ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.45 by root, Mon Dec 26 19:20:26 2005 UTC vs.
Revision 1.209 by root, Tue Sep 27 00:41:51 2011 UTC

…		…
4		4
5	=head1 SYNOPSIS	5	=head1 SYNOPSIS
6		6
7	use IO::AIO;	7	use IO::AIO;
8		8
9	aio_open "/etc/passwd", O_RDONLY, 0, sub {	9	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
10	my ($fh) = @_;	10	my $fh = shift
		11	or die "/etc/passwd: $!";
11	...	12	...
12	};	13	};
13		14
14	aio_unlink "/tmp/file", sub { };	15	aio_unlink "/tmp/file", sub { };
15		16
16	aio_read $fh, 30000, 1024, $buffer, 0, sub {	17	aio_read $fh, 30000, 1024, $buffer, 0, sub {
17	$_[0] > 0 or die "read error: $!";	18	$_[0] > 0 or die "read error: $!";
18	};	19	};
19		20
20	# AnyEvent	21	# version 2+ has request and group objects
21	open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";	22	use IO::AIO 2;
22	my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
23		23
24	# Event	24	aioreq_pri 4; # give next request a very high priority
25	Event->io (fd => IO::AIO::poll_fileno,	25	my $req = aio_unlink "/tmp/file", sub { };
26	poll => 'r',	26	$req->cancel; # cancel request if still in queue
27	cb => \&IO::AIO::poll_cb);
28		27
29	# Glib/Gtk2	28	my $grp = aio_group sub { print "all stats done\n" };
30	add_watch Glib::IO IO::AIO::poll_fileno,	29	add $grp aio_stat "..." for ...;
31	in => sub { IO::AIO::poll_cb; 1 };
32
33	# Tk
34	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
35	readable => \&IO::AIO::poll_cb);
36
37	# Danga::Socket
38	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
39	\&IO::AIO::poll_cb);
40
41		30
42	=head1 DESCRIPTION	31	=head1 DESCRIPTION
43		32
44	This module implements asynchronous I/O using whatever means your	33	This module implements asynchronous I/O using whatever means your
45	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>).
46		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
47	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
48	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
49	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
50	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
51	aio functions available on many operating systems. However, they are often	57	functions available on many operating systems. However, they are often
52	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
53	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
54	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.
55		62
56	Although the module will work with in the presence of other threads, it is	63	Although the module will work in the presence of other (Perl-) threads,
57	currently not reentrant, so use appropriate locking yourself, always call	64	it is currently not reentrant in any way, so use appropriate locking
58	C<poll_cb> from within the same thread, or never call C<poll_cb> (or other	65	yourself, always call C<poll_cb> from within the same thread, or never
59	C<aio_> functions) recursively.	66	call C<poll_cb> (or other C<aio_> functions) recursively.
		67
		68	=head2 EXAMPLE
		69
		70	This is a simple example that uses the EV module and loads
		71	F</etc/passwd> asynchronously:
		72
		73	use Fcntl;
		74	use EV;
		75	use IO::AIO;
		76
		77	# register the IO::AIO callback with EV
		78	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		79
		80	# queue the request to open /etc/passwd
		81	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
		82	my $fh = shift
		83	or die "error while opening: $!";
		84
		85	# stat'ing filehandles is generally non-blocking
		86	my $size = -s $fh;
		87
		88	# queue a request to read the file
		89	my $contents;
		90	aio_read $fh, 0, $size, $contents, 0, sub {
		91	$_[0] == $size
		92	or die "short read: $!";
		93
		94	close $fh;
		95
		96	# file contents now in $contents
		97	print $contents;
		98
		99	# exit event loop and program
		100	EV::unloop;
		101	};
		102	};
		103
		104	# possibly queue up other requests, or open GUI windows,
		105	# check for sockets etc. etc.
		106
		107	# process events as long as there are some:
		108	EV::loop;
		109
		110	=head1 REQUEST ANATOMY AND LIFETIME
		111
		112	Every C<aio_*> function creates a request. which is a C data structure not
		113	directly visible to Perl.
		114
		115	If called in non-void context, every request function returns a Perl
		116	object representing the request. In void context, nothing is returned,
		117	which saves a bit of memory.
		118
		119	The perl object is a fairly standard ref-to-hash object. The hash contents
		120	are not used by IO::AIO so you are free to store anything you like in it.
		121
		122	During their existance, aio requests travel through the following states,
		123	in order:
		124
		125	=over 4
		126
		127	=item ready
		128
		129	Immediately after a request is created it is put into the ready state,
		130	waiting for a thread to execute it.
		131
		132	=item execute
		133
		134	A thread has accepted the request for processing and is currently
		135	executing it (e.g. blocking in read).
		136
		137	=item pending
		138
		139	The request has been executed and is waiting for result processing.
		140
		141	While request submission and execution is fully asynchronous, result
		142	processing is not and relies on the perl interpreter calling C<poll_cb>
		143	(or another function with the same effect).
		144
		145	=item result
		146
		147	The request results are processed synchronously by C<poll_cb>.
		148
		149	The C<poll_cb> function will process all outstanding aio requests by
		150	calling their callbacks, freeing memory associated with them and managing
		151	any groups they are contained in.
		152
		153	=item done
		154
		155	Request has reached the end of its lifetime and holds no resources anymore
		156	(except possibly for the Perl object, but its connection to the actual
		157	aio request is severed and calling its methods will either do nothing or
		158	result in a runtime error).
		159
		160	=back
60		161
61	=cut	162	=cut
62		163
63	package IO::AIO;	164	package IO::AIO;
64		165
65	no warnings;	166	use Carp ();
		167
		168	use common::sense;
66		169
67	use base 'Exporter';	170	use base 'Exporter';
68		171
69	use Fcntl ();
70
71	BEGIN {	172	BEGIN {
72	$VERSION = '1.71';	173	our $VERSION = '4.0';
73		174
74	@EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat	175	our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close
75	aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink	176	aio_stat aio_lstat aio_unlink aio_rmdir aio_readdir aio_readdirx
76	aio_fsync aio_fdatasync aio_readahead);	177	aio_scandir aio_symlink aio_readlink aio_realpath aio_sync
77	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel	178	aio_fsync aio_syncfs aio_fdatasync aio_sync_file_range aio_fallocate
78	max_outstanding nreqs);	179	aio_pathsync aio_readahead
		180	aio_rename aio_link aio_move aio_copy aio_group
		181	aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
		182	aio_chmod aio_utime aio_truncate
		183	aio_msync aio_mtouch aio_mlock aio_mlockall
		184	aio_statvfs
		185	aio_wd);
		186
		187	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		188	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		189	min_parallel max_parallel max_idle idle_timeout
		190	nreqs nready npending nthreads
		191	max_poll_time max_poll_reqs
		192	sendfile fadvise madvise
		193	mmap munmap munlock munlockall);
		194
		195	push @AIO_REQ, qw(aio_busy); # not exported
		196
		197	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
79		198
80	require XSLoader;	199	require XSLoader;
81	XSLoader::load IO::AIO, $VERSION;	200	XSLoader::load ("IO::AIO", $VERSION);
82	}	201	}
83		202
84	=head1 FUNCTIONS	203	=head1 FUNCTIONS
85		204
		205	=head2 QUICK OVERVIEW
		206
		207	This section simply lists the prototypes of the most important functions
		208	for quick reference. See the following sections for function-by-function
		209	documentation.
		210
		211	aio_wd $pathname, $callback->($wd)
		212	aio_open $pathname, $flags, $mode, $callback->($fh)
		213	aio_close $fh, $callback->($status)
		214	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		215	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		216	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		217	aio_readahead $fh,$offset,$length, $callback->($retval)
		218	aio_stat $fh_or_path, $callback->($status)
		219	aio_lstat $fh, $callback->($status)
		220	aio_statvfs $fh_or_path, $callback->($statvfs)
		221	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		222	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		223	aio_truncate $fh_or_path, $offset, $callback->($status)
		224	aio_chmod $fh_or_path, $mode, $callback->($status)
		225	aio_unlink $pathname, $callback->($status)
		226	aio_mknod $pathname, $mode, $dev, $callback->($status)
		227	aio_link $srcpath, $dstpath, $callback->($status)
		228	aio_symlink $srcpath, $dstpath, $callback->($status)
		229	aio_readlink $pathname, $callback->($link)
		230	aio_realpath $pathname, $callback->($link)
		231	aio_rename $srcpath, $dstpath, $callback->($status)
		232	aio_mkdir $pathname, $mode, $callback->($status)
		233	aio_rmdir $pathname, $callback->($status)
		234	aio_readdir $pathname, $callback->($entries)
		235	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		236	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		237	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		238	aio_load $pathname, $data, $callback->($status)
		239	aio_copy $srcpath, $dstpath, $callback->($status)
		240	aio_move $srcpath, $dstpath, $callback->($status)
		241	aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
		242	aio_rmtree $pathname, $callback->($status)
		243	aio_sync $callback->($status)
		244	aio_syncfs $fh, $callback->($status)
		245	aio_fsync $fh, $callback->($status)
		246	aio_fdatasync $fh, $callback->($status)
		247	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		248	aio_pathsync $pathname, $callback->($status)
		249	aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		250	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		251	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		252	aio_mlockall $flags, $callback->($status)
		253	aio_group $callback->(...)
		254	aio_nop $callback->()
		255
		256	$prev_pri = aioreq_pri [$pri]
		257	aioreq_nice $pri_adjust
		258
		259	IO::AIO::poll_wait
		260	IO::AIO::poll_cb
		261	IO::AIO::poll
		262	IO::AIO::flush
		263	IO::AIO::max_poll_reqs $nreqs
		264	IO::AIO::max_poll_time $seconds
		265	IO::AIO::min_parallel $nthreads
		266	IO::AIO::max_parallel $nthreads
		267	IO::AIO::max_idle $nthreads
		268	IO::AIO::idle_timeout $seconds
		269	IO::AIO::max_outstanding $maxreqs
		270	IO::AIO::nreqs
		271	IO::AIO::nready
		272	IO::AIO::npending
		273
		274	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		275	IO::AIO::fadvise $fh, $offset, $len, $advice
		276	IO::AIO::madvise $scalar, $offset, $length, $advice
		277	IO::AIO::mprotect $scalar, $offset, $length, $protect
		278	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		279	IO::AIO::munlockall
		280
86	=head2 AIO FUNCTIONS	281	=head2 AIO REQUEST FUNCTIONS
87		282
88	All the C<aio_*> calls are more or less thin wrappers around the syscall	283	All the C<aio_*> calls are more or less thin wrappers around the syscall
89	with the same name (sans C<aio_>). The arguments are similar or identical,	284	with the same name (sans C<aio_>). The arguments are similar or identical,
90	and they all accept an additional (and optional) C<$callback> argument	285	and they all accept an additional (and optional) C<$callback> argument
91	which must be a code reference. This code reference will get called with	286	which must be a code reference. This code reference will get called with
92	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	287	the syscall return code (e.g. most syscalls return C<-1> on error, unlike
93	perl, which usually delivers "false") as it's sole argument when the given	288	perl, which usually delivers "false") as its sole argument after the given
94	syscall has been executed asynchronously.	289	syscall has been executed asynchronously.
95		290
96	All functions expecting a filehandle keep a copy of the filehandle	291	All functions expecting a filehandle keep a copy of the filehandle
97	internally until the request has finished.	292	internally until the request has finished.
98		293
		294	All functions return request objects of type L<IO::AIO::REQ> that allow
		295	further manipulation of those requests while they are in-flight.
		296
99	The pathnames you pass to these routines I<must> be absolute and	297	The pathnames you pass to these routines I<should> be absolute. The
100	encoded in byte form. The reason for the former is that at the time the	298	reason for this is that at the time the request is being executed, the
101	request is being executed, the current working directory could have	299	current working directory could have changed. Alternatively, you can make
102	changed. Alternatively, you can make sure that you never change the	300	sure that you never change the current working directory anywhere in
103	current working directory.	301	the program and then use relative paths. Lastly, you can take advantage
		302	of IO::AIOs working directory abstraction - see the description of the
		303	C<IO::AIO::WD> class later in this document.
104		304
105	To encode pathnames to byte form, either make sure you either: a)	305	To encode pathnames as octets, either make sure you either: a) always pass
106	always pass in filenames you got from outside (command line, readdir	306	in filenames you got from outside (command line, readdir etc.) without
107	etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode	307	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
108	your pathnames to the locale (or other) encoding in effect in the user	308	your pathnames to the locale (or other) encoding in effect in the user
109	environment, d) use Glib::filename_from_unicode on unicode filenames or e)	309	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
110	use something else.	310	use something else to ensure your scalar has the correct contents.
		311
		312	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		313	handles correctly whether it is set or not.
111		314
112	=over 4	315	=over 4
		316
		317	=item $prev_pri = aioreq_pri [$pri]
		318
		319	Returns the priority value that would be used for the next request and, if
		320	C<$pri> is given, sets the priority for the next aio request.
		321
		322	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		323	and C<4>, respectively. Requests with higher priority will be serviced
		324	first.
		325
		326	The priority will be reset to C<0> after each call to one of the C<aio_*>
		327	functions.
		328
		329	Example: open a file with low priority, then read something from it with
		330	higher priority so the read request is serviced before other low priority
		331	open requests (potentially spamming the cache):
		332
		333	aioreq_pri -3;
		334	aio_open ..., sub {
		335	return unless $_[0];
		336
		337	aioreq_pri -2;
		338	aio_read $_[0], ..., sub {
		339	...
		340	};
		341	};
		342
		343
		344	=item aioreq_nice $pri_adjust
		345
		346	Similar to C<aioreq_pri>, but subtracts the given value from the current
		347	priority, so the effect is cumulative.
		348
113		349
114	=item aio_open $pathname, $flags, $mode, $callback->($fh)	350	=item aio_open $pathname, $flags, $mode, $callback->($fh)
115		351
116	Asynchronously open or create a file and call the callback with a newly	352	Asynchronously open or create a file and call the callback with a newly
117	created filehandle for the file.	353	created filehandle for the file.
…		…
123	list. They are the same as used by C<sysopen>.	359	list. They are the same as used by C<sysopen>.
124		360
125	Likewise, C<$mode> specifies the mode of the newly created file, if it	361	Likewise, C<$mode> specifies the mode of the newly created file, if it
126	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,	362	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
127	except that it is mandatory (i.e. use C<0> if you don't create new files,	363	except that it is mandatory (i.e. use C<0> if you don't create new files,
128	and C<0666> or C<0777> if you do).	364	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		365	by the umask in effect then the request is being executed, so better never
		366	change the umask.
129		367
130	Example:	368	Example:
131		369
132	aio_open "/etc/passwd", O_RDONLY, 0, sub {	370	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
133	if ($_[0]) {	371	if ($_[0]) {
134	print "open successful, fh is $_[0]\n";	372	print "open successful, fh is $_[0]\n";
135	...	373	...
136	} else {	374	} else {
137	die "open failed: $!\n";	375	die "open failed: $!\n";
138	}	376	}
139	};	377	};
140		378
		379	In addition to all the common open modes/flags (C<O_RDONLY>, C<O_WRONLY>,
		380	C<O_RDWR>, C<O_CREAT>, C<O_TRUNC>, C<O_EXCL> and C<O_APPEND>), the
		381	following POSIX and non-POSIX constants are available (missing ones on
		382	your system are, as usual, C<0>):
		383
		384	C<O_ASYNC>, C<O_DIRECT>, C<O_NOATIME>, C<O_CLOEXEC>, C<O_NOCTTY>, C<O_NOFOLLOW>,
		385	C<O_NONBLOCK>, C<O_EXEC>, C<O_SEARCH>, C<O_DIRECTORY>, C<O_DSYNC>,
		386	C<O_RSYNC>, C<O_SYNC> and C<O_TTY_INIT>.
		387
		388
141	=item aio_close $fh, $callback->($status)	389	=item aio_close $fh, $callback->($status)
142		390
143	Asynchronously close a file and call the callback with the result	391	Asynchronously close a file and call the callback with the result
144	code. I<WARNING:> although accepted, you should not pass in a perl	392	code.
145	filehandle here, as perl will likely close the file descriptor another
146	time when the filehandle is destroyed. Normally, you can safely call perls
147	C<close> or just let filehandles go out of scope.
148		393
149	This is supposed to be a bug in the API, so that might change. It's	394	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
150	therefore best to avoid this function.	395	closing the file descriptor associated with the filehandle itself.
		396
		397	Therefore, C<aio_close> will not close the filehandle - instead it will
		398	use dup2 to overwrite the file descriptor with the write-end of a pipe
		399	(the pipe fd will be created on demand and will be cached).
		400
		401	Or in other words: the file descriptor will be closed, but it will not be
		402	free for reuse until the perl filehandle is closed.
		403
		404	=cut
151		405
152	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	406	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
153		407
154	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	408	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
155		409
156	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	410	Reads or writes C<$length> bytes from or to the specified C<$fh> and
157	into the scalar given by C<data> and offset C<dataoffset> and calls the	411	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset>
158	callback without the actual number of bytes read (or -1 on error, just	412	and calls the callback without the actual number of bytes read (or -1 on
159	like the syscall).	413	error, just like the syscall).
		414
		415	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		416	offset plus the actual number of bytes read.
		417
		418	If C<$offset> is undefined, then the current file descriptor offset will
		419	be used (and updated), otherwise the file descriptor offset will not be
		420	changed by these calls.
		421
		422	If C<$length> is undefined in C<aio_write>, use the remaining length of
		423	C<$data>.
		424
		425	If C<$dataoffset> is less than zero, it will be counted from the end of
		426	C<$data>.
160		427
161	The C<$data> scalar I<MUST NOT> be modified in any way while the request	428	The C<$data> scalar I<MUST NOT> be modified in any way while the request
162	is outstanding. Modifying it can result in segfaults or WW3 (if the	429	is outstanding. Modifying it can result in segfaults or World War III (if
163	necessary/optional hardware is installed).	430	the necessary/optional hardware is installed).
164		431
165	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	432	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
166	offset C<0> within the scalar:	433	offset C<0> within the scalar:
167		434
168	aio_read $fh, 7, 15, $buffer, 0, sub {	435	aio_read $fh, 7, 15, $buffer, 0, sub {
169	$_[0] > 0 or die "read error: $!";	436	$_[0] > 0 or die "read error: $!";
170	print "read $_[0] bytes: <$buffer>\n";	437	print "read $_[0] bytes: <$buffer>\n";
171	};	438	};
172		439
		440
173	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)	441	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
174		442
175	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts	443	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
176	reading at byte offset C<$in_offset>, and starts writing at the current	444	reading at byte offset C<$in_offset>, and starts writing at the current
177	file offset of C<$out_fh>. Because of that, it is not safe to issue more	445	file offset of C<$out_fh>. Because of that, it is not safe to issue more
178	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each	446	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
179	other.	447	other. The same C<$in_fh> works fine though, as this function does not
		448	move or use the file offset of C<$in_fh>.
180		449
		450	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		451	are written, and there is no way to find out how many more bytes have been
		452	read from C<aio_sendfile> alone, as C<aio_sendfile> only provides the
		453	number of bytes written to C<$out_fh>. Only if the result value equals
		454	C<$length> one can assume that C<$length> bytes have been read.
		455
		456	Unlike with other C<aio_> functions, it makes a lot of sense to use
		457	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		458	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		459	the socket I/O will be non-blocking. Note, however, that you can run
		460	into a trap where C<aio_sendfile> reads some data with readahead, then
		461	fails to write all data, and when the socket is ready the next time, the
		462	data in the cache is already lost, forcing C<aio_sendfile> to again hit
		463	the disk. Explicit C<aio_read> + C<aio_write> let's you better control
		464	resource usage.
		465
181	This call tries to make use of a native C<sendfile> syscall to provide	466	This call tries to make use of a native C<sendfile>-like syscall to
182	zero-copy operation. For this to work, C<$out_fh> should refer to a	467	provide zero-copy operation. For this to work, C<$out_fh> should refer to
183	socket, and C<$in_fh> should refer to mmap'able file.	468	a socket, and C<$in_fh> should refer to an mmap'able file.
184		469
185	If the native sendfile call fails or is not implemented, it will be	470	If a native sendfile cannot be found or it fails with C<ENOSYS>,
186	emulated, so you can call C<aio_sendfile> on any type of filehandle	471	C<EINVAL>, C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or
		472	C<ENOTSOCK>, it will be emulated, so you can call C<aio_sendfile> on any
187	regardless of the limitations of the operating system.	473	type of filehandle regardless of the limitations of the operating system.
188		474
189	Please note, however, that C<aio_sendfile> can read more bytes from	475	As native sendfile syscalls (as practically any non-POSIX interface hacked
190	C<$in_fh> than are written, and there is no way to find out how many	476	together in a hurry to improve benchmark numbers) tend to be rather buggy
191	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only	477	on many systems, this implementation tries to work around some known bugs
192	provides the number of bytes written to C<$out_fh>. Only if the result	478	in Linux and FreeBSD kernels (probably others, too), but that might fail,
193	value equals C<$length> one can assume that C<$length> bytes have been	479	so you really really should check the return value of C<aio_sendfile> -
194	read.	480	fewre bytes than expected might have been transferred.
		481
195		482
196	=item aio_readahead $fh,$offset,$length, $callback->($retval)	483	=item aio_readahead $fh,$offset,$length, $callback->($retval)
197		484
198	C<aio_readahead> populates the page cache with data from a file so that	485	C<aio_readahead> populates the page cache with data from a file so that
199	subsequent reads from that file will not block on disk I/O. The C<$offset>	486	subsequent reads from that file will not block on disk I/O. The C<$offset>
…		…
205	file. The current file offset of the file is left unchanged.	492	file. The current file offset of the file is left unchanged.
206		493
207	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	494	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
208	emulated by simply reading the data, which would have a similar effect.	495	emulated by simply reading the data, which would have a similar effect.
209		496
		497
210	=item aio_stat $fh_or_path, $callback->($status)	498	=item aio_stat $fh_or_path, $callback->($status)
211		499
212	=item aio_lstat $fh, $callback->($status)	500	=item aio_lstat $fh, $callback->($status)
213		501
214	Works like perl's C<stat> or C<lstat> in void context. The callback will	502	Works like perl's C<stat> or C<lstat> in void context. The callback will
…		…
219	for an explanation.	507	for an explanation.
220		508
221	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	509	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
222	error when stat'ing a large file, the results will be silently truncated	510	error when stat'ing a large file, the results will be silently truncated
223	unless perl itself is compiled with large file support.	511	unless perl itself is compiled with large file support.
		512
		513	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		514	following constants and functions (if not implemented, the constants will
		515	be C<0> and the functions will either C<croak> or fall back on traditional
		516	behaviour).
		517
		518	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		519	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		520	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
224		521
225	Example: Print the length of F</etc/passwd>:	522	Example: Print the length of F</etc/passwd>:
226		523
227	aio_stat "/etc/passwd", sub {	524	aio_stat "/etc/passwd", sub {
228	$_[0] and die "stat failed: $!";	525	$_[0] and die "stat failed: $!";
229	print "size is ", -s _, "\n";	526	print "size is ", -s _, "\n";
230	};	527	};
231		528
		529
		530	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		531
		532	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		533	whether a file handle or path was passed.
		534
		535	On success, the callback is passed a hash reference with the following
		536	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		537	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		538	is passed.
		539
		540	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		541	C<ST_NOSUID>.
		542
		543	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		544	their correct value when available, or to C<0> on systems that do
		545	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		546	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		547	C<ST_NODIRATIME> and C<ST_RELATIME>.
		548
		549	Example: stat C</wd> and dump out the data if successful.
		550
		551	aio_statvfs "/wd", sub {
		552	my $f = $_[0]
		553	or die "statvfs: $!";
		554
		555	use Data::Dumper;
		556	say Dumper $f;
		557	};
		558
		559	# result:
		560	{
		561	bsize => 1024,
		562	bfree => 4333064312,
		563	blocks => 10253828096,
		564	files => 2050765568,
		565	flag => 4096,
		566	favail => 2042092649,
		567	bavail => 4333064312,
		568	ffree => 2042092649,
		569	namemax => 255,
		570	frsize => 1024,
		571	fsid => 1810
		572	}
		573
		574
		575	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		576
		577	Works like perl's C<utime> function (including the special case of $atime
		578	and $mtime being undef). Fractional times are supported if the underlying
		579	syscalls support them.
		580
		581	When called with a pathname, uses utimes(2) if available, otherwise
		582	utime(2). If called on a file descriptor, uses futimes(2) if available,
		583	otherwise returns ENOSYS, so this is not portable.
		584
		585	Examples:
		586
		587	# set atime and mtime to current time (basically touch(1)):
		588	aio_utime "path", undef, undef;
		589	# set atime to current time and mtime to beginning of the epoch:
		590	aio_utime "path", time, undef; # undef==0
		591
		592
		593	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		594
		595	Works like perl's C<chown> function, except that C<undef> for either $uid
		596	or $gid is being interpreted as "do not change" (but -1 can also be used).
		597
		598	Examples:
		599
		600	# same as "chown root path" in the shell:
		601	aio_chown "path", 0, -1;
		602	# same as above:
		603	aio_chown "path", 0, undef;
		604
		605
		606	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		607
		608	Works like truncate(2) or ftruncate(2).
		609
		610
		611	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		612
		613	Works like perl's C<chmod> function.
		614
		615
232	=item aio_unlink $pathname, $callback->($status)	616	=item aio_unlink $pathname, $callback->($status)
233		617
234	Asynchronously unlink (delete) a file and call the callback with the	618	Asynchronously unlink (delete) a file and call the callback with the
235	result code.	619	result code.
236		620
		621
		622	=item aio_mknod $pathname, $mode, $dev, $callback->($status)
		623
		624	[EXPERIMENTAL]
		625
		626	Asynchronously create a device node (or fifo). See mknod(2).
		627
		628	The only (POSIX-) portable way of calling this function is:
		629
		630	aio_mknod $pathname, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		631
		632	See C<aio_stat> for info about some potentially helpful extra constants
		633	and functions.
		634
		635	=item aio_link $srcpath, $dstpath, $callback->($status)
		636
		637	Asynchronously create a new link to the existing object at C<$srcpath> at
		638	the path C<$dstpath> and call the callback with the result code.
		639
		640
		641	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		642
		643	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		644	the path C<$dstpath> and call the callback with the result code.
		645
		646
		647	=item aio_readlink $pathname, $callback->($link)
		648
		649	Asynchronously read the symlink specified by C<$path> and pass it to
		650	the callback. If an error occurs, nothing or undef gets passed to the
		651	callback.
		652
		653
		654	=item aio_realpath $pathname, $callback->($path)
		655
		656	Asynchronously make the path absolute and resolve any symlinks in
		657	C<$path>. The resulting path only consists of directories (Same as
		658	L<Cwd::realpath>).
		659
		660	This request can be used to get the absolute path of the current working
		661	directory by passing it a path of F<.> (a single dot).
		662
		663
		664	=item aio_rename $srcpath, $dstpath, $callback->($status)
		665
		666	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		667	rename(2) and call the callback with the result code.
		668
		669
		670	=item aio_mkdir $pathname, $mode, $callback->($status)
		671
		672	Asynchronously mkdir (create) a directory and call the callback with
		673	the result code. C<$mode> will be modified by the umask at the time the
		674	request is executed, so do not change your umask.
		675
		676
237	=item aio_rmdir $pathname, $callback->($status)	677	=item aio_rmdir $pathname, $callback->($status)
238		678
239	Asynchronously rmdir (delete) a directory and call the callback with the	679	Asynchronously rmdir (delete) a directory and call the callback with the
240	result code.	680	result code.
241		681
		682
242	=item aio_readdir $pathname $callback->($entries)	683	=item aio_readdir $pathname, $callback->($entries)
243		684
244	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire	685	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
245	directory (i.e. opendir + readdir + closedir). The entries will not be	686	directory (i.e. opendir + readdir + closedir). The entries will not be
246	sorted, and will B<NOT> include the C<.> and C<..> entries.	687	sorted, and will B<NOT> include the C<.> and C<..> entries.
247		688
248	The callback a single argument which is either C<undef> or an array-ref	689	The callback is passed a single argument which is either C<undef> or an
249	with the filenames.	690	array-ref with the filenames.
250		691
		692
		693	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		694
		695	Quite similar to C<aio_readdir>, but the C<$flags> argument allows one to
		696	tune behaviour and output format. In case of an error, C<$entries> will be
		697	C<undef>.
		698
		699	The flags are a combination of the following constants, ORed together (the
		700	flags will also be passed to the callback, possibly modified):
		701
		702	=over 4
		703
		704	=item IO::AIO::READDIR_DENTS
		705
		706	When this flag is off, then the callback gets an arrayref consisting of
		707	names only (as with C<aio_readdir>), otherwise it gets an arrayref with
		708	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		709	entry in more detail.
		710
		711	C<$name> is the name of the entry.
		712
		713	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		714
		715	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		716	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		717	C<IO::AIO::DT_WHT>.
		718
		719	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		720	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		721	scalars are read-only: you can not modify them.
		722
		723	C<$inode> is the inode number (which might not be exact on systems with 64
		724	bit inode numbers and 32 bit perls). This field has unspecified content on
		725	systems that do not deliver the inode information.
		726
		727	=item IO::AIO::READDIR_DIRS_FIRST
		728
		729	When this flag is set, then the names will be returned in an order where
		730	likely directories come first, in optimal stat order. This is useful when
		731	you need to quickly find directories, or you want to find all directories
		732	while avoiding to stat() each entry.
		733
		734	If the system returns type information in readdir, then this is used
		735	to find directories directly. Otherwise, likely directories are names
		736	beginning with ".", or otherwise names with no dots, of which names with
		737	short names are tried first.
		738
		739	=item IO::AIO::READDIR_STAT_ORDER
		740
		741	When this flag is set, then the names will be returned in an order
		742	suitable for stat()'ing each one. That is, when you plan to stat()
		743	all files in the given directory, then the returned order will likely
		744	be fastest.
		745
		746	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		747	the likely dirs come first, resulting in a less optimal stat order.
		748
		749	=item IO::AIO::READDIR_FOUND_UNKNOWN
		750
		751	This flag should not be set when calling C<aio_readdirx>. Instead, it
		752	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		753	C<IO::AIO::DT_UNKNOWN>. The absence of this flag therefore indicates that all
		754	C<$type>'s are known, which can be used to speed up some algorithms.
		755
		756	=back
		757
		758
		759	=item aio_load $pathname, $data, $callback->($status)
		760
		761	This is a composite request that tries to fully load the given file into
		762	memory. Status is the same as with aio_read.
		763
		764	=cut
		765
		766	sub aio_load($$;$) {
		767	my ($path, undef, $cb) = @_;
		768	my $data = \$_[1];
		769
		770	my $pri = aioreq_pri;
		771	my $grp = aio_group $cb;
		772
		773	aioreq_pri $pri;
		774	add $grp aio_open $path, O_RDONLY, 0, sub {
		775	my $fh = shift
		776	or return $grp->result (-1);
		777
		778	aioreq_pri $pri;
		779	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		780	$grp->result ($_[0]);
		781	};
		782	};
		783
		784	$grp
		785	}
		786
		787	=item aio_copy $srcpath, $dstpath, $callback->($status)
		788
		789	Try to copy the I<file> (directories not supported as either source or
		790	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		791	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		792
		793	This is a composite request that creates the destination file with
		794	mode 0200 and copies the contents of the source file into it using
		795	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		796	uid/gid, in that order.
		797
		798	If an error occurs, the partial destination file will be unlinked, if
		799	possible, except when setting atime, mtime, access mode and uid/gid, where
		800	errors are being ignored.
		801
		802	=cut
		803
		804	sub aio_copy($$;$) {
		805	my ($src, $dst, $cb) = @_;
		806
		807	my $pri = aioreq_pri;
		808	my $grp = aio_group $cb;
		809
		810	aioreq_pri $pri;
		811	add $grp aio_open $src, O_RDONLY, 0, sub {
		812	if (my $src_fh = $_[0]) {
		813	my @stat = stat $src_fh; # hmm, might block over nfs?
		814
		815	aioreq_pri $pri;
		816	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		817	if (my $dst_fh = $_[0]) {
		818	aioreq_pri $pri;
		819	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		820	if ($_[0] == $stat[7]) {
		821	$grp->result (0);
		822	close $src_fh;
		823
		824	my $ch = sub {
		825	aioreq_pri $pri;
		826	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		827	aioreq_pri $pri;
		828	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		829	aioreq_pri $pri;
		830	add $grp aio_close $dst_fh;
		831	}
		832	};
		833	};
		834
		835	aioreq_pri $pri;
		836	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		837	if ($_[0] < 0 && $! == ENOSYS) {
		838	aioreq_pri $pri;
		839	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		840	} else {
		841	$ch->();
		842	}
		843	};
		844	} else {
		845	$grp->result (-1);
		846	close $src_fh;
		847	close $dst_fh;
		848
		849	aioreq $pri;
		850	add $grp aio_unlink $dst;
		851	}
		852	};
		853	} else {
		854	$grp->result (-1);
		855	}
		856	},
		857
		858	} else {
		859	$grp->result (-1);
		860	}
		861	};
		862
		863	$grp
		864	}
		865
		866	=item aio_move $srcpath, $dstpath, $callback->($status)
		867
		868	Try to move the I<file> (directories not supported as either source or
		869	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		870	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		871
		872	This is a composite request that tries to rename(2) the file first; if
		873	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		874	that is successful, unlinks the C<$srcpath>.
		875
		876	=cut
		877
		878	sub aio_move($$;$) {
		879	my ($src, $dst, $cb) = @_;
		880
		881	my $pri = aioreq_pri;
		882	my $grp = aio_group $cb;
		883
		884	aioreq_pri $pri;
		885	add $grp aio_rename $src, $dst, sub {
		886	if ($_[0] && $! == EXDEV) {
		887	aioreq_pri $pri;
		888	add $grp aio_copy $src, $dst, sub {
		889	$grp->result ($_[0]);
		890
		891	unless ($_[0]) {
		892	aioreq_pri $pri;
		893	add $grp aio_unlink $src;
		894	}
		895	};
		896	} else {
		897	$grp->result ($_[0]);
		898	}
		899	};
		900
		901	$grp
		902	}
		903
251	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)	904	=item aio_scandir $pathname, $maxreq, $callback->($dirs, $nondirs)
252		905
253	Scans a directory (similar to C<aio_readdir>) and tries to separate the	906	Scans a directory (similar to C<aio_readdir>) but additionally tries to
254	entries of directory C<$path> into two sets of names, ones you can recurse	907	efficiently separate the entries of directory C<$path> into two sets of
255	into (directories), and ones you cannot recurse into (everything else).	908	names, directories you can recurse into (directories), and ones you cannot
		909	recurse into (everything else, including symlinks to directories).
256		910
257	C<aio_scandir> is a composite request that consists of many	911	C<aio_scandir> is a composite request that creates of many sub requests_
258	aio-primitives. C<$maxreq> specifies the maximum number of outstanding	912	C<$maxreq> specifies the maximum number of outstanding aio requests that
259	aio requests that this function generates. If it is C<< <= 0 >>, then a	913	this function generates. If it is C<< <= 0 >>, then a suitable default
260	suitable default will be chosen (currently 8).	914	will be chosen (currently 4).
261		915
262	On error, the callback is called without arguments, otherwise it receives	916	On error, the callback is called without arguments, otherwise it receives
263	two array-refs with path-relative entry names.	917	two array-refs with path-relative entry names.
264		918
265	Example:	919	Example:
…		…
272		926
273	Implementation notes.	927	Implementation notes.
274		928
275	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.	929	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
276		930
		931	If readdir returns file type information, then this is used directly to
		932	find directories.
		933
277	After reading the directory, the modification time, size etc. of the	934	Otherwise, after reading the directory, the modification time, size etc.
278	directory before and after the readdir is checked, and if they match, the	935	of the directory before and after the readdir is checked, and if they
279	link count will be used to decide how many entries are directories (if	936	match (and isn't the current time), the link count will be used to decide
280	>= 2). Otherwise, no knowledge of the number of subdirectories will be	937	how many entries are directories (if >= 2). Otherwise, no knowledge of the
281	assumed.	938	number of subdirectories will be assumed.
282		939
283	Then entires will be sorted into likely directories (everything without a	940	Then entries will be sorted into likely directories a non-initial dot
284	non-initial dot) and likely non-directories (everything else). Then every	941	currently) and likely non-directories (see C<aio_readdirx>). Then every
285	entry + C</.> will be C<stat>'ed, likely directories first. This is often	942	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
		943	in order of their inode numbers. If that succeeds, it assumes that the
		944	entry is a directory or a symlink to directory (which will be checked
		945	separately). This is often faster than stat'ing the entry itself because
286	faster because filesystems might detect the type of the entry without	946	filesystems might detect the type of the entry without reading the inode
287	reading the inode data (e.g. ext2s filetype feature). If that succeeds,	947	data (e.g. ext2fs filetype feature), even on systems that cannot return
288	it assumes that the entry is a directory or a symlink to directory (which	948	the filetype information on readdir.
289	will be checked seperately).
290		949
291	If the known number of directories has been reached, the rest of the	950	If the known number of directories (link count - 2) has been reached, the
292	entries is assumed to be non-directories.	951	rest of the entries is assumed to be non-directories.
		952
		953	This only works with certainty on POSIX (= UNIX) filesystems, which
		954	fortunately are the vast majority of filesystems around.
		955
		956	It will also likely work on non-POSIX filesystems with reduced efficiency
		957	as those tend to return 0 or 1 as link counts, which disables the
		958	directory counting heuristic.
293		959
294	=cut	960	=cut
295		961
296	sub aio_scandir($$$) {	962	sub aio_scandir($$;$) {
297	my ($path, $maxreq, $cb) = @_;	963	my ($path, $maxreq, $cb) = @_;
298		964
		965	my $pri = aioreq_pri;
		966
		967	my $grp = aio_group $cb;
		968
299	$maxreq = 8 if $maxreq <= 0;	969	$maxreq = 4 if $maxreq <= 0;
300		970
301	# stat once	971	# stat once
		972	aioreq_pri $pri;
302	aio_stat $path, sub {	973	add $grp aio_stat $path, sub {
303	$cb->() if $_[0];	974	return $grp->result () if $_[0];
		975	my $now = time;
304	my $hash1 = join ":", (stat _)[0,1,3,7,9];	976	my $hash1 = join ":", (stat _)[0,1,3,7,9];
305		977
306	# read the directory entries	978	# read the directory entries
307	aio_readdir $path, sub {	979	aioreq_pri $pri;
		980	add $grp aio_readdirx $path, READDIR_DIRS_FIRST, sub {
308	my $entries = shift	981	my $entries = shift
309	or return $cb->();	982	or return $grp->result ();
310		983
311	# stat the dir another time	984	# stat the dir another time
		985	aioreq_pri $pri;
312	aio_stat $path, sub {	986	add $grp aio_stat $path, sub {
313	my $hash2 = join ":", (stat _)[0,1,3,7,9];	987	my $hash2 = join ":", (stat _)[0,1,3,7,9];
314		988
315	my $ndirs;	989	my $ndirs;
316		990
317	# take the slow route if anything looks fishy	991	# take the slow route if anything looks fishy
318	if ($hash1 ne $hash2) {	992	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
319	$ndirs = -1;	993	$ndirs = -1;
320	} else {	994	} else {
321	# if nlink == 2, we are finished	995	# if nlink == 2, we are finished
322	# on non-posix-fs's, we rely on nlink < 2	996	# for non-posix-fs's, we rely on nlink < 2
323	$ndirs = (stat _)[3] - 2	997	$ndirs = (stat _)[3] - 2
324	or return $cb->([], $entries);	998	or return $grp->result ([], $entries);
325	}	999	}
326		1000
327	# sort into likely dirs and likely nondirs
328	# dirs == files without ".", short entries first
329	$entries = [map $_->[0],
330	sort { $b->[1] cmp $a->[1] }
331	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
332	@$entries];
333
334	my (@dirs, @nondirs);	1001	my (@dirs, @nondirs);
335		1002
336	my ($statcb, $schedcb);	1003	my $statgrp = add $grp aio_group sub {
337	my $nreq = 0;	1004	$grp->result (\@dirs, \@nondirs);
		1005	};
338		1006
339	$schedcb = sub {	1007	limit $statgrp $maxreq;
340	if (@$entries) {	1008	feed $statgrp sub {
341	if ($nreq < $maxreq) {	1009	return unless @$entries;
342	my $ent = pop @$entries;	1010	my $entry = shift @$entries;
		1011
		1012	aioreq_pri $pri;
		1013	add $statgrp aio_stat "$path/$entry/.", sub {
		1014	if ($_[0] < 0) {
		1015	push @nondirs, $entry;
		1016	} else {
		1017	# need to check for real directory
		1018	aioreq_pri $pri;
		1019	add $statgrp aio_lstat "$path/$entry", sub {
		1020	if (-d _) {
		1021	push @dirs, $entry;
		1022
		1023	unless (--$ndirs) {
		1024	push @nondirs, @$entries;
		1025	feed $statgrp;
		1026	}
		1027	} else {
		1028	push @nondirs, $entry;
		1029	}
343	$nreq++;	1030	}
344	aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) };
345	}	1031	}
346	} elsif (!$nreq) {
347	# finished
348	undef $statcb;
349	undef $schedcb;
350	$cb->(\@dirs, \@nondirs) if $cb;
351	undef $cb;
352	}	1032	};
353	};	1033	};
354	$statcb = sub {
355	my ($status, $entry) = @_;
356
357	if ($status < 0) {
358	$nreq--;
359	push @nondirs, $entry;
360	&$schedcb;
361	} else {
362	# need to check for real directory
363	aio_lstat "$path/$entry", sub {
364	$nreq--;
365
366	if (-d _) {
367	push @dirs, $entry;
368
369	if (!--$ndirs) {
370	push @nondirs, @$entries;
371	$entries = [];
372	}
373	} else {
374	push @nondirs, $entry;
375	}
376
377	&$schedcb;
378	}
379	}
380	};
381
382	&$schedcb while @$entries && $nreq < $maxreq;
383	};	1034	};
384	};	1035	};
385	};	1036	};
		1037
		1038	$grp
386	}	1039	}
		1040
		1041	=item aio_rmtree $pathname, $callback->($status)
		1042
		1043	Delete a directory tree starting (and including) C<$path>, return the
		1044	status of the final C<rmdir> only. This is a composite request that
		1045	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1046	everything else.
		1047
		1048	=cut
		1049
		1050	sub aio_rmtree;
		1051	sub aio_rmtree($;$) {
		1052	my ($path, $cb) = @_;
		1053
		1054	my $pri = aioreq_pri;
		1055	my $grp = aio_group $cb;
		1056
		1057	aioreq_pri $pri;
		1058	add $grp aio_scandir $path, 0, sub {
		1059	my ($dirs, $nondirs) = @_;
		1060
		1061	my $dirgrp = aio_group sub {
		1062	add $grp aio_rmdir $path, sub {
		1063	$grp->result ($_[0]);
		1064	};
		1065	};
		1066
		1067	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1068	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1069
		1070	add $grp $dirgrp;
		1071	};
		1072
		1073	$grp
		1074	}
		1075
		1076	=item aio_sync $callback->($status)
		1077
		1078	Asynchronously call sync and call the callback when finished.
387		1079
388	=item aio_fsync $fh, $callback->($status)	1080	=item aio_fsync $fh, $callback->($status)
389		1081
390	Asynchronously call fsync on the given filehandle and call the callback	1082	Asynchronously call fsync on the given filehandle and call the callback
391	with the fsync result code.	1083	with the fsync result code.
…		…
396	callback with the fdatasync result code.	1088	callback with the fdatasync result code.
397		1089
398	If this call isn't available because your OS lacks it or it couldn't be	1090	If this call isn't available because your OS lacks it or it couldn't be
399	detected, it will be emulated by calling C<fsync> instead.	1091	detected, it will be emulated by calling C<fsync> instead.
400		1092
		1093	=item aio_syncfs $fh, $callback->($status)
		1094
		1095	Asynchronously call the syncfs syscall to sync the filesystem associated
		1096	to the given filehandle and call the callback with the syncfs result
		1097	code. If syncfs is not available, calls sync(), but returns C<-1> and sets
		1098	errno to C<ENOSYS> nevertheless.
		1099
		1100	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1101
		1102	Sync the data portion of the file specified by C<$offset> and C<$length>
		1103	to disk (but NOT the metadata), by calling the Linux-specific
		1104	sync_file_range call. If sync_file_range is not available or it returns
		1105	ENOSYS, then fdatasync or fsync is being substituted.
		1106
		1107	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1108	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1109	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1110	manpage for details.
		1111
		1112	=item aio_pathsync $pathname, $callback->($status)
		1113
		1114	This request tries to open, fsync and close the given path. This is a
		1115	composite request intended to sync directories after directory operations
		1116	(E.g. rename). This might not work on all operating systems or have any
		1117	specific effect, but usually it makes sure that directory changes get
		1118	written to disc. It works for anything that can be opened for read-only,
		1119	not just directories.
		1120
		1121	Future versions of this function might fall back to other methods when
		1122	C<fsync> on the directory fails (such as calling C<sync>).
		1123
		1124	Passes C<0> when everything went ok, and C<-1> on error.
		1125
		1126	=cut
		1127
		1128	sub aio_pathsync($;$) {
		1129	my ($path, $cb) = @_;
		1130
		1131	my $pri = aioreq_pri;
		1132	my $grp = aio_group $cb;
		1133
		1134	aioreq_pri $pri;
		1135	add $grp aio_open $path, O_RDONLY, 0, sub {
		1136	my ($fh) = @_;
		1137	if ($fh) {
		1138	aioreq_pri $pri;
		1139	add $grp aio_fsync $fh, sub {
		1140	$grp->result ($_[0]);
		1141
		1142	aioreq_pri $pri;
		1143	add $grp aio_close $fh;
		1144	};
		1145	} else {
		1146	$grp->result (-1);
		1147	}
		1148	};
		1149
		1150	$grp
		1151	}
		1152
		1153	=item aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1154
		1155	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1156	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1157	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1158	scalar must only be modified in-place while an aio operation is pending on
		1159	it).
		1160
		1161	It calls the C<msync> function of your OS, if available, with the memory
		1162	area starting at C<$offset> in the string and ending C<$length> bytes
		1163	later. If C<$length> is negative, counts from the end, and if C<$length>
		1164	is C<undef>, then it goes till the end of the string. The flags can be
		1165	a combination of C<IO::AIO::MS_ASYNC>, C<IO::AIO::MS_INVALIDATE> and
		1166	C<IO::AIO::MS_SYNC>.
		1167
		1168	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1169
		1170	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1171	scalars.
		1172
		1173	It touches (reads or writes) all memory pages in the specified
		1174	range inside the scalar. All caveats and parameters are the same
		1175	as for C<aio_msync>, above, except for flags, which must be either
		1176	C<0> (which reads all pages and ensures they are instantiated) or
		1177	C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and
		1178	writing an octet from it, which dirties the page).
		1179
		1180	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1181
		1182	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1183	scalars.
		1184
		1185	It reads in all the pages of the underlying storage into memory (if any)
		1186	and locks them, so they are not getting swapped/paged out or removed.
		1187
		1188	If C<$length> is undefined, then the scalar will be locked till the end.
		1189
		1190	On systems that do not implement C<mlock>, this function returns C<-1>
		1191	and sets errno to C<ENOSYS>.
		1192
		1193	Note that the corresponding C<munlock> is synchronous and is
		1194	documented under L<MISCELLANEOUS FUNCTIONS>.
		1195
		1196	Example: open a file, mmap and mlock it - both will be undone when
		1197	C<$data> gets destroyed.
		1198
		1199	open my $fh, "<", $path or die "$path: $!";
		1200	my $data;
		1201	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1202	aio_mlock $data; # mlock in background
		1203
		1204	=item aio_mlockall $flags, $callback->($status)
		1205
		1206	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1207	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1208
		1209	On systems that do not implement C<mlockall>, this function returns C<-1>
		1210	and sets errno to C<ENOSYS>.
		1211
		1212	Note that the corresponding C<munlockall> is synchronous and is
		1213	documented under L<MISCELLANEOUS FUNCTIONS>.
		1214
		1215	Example: asynchronously lock all current and future pages into memory.
		1216
		1217	aio_mlockall IO::AIO::MCL_FUTURE;
		1218
		1219	=item aio_group $callback->(...)
		1220
		1221	This is a very special aio request: Instead of doing something, it is a
		1222	container for other aio requests, which is useful if you want to bundle
		1223	many requests into a single, composite, request with a definite callback
		1224	and the ability to cancel the whole request with its subrequests.
		1225
		1226	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		1227	for more info.
		1228
		1229	Example:
		1230
		1231	my $grp = aio_group sub {
		1232	print "all stats done\n";
		1233	};
		1234
		1235	add $grp
		1236	(aio_stat ...),
		1237	(aio_stat ...),
		1238	...;
		1239
		1240	=item aio_nop $callback->()
		1241
		1242	This is a special request - it does nothing in itself and is only used for
		1243	side effects, such as when you want to add a dummy request to a group so
		1244	that finishing the requests in the group depends on executing the given
		1245	code.
		1246
		1247	While this request does nothing, it still goes through the execution
		1248	phase and still requires a worker thread. Thus, the callback will not
		1249	be executed immediately but only after other requests in the queue have
		1250	entered their execution phase. This can be used to measure request
		1251	latency.
		1252
		1253	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		1254
		1255	Mainly used for debugging and benchmarking, this aio request puts one of
		1256	the request workers to sleep for the given time.
		1257
		1258	While it is theoretically handy to have simple I/O scheduling requests
		1259	like sleep and file handle readable/writable, the overhead this creates is
		1260	immense (it blocks a thread for a long time) so do not use this function
		1261	except to put your application under artificial I/O pressure.
		1262
401	=back	1263	=back
402		1264
		1265
		1266	=head2 IO::AIO::WD - multiple working directories
		1267
		1268	Your process only has one current working directory, which is used by all
		1269	threads. This makes it hard to use relative paths (some other component
		1270	could call C<chdir> at any time, and it is hard to control when the path
		1271	will be used by IO::AIO).
		1272
		1273	One solution for this is to always use absolute paths. This usually works,
		1274	but can be quite slow (the kernel has to walk the whole path on every
		1275	access), and can also be a hassle to implement.
		1276
		1277	Newer POSIX systems have a number of functions (openat, fdopendir,
		1278	futimensat and so on) that make it possible to specify working directories
		1279	per operation.
		1280
		1281	For portability, and because the clowns who "designed", or shall I write,
		1282	perpetrated this new interface were obviously half-drunk, this abstraction
		1283	cannot be perfect, though.
		1284
		1285	IO::AIO allows you to convert directory paths into a so-called IO::AIO::WD
		1286	object. This object stores the canonicalised, absolute version of the
		1287	path, and on systems that allow it, also a directory file descriptor.
		1288
		1289	Everywhere where a pathname is accepted by IO::AIO (e.g. in C<aio_stat>
		1290	or C<aio_unlink>), one can specify an array reference with an IO::AIO::WD
		1291	object and a pathname instead. If the pathname is absolute, the
		1292	IO::AIO::WD objetc is ignored, otherwise the pathname is resolved relative
		1293	to that IO::AIO::WD object.
		1294
		1295	For example, to get a wd object for F</etc> and then stat F<passwd>
		1296	inside, you would write:
		1297
		1298	aio_wd "/etc", sub {
		1299	my $etcdir = shift;
		1300
		1301	# although $etcdir can be undef on error, there is generally no reason
		1302	# to check for errors here, as aio_stat will fail with ENOENT
		1303	# when $etcdir is undef.
		1304
		1305	aio_stat [$etcdir, "passwd"], sub {
		1306	# yay
		1307	};
		1308	};
		1309
		1310	This shows that creating an IO::AIO::WD object is itself a potentially
		1311	blocking operation, which is why it is done asynchronously.
		1312
		1313	As with normal pathnames, IO::AIO keeps a copy of the working directory
		1314	object and the pathname string, so you could write the following without
		1315	causing any issues due to C<$path> getting reused:
		1316
		1317	my $path = [$wd, undef];
		1318
		1319	for my $name (qw(abc def ghi)) {
		1320	$path->[1] = $name;
		1321	aio_stat $path, sub {
		1322	# ...
		1323	};
		1324	}
		1325
		1326	There are some caveats: when directories get renamed (or deleted), the
		1327	pathname string doesn't change, so will point to the new directory (or
		1328	nowhere at all), while the directory fd, if available on the system,
		1329	will still point to the original directory. Most functions accepting a
		1330	pathname will use the directory fd on newer systems, and the string on
		1331	older systems. Some functions (such as realpath) will always rely on the
		1332	string form of the pathname.
		1333
		1334	So this fucntionality is mainly useful to get some protection against
		1335	C<chdir>, to easily get an absolute path out of a relative path for future
		1336	reference, and to speed up doing many operations in the same directory
		1337	(e.g. when stat'ing all files in a directory).
		1338
		1339	The following functions implement this working directory abstraction:
		1340
		1341	=over 4
		1342
		1343	=item aio_wd $pathname, $callback->($wd)
		1344
		1345	Asynchonously canonicalise the given pathname and convert it to an
		1346	IO::AIO::WD object representing it. If possible and supported on the
		1347	system, also open a directory fd to speed up pathname resolution relative
		1348	to this working directory.
		1349
		1350	If something goes wrong, then C<undef> is passwd to the callback instead
		1351	of a working directory object and C<$!> is set appropriately. Since
		1352	passing C<undef> as working directory component of a pathname fails the
		1353	request with C<ENOENT>, there is often no need for error checking in the
		1354	C<aio_wd> callback, as future requests using the value will fail in the
		1355	expected way.
		1356
		1357	If this call isn't available because your OS lacks it or it couldn't be
		1358	detected, it will be emulated by calling C<fsync> instead.
		1359
		1360	=item IO::AIO::CWD
		1361
		1362	This is a compiletime constant (object) that represents the process
		1363	current working directory.
		1364
		1365	Specifying this object as working directory object for a pathname is as
		1366	if the pathname would be specified directly, without a directory object,
		1367	e.g., these calls are functionally identical:
		1368
		1369	aio_stat "somefile", sub { ... };
		1370	aio_stat [IO::AIO::CWD, "somefile"], sub { ... };
		1371
		1372	=back
		1373
		1374
		1375	=head2 IO::AIO::REQ CLASS
		1376
		1377	All non-aggregate C<aio_*> functions return an object of this class when
		1378	called in non-void context.
		1379
		1380	=over 4
		1381
		1382	=item cancel $req
		1383
		1384	Cancels the request, if possible. Has the effect of skipping execution
		1385	when entering the B<execute> state and skipping calling the callback when
		1386	entering the the B<result> state, but will leave the request otherwise
		1387	untouched (with the exception of readdir). That means that requests that
		1388	currently execute will not be stopped and resources held by the request
		1389	will not be freed prematurely.
		1390
		1391	=item cb $req $callback->(...)
		1392
		1393	Replace (or simply set) the callback registered to the request.
		1394
		1395	=back
		1396
		1397	=head2 IO::AIO::GRP CLASS
		1398
		1399	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		1400	objects of this class, too.
		1401
		1402	A IO::AIO::GRP object is a special request that can contain multiple other
		1403	aio requests.
		1404
		1405	You create one by calling the C<aio_group> constructing function with a
		1406	callback that will be called when all contained requests have entered the
		1407	C<done> state:
		1408
		1409	my $grp = aio_group sub {
		1410	print "all requests are done\n";
		1411	};
		1412
		1413	You add requests by calling the C<add> method with one or more
		1414	C<IO::AIO::REQ> objects:
		1415
		1416	$grp->add (aio_unlink "...");
		1417
		1418	add $grp aio_stat "...", sub {
		1419	$_[0] or return $grp->result ("error");
		1420
		1421	# add another request dynamically, if first succeeded
		1422	add $grp aio_open "...", sub {
		1423	$grp->result ("ok");
		1424	};
		1425	};
		1426
		1427	This makes it very easy to create composite requests (see the source of
		1428	C<aio_move> for an application) that work and feel like simple requests.
		1429
		1430	=over 4
		1431
		1432	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		1433	C<IO::AIO::poll_cb>, just like any other request.
		1434
		1435	=item * They can be canceled like any other request. Canceling will cancel not
		1436	only the request itself, but also all requests it contains.
		1437
		1438	=item * They can also can also be added to other IO::AIO::GRP objects.
		1439
		1440	=item * You must not add requests to a group from within the group callback (or
		1441	any later time).
		1442
		1443	=back
		1444
		1445	Their lifetime, simplified, looks like this: when they are empty, they
		1446	will finish very quickly. If they contain only requests that are in the
		1447	C<done> state, they will also finish. Otherwise they will continue to
		1448	exist.
		1449
		1450	That means after creating a group you have some time to add requests
		1451	(precisely before the callback has been invoked, which is only done within
		1452	the C<poll_cb>). And in the callbacks of those requests, you can add
		1453	further requests to the group. And only when all those requests have
		1454	finished will the the group itself finish.
		1455
		1456	=over 4
		1457
		1458	=item add $grp ...
		1459
		1460	=item $grp->add (...)
		1461
		1462	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1463	be added, including other groups, as long as you do not create circular
		1464	dependencies.
		1465
		1466	Returns all its arguments.
		1467
		1468	=item $grp->cancel_subs
		1469
		1470	Cancel all subrequests and clears any feeder, but not the group request
		1471	itself. Useful when you queued a lot of events but got a result early.
		1472
		1473	The group request will finish normally (you cannot add requests to the
		1474	group).
		1475
		1476	=item $grp->result (...)
		1477
		1478	Set the result value(s) that will be passed to the group callback when all
		1479	subrequests have finished and set the groups errno to the current value
		1480	of errno (just like calling C<errno> without an error number). By default,
		1481	no argument will be passed and errno is zero.
		1482
		1483	=item $grp->errno ([$errno])
		1484
		1485	Sets the group errno value to C<$errno>, or the current value of errno
		1486	when the argument is missing.
		1487
		1488	Every aio request has an associated errno value that is restored when
		1489	the callback is invoked. This method lets you change this value from its
		1490	default (0).
		1491
		1492	Calling C<result> will also set errno, so make sure you either set C<$!>
		1493	before the call to C<result>, or call c<errno> after it.
		1494
		1495	=item feed $grp $callback->($grp)
		1496
		1497	Sets a feeder/generator on this group: every group can have an attached
		1498	generator that generates requests if idle. The idea behind this is that,
		1499	although you could just queue as many requests as you want in a group,
		1500	this might starve other requests for a potentially long time. For example,
		1501	C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests,
		1502	delaying any later requests for a long time.
		1503
		1504	To avoid this, and allow incremental generation of requests, you can
		1505	instead a group and set a feeder on it that generates those requests. The
		1506	feed callback will be called whenever there are few enough (see C<limit>,
		1507	below) requests active in the group itself and is expected to queue more
		1508	requests.
		1509
		1510	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1511	not impose any limits).
		1512
		1513	If the feed does not queue more requests when called, it will be
		1514	automatically removed from the group.
		1515
		1516	If the feed limit is C<0> when this method is called, it will be set to
		1517	C<2> automatically.
		1518
		1519	Example:
		1520
		1521	# stat all files in @files, but only ever use four aio requests concurrently:
		1522
		1523	my $grp = aio_group sub { print "finished\n" };
		1524	limit $grp 4;
		1525	feed $grp sub {
		1526	my $file = pop @files
		1527	or return;
		1528
		1529	add $grp aio_stat $file, sub { ... };
		1530	};
		1531
		1532	=item limit $grp $num
		1533
		1534	Sets the feeder limit for the group: The feeder will be called whenever
		1535	the group contains less than this many requests.
		1536
		1537	Setting the limit to C<0> will pause the feeding process.
		1538
		1539	The default value for the limit is C<0>, but note that setting a feeder
		1540	automatically bumps it up to C<2>.
		1541
		1542	=back
		1543
403	=head2 SUPPORT FUNCTIONS	1544	=head2 SUPPORT FUNCTIONS
404		1545
		1546	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1547
405	=over 4	1548	=over 4
406		1549
407	=item $fileno = IO::AIO::poll_fileno	1550	=item $fileno = IO::AIO::poll_fileno
408		1551
409	Return the I<request result pipe file descriptor>. This filehandle must be	1552	Return the I<request result pipe file descriptor>. This filehandle must be
410	polled for reading by some mechanism outside this module (e.g. Event or	1553	polled for reading by some mechanism outside this module (e.g. EV, Glib,
411	select, see below or the SYNOPSIS). If the pipe becomes readable you have	1554	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
412	to call C<poll_cb> to check the results.	1555	you have to call C<poll_cb> to check the results.
413		1556
414	See C<poll_cb> for an example.	1557	See C<poll_cb> for an example.
415		1558
416	=item IO::AIO::poll_cb	1559	=item IO::AIO::poll_cb
417		1560
418	Process all outstanding events on the result pipe. You have to call this	1561	Process some outstanding events on the result pipe. You have to call
419	regularly. Returns the number of events processed. Returns immediately	1562	this regularly. Returns C<0> if all events could be processed (or there
420	when no events are outstanding.	1563	were no events to process), or C<-1> if it returned earlier for whatever
		1564	reason. Returns immediately when no events are outstanding. The amount of
		1565	events processed depends on the settings of C<IO::AIO::max_poll_req> and
		1566	C<IO::AIO::max_poll_time>.
		1567
		1568	If not all requests were processed for whatever reason, the filehandle
		1569	will still be ready when C<poll_cb> returns, so normally you don't have to
		1570	do anything special to have it called later.
		1571
		1572	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1573	ready, it can be beneficial to call this function from loops which submit
		1574	a lot of requests, to make sure the results get processed when they become
		1575	available and not just when the loop is finished and the event loop takes
		1576	over again. This function returns very fast when there are no outstanding
		1577	requests.
421		1578
422	Example: Install an Event watcher that automatically calls	1579	Example: Install an Event watcher that automatically calls
423	IO::AIO::poll_cb with high priority:	1580	IO::AIO::poll_cb with high priority (more examples can be found in the
		1581	SYNOPSIS section, at the top of this document):
424		1582
425	Event->io (fd => IO::AIO::poll_fileno,	1583	Event->io (fd => IO::AIO::poll_fileno,
426	poll => 'r', async => 1,	1584	poll => 'r', async => 1,
427	cb => \&IO::AIO::poll_cb);	1585	cb => \&IO::AIO::poll_cb);
428		1586
429	=item IO::AIO::poll_wait	1587	=item IO::AIO::poll_wait
430		1588
		1589	If there are any outstanding requests and none of them in the result
431	Wait till the result filehandle becomes ready for reading (simply does a	1590	phase, wait till the result filehandle becomes ready for reading (simply
432	C<select> on the filehandle. This is useful if you want to synchronously wait	1591	does a C<select> on the filehandle. This is useful if you want to
433	for some requests to finish).	1592	synchronously wait for some requests to finish).
434		1593
435	See C<nreqs> for an example.	1594	See C<nreqs> for an example.
436		1595
		1596	=item IO::AIO::poll
		1597
		1598	Waits until some requests have been handled.
		1599
		1600	Returns the number of requests processed, but is otherwise strictly
		1601	equivalent to:
		1602
		1603	IO::AIO::poll_wait, IO::AIO::poll_cb
		1604
437	=item IO::AIO::nreqs	1605	=item IO::AIO::flush
438		1606
439	Returns the number of requests currently outstanding (i.e. for which their	1607	Wait till all outstanding AIO requests have been handled.
440	callback has not been invoked yet).
441		1608
442	Example: wait till there are no outstanding requests anymore:	1609	Strictly equivalent to:
443		1610
444	IO::AIO::poll_wait, IO::AIO::poll_cb	1611	IO::AIO::poll_wait, IO::AIO::poll_cb
445	while IO::AIO::nreqs;	1612	while IO::AIO::nreqs;
446		1613
447	=item IO::AIO::flush	1614	=item IO::AIO::max_poll_reqs $nreqs
448		1615
449	Wait till all outstanding AIO requests have been handled.	1616	=item IO::AIO::max_poll_time $seconds
450		1617
451	Strictly equivalent to:	1618	These set the maximum number of requests (default C<0>, meaning infinity)
		1619	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1620	the maximum amount of time (default C<0>, meaning infinity) spent in
		1621	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1622	of time C<poll_cb> is allowed to use).
452		1623
453	IO::AIO::poll_wait, IO::AIO::poll_cb	1624	Setting C<max_poll_time> to a non-zero value creates an overhead of one
454	while IO::AIO::nreqs;	1625	syscall per request processed, which is not normally a problem unless your
		1626	callbacks are really really fast or your OS is really really slow (I am
		1627	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
455		1628
456	=item IO::AIO::poll	1629	Setting these is useful if you want to ensure some level of
		1630	interactiveness when perl is not fast enough to process all requests in
		1631	time.
457		1632
458	Waits until some requests have been handled.	1633	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
459		1634
460	Strictly equivalent to:	1635	Example: Install an Event watcher that automatically calls
		1636	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1637	program get the CPU sometimes even under high AIO load.
461		1638
462	IO::AIO::poll_wait, IO::AIO::poll_cb	1639	# try not to spend much more than 0.1s in poll_cb
463	if IO::AIO::nreqs;	1640	IO::AIO::max_poll_time 0.1;
		1641
		1642	# use a low priority so other tasks have priority
		1643	Event->io (fd => IO::AIO::poll_fileno,
		1644	poll => 'r', nice => 1,
		1645	cb => &IO::AIO::poll_cb);
		1646
		1647	=back
		1648
		1649	=head3 CONTROLLING THE NUMBER OF THREADS
		1650
		1651	=over
464		1652
465	=item IO::AIO::min_parallel $nthreads	1653	=item IO::AIO::min_parallel $nthreads
466		1654
467	Set the minimum number of AIO threads to C<$nthreads>. The current default	1655	Set the minimum number of AIO threads to C<$nthreads>. The current
468	is C<4>, which means four asynchronous operations can be done at one time	1656	default is C<8>, which means eight asynchronous operations can execute
469	(the number of outstanding operations, however, is unlimited).	1657	concurrently at any one time (the number of outstanding requests,
		1658	however, is unlimited).
470		1659
471	IO::AIO starts threads only on demand, when an AIO request is queued and	1660	IO::AIO starts threads only on demand, when an AIO request is queued and
472	no free thread exists.	1661	no free thread exists. Please note that queueing up a hundred requests can
		1662	create demand for a hundred threads, even if it turns out that everything
		1663	is in the cache and could have been processed faster by a single thread.
473		1664
474	It is recommended to keep the number of threads low, as some Linux	1665	It is recommended to keep the number of threads relatively low, as some
475	kernel versions will scale negatively with the number of threads (higher	1666	Linux kernel versions will scale negatively with the number of threads
476	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1667	(higher parallelity => MUCH higher latency). With current Linux 2.6
477	threads should be fine.	1668	versions, 4-32 threads should be fine.
478		1669
479	Under most circumstances you don't need to call this function, as the	1670	Under most circumstances you don't need to call this function, as the
480	module selects a default that is suitable for low to moderate load.	1671	module selects a default that is suitable for low to moderate load.
481		1672
482	=item IO::AIO::max_parallel $nthreads	1673	=item IO::AIO::max_parallel $nthreads
…		…
491	This module automatically runs C<max_parallel 0> at program end, to ensure	1682	This module automatically runs C<max_parallel 0> at program end, to ensure
492	that all threads are killed and that there are no outstanding requests.	1683	that all threads are killed and that there are no outstanding requests.
493		1684
494	Under normal circumstances you don't need to call this function.	1685	Under normal circumstances you don't need to call this function.
495		1686
		1687	=item IO::AIO::max_idle $nthreads
		1688
		1689	Limit the number of threads (default: 4) that are allowed to idle
		1690	(i.e., threads that did not get a request to process within the idle
		1691	timeout (default: 10 seconds). That means if a thread becomes idle while
		1692	C<$nthreads> other threads are also idle, it will free its resources and
		1693	exit.
		1694
		1695	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1696	to allow for extremely high load situations, but want to free resources
		1697	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1698
		1699	The default is probably ok in most situations, especially if thread
		1700	creation is fast. If thread creation is very slow on your system you might
		1701	want to use larger values.
		1702
		1703	=item IO::AIO::idle_timeout $seconds
		1704
		1705	Sets the minimum idle timeout (default 10) after which worker threads are
		1706	allowed to exit. SEe C<IO::AIO::max_idle>.
		1707
496	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1708	=item IO::AIO::max_outstanding $maxreqs
497		1709
498	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1710	Sets the maximum number of outstanding requests to C<$nreqs>. If
499	try to queue up more than this number of requests, the caller will block until	1711	you do queue up more than this number of requests, the next call to
500	some requests have been handled.	1712	C<IO::AIO::poll_cb> (and other functions calling C<poll_cb>, such as
		1713	C<IO::AIO::flush> or C<IO::AIO::poll>) will block until the limit is no
		1714	longer exceeded.
501		1715
502	The default is very large, so normally there is no practical limit. If you	1716	In other words, this setting does not enforce a queue limit, but can be
503	queue up many requests in a loop it often improves speed if you set	1717	used to make poll functions block if the limit is exceeded.
504	this to a relatively low number, such as C<100>.
505		1718
506	Under normal circumstances you don't need to call this function.	1719	This is a very bad function to use in interactive programs because it
		1720	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1721	use an C<aio_group> together with a feed callback.
		1722
		1723	It's main use is in scripts without an event loop - when you want to stat
		1724	a lot of files, you can write somehting like this:
		1725
		1726	IO::AIO::max_outstanding 32;
		1727
		1728	for my $path (...) {
		1729	aio_stat $path , ...;
		1730	IO::AIO::poll_cb;
		1731	}
		1732
		1733	IO::AIO::flush;
		1734
		1735	The call to C<poll_cb> inside the loop will normally return instantly, but
		1736	as soon as more thna C<32> reqeusts are in-flight, it will block until
		1737	some requests have been handled. This keeps the loop from pushing a large
		1738	number of C<aio_stat> requests onto the queue.
		1739
		1740	The default value for C<max_outstanding> is very large, so there is no
		1741	practical limit on the number of outstanding requests.
507		1742
508	=back	1743	=back
509		1744
		1745	=head3 STATISTICAL INFORMATION
		1746
		1747	=over
		1748
		1749	=item IO::AIO::nreqs
		1750
		1751	Returns the number of requests currently in the ready, execute or pending
		1752	states (i.e. for which their callback has not been invoked yet).
		1753
		1754	Example: wait till there are no outstanding requests anymore:
		1755
		1756	IO::AIO::poll_wait, IO::AIO::poll_cb
		1757	while IO::AIO::nreqs;
		1758
		1759	=item IO::AIO::nready
		1760
		1761	Returns the number of requests currently in the ready state (not yet
		1762	executed).
		1763
		1764	=item IO::AIO::npending
		1765
		1766	Returns the number of requests currently in the pending state (executed,
		1767	but not yet processed by poll_cb).
		1768
		1769	=back
		1770
		1771	=head3 MISCELLANEOUS FUNCTIONS
		1772
		1773	IO::AIO implements some functions that might be useful, but are not
		1774	asynchronous.
		1775
		1776	=over 4
		1777
		1778	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		1779
		1780	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		1781	but is blocking (this makes most sense if you know the input data is
		1782	likely cached already and the output filehandle is set to non-blocking
		1783	operations).
		1784
		1785	Returns the number of bytes copied, or C<-1> on error.
		1786
		1787	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		1788
		1789	Simply calls the C<posix_fadvise> function (see its
		1790	manpage for details). The following advice constants are
		1791	available: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		1792	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		1793	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		1794
		1795	On systems that do not implement C<posix_fadvise>, this function returns
		1796	ENOSYS, otherwise the return value of C<posix_fadvise>.
		1797
		1798	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		1799
		1800	Simply calls the C<posix_madvise> function (see its
		1801	manpage for details). The following advice constants are
		1802	available: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		1803	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>.
		1804
		1805	On systems that do not implement C<posix_madvise>, this function returns
		1806	ENOSYS, otherwise the return value of C<posix_madvise>.
		1807
		1808	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		1809
		1810	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		1811	$scalar (see its manpage for details). The following protect
		1812	constants are available: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		1813	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		1814
		1815	On systems that do not implement C<mprotect>, this function returns
		1816	ENOSYS, otherwise the return value of C<mprotect>.
		1817
		1818	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		1819
		1820	Memory-maps a file (or anonymous memory range) and attaches it to the
		1821	given C<$scalar>, which will act like a string scalar.
		1822
		1823	The only operations allowed on the scalar are C<substr>/C<vec> that don't
		1824	change the string length, and most read-only operations such as copying it
		1825	or searching it with regexes and so on.
		1826
		1827	Anything else is unsafe and will, at best, result in memory leaks.
		1828
		1829	The memory map associated with the C<$scalar> is automatically removed
		1830	when the C<$scalar> is destroyed, or when the C<IO::AIO::mmap> or
		1831	C<IO::AIO::munmap> functions are called.
		1832
		1833	This calls the C<mmap>(2) function internally. See your system's manual
		1834	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		1835
		1836	The C<$length> must be larger than zero and smaller than the actual
		1837	filesize.
		1838
		1839	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		1840	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		1841
		1842	C<$flags> can be a combination of C<IO::AIO::MAP_SHARED> or
		1843	C<IO::AIO::MAP_PRIVATE>, or a number of system-specific flags (when
		1844	not available, the are defined as 0): C<IO::AIO::MAP_ANONYMOUS>
		1845	(which is set to C<MAP_ANON> if your system only provides this
		1846	constant), C<IO::AIO::MAP_HUGETLB>, C<IO::AIO::MAP_LOCKED>,
		1847	C<IO::AIO::MAP_NORESERVE>, C<IO::AIO::MAP_POPULATE> or
		1848	C<IO::AIO::MAP_NONBLOCK>
		1849
		1850	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		1851
		1852	C<$offset> is the offset from the start of the file - it generally must be
		1853	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		1854
		1855	Example:
		1856
		1857	use Digest::MD5;
		1858	use IO::AIO;
		1859
		1860	open my $fh, "<verybigfile"
		1861	or die "$!";
		1862
		1863	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		1864	or die "verybigfile: $!";
		1865
		1866	my $fast_md5 = md5 $data;
		1867
		1868	=item IO::AIO::munmap $scalar
		1869
		1870	Removes a previous mmap and undefines the C<$scalar>.
		1871
		1872	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		1873
		1874	Calls the C<munlock> function, undoing the effects of a previous
		1875	C<aio_mlock> call (see its description for details).
		1876
		1877	=item IO::AIO::munlockall
		1878
		1879	Calls the C<munlockall> function.
		1880
		1881	On systems that do not implement C<munlockall>, this function returns
		1882	ENOSYS, otherwise the return value of C<munlockall>.
		1883
		1884	=back
		1885
510	=cut	1886	=cut
511		1887
512	# support function to convert a fd into a perl filehandle
513	sub _fd2fh {
514	return undef if $_[0] < 0;
515
516	# try to generate nice filehandles
517	my $sym = "IO::AIO::fd#$_[0]";
518	local *$sym;
519
520	open *$sym, "+<&=$_[0]" # usually works under any unix
521	or open *$sym, "<&=$_[0]" # cygwin needs this
522	or open *$sym, ">&=$_[0]" # or this
523	or return undef;
524
525	*$sym
526	}
527
528	min_parallel 4;	1888	min_parallel 8;
529		1889
530	END {	1890	END { flush }
531	max_parallel 0;
532	}
533		1891
534	1;	1892	1;
535		1893
		1894	=head1 EVENT LOOP INTEGRATION
		1895
		1896	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		1897	automatically into many event loops:
		1898
		1899	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		1900	use AnyEvent::AIO;
		1901
		1902	You can also integrate IO::AIO manually into many event loops, here are
		1903	some examples of how to do this:
		1904
		1905	# EV integration
		1906	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		1907
		1908	# Event integration
		1909	Event->io (fd => IO::AIO::poll_fileno,
		1910	poll => 'r',
		1911	cb => \&IO::AIO::poll_cb);
		1912
		1913	# Glib/Gtk2 integration
		1914	add_watch Glib::IO IO::AIO::poll_fileno,
		1915	in => sub { IO::AIO::poll_cb; 1 };
		1916
		1917	# Tk integration
		1918	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		1919	readable => \&IO::AIO::poll_cb);
		1920
		1921	# Danga::Socket integration
		1922	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		1923	\&IO::AIO::poll_cb);
		1924
536	=head2 FORK BEHAVIOUR	1925	=head2 FORK BEHAVIOUR
537		1926
538	Before the fork, IO::AIO enters a quiescent state where no requests	1927	Usage of pthreads in a program changes the semantics of fork
539	can be added in other threads and no results will be processed. After	1928	considerably. Specifically, only async-safe functions can be called after
540	the fork the parent simply leaves the quiescent state and continues	1929	fork. Perl doesn't know about this, so in general, you cannot call fork
541	request/result processing, while the child clears the request/result	1930	with defined behaviour in perl if pthreads are involved. IO::AIO uses
542	queue (so the requests started before the fork will only be handled in	1931	pthreads, so this applies, but many other extensions and (for inexplicable
543	the parent). Threats will be started on demand until the limit ste in the	1932	reasons) perl itself often is linked against pthreads, so this limitation
544	parent process has been reached again.	1933	applies to quite a lot of perls.
		1934
		1935	This module no longer tries to fight your OS, or POSIX. That means IO::AIO
		1936	only works in the process that loaded it. Forking is fully supported, but
		1937	using IO::AIO in the child is not.
		1938
		1939	You might get around by not I<using> IO::AIO before (or after)
		1940	forking. You could also try to call the L<IO::AIO::reinit> function in the
		1941	child:
		1942
		1943	=over 4
		1944
		1945	=item IO::AIO::reinit
		1946
		1947	Abandons all current requests and I/O threads and simply reinitialises all
		1948	data structures. This is not an operation supported by any standards, but
		1949	happens to work on GNU/Linux and some newer BSD systems.
		1950
		1951	The only reasonable use for this function is to call it after forking, if
		1952	C<IO::AIO> was used in the parent. Calling it while IO::AIO is active in
		1953	the process will result in undefined behaviour. Calling it at any time
		1954	will also result in any undefined (by POSIX) behaviour.
		1955
		1956	=back
		1957
		1958	=head2 MEMORY USAGE
		1959
		1960	Per-request usage:
		1961
		1962	Each aio request uses - depending on your architecture - around 100-200
		1963	bytes of memory. In addition, stat requests need a stat buffer (possibly
		1964	a few hundred bytes), readdir requires a result buffer and so on. Perl
		1965	scalars and other data passed into aio requests will also be locked and
		1966	will consume memory till the request has entered the done state.
		1967
		1968	This is not awfully much, so queuing lots of requests is not usually a
		1969	problem.
		1970
		1971	Per-thread usage:
		1972
		1973	In the execution phase, some aio requests require more memory for
		1974	temporary buffers, and each thread requires a stack and other data
		1975	structures (usually around 16k-128k, depending on the OS).
		1976
		1977	=head1 KNOWN BUGS
		1978
		1979	Known bugs will be fixed in the next release.
545		1980
546	=head1 SEE ALSO	1981	=head1 SEE ALSO
547		1982
548	L<Coro>, L<Linux::AIO>.	1983	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		1984	more natural syntax.
549		1985
550	=head1 AUTHOR	1986	=head1 AUTHOR
551		1987
552	Marc Lehmann <schmorp@schmorp.de>	1988	Marc Lehmann <schmorp@schmorp.de>
553	http://home.schmorp.de/	1989	http://home.schmorp.de/

Diff Legend

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