821	request is executed, so do not change your umask.
		822
		823
		824	=item aio_rmdir $pathname, $callback->($status)
		825
		826	Asynchronously rmdir (delete) a directory and call the callback with the
		827	result code.
		828
		829	On systems that support the AIO::WD working directory abstraction
		830	natively, the case C<[$wd, "."]> is specialcased - instead of failing,
		831	C<rmdir> is called on the absolute path of C<$wd>.
		832
		833
		834	=item aio_readdir $pathname, $callback->($entries)
		835
		836	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
		837	directory (i.e. opendir + readdir + closedir). The entries will not be
		838	sorted, and will B<NOT> include the C<.> and C<..> entries.
		839
		840	The callback is passed a single argument which is either C<undef> or an
		841	array-ref with the filenames.
		842
		843
		844	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		845
		846	Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
		847	tune behaviour and output format. In case of an error, C<$entries> will be
		848	C<undef>.
		849
		850	The flags are a combination of the following constants, ORed together (the
		851	flags will also be passed to the callback, possibly modified):
		852
		853	=over 4
		854
		855	=item IO::AIO::READDIR_DENTS
		856
		857	When this flag is off, then the callback gets an arrayref consisting of
		858	names only (as with C<aio_readdir>), otherwise it gets an arrayref with
		859	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		860	entry in more detail.
		861
		862	C<$name> is the name of the entry.
		863
		864	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		865
		866	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		867	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		868	C<IO::AIO::DT_WHT>.
		869
		870	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		871	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		872	scalars are read-only: you can not modify them.
		873
		874	C<$inode> is the inode number (which might not be exact on systems with 64
		875	bit inode numbers and 32 bit perls). This field has unspecified content on
		876	systems that do not deliver the inode information.
		877
		878	=item IO::AIO::READDIR_DIRS_FIRST
		879
		880	When this flag is set, then the names will be returned in an order where
		881	likely directories come first, in optimal stat order. This is useful when
		882	you need to quickly find directories, or you want to find all directories
		883	while avoiding to stat() each entry.
		884
		885	If the system returns type information in readdir, then this is used
		886	to find directories directly. Otherwise, likely directories are names
		887	beginning with ".", or otherwise names with no dots, of which names with
		888	short names are tried first.
		889
		890	=item IO::AIO::READDIR_STAT_ORDER
		891
		892	When this flag is set, then the names will be returned in an order
		893	suitable for stat()'ing each one. That is, when you plan to stat()
		894	all files in the given directory, then the returned order will likely
		895	be fastest.
		896
		897	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		898	the likely dirs come first, resulting in a less optimal stat order.
		899
		900	=item IO::AIO::READDIR_FOUND_UNKNOWN
		901
		902	This flag should not be set when calling C<aio_readdirx>. Instead, it
		903	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		904	C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
		905	C<$type>'s are known, which can be used to speed up some algorithms.
		906
		907	=back
		908
		909
		910	=item aio_load $pathname, $data, $callback->($status)
		911
		912	This is a composite request that tries to fully load the given file into
		913	memory. Status is the same as with aio_read.
		914
		915	=cut
		916
		917	sub aio_load($$;$) {
		918	my ($path, undef, $cb) = @_;
		919	my $data = \$_[1];
		920
		921	my $pri = aioreq_pri;
		922	my $grp = aio_group $cb;
		923
		924	aioreq_pri $pri;
		925	add $grp aio_open $path, O_RDONLY, 0, sub {
		926	my $fh = shift
		927	or return $grp->result (-1);
		928
		929	aioreq_pri $pri;
		930	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		931	$grp->result ($_[0]);
		932	};
		933	};
		934
		935	$grp
		936	}
		937
		938	=item aio_copy $srcpath, $dstpath, $callback->($status)
		939
		940	Try to copy the I<file> (directories not supported as either source or
		941	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		942	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		943
		944	This is a composite request that creates the destination file with
		945	mode 0200 and copies the contents of the source file into it using
		946	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		947	uid/gid, in that order.
		948
		949	If an error occurs, the partial destination file will be unlinked, if
		950	possible, except when setting atime, mtime, access mode and uid/gid, where
		951	errors are being ignored.
		952
		953	=cut
		954
		955	sub aio_copy($$;$) {
		956	my ($src, $dst, $cb) = @_;
		957
		958	my $pri = aioreq_pri;
		959	my $grp = aio_group $cb;
		960
		961	aioreq_pri $pri;
		962	add $grp aio_open $src, O_RDONLY, 0, sub {
		963	if (my $src_fh = $_[0]) {
		964	my @stat = stat $src_fh; # hmm, might block over nfs?
		965
		966	aioreq_pri $pri;
		967	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		968	if (my $dst_fh = $_[0]) {
		969	aioreq_pri $pri;
		970	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		971	if ($_[0] == $stat[7]) {
		972	$grp->result (0);
		973	close $src_fh;
		974
		975	my $ch = sub {
		976	aioreq_pri $pri;
		977	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		978	aioreq_pri $pri;
		979	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		980	aioreq_pri $pri;
		981	add $grp aio_close $dst_fh;
		982	}
		983	};
		984	};
		985
		986	aioreq_pri $pri;
		987	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		988	if ($_[0] < 0 && $! == ENOSYS) {
		989	aioreq_pri $pri;
		990	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		991	} else {
		992	$ch->();
		993	}
		994	};
		995	} else {
		996	$grp->result (-1);
		997	close $src_fh;
		998	close $dst_fh;
		999
		1000	aioreq $pri;
		1001	add $grp aio_unlink $dst;
		1002	}
		1003	};
		1004	} else {
		1005	$grp->result (-1);
		1006	}
		1007	},
		1008
		1009	} else {
		1010	$grp->result (-1);
		1011	}
		1012	};
		1013
		1014	$grp
		1015	}
		1016
		1017	=item aio_move $srcpath, $dstpath, $callback->($status)
		1018
		1019	Try to move the I<file> (directories not supported as either source or
		1020	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		1021	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		1022
		1023	This is a composite request that tries to rename(2) the file first; if
		1024	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		1025	that is successful, unlinks the C<$srcpath>.
		1026
		1027	=cut
		1028
		1029	sub aio_move($$;$) {
		1030	my ($src, $dst, $cb) = @_;
		1031
		1032	my $pri = aioreq_pri;
		1033	my $grp = aio_group $cb;
		1034
		1035	aioreq_pri $pri;
		1036	add $grp aio_rename $src, $dst, sub {
		1037	if ($_[0] && $! == EXDEV) {
		1038	aioreq_pri $pri;
		1039	add $grp aio_copy $src, $dst, sub {
		1040	$grp->result ($_[0]);
		1041
		1042	unless ($_[0]) {
		1043	aioreq_pri $pri;
		1044	add $grp aio_unlink $src;
		1045	}
		1046	};
		1047	} else {
		1048	$grp->result ($_[0]);
		1049	}
		1050	};
		1051
		1052	$grp
		1053	}
		1054
		1055	=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		1056
		1057	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		1058	efficiently separate the entries of directory C<$path> into two sets of
		1059	names, directories you can recurse into (directories), and ones you cannot
		1060	recurse into (everything else, including symlinks to directories).
		1061
		1062	C<aio_scandir> is a composite request that creates of many sub requests_
		1063	C<$maxreq> specifies the maximum number of outstanding aio requests that
		1064	this function generates. If it is C<< <= 0 >>, then a suitable default
		1065	will be chosen (currently 4).
		1066
		1067	On error, the callback is called without arguments, otherwise it receives
		1068	two array-refs with path-relative entry names.
		1069
		1070	Example:
		1071
		1072	aio_scandir $dir, 0, sub {
		1073	my ($dirs, $nondirs) = @_;
		1074	print "real directories: @$dirs\n";
		1075	print "everything else: @$nondirs\n";
		1076	};
		1077
		1078	Implementation notes.
		1079
		1080	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		1081
		1082	If readdir returns file type information, then this is used directly to
		1083	find directories.
		1084
		1085	Otherwise, after reading the directory, the modification time, size etc.
		1086	of the directory before and after the readdir is checked, and if they
		1087	match (and isn't the current time), the link count will be used to decide
		1088	how many entries are directories (if >= 2). Otherwise, no knowledge of the
		1089	number of subdirectories will be assumed.
		1090
		1091	Then entries will be sorted into likely directories a non-initial dot
		1092	currently) and likely non-directories (see C<aio_readdirx>). Then every
		1093	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
		1094	in order of their inode numbers. If that succeeds, it assumes that the
		1095	entry is a directory or a symlink to directory (which will be checked
		1096	separately). This is often faster than stat'ing the entry itself because
		1097	filesystems might detect the type of the entry without reading the inode
		1098	data (e.g. ext2fs filetype feature), even on systems that cannot return
		1099	the filetype information on readdir.
		1100
		1101	If the known number of directories (link count - 2) has been reached, the
		1102	rest of the entries is assumed to be non-directories.
		1103
		1104	This only works with certainty on POSIX (= UNIX) filesystems, which
		1105	fortunately are the vast majority of filesystems around.
		1106
		1107	It will also likely work on non-POSIX filesystems with reduced efficiency
		1108	as those tend to return 0 or 1 as link counts, which disables the
		1109	directory counting heuristic.
		1110
		1111	=cut
		1112
		1113	sub aio_scandir($$;$) {
		1114	my ($path, $maxreq, $cb) = @_;
		1115
		1116	my $pri = aioreq_pri;
		1117
		1118	my $grp = aio_group $cb;
		1119
		1120	$maxreq = 4 if $maxreq <= 0;
		1121
		1122	# get a wd object
		1123	aioreq_pri $pri;
		1124	add $grp aio_wd $path, sub {
		1125	$_[0]
		1126	or return $grp->result ();
		1127
		1128	my $wd = [shift, "."];
		1129
		1130	# stat once
		1131	aioreq_pri $pri;
		1132	add $grp aio_stat $wd, sub {
		1133	return $grp->result () if $_[0];
		1134	my $now = time;
		1135	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		1136
		1137	# read the directory entries
		1138	aioreq_pri $pri;
		1139	add $grp aio_readdirx $wd, READDIR_DIRS_FIRST, sub {
		1140	my $entries = shift
		1141	or return $grp->result ();
		1142
		1143	# stat the dir another time
		1144	aioreq_pri $pri;
		1145	add $grp aio_stat $wd, sub {
		1146	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		1147
		1148	my $ndirs;
		1149
		1150	# take the slow route if anything looks fishy
		1151	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		1152	$ndirs = -1;
		1153	} else {
		1154	# if nlink == 2, we are finished
		1155	# for non-posix-fs's, we rely on nlink < 2
		1156	$ndirs = (stat _)[3] - 2
		1157	or return $grp->result ([], $entries);
		1158	}
		1159
		1160	my (@dirs, @nondirs);
		1161
		1162	my $statgrp = add $grp aio_group sub {
		1163	$grp->result (\@dirs, \@nondirs);
		1164	};
		1165
		1166	limit $statgrp $maxreq;
		1167	feed $statgrp sub {
		1168	return unless @$entries;
		1169	my $entry = shift @$entries;
		1170
		1171	aioreq_pri $pri;
		1172	$wd->[1] = "$entry/.";
		1173	add $statgrp aio_stat $wd, sub {
		1174	if ($_[0] < 0) {
		1175	push @nondirs, $entry;
		1176	} else {
		1177	# need to check for real directory
		1178	aioreq_pri $pri;
		1179	$wd->[1] = $entry;
		1180	add $statgrp aio_lstat $wd, sub {
		1181	if (-d _) {
		1182	push @dirs, $entry;
		1183
		1184	unless (--$ndirs) {
		1185	push @nondirs, @$entries;
		1186	feed $statgrp;
		1187	}
		1188	} else {
		1189	push @nondirs, $entry;
		1190	}
		1191	}
		1192	}
		1193	};
		1194	};
		1195	};
		1196	};
		1197	};
		1198	};
		1199
		1200	$grp
		1201	}
		1202
		1203	=item aio_rmtree $pathname, $callback->($status)
		1204
		1205	Delete a directory tree starting (and including) C<$path>, return the
		1206	status of the final C<rmdir> only. This is a composite request that
		1207	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1208	everything else.
		1209
		1210	=cut
		1211
		1212	sub aio_rmtree;
		1213	sub aio_rmtree($;$) {
		1214	my ($path, $cb) = @_;
		1215
		1216	my $pri = aioreq_pri;
		1217	my $grp = aio_group $cb;
		1218
		1219	aioreq_pri $pri;
		1220	add $grp aio_scandir $path, 0, sub {
		1221	my ($dirs, $nondirs) = @_;
		1222
		1223	my $dirgrp = aio_group sub {
		1224	add $grp aio_rmdir $path, sub {
		1225	$grp->result ($_[0]);
		1226	};
		1227	};
		1228
		1229	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1230	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1231
		1232	add $grp $dirgrp;
		1233	};
		1234
		1235	$grp
		1236	}
		1237
		1238	=item aio_fcntl $fh, $cmd, $arg, $callback->($status)
		1239
		1240	=item aio_ioctl $fh, $request, $buf, $callback->($status)
		1241
		1242	These work just like the C<fcntl> and C<ioctl> built-in functions, except
		1243	they execute asynchronously and pass the return value to the callback.
		1244
		1245	Both calls can be used for a lot of things, some of which make more sense
		1246	to run asynchronously in their own thread, while some others make less
		1247	sense. For example, calls that block waiting for external events, such
		1248	as locking, will also lock down an I/O thread while it is waiting, which
		1249	can deadlock the whole I/O system. At the same time, there might be no
		1250	alternative to using a thread to wait.
		1251
		1252	So in general, you should only use these calls for things that do
		1253	(filesystem) I/O, not for things that wait for other events (network,
		1254	other processes), although if you are careful and know what you are doing,
		1255	you still can.
		1256
		1257	The following constants are available (missing ones are, as usual C<0>):
		1258
		1259	C<FIFREEZE>, C<FITHAW>, C<FITRIM>, C<FICLONE>, C<FICLONERANGE>, C<FIDEDUPERANGE>.
		1260
		1261	C<FS_IOC_GETFLAGS>, C<FS_IOC_SETFLAGS>, C<FS_IOC_GETVERSION>, C<FS_IOC_SETVERSION>,
		1262	C<FS_IOC_FIEMAP>.
		1263
		1264	C<FS_IOC_FSGETXATTR>, C<FS_IOC_FSSETXATTR>, C<FS_IOC_SET_ENCRYPTION_POLICY>,
		1265	C<FS_IOC_GET_ENCRYPTION_PWSALT>, C<FS_IOC_GET_ENCRYPTION_POLICY>, C<FS_KEY_DESCRIPTOR_SIZE>.
		1266
		1267	C<FS_SECRM_FL>, C<FS_UNRM_FL>, C<FS_COMPR_FL>, C<FS_SYNC_FL>, C<FS_IMMUTABLE_FL>,
		1268	C<FS_APPEND_FL>, C<FS_NODUMP_FL>, C<FS_NOATIME_FL>, C<FS_DIRTY_FL>,
		1269	C<FS_COMPRBLK_FL>, C<FS_NOCOMP_FL>, C<FS_ENCRYPT_FL>, C<FS_BTREE_FL>,
		1270	C<FS_INDEX_FL>, C<FS_JOURNAL_DATA_FL>, C<FS_NOTAIL_FL>, C<FS_DIRSYNC_FL>, C<FS_TOPDIR_FL>,
		1271	C<FS_FL_USER_MODIFIABLE>.
		1272
		1273	C<FS_XFLAG_REALTIME>, C<FS_XFLAG_PREALLOC>, C<FS_XFLAG_IMMUTABLE>, C<FS_XFLAG_APPEND>,
		1274	C<FS_XFLAG_SYNC>, C<FS_XFLAG_NOATIME>, C<FS_XFLAG_NODUMP>, C<FS_XFLAG_RTINHERIT>,
		1275	C<FS_XFLAG_PROJINHERIT>, C<FS_XFLAG_NOSYMLINKS>, C<FS_XFLAG_EXTSIZE>, C<FS_XFLAG_EXTSZINHERIT>,
		1276	C<FS_XFLAG_NODEFRAG>, C<FS_XFLAG_FILESTREAM>, C<FS_XFLAG_DAX>, C<FS_XFLAG_HASATTR>,
		1277
		1278	=item aio_sync $callback->($status)
		1279
		1280	Asynchronously call sync and call the callback when finished.
		1281
179	=item aio_fsync $fh, $callback	1282	=item aio_fsync $fh, $callback->($status)
180		1283
181	Asynchronously call fsync on the given filehandle and call the callback	1284	Asynchronously call fsync on the given filehandle and call the callback
182	with the fsync result code.	1285	with the fsync result code.
183		1286
184	=item aio_fdatasync $fh, $callback	1287	=item aio_fdatasync $fh, $callback->($status)
185		1288
186	Asynchronously call fdatasync on the given filehandle and call the	1289	Asynchronously call fdatasync on the given filehandle and call the
187	callback with the fdatasync result code.	1290	callback with the fdatasync result code.
188		1291
		1292	If this call isn't available because your OS lacks it or it couldn't be
		1293	detected, it will be emulated by calling C<fsync> instead.
		1294
		1295	=item aio_syncfs $fh, $callback->($status)
		1296
		1297	Asynchronously call the syncfs syscall to sync the filesystem associated
		1298	to the given filehandle and call the callback with the syncfs result
		1299	code. If syncfs is not available, calls sync(), but returns C<-1> and sets
		1300	errno to C<ENOSYS> nevertheless.
		1301
		1302	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1303
		1304	Sync the data portion of the file specified by C<$offset> and C<$length>
		1305	to disk (but NOT the metadata), by calling the Linux-specific
		1306	sync_file_range call. If sync_file_range is not available or it returns
		1307	ENOSYS, then fdatasync or fsync is being substituted.
		1308
		1309	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1310	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1311	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1312	manpage for details.
		1313
		1314	=item aio_pathsync $pathname, $callback->($status)
		1315
		1316	This request tries to open, fsync and close the given path. This is a
		1317	composite request intended to sync directories after directory operations
		1318	(E.g. rename). This might not work on all operating systems or have any
		1319	specific effect, but usually it makes sure that directory changes get
		1320	written to disc. It works for anything that can be opened for read-only,
		1321	not just directories.
		1322
		1323	Future versions of this function might fall back to other methods when
		1324	C<fsync> on the directory fails (such as calling C<sync>).
		1325
		1326	Passes C<0> when everything went ok, and C<-1> on error.
		1327
		1328	=cut
		1329
		1330	sub aio_pathsync($;$) {
		1331	my ($path, $cb) = @_;
		1332
		1333	my $pri = aioreq_pri;
		1334	my $grp = aio_group $cb;
		1335
		1336	aioreq_pri $pri;
		1337	add $grp aio_open $path, O_RDONLY, 0, sub {
		1338	my ($fh) = @_;
		1339	if ($fh) {
		1340	aioreq_pri $pri;
		1341	add $grp aio_fsync $fh, sub {
		1342	$grp->result ($_[0]);
		1343
		1344	aioreq_pri $pri;
		1345	add $grp aio_close $fh;
		1346	};
		1347	} else {
		1348	$grp->result (-1);
		1349	}
		1350	};
		1351
		1352	$grp
		1353	}
		1354
		1355	=item aio_msync $scalar, $offset = 0, $length = undef, flags = MS_SYNC, $callback->($status)
		1356
		1357	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1358	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1359	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1360	scalar must only be modified in-place while an aio operation is pending on
		1361	it).
		1362
		1363	It calls the C<msync> function of your OS, if available, with the memory
		1364	area starting at C<$offset> in the string and ending C<$length> bytes
		1365	later. If C<$length> is negative, counts from the end, and if C<$length>
		1366	is C<undef>, then it goes till the end of the string. The flags can be
		1367	either C<IO::AIO::MS_ASYNC> or C<IO::AIO::MS_SYNC>, plus an optional
		1368	C<IO::AIO::MS_INVALIDATE>.
		1369
		1370	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1371
		1372	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1373	scalars.
		1374
		1375	It touches (reads or writes) all memory pages in the specified
		1376	range inside the scalar. All caveats and parameters are the same
		1377	as for C<aio_msync>, above, except for flags, which must be either
		1378	C<0> (which reads all pages and ensures they are instantiated) or
		1379	C<IO::AIO::MT_MODIFY>, which modifies the memory pages (by reading and
		1380	writing an octet from it, which dirties the page).
		1381
		1382	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1383
		1384	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1385	scalars.
		1386
		1387	It reads in all the pages of the underlying storage into memory (if any)
		1388	and locks them, so they are not getting swapped/paged out or removed.
		1389
		1390	If C<$length> is undefined, then the scalar will be locked till the end.
		1391
		1392	On systems that do not implement C<mlock>, this function returns C<-1>
		1393	and sets errno to C<ENOSYS>.
		1394
		1395	Note that the corresponding C<munlock> is synchronous and is
		1396	documented under L<MISCELLANEOUS FUNCTIONS>.
		1397
		1398	Example: open a file, mmap and mlock it - both will be undone when
		1399	C<$data> gets destroyed.
		1400
		1401	open my $fh, "<", $path or die "$path: $!";
		1402	my $data;
		1403	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1404	aio_mlock $data; # mlock in background
		1405
		1406	=item aio_mlockall $flags, $callback->($status)
		1407
		1408	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1409	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1410
		1411	On systems that do not implement C<mlockall>, this function returns C<-1>
		1412	and sets errno to C<ENOSYS>.
		1413
		1414	Note that the corresponding C<munlockall> is synchronous and is
		1415	documented under L<MISCELLANEOUS FUNCTIONS>.
		1416
		1417	Example: asynchronously lock all current and future pages into memory.
		1418
		1419	aio_mlockall IO::AIO::MCL_FUTURE;
		1420
		1421	=item aio_fiemap $fh, $start, $length, $flags, $count, $cb->(\@extents)
		1422
		1423	Queries the extents of the given file (by calling the Linux C<FIEMAP>
		1424	ioctl, see L<http://cvs.schmorp.de/IO-AIO/doc/fiemap.txt> for details). If
		1425	the ioctl is not available on your OS, then this request will fail with
		1426	C<ENOSYS>.
		1427
		1428	C<$start> is the starting offset to query extents for, C<$length> is the
		1429	size of the range to query - if it is C<undef>, then the whole file will
		1430	be queried.
		1431
		1432	C<$flags> is a combination of flags (C<IO::AIO::FIEMAP_FLAG_SYNC> or
		1433	C<IO::AIO::FIEMAP_FLAG_XATTR> - C<IO::AIO::FIEMAP_FLAGS_COMPAT> is also
		1434	exported), and is normally C<0> or C<IO::AIO::FIEMAP_FLAG_SYNC> to query
		1435	the data portion.
		1436
		1437	C<$count> is the maximum number of extent records to return. If it is
		1438	C<undef>, then IO::AIO queries all extents of the range. As a very special
		1439	case, if it is C<0>, then the callback receives the number of extents
		1440	instead of the extents themselves (which is unreliable, see below).
		1441
		1442	If an error occurs, the callback receives no arguments. The special
		1443	C<errno> value C<IO::AIO::EBADR> is available to test for flag errors.
		1444
		1445	Otherwise, the callback receives an array reference with extent
		1446	structures. Each extent structure is an array reference itself, with the
		1447	following members:
		1448
		1449	[$logical, $physical, $length, $flags]
		1450
		1451	Flags is any combination of the following flag values (typically either C<0>
		1452	or C<IO::AIO::FIEMAP_EXTENT_LAST> (1)):
		1453
		1454	C<IO::AIO::FIEMAP_EXTENT_LAST>, C<IO::AIO::FIEMAP_EXTENT_UNKNOWN>,
		1455	C<IO::AIO::FIEMAP_EXTENT_DELALLOC>, C<IO::AIO::FIEMAP_EXTENT_ENCODED>,
		1456	C<IO::AIO::FIEMAP_EXTENT_DATA_ENCRYPTED>, C<IO::AIO::FIEMAP_EXTENT_NOT_ALIGNED>,
		1457	C<IO::AIO::FIEMAP_EXTENT_DATA_INLINE>, C<IO::AIO::FIEMAP_EXTENT_DATA_TAIL>,
		1458	C<IO::AIO::FIEMAP_EXTENT_UNWRITTEN>, C<IO::AIO::FIEMAP_EXTENT_MERGED> or
		1459	C<IO::AIO::FIEMAP_EXTENT_SHARED>.
		1460
		1461	At the time of this writing (Linux 3.2), this requets is unreliable unless
		1462	C<$count> is C<undef>, as the kernel has all sorts of bugs preventing
		1463	it to return all extents of a range for files with large number of
		1464	extents. The code works around all these issues if C<$count> is undef.
		1465
		1466	=item aio_group $callback->(...)
		1467
		1468	This is a very special aio request: Instead of doing something, it is a
		1469	container for other aio requests, which is useful if you want to bundle
		1470	many requests into a single, composite, request with a definite callback
		1471	and the ability to cancel the whole request with its subrequests.
		1472
		1473	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		1474	for more info.
		1475
		1476	Example:
		1477
		1478	my $grp = aio_group sub {
		1479	print "all stats done\n";
		1480	};
		1481
		1482	add $grp
		1483	(aio_stat ...),
		1484	(aio_stat ...),
		1485	...;
		1486
		1487	=item aio_nop $callback->()
		1488
		1489	This is a special request - it does nothing in itself and is only used for
		1490	side effects, such as when you want to add a dummy request to a group so
		1491	that finishing the requests in the group depends on executing the given
		1492	code.
		1493
		1494	While this request does nothing, it still goes through the execution
		1495	phase and still requires a worker thread. Thus, the callback will not
		1496	be executed immediately but only after other requests in the queue have
		1497	entered their execution phase. This can be used to measure request
		1498	latency.
		1499
		1500	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		1501
		1502	Mainly used for debugging and benchmarking, this aio request puts one of
		1503	the request workers to sleep for the given time.
		1504
		1505	While it is theoretically handy to have simple I/O scheduling requests
		1506	like sleep and file handle readable/writable, the overhead this creates is
		1507	immense (it blocks a thread for a long time) so do not use this function
		1508	except to put your application under artificial I/O pressure.
		1509
189	=back	1510	=back
190		1511
		1512
		1513	=head2 IO::AIO::WD - multiple working directories
		1514
		1515	Your process only has one current working directory, which is used by all
		1516	threads. This makes it hard to use relative paths (some other component
		1517	could call C<chdir> at any time, and it is hard to control when the path
		1518	will be used by IO::AIO).
		1519
		1520	One solution for this is to always use absolute paths. This usually works,
		1521	but can be quite slow (the kernel has to walk the whole path on every
		1522	access), and can also be a hassle to implement.
		1523
		1524	Newer POSIX systems have a number of functions (openat, fdopendir,
		1525	futimensat and so on) that make it possible to specify working directories
		1526	per operation.
		1527
		1528	For portability, and because the clowns who "designed", or shall I write,
		1529	perpetrated this new interface were obviously half-drunk, this abstraction
		1530	cannot be perfect, though.
		1531
		1532	IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
		1533	object. This object stores the canonicalised, absolute version of the
		1534	path, and on systems that allow it, also a directory file descriptor.
		1535
		1536	Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
		1537	or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
		1538	object and a pathname instead (or the IO::AIO::WD object alone, which
		1539	gets interpreted as C<[$wd, "."]>). If the pathname is absolute, the
		1540	IO::AIO::WD object is ignored, otherwise the pathname is resolved relative
		1541	to that IO::AIO::WD object.
		1542
		1543	For example, to get a wd object for F</etc> and then stat F<passwd>
		1544	inside, you would write:
		1545
		1546	aio_wd "/etc", sub {
		1547	my $etcdir = shift;
		1548
		1549	# although $etcdir can be undef on error, there is generally no reason
		1550	# to check for errors here, as aio_stat will fail with ENOENT
		1551	# when $etcdir is undef.
		1552
		1553	aio_stat [$etcdir, "passwd"], sub {
		1554	# yay
		1555	};
		1556	};
		1557
		1558	The fact that C<aio_wd> is a request and not a normal function shows that
		1559	creating an IO::AIO::WD object is itself a potentially blocking operation,
		1560	which is why it is done asynchronously.
		1561
		1562	To stat the directory obtained with C<aio_wd> above, one could write
		1563	either of the following three request calls:
		1564
		1565	aio_lstat "/etc" , sub { ... # pathname as normal string
		1566	aio_lstat [$wd, "."], sub { ... # "." relative to $wd (i.e. $wd itself)
		1567	aio_lstat $wd , sub { ... # shorthand for the previous
		1568
		1569	As with normal pathnames, IO::AIO keeps a copy of the working directory
		1570	object and the pathname string, so you could write the following without
		1571	causing any issues due to C<$path> getting reused:
		1572
		1573	my $path = [$wd, undef];
		1574
		1575	for my $name (qw(abc def ghi)) {
		1576	$path->[1] = $name;
		1577	aio_stat $path, sub {
		1578	# ...
		1579	};
		1580	}
		1581
		1582	There are some caveats: when directories get renamed (or deleted), the
		1583	pathname string doesn't change, so will point to the new directory (or
		1584	nowhere at all), while the directory fd, if available on the system,
		1585	will still point to the original directory. Most functions accepting a
		1586	pathname will use the directory fd on newer systems, and the string on
		1587	older systems. Some functions (such as realpath) will always rely on the
		1588	string form of the pathname.
		1589
		1590	So this functionality is mainly useful to get some protection against
		1591	C<chdir>, to easily get an absolute path out of a relative path for future
		1592	reference, and to speed up doing many operations in the same directory
		1593	(e.g. when stat'ing all files in a directory).
		1594
		1595	The following functions implement this working directory abstraction:
		1596
		1597	=over 4
		1598
		1599	=item aio_wd $pathname, $callback->($wd)
		1600
		1601	Asynchonously canonicalise the given pathname and convert it to an
		1602	IO::AIO::WD object representing it. If possible and supported on the
		1603	system, also open a directory fd to speed up pathname resolution relative
		1604	to this working directory.
		1605
		1606	If something goes wrong, then C<undef> is passwd to the callback instead
		1607	of a working directory object and C<$!> is set appropriately. Since
		1608	passing C<undef> as working directory component of a pathname fails the
		1609	request with C<ENOENT>, there is often no need for error checking in the
		1610	C<aio_wd> callback, as future requests using the value will fail in the
		1611	expected way.
		1612
		1613	=item IO::AIO::CWD
		1614
		1615	This is a compiletime constant (object) that represents the process
		1616	current working directory.
		1617
		1618	Specifying this object as working directory object for a pathname is as if
		1619	the pathname would be specified directly, without a directory object. For
		1620	example, these calls are functionally identical:
		1621
		1622	aio_stat "somefile", sub { ... };
		1623	aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
		1624
		1625	=back
		1626
		1627	To recover the path associated with an IO::AIO::WD object, you can use
		1628	C<aio_realpath>:
		1629
		1630	aio_realpath $wd, sub {
		1631	warn "path is $_[0]\n";
		1632	};
		1633
		1634	Currently, C<aio_statvfs> always, and C<aio_rename> and C<aio_rmdir>
		1635	sometimes, fall back to using an absolue path.
		1636
		1637	=head2 IO::AIO::REQ CLASS
		1638
		1639	All non-aggregate C<aio_*> functions return an object of this class when
		1640	called in non-void context.
		1641
		1642	=over 4
		1643
		1644	=item cancel $req
		1645
		1646	Cancels the request, if possible. Has the effect of skipping execution
		1647	when entering the B<execute> state and skipping calling the callback when
		1648	entering the the B<result> state, but will leave the request otherwise
		1649	untouched (with the exception of readdir). That means that requests that
		1650	currently execute will not be stopped and resources held by the request
		1651	will not be freed prematurely.
		1652
		1653	=item cb $req $callback->(...)
		1654
		1655	Replace (or simply set) the callback registered to the request.
		1656
		1657	=back
		1658
		1659	=head2 IO::AIO::GRP CLASS
		1660
		1661	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		1662	objects of this class, too.
		1663
		1664	A IO::AIO::GRP object is a special request that can contain multiple other
		1665	aio requests.
		1666
		1667	You create one by calling the C<aio_group> constructing function with a
		1668	callback that will be called when all contained requests have entered the
		1669	C<done> state:
		1670
		1671	my $grp = aio_group sub {
		1672	print "all requests are done\n";
		1673	};
		1674
		1675	You add requests by calling the C<add> method with one or more
		1676	C<IO::AIO::REQ> objects:
		1677
		1678	$grp->add (aio_unlink "...");
		1679
		1680	add $grp aio_stat "...", sub {
		1681	$_[0] or return $grp->result ("error");
		1682
		1683	# add another request dynamically, if first succeeded
		1684	add $grp aio_open "...", sub {
		1685	$grp->result ("ok");
		1686	};
		1687	};
		1688
		1689	This makes it very easy to create composite requests (see the source of
		1690	C<aio_move> for an application) that work and feel like simple requests.
		1691
		1692	=over 4
		1693
		1694	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		1695	C<IO::AIO::poll_cb>, just like any other request.
		1696
		1697	=item * They can be canceled like any other request. Canceling will cancel not
		1698	only the request itself, but also all requests it contains.
		1699
		1700	=item * They can also can also be added to other IO::AIO::GRP objects.
		1701
		1702	=item * You must not add requests to a group from within the group callback (or
		1703	any later time).
		1704
		1705	=back
		1706
		1707	Their lifetime, simplified, looks like this: when they are empty, they
		1708	will finish very quickly. If they contain only requests that are in the
		1709	C<done> state, they will also finish. Otherwise they will continue to
		1710	exist.
		1711
		1712	That means after creating a group you have some time to add requests
		1713	(precisely before the callback has been invoked, which is only done within
		1714	the C<poll_cb>). And in the callbacks of those requests, you can add
		1715	further requests to the group. And only when all those requests have
		1716	finished will the the group itself finish.
		1717
		1718	=over 4
		1719
		1720	=item add $grp ...
		1721
		1722	=item $grp->add (...)
		1723
		1724	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1725	be added, including other groups, as long as you do not create circular
		1726	dependencies.
		1727
		1728	Returns all its arguments.
		1729
		1730	=item $grp->cancel_subs
		1731
		1732	Cancel all subrequests and clears any feeder, but not the group request
		1733	itself. Useful when you queued a lot of events but got a result early.
		1734
		1735	The group request will finish normally (you cannot add requests to the
		1736	group).
		1737
		1738	=item $grp->result (...)
		1739
		1740	Set the result value(s) that will be passed to the group callback when all
		1741	subrequests have finished and set the groups errno to the current value
		1742	of errno (just like calling C<errno> without an error number). By default,
		1743	no argument will be passed and errno is zero.
		1744
		1745	=item $grp->errno ([$errno])
		1746
		1747	Sets the group errno value to C<$errno>, or the current value of errno
		1748	when the argument is missing.
		1749
		1750	Every aio request has an associated errno value that is restored when
		1751	the callback is invoked. This method lets you change this value from its
		1752	default (0).
		1753
		1754	Calling C<result> will also set errno, so make sure you either set C<$!>
		1755	before the call to C<result>, or call c<errno> after it.
		1756
		1757	=item feed $grp $callback->($grp)
		1758
		1759	Sets a feeder/generator on this group: every group can have an attached
		1760	generator that generates requests if idle. The idea behind this is that,
		1761	although you could just queue as many requests as you want in a group,
		1762	this might starve other requests for a potentially long time. For example,
		1763	C<aio_scandir> might generate hundreds of thousands of C<aio_stat>
		1764	requests, delaying any later requests for a long time.
		1765
		1766	To avoid this, and allow incremental generation of requests, you can
		1767	instead a group and set a feeder on it that generates those requests. The
		1768	feed callback will be called whenever there are few enough (see C<limit>,
		1769	below) requests active in the group itself and is expected to queue more
		1770	requests.
		1771
		1772	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1773	not impose any limits).
		1774
		1775	If the feed does not queue more requests when called, it will be
		1776	automatically removed from the group.
		1777
		1778	If the feed limit is C<0> when this method is called, it will be set to
		1779	C<2> automatically.
		1780
		1781	Example:
		1782
		1783	# stat all files in @files, but only ever use four aio requests concurrently:
		1784
		1785	my $grp = aio_group sub { print "finished\n" };
		1786	limit $grp 4;
		1787	feed $grp sub {
		1788	my $file = pop @files
		1789	or return;
		1790
		1791	add $grp aio_stat $file, sub { ... };
		1792	};
		1793
		1794	=item limit $grp $num
		1795
		1796	Sets the feeder limit for the group: The feeder will be called whenever
		1797	the group contains less than this many requests.
		1798
		1799	Setting the limit to C<0> will pause the feeding process.
		1800
		1801	The default value for the limit is C<0>, but note that setting a feeder
		1802	automatically bumps it up to C<2>.
		1803
		1804	=back
		1805
191	=head2 SUPPORT FUNCTIONS	1806	=head2 SUPPORT FUNCTIONS
192		1807
		1808	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1809
193	=over 4	1810	=over 4
194		1811
195	=item $fileno = IO::AIO::poll_fileno	1812	=item $fileno = IO::AIO::poll_fileno
196		1813
197	Return the I<request result pipe filehandle>. This filehandle must be	1814	Return the I<request result pipe file descriptor>. This filehandle must be
198	polled for reading by some mechanism outside this module (e.g. Event	1815	polled for reading by some mechanism outside this module (e.g. EV, Glib,
199	or select, see below). If the pipe becomes readable you have to call	1816	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
200	C<poll_cb> to check the results.	1817	you have to call C<poll_cb> to check the results.
201		1818
202	See C<poll_cb> for an example.	1819	See C<poll_cb> for an example.
203		1820
204	=item IO::AIO::poll_cb	1821	=item IO::AIO::poll_cb
205		1822
206	Process all outstanding events on the result pipe. You have to call this	1823	Process some requests that have reached the result phase (i.e. they have
207	regularly. Returns the number of events processed. Returns immediately	1824	been executed but the results are not yet reported). You have to call
208	when no events are outstanding.	1825	this "regularly" to finish outstanding requests.
209		1826
210	You can use Event to multiplex, e.g.:	1827	Returns C<0> if all events could be processed (or there were no
		1828	events to process), or C<-1> if it returned earlier for whatever
		1829	reason. Returns immediately when no events are outstanding. The amount
		1830	of events processed depends on the settings of C<IO::AIO::max_poll_req>,
		1831	C<IO::AIO::max_poll_time> and C<IO::AIO::max_outstanding>.
		1832
		1833	If not all requests were processed for whatever reason, the poll file
		1834	descriptor will still be ready when C<poll_cb> returns, so normally you
		1835	don't have to do anything special to have it called later.
		1836
		1837	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1838	ready, it can be beneficial to call this function from loops which submit
		1839	a lot of requests, to make sure the results get processed when they become
		1840	available and not just when the loop is finished and the event loop takes
		1841	over again. This function returns very fast when there are no outstanding
		1842	requests.
		1843
		1844	Example: Install an Event watcher that automatically calls
		1845	IO::AIO::poll_cb with high priority (more examples can be found in the
		1846	SYNOPSIS section, at the top of this document):
211		1847
212	Event->io (fd => IO::AIO::poll_fileno,	1848	Event->io (fd => IO::AIO::poll_fileno,
213	poll => 'r', async => 1,	1849	poll => 'r', async => 1,
214	cb => \&IO::AIO::poll_cb);	1850	cb => \&IO::AIO::poll_cb);
215		1851
216	=item IO::AIO::poll_wait	1852	=item IO::AIO::poll_wait
217		1853
218	Wait till the result filehandle becomes ready for reading (simply does a	1854	Wait until either at least one request is in the result phase or no
219	select on the filehandle. This is useful if you want to synchronously wait	1855	requests are outstanding anymore.
220	for some requests to finish).	1856
		1857	This is useful if you want to synchronously wait for some requests to
		1858	become ready, without actually handling them.
221		1859
222	See C<nreqs> for an example.	1860	See C<nreqs> for an example.
223		1861
		1862	=item IO::AIO::poll
		1863
		1864	Waits until some requests have been handled.
		1865
		1866	Returns the number of requests processed, but is otherwise strictly
		1867	equivalent to:
		1868
		1869	IO::AIO::poll_wait, IO::AIO::poll_cb
		1870
224	=item IO::AIO::nreqs	1871	=item IO::AIO::flush
225		1872
226	Returns the number of requests currently outstanding.	1873	Wait till all outstanding AIO requests have been handled.
227		1874
228	Example: wait till there are no outstanding requests anymore:	1875	Strictly equivalent to:
229		1876
230	IO::AIO::poll_wait, IO::AIO::poll_cb	1877	IO::AIO::poll_wait, IO::AIO::poll_cb
231	while IO::AIO::nreqs;	1878	while IO::AIO::nreqs;
232		1879
		1880	=item IO::AIO::max_poll_reqs $nreqs
		1881
		1882	=item IO::AIO::max_poll_time $seconds
		1883
		1884	These set the maximum number of requests (default C<0>, meaning infinity)
		1885	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1886	the maximum amount of time (default C<0>, meaning infinity) spent in
		1887	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1888	of time C<poll_cb> is allowed to use).
		1889
		1890	Setting C<max_poll_time> to a non-zero value creates an overhead of one
		1891	syscall per request processed, which is not normally a problem unless your
		1892	callbacks are really really fast or your OS is really really slow (I am
		1893	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
		1894
		1895	Setting these is useful if you want to ensure some level of
		1896	interactiveness when perl is not fast enough to process all requests in
		1897	time.
		1898
		1899	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
		1900
		1901	Example: Install an Event watcher that automatically calls
		1902	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1903	program get the CPU sometimes even under high AIO load.
		1904
		1905	# try not to spend much more than 0.1s in poll_cb
		1906	IO::AIO::max_poll_time 0.1;
		1907
		1908	# use a low priority so other tasks have priority
		1909	Event->io (fd => IO::AIO::poll_fileno,
		1910	poll => 'r', nice => 1,
		1911	cb => &IO::AIO::poll_cb);
		1912
		1913	=back
		1914
		1915	=head3 CONTROLLING THE NUMBER OF THREADS
		1916
		1917	=over
		1918
233	=item IO::AIO::min_parallel $nthreads	1919	=item IO::AIO::min_parallel $nthreads
234		1920
235	Set the minimum number of AIO threads to C<$nthreads>. The default is	1921	Set the minimum number of AIO threads to C<$nthreads>. The current
236	C<1>, which means a single asynchronous operation can be done at one time	1922	default is C<8>, which means eight asynchronous operations can execute
237	(the number of outstanding operations, however, is unlimited).	1923	concurrently at any one time (the number of outstanding requests,
		1924	however, is unlimited).
238		1925
		1926	IO::AIO starts threads only on demand, when an AIO request is queued and
		1927	no free thread exists. Please note that queueing up a hundred requests can
		1928	create demand for a hundred threads, even if it turns out that everything
		1929	is in the cache and could have been processed faster by a single thread.
		1930
239	It is recommended to keep the number of threads low, as some Linux	1931	It is recommended to keep the number of threads relatively low, as some
240	kernel versions will scale negatively with the number of threads (higher	1932	Linux kernel versions will scale negatively with the number of threads
241	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1933	(higher parallelity => MUCH higher latency). With current Linux 2.6
242	threads should be fine.	1934	versions, 4-32 threads should be fine.
243		1935
244	Under normal circumstances you don't need to call this function, as this	1936	Under most circumstances you don't need to call this function, as the
245	module automatically starts some threads (the exact number might change,	1937	module selects a default that is suitable for low to moderate load.
246	and is currently 4).
247		1938
248	=item IO::AIO::max_parallel $nthreads	1939	=item IO::AIO::max_parallel $nthreads
249		1940
250	Sets the maximum number of AIO threads to C<$nthreads>. If more than	1941	Sets the maximum number of AIO threads to C<$nthreads>. If more than the
251	the specified number of threads are currently running, kill them. This	1942	specified number of threads are currently running, this function kills
252	function blocks until the limit is reached.	1943	them. This function blocks until the limit is reached.
		1944
		1945	While C<$nthreads> are zero, aio requests get queued but not executed
		1946	until the number of threads has been increased again.
253		1947
254	This module automatically runs C<max_parallel 0> at program end, to ensure	1948	This module automatically runs C<max_parallel 0> at program end, to ensure
255	that all threads are killed and that there are no outstanding requests.	1949	that all threads are killed and that there are no outstanding requests.
256		1950
257	Under normal circumstances you don't need to call this function.	1951	Under normal circumstances you don't need to call this function.
258		1952
		1953	=item IO::AIO::max_idle $nthreads
		1954
		1955	Limit the number of threads (default: 4) that are allowed to idle
		1956	(i.e., threads that did not get a request to process within the idle
		1957	timeout (default: 10 seconds). That means if a thread becomes idle while
		1958	C<$nthreads> other threads are also idle, it will free its resources and
		1959	exit.
		1960
		1961	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1962	to allow for extremely high load situations, but want to free resources
		1963	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1964
		1965	The default is probably ok in most situations, especially if thread
		1966	creation is fast. If thread creation is very slow on your system you might
		1967	want to use larger values.
		1968
		1969	=item IO::AIO::idle_timeout $seconds
		1970
		1971	Sets the minimum idle timeout (default 10) after which worker threads are
		1972	allowed to exit. SEe C<IO::AIO::max_idle>.
		1973
259	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1974	=item IO::AIO::max_outstanding $maxreqs
260		1975
261	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1976	Sets the maximum number of outstanding requests to C<$nreqs>. If
262	try to queue up more than this number of requests, the caller will block until	1977	you do queue up more than this number of requests, the next call to
263	some requests have been handled.	1978	C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
		1979	C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
		1980	longer exceeded.
264		1981
265	The default is very large, so normally there is no practical limit. If you	1982	In other words, this setting does not enforce a queue limit, but can be
266	queue up many requests in a loop it it often improves speed if you set	1983	used to make poll functions block if the limit is exceeded.
267	this to a relatively low number, such as C<100>.
268		1984
269	Under normal circumstances you don't need to call this function.	1985	This is a very bad function to use in interactive programs because it
		1986	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1987	use an C<aio_group> together with a feed callback.
		1988
		1989	Its main use is in scripts without an event loop - when you want to stat
		1990	a lot of files, you can write somehting like this:
		1991
		1992	IO::AIO::max_outstanding 32;
		1993
		1994	for my $path (...) {
		1995	aio_stat $path , ...;
		1996	IO::AIO::poll_cb;
		1997	}
		1998
		1999	IO::AIO::flush;
		2000
		2001	The call to C<poll_cb> inside the loop will normally return instantly, but
		2002	as soon as more thna C<32> reqeusts are in-flight, it will block until
		2003	some requests have been handled. This keeps the loop from pushing a large
		2004	number of C<aio_stat> requests onto the queue.
		2005
		2006	The default value for C<max_outstanding> is very large, so there is no
		2007	practical limit on the number of outstanding requests.
270		2008
271	=back	2009	=back
272		2010
		2011	=head3 STATISTICAL INFORMATION
		2012
		2013	=over
		2014
		2015	=item IO::AIO::nreqs
		2016
		2017	Returns the number of requests currently in the ready, execute or pending
		2018	states (i.e. for which their callback has not been invoked yet).
		2019
		2020	Example: wait till there are no outstanding requests anymore:
		2021
		2022	IO::AIO::poll_wait, IO::AIO::poll_cb
		2023	while IO::AIO::nreqs;
		2024
		2025	=item IO::AIO::nready
		2026
		2027	Returns the number of requests currently in the ready state (not yet
		2028	executed).
		2029
		2030	=item IO::AIO::npending
		2031
		2032	Returns the number of requests currently in the pending state (executed,
		2033	but not yet processed by poll_cb).
		2034
		2035	=back
		2036
		2037	=head3 MISCELLANEOUS FUNCTIONS
		2038
		2039	IO::AIO implements some functions that are useful when you want to use
		2040	some "Advanced I/O" function not available to in Perl, without going the
		2041	"Asynchronous I/O" route. Many of these have an asynchronous C<aio_*>
		2042	counterpart.
		2043
		2044	=over 4
		2045
		2046	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		2047
		2048	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		2049	but is blocking (this makes most sense if you know the input data is
		2050	likely cached already and the output filehandle is set to non-blocking
		2051	operations).
		2052
		2053	Returns the number of bytes copied, or C<-1> on error.
		2054
		2055	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		2056
		2057	Simply calls the C<posix_fadvise> function (see its
		2058	manpage for details). The following advice constants are
		2059	available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		2060	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		2061	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		2062
		2063	On systems that do not implement C<posix_fadvise>, this function returns
		2064	ENOSYS, otherwise the return value of C<posix_fadvise>.
		2065
		2066	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		2067
		2068	Simply calls the C<posix_madvise> function (see its
		2069	manpage for details). The following advice constants are
		2070	available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		2071	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>,
		2072	C<IO::AIO::MADV_FREE>.
		2073
		2074	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2075	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2076	will be reduced to fit into the C<$scalar>.
		2077
		2078	On systems that do not implement C<posix_madvise>, this function returns
		2079	ENOSYS, otherwise the return value of C<posix_madvise>.
		2080
		2081	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		2082
		2083	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		2084	$scalar (see its manpage for details). The following protect
		2085	constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		2086	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		2087
		2088	If C<$offset> is negative, counts from the end. If C<$length> is negative,
		2089	the remaining length of the C<$scalar> is used. If possible, C<$length>
		2090	will be reduced to fit into the C<$scalar>.
		2091
		2092	On systems that do not implement C<mprotect>, this function returns
		2093	ENOSYS, otherwise the return value of C<mprotect>.
		2094
		2095	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		2096
		2097	Memory-maps a file (or anonymous memory range) and attaches it to the
		2098	given C<$scalar>, which will act like a string scalar. Returns true on
		2099	success, and false otherwise.
		2100
		2101	The scalar must exist, but its contents do not matter - this means you
		2102	cannot use a nonexistant array or hash element. When in doubt, C<undef>
		2103	the scalar first.
		2104
		2105	The only operations allowed on the mmapped scalar are C<substr>/C<vec>,
		2106	which don't change the string length, and most read-only operations such
		2107	as copying it or searching it with regexes and so on.
		2108
		2109	Anything else is unsafe and will, at best, result in memory leaks.
		2110
		2111	The memory map associated with the C<$scalar> is automatically removed
		2112	when the C<$scalar> is undef'd or destroyed, or when the C<IO::AIO::mmap>
		2113	or C<IO::AIO::munmap> functions are called on it.
		2114
		2115	This calls the C<mmap>(2) function internally. See your system's manual
		2116	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		2117
		2118	The C<$length> must be larger than zero and smaller than the actual
		2119	filesize.
		2120
		2121	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		2122	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		2123
		2124	C<$flags> can be a combination of
		2125	C<IO::AIO::MAP_SHARED> or
		2126	C<IO::AIO::MAP_PRIVATE>,
		2127	or a number of system-specific flags (when not available, the are C<0>):
		2128	C<IO::AIO::MAP_ANONYMOUS> (which is set to C<MAP_ANON> if your system only provides this constant),
		2129	C<IO::AIO::MAP_LOCKED>,
		2130	C<IO::AIO::MAP_NORESERVE>,
		2131	C<IO::AIO::MAP_POPULATE>,
		2132	C<IO::AIO::MAP_NONBLOCK>,
		2133	C<IO::AIO::MAP_FIXED>,
		2134	C<IO::AIO::MAP_GROWSDOWN>,
		2135	C<IO::AIO::MAP_32BIT>,
		2136	C<IO::AIO::MAP_HUGETLB> or
		2137	C<IO::AIO::MAP_STACK>.
		2138
		2139	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		2140
		2141	C<$offset> is the offset from the start of the file - it generally must be
		2142	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		2143
		2144	Example:
		2145
		2146	use Digest::MD5;
		2147	use IO::AIO;
		2148
		2149	open my $fh, "<verybigfile"
		2150	or die "$!";
		2151
		2152	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		2153	or die "verybigfile: $!";
		2154
		2155	my $fast_md5 = md5 $data;
		2156
		2157	=item IO::AIO::munmap $scalar
		2158
		2159	Removes a previous mmap and undefines the C<$scalar>.
		2160
		2161	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		2162
		2163	Calls the C<munlock> function, undoing the effects of a previous
		2164	C<aio_mlock> call (see its description for details).
		2165
		2166	=item IO::AIO::munlockall
		2167
		2168	Calls the C<munlockall> function.
		2169
		2170	On systems that do not implement C<munlockall>, this function returns
		2171	ENOSYS, otherwise the return value of C<munlockall>.
		2172
		2173	=item IO::AIO::splice $r_fh, $r_off, $w_fh, $w_off, $length, $flags
		2174
		2175	Calls the GNU/Linux C<splice(2)> syscall, if available. If C<$r_off> or
		2176	C<$w_off> are C<undef>, then C<NULL> is passed for these, otherwise they
		2177	should be the file offset.
		2178
		2179	C<$r_fh> and C<$w_fh> should not refer to the same file, as splice might
		2180	silently corrupt the data in this case.
		2181
		2182	The following symbol flag values are available: C<IO::AIO::SPLICE_F_MOVE>,
		2183	C<IO::AIO::SPLICE_F_NONBLOCK>, C<IO::AIO::SPLICE_F_MORE> and
		2184	C<IO::AIO::SPLICE_F_GIFT>.
		2185
		2186	See the C<splice(2)> manpage for details.
		2187
		2188	=item IO::AIO::tee $r_fh, $w_fh, $length, $flags
		2189
		2190	Calls the GNU/Linux C<tee(2)> syscall, see its manpage and the
		2191	description for C<IO::AIO::splice> above for details.
		2192
		2193	=item $actual_size = IO::AIO::pipesize $r_fh[, $new_size]
		2194
		2195	Attempts to query or change the pipe buffer size. Obviously works only
		2196	on pipes, and currently works only on GNU/Linux systems, and fails with
		2197	C<-1>/C<ENOSYS> everywhere else. If anybody knows how to influence pipe buffer
		2198	size on other systems, drop me a note.
		2199
		2200	=item ($rfh, $wfh) = IO::AIO::pipe2 [$flags]
		2201
		2202	This is a direct interface to the Linux L<pipe2(2)> system call. If
		2203	C<$flags> is missing or C<0>, then this should be the same as a call to
		2204	perl's built-in C<pipe> function and create a new pipe, and works on
		2205	systems that lack the pipe2 syscall. On win32, this case invokes C<_pipe
		2206	(..., 4096, O_BINARY)>.
		2207
		2208	If C<$flags> is non-zero, it tries to invoke the pipe2 system call with
		2209	the given flags (Linux 2.6.27, glibc 2.9).
		2210
		2211	On success, the read and write file handles are returned.
		2212
		2213	On error, nothing will be returned. If the pipe2 syscall is missing and
		2214	C<$flags> is non-zero, fails with C<ENOSYS>.
		2215
		2216	Please refer to L<pipe2(2)> for more info on the C<$flags>, but at the
		2217	time of this writing, C<IO::AIO::O_CLOEXEC>, C<IO::AIO::O_NONBLOCK> and
		2218	C<IO::AIO::O_DIRECT> (Linux 3.4, for packet-based pipes) were supported.
		2219
		2220	=back
		2221
273	=cut	2222	=cut
274		2223
275	# support function to convert a fd into a perl filehandle
276	sub _fd2fh {
277	return undef if $_[0] < 0;
278
279	# try to be perl5.6-compatible
280	local *AIO_FH;
281	open AIO_FH, "+<&=$_[0]"
282	or return undef;
283
284	*AIO_FH
285	}
286
287	min_parallel 4;	2224	min_parallel 8;
288		2225
289	END {	2226	END { flush }
290	max_parallel 0;
291	}
292		2227
293	1;	2228	1;
294		2229
		2230	=head1 EVENT LOOP INTEGRATION
		2231
		2232	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		2233	automatically into many event loops:
		2234
		2235	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		2236	use AnyEvent::AIO;
		2237
		2238	You can also integrate IO::AIO manually into many event loops, here are
		2239	some examples of how to do this:
		2240
		2241	# EV integration
		2242	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		2243
		2244	# Event integration
		2245	Event->io (fd => IO::AIO::poll_fileno,
		2246	poll => 'r',
		2247	cb => \&IO::AIO::poll_cb);
		2248
		2249	# Glib/Gtk2 integration
		2250	add_watch Glib::IO IO::AIO::poll_fileno,
		2251	in => sub { IO::AIO::poll_cb; 1 };
		2252
		2253	# Tk integration
		2254	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		2255	readable => \&IO::AIO::poll_cb);
		2256
		2257	# Danga::Socket integration
		2258	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		2259	\&IO::AIO::poll_cb);
		2260
		2261	=head2 FORK BEHAVIOUR
		2262
		2263	Usage of pthreads in a program changes the semantics of fork
		2264	considerably. Specifically, only async-safe functions can be called after
		2265	fork. Perl doesn't know about this, so in general, you cannot call fork
		2266	with defined behaviour in perl if pthreads are involved. IO::AIO uses
		2267	pthreads, so this applies, but many other extensions and (for inexplicable
		2268	reasons) perl itself often is linked against pthreads, so this limitation
		2269	applies to quite a lot of perls.
		2270
		2271	This module no longer tries to fight your OS, or POSIX. That means IO::AIO
		2272	only works in the process that loaded it. Forking is fully supported, but
		2273	using IO::AIO in the child is not.
		2274
		2275	You might get around by not I<using> IO::AIO before (or after)
		2276	forking. You could also try to call the L<IO::AIO::reinit> function in the
		2277	child:
		2278
		2279	=over 4
		2280
		2281	=item IO::AIO::reinit
		2282
		2283	Abandons all current requests and I/O threads and simply reinitialises all
		2284	data structures. This is not an operation supported by any standards, but
		2285	happens to work on GNU/Linux and some newer BSD systems.
		2286
		2287	The only reasonable use for this function is to call it after forking, if
		2288	C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
		2289	the process will result in undefined behaviour. Calling it at any time
		2290	will also result in any undefined (by POSIX) behaviour.
		2291
		2292	=back
		2293
		2294	=head2 MEMORY USAGE
		2295
		2296	Per-request usage:
		2297
		2298	Each aio request uses - depending on your architecture - around 100-200
		2299	bytes of memory. In addition, stat requests need a stat buffer (possibly
		2300	a few hundred bytes), readdir requires a result buffer and so on. Perl
		2301	scalars and other data passed into aio requests will also be locked and
		2302	will consume memory till the request has entered the done state.
		2303
		2304	This is not awfully much, so queuing lots of requests is not usually a
		2305	problem.
		2306
		2307	Per-thread usage:
		2308
		2309	In the execution phase, some aio requests require more memory for
		2310	temporary buffers, and each thread requires a stack and other data
		2311	structures (usually around 16k-128k, depending on the OS).
		2312
		2313	=head1 KNOWN BUGS
		2314
		2315	Known bugs will be fixed in the next release.
		2316
295	=head1 SEE ALSO	2317	=head1 SEE ALSO
296		2318
297	L<Coro>, L<Linux::AIO>.	2319	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		2320	more natural syntax.
298		2321
299	=head1 AUTHOR	2322	=head1 AUTHOR
300		2323
301	Marc Lehmann <schmorp@schmorp.de>	2324	Marc Lehmann <schmorp@schmorp.de>
302	http://home.schmorp.de/	2325	http://home.schmorp.de/

Diff Legend

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