ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.55 by root, Sun Oct 22 00:49:29 2006 UTC vs.
Revision 1.193 by root, Thu May 26 04:15:37 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	use IO::AIO 2; # version has aio objects	21	# version 2+ has request and group objects
		22	use IO::AIO 2;
21		23
		24	aioreq_pri 4; # give next request a very high priority
22	my $req = aio_unlink "/tmp/file", sub { };	25	my $req = aio_unlink "/tmp/file", sub { };
23	$req->cancel; # cancel request if still in queue	26	$req->cancel; # cancel request if still in queue
24		27
25	# AnyEvent	28	my $grp = aio_group sub { print "all stats done\n" };
26	open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";	29	add $grp aio_stat "..." for ...;
27	my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
28
29	# Event
30	Event->io (fd => IO::AIO::poll_fileno,
31	poll => 'r',
32	cb => \&IO::AIO::poll_cb);
33
34	# Glib/Gtk2
35	add_watch Glib::IO IO::AIO::poll_fileno,
36	in => sub { IO::AIO::poll_cb; 1 };
37
38	# Tk
39	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
40	readable => \&IO::AIO::poll_cb);
41
42	# Danga::Socket
43	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
44	\&IO::AIO::poll_cb);
45
46		30
47	=head1 DESCRIPTION	31	=head1 DESCRIPTION
48		32
49	This module implements asynchronous I/O using whatever means your	33	This module implements asynchronous I/O using whatever means your
50	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>).
51		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
52	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
53	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
54	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
55	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
56	aio functions available on many operating systems. However, they are often	57	functions available on many operating systems. However, they are often
57	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
58	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
59	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.
60		62
61	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,
62	currently not reentrant, so use appropriate locking yourself, always call	64	it is currently not reentrant in any way, so use appropriate locking
63	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
64	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
65		161
66	=cut	162	=cut
67		163
68	package IO::AIO;	164	package IO::AIO;
69		165
70	no warnings;	166	use Carp ();
71	use strict 'vars';	167
		168	use common::sense;
72		169
73	use base 'Exporter';	170	use base 'Exporter';
74		171
75	BEGIN {	172	BEGIN {
76	our $VERSION = '2.0';	173	our $VERSION = '3.8';
77		174
78	our @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
79	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
80	aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move	177	aio_scandir aio_symlink aio_readlink aio_sync aio_fsync
		178	aio_fdatasync aio_sync_file_range aio_pathsync aio_readahead
		179	aio_rename aio_link aio_move aio_copy aio_group
		180	aio_nop aio_mknod aio_load aio_rmtree aio_mkdir aio_chown
		181	aio_chmod aio_utime aio_truncate
		182	aio_msync aio_mtouch aio_mlock aio_mlockall
81	aio_group);	183	aio_statvfs);
82	our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);	184
		185	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		186	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		187	min_parallel max_parallel max_idle idle_timeout
		188	nreqs nready npending nthreads
		189	max_poll_time max_poll_reqs
		190	sendfile fadvise madvise
		191	mmap munmap munlock munlockall);
		192
		193	push @AIO_REQ, qw(aio_busy); # not exported
83		194
84	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';	195	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
85		196
86	require XSLoader;	197	require XSLoader;
87	XSLoader::load ("IO::AIO", $VERSION);	198	XSLoader::load ("IO::AIO", $VERSION);
88	}	199	}
89		200
90	=head1 FUNCTIONS	201	=head1 FUNCTIONS
91		202
		203	=head2 QUICK OVERVIEW
		204
		205	This section simply lists the prototypes of the most important functions
		206	for quick reference. See the following sections for function-by-function
		207	documentation.
		208
		209	aio_open $pathname, $flags, $mode, $callback->($fh)
		210	aio_close $fh, $callback->($status)
		211	aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		212	aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
		213	aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		214	aio_readahead $fh,$offset,$length, $callback->($retval)
		215	aio_stat $fh_or_path, $callback->($status)
		216	aio_lstat $fh, $callback->($status)
		217	aio_statvfs $fh_or_path, $callback->($statvfs)
		218	aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		219	aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		220	aio_truncate $fh_or_path, $offset, $callback->($status)
		221	aio_chmod $fh_or_path, $mode, $callback->($status)
		222	aio_unlink $pathname, $callback->($status)
		223	aio_mknod $path, $mode, $dev, $callback->($status)
		224	aio_link $srcpath, $dstpath, $callback->($status)
		225	aio_symlink $srcpath, $dstpath, $callback->($status)
		226	aio_readlink $path, $callback->($link)
		227	aio_rename $srcpath, $dstpath, $callback->($status)
		228	aio_mkdir $pathname, $mode, $callback->($status)
		229	aio_rmdir $pathname, $callback->($status)
		230	aio_readdir $pathname, $callback->($entries)
		231	aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		232	IO::AIO::READDIR_DENTS IO::AIO::READDIR_DIRS_FIRST
		233	IO::AIO::READDIR_STAT_ORDER IO::AIO::READDIR_FOUND_UNKNOWN
		234	aio_load $path, $data, $callback->($status)
		235	aio_copy $srcpath, $dstpath, $callback->($status)
		236	aio_move $srcpath, $dstpath, $callback->($status)
		237	aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
		238	aio_rmtree $path, $callback->($status)
		239	aio_sync $callback->($status)
		240	aio_fsync $fh, $callback->($status)
		241	aio_fdatasync $fh, $callback->($status)
		242	aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		243	aio_pathsync $path, $callback->($status)
		244	aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		245	aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		246	aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		247	aio_mlockall $flags, $callback->($status)
		248	aio_group $callback->(...)
		249	aio_nop $callback->()
		250
		251	$prev_pri = aioreq_pri [$pri]
		252	aioreq_nice $pri_adjust
		253
		254	IO::AIO::poll_wait
		255	IO::AIO::poll_cb
		256	IO::AIO::poll
		257	IO::AIO::flush
		258	IO::AIO::max_poll_reqs $nreqs
		259	IO::AIO::max_poll_time $seconds
		260	IO::AIO::min_parallel $nthreads
		261	IO::AIO::max_parallel $nthreads
		262	IO::AIO::max_idle $nthreads
		263	IO::AIO::idle_timeout $seconds
		264	IO::AIO::max_outstanding $maxreqs
		265	IO::AIO::nreqs
		266	IO::AIO::nready
		267	IO::AIO::npending
		268
		269	IO::AIO::sendfile $ofh, $ifh, $offset, $count
		270	IO::AIO::fadvise $fh, $offset, $len, $advice
		271	IO::AIO::madvise $scalar, $offset, $length, $advice
		272	IO::AIO::mprotect $scalar, $offset, $length, $protect
		273	IO::AIO::munlock $scalar, $offset = 0, $length = undef
		274	IO::AIO::munlockall
		275
92	=head2 AIO FUNCTIONS	276	=head2 AIO REQUEST FUNCTIONS
93		277
94	All the C<aio_*> calls are more or less thin wrappers around the syscall	278	All the C<aio_*> calls are more or less thin wrappers around the syscall
95	with the same name (sans C<aio_>). The arguments are similar or identical,	279	with the same name (sans C<aio_>). The arguments are similar or identical,
96	and they all accept an additional (and optional) C<$callback> argument	280	and they all accept an additional (and optional) C<$callback> argument
97	which must be a code reference. This code reference will get called with	281	which must be a code reference. This code reference will get called with
98	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	282	the syscall return code (e.g. most syscalls return C<-1> on error, unlike
99	perl, which usually delivers "false") as it's sole argument when the given	283	perl, which usually delivers "false") as its sole argument after the given
100	syscall has been executed asynchronously.	284	syscall has been executed asynchronously.
101		285
102	All functions expecting a filehandle keep a copy of the filehandle	286	All functions expecting a filehandle keep a copy of the filehandle
103	internally until the request has finished.	287	internally until the request has finished.
104		288
105	All requests return objects of type L<IO::AIO::REQ> that allow further	289	All functions return request objects of type L<IO::AIO::REQ> that allow
106	manipulation of those requests while they are in-flight.	290	further manipulation of those requests while they are in-flight.
107		291
108	The pathnames you pass to these routines I<must> be absolute and	292	The pathnames you pass to these routines I<must> be absolute and
109	encoded in byte form. The reason for the former is that at the time the	293	encoded as octets. The reason for the former is that at the time the
110	request is being executed, the current working directory could have	294	request is being executed, the current working directory could have
111	changed. Alternatively, you can make sure that you never change the	295	changed. Alternatively, you can make sure that you never change the
112	current working directory.	296	current working directory anywhere in the program and then use relative
		297	paths.
113		298
114	To encode pathnames to byte form, either make sure you either: a)	299	To encode pathnames as octets, either make sure you either: a) always pass
115	always pass in filenames you got from outside (command line, readdir	300	in filenames you got from outside (command line, readdir etc.) without
116	etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode	301	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
117	your pathnames to the locale (or other) encoding in effect in the user	302	your pathnames to the locale (or other) encoding in effect in the user
118	environment, d) use Glib::filename_from_unicode on unicode filenames or e)	303	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
119	use something else.	304	use something else to ensure your scalar has the correct contents.
		305
		306	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		307	handles correctly whether it is set or not.
120		308
121	=over 4	309	=over 4
		310
		311	=item $prev_pri = aioreq_pri [$pri]
		312
		313	Returns the priority value that would be used for the next request and, if
		314	C<$pri> is given, sets the priority for the next aio request.
		315
		316	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		317	and C<4>, respectively. Requests with higher priority will be serviced
		318	first.
		319
		320	The priority will be reset to C<0> after each call to one of the C<aio_*>
		321	functions.
		322
		323	Example: open a file with low priority, then read something from it with
		324	higher priority so the read request is serviced before other low priority
		325	open requests (potentially spamming the cache):
		326
		327	aioreq_pri -3;
		328	aio_open ..., sub {
		329	return unless $_[0];
		330
		331	aioreq_pri -2;
		332	aio_read $_[0], ..., sub {
		333	...
		334	};
		335	};
		336
		337
		338	=item aioreq_nice $pri_adjust
		339
		340	Similar to C<aioreq_pri>, but subtracts the given value from the current
		341	priority, so the effect is cumulative.
		342
122		343
123	=item aio_open $pathname, $flags, $mode, $callback->($fh)	344	=item aio_open $pathname, $flags, $mode, $callback->($fh)
124		345
125	Asynchronously open or create a file and call the callback with a newly	346	Asynchronously open or create a file and call the callback with a newly
126	created filehandle for the file.	347	created filehandle for the file.
…		…
132	list. They are the same as used by C<sysopen>.	353	list. They are the same as used by C<sysopen>.
133		354
134	Likewise, C<$mode> specifies the mode of the newly created file, if it	355	Likewise, C<$mode> specifies the mode of the newly created file, if it
135	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,	356	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
136	except that it is mandatory (i.e. use C<0> if you don't create new files,	357	except that it is mandatory (i.e. use C<0> if you don't create new files,
137	and C<0666> or C<0777> if you do).	358	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		359	by the umask in effect then the request is being executed, so better never
		360	change the umask.
138		361
139	Example:	362	Example:
140		363
141	aio_open "/etc/passwd", O_RDONLY, 0, sub {	364	aio_open "/etc/passwd", IO::AIO::O_RDONLY, 0, sub {
142	if ($_[0]) {	365	if ($_[0]) {
143	print "open successful, fh is $_[0]\n";	366	print "open successful, fh is $_[0]\n";
144	...	367	...
145	} else {	368	} else {
146	die "open failed: $!\n";	369	die "open failed: $!\n";
147	}	370	}
148	};	371	};
149		372
		373
150	=item aio_close $fh, $callback->($status)	374	=item aio_close $fh, $callback->($status)
151		375
152	Asynchronously close a file and call the callback with the result	376	Asynchronously close a file and call the callback with the result
153	code. I<WARNING:> although accepted, you should not pass in a perl	377	code.
154	filehandle here, as perl will likely close the file descriptor another
155	time when the filehandle is destroyed. Normally, you can safely call perls
156	C<close> or just let filehandles go out of scope.
157		378
158	This is supposed to be a bug in the API, so that might change. It's	379	Unfortunately, you can't do this to perl. Perl I<insists> very strongly on
159	therefore best to avoid this function.	380	closing the file descriptor associated with the filehandle itself.
		381
		382	Therefore, C<aio_close> will not close the filehandle - instead it will
		383	use dup2 to overwrite the file descriptor with the write-end of a pipe
		384	(the pipe fd will be created on demand and will be cached).
		385
		386	Or in other words: the file descriptor will be closed, but it will not be
		387	free for reuse until the perl filehandle is closed.
		388
		389	=cut
160		390
161	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	391	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
162		392
163	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)	393	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
164		394
165	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	395	Reads or writes C<$length> bytes from or to the specified C<$fh> and
166	into the scalar given by C<data> and offset C<dataoffset> and calls the	396	C<$offset> into the scalar given by C<$data> and offset C<$dataoffset>
167	callback without the actual number of bytes read (or -1 on error, just	397	and calls the callback without the actual number of bytes read (or -1 on
168	like the syscall).	398	error, just like the syscall).
		399
		400	C<aio_read> will, like C<sysread>, shrink or grow the C<$data> scalar to
		401	offset plus the actual number of bytes read.
		402
		403	If C<$offset> is undefined, then the current file descriptor offset will
		404	be used (and updated), otherwise the file descriptor offset will not be
		405	changed by these calls.
		406
		407	If C<$length> is undefined in C<aio_write>, use the remaining length of
		408	C<$data>.
		409
		410	If C<$dataoffset> is less than zero, it will be counted from the end of
		411	C<$data>.
169		412
170	The C<$data> scalar I<MUST NOT> be modified in any way while the request	413	The C<$data> scalar I<MUST NOT> be modified in any way while the request
171	is outstanding. Modifying it can result in segfaults or WW3 (if the	414	is outstanding. Modifying it can result in segfaults or World War III (if
172	necessary/optional hardware is installed).	415	the necessary/optional hardware is installed).
173		416
174	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	417	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
175	offset C<0> within the scalar:	418	offset C<0> within the scalar:
176		419
177	aio_read $fh, 7, 15, $buffer, 0, sub {	420	aio_read $fh, 7, 15, $buffer, 0, sub {
178	$_[0] > 0 or die "read error: $!";	421	$_[0] > 0 or die "read error: $!";
179	print "read $_[0] bytes: <$buffer>\n";	422	print "read $_[0] bytes: <$buffer>\n";
180	};	423	};
181		424
182	=item aio_move $srcpath, $dstpath, $callback->($status)
183
184	Try to move the I<file> (directories not supported as either source or
185	destination) from C<$srcpath> to C<$dstpath> and call the callback with
186	the C<0> (error) or C<-1> ok.
187
188	This is a composite request that tries to rename(2) the file first. If
189	rename files with C<EXDEV>, it creates the destination file with mode 0200
190	and copies the contents of the source file into it using C<aio_sendfile>,
191	followed by restoring atime, mtime, access mode and uid/gid, in that
192	order, and unlinking the C<$srcpath>.
193
194	If an error occurs, the partial destination file will be unlinked, if
195	possible, except when setting atime, mtime, access mode and uid/gid, where
196	errors are being ignored.
197
198	=cut
199
200	sub aio_move($$$) {
201	my ($src, $dst, $cb) = @_;
202
203	my $grp = aio_group;
204
205	add $grp aio_rename $src, $dst, sub {
206	if ($_[0] && $! == EXDEV) {
207	add $grp aio_open $src, O_RDONLY, 0, sub {
208	if (my $src_fh = $_[0]) {
209	my @stat = stat $src_fh;
210
211	add $grp aio_open $dst, O_WRONLY, 0200, sub {
212	if (my $dst_fh = $_[0]) {
213	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
214	close $src_fh;
215
216	if ($_[0] == $stat[7]) {
217	utime $stat[8], $stat[9], $dst;
218	chmod $stat[2] & 07777, $dst_fh;
219	chown $stat[4], $stat[5], $dst_fh;
220	close $dst_fh;
221
222	add $grp aio_unlink $src, sub {
223	$cb->($_[0]);
224	};
225	} else {
226	my $errno = $!;
227	add $grp aio_unlink $dst, sub {
228	$! = $errno;
229	$cb->(-1);
230	};
231	}
232	};
233	} else {
234	$cb->(-1);
235	}
236	},
237
238	} else {
239	$cb->(-1);
240	}
241	};
242	} else {
243	$cb->($_[0]);
244	}
245	};
246
247	$grp
248	}
249		425
250	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)	426	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
251		427
252	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts	428	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
253	reading at byte offset C<$in_offset>, and starts writing at the current	429	reading at byte offset C<$in_offset>, and starts writing at the current
254	file offset of C<$out_fh>. Because of that, it is not safe to issue more	430	file offset of C<$out_fh>. Because of that, it is not safe to issue more
255	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each	431	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
256	other.	432	other.
257		433
		434	Please note that C<aio_sendfile> can read more bytes from C<$in_fh> than
		435	are written, and there is no way to find out how many bytes have been read
		436	from C<aio_sendfile> alone, as C<aio_sendfile> only provides the number of
		437	bytes written to C<$out_fh>. Only if the result value equals C<$length>
		438	one can assume that C<$length> bytes have been read.
		439
		440	Unlike with other C<aio_> functions, it makes a lot of sense to use
		441	C<aio_sendfile> on non-blocking sockets, as long as one end (typically
		442	the C<$in_fh>) is a file - the file I/O will then be asynchronous, while
		443	the socket I/O will be non-blocking. Note, however, that you can run into
		444	a trap where C<aio_sendfile> reads some data with readahead, then fails
		445	to write all data, and when the socket is ready the next time, the data
		446	in the cache is already lost, forcing C<aio_sendfile> to again hit the
		447	disk. Explicit C<aio_read> + C<aio_write> let's you control resource usage
		448	much better.
		449
258	This call tries to make use of a native C<sendfile> syscall to provide	450	This call tries to make use of a native C<sendfile> syscall to provide
259	zero-copy operation. For this to work, C<$out_fh> should refer to a	451	zero-copy operation. For this to work, C<$out_fh> should refer to a
260	socket, and C<$in_fh> should refer to mmap'able file.	452	socket, and C<$in_fh> should refer to an mmap'able file.
261		453
262	If the native sendfile call fails or is not implemented, it will be	454	If a native sendfile cannot be found or it fails with C<ENOSYS>,
		455	C<ENOTSUP>, C<EOPNOTSUPP>, C<EAFNOSUPPORT>, C<EPROTOTYPE> or C<ENOTSOCK>,
263	emulated, so you can call C<aio_sendfile> on any type of filehandle	456	it will be emulated, so you can call C<aio_sendfile> on any type of
264	regardless of the limitations of the operating system.	457	filehandle regardless of the limitations of the operating system.
265		458
266	Please note, however, that C<aio_sendfile> can read more bytes from
267	C<$in_fh> than are written, and there is no way to find out how many
268	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
269	provides the number of bytes written to C<$out_fh>. Only if the result
270	value equals C<$length> one can assume that C<$length> bytes have been
271	read.
272		459
273	=item aio_readahead $fh,$offset,$length, $callback->($retval)	460	=item aio_readahead $fh,$offset,$length, $callback->($retval)
274		461
275	C<aio_readahead> populates the page cache with data from a file so that	462	C<aio_readahead> populates the page cache with data from a file so that
276	subsequent reads from that file will not block on disk I/O. The C<$offset>	463	subsequent reads from that file will not block on disk I/O. The C<$offset>
…		…
282	file. The current file offset of the file is left unchanged.	469	file. The current file offset of the file is left unchanged.
283		470
284	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	471	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
285	emulated by simply reading the data, which would have a similar effect.	472	emulated by simply reading the data, which would have a similar effect.
286		473
		474
287	=item aio_stat $fh_or_path, $callback->($status)	475	=item aio_stat $fh_or_path, $callback->($status)
288		476
289	=item aio_lstat $fh, $callback->($status)	477	=item aio_lstat $fh, $callback->($status)
290		478
291	Works like perl's C<stat> or C<lstat> in void context. The callback will	479	Works like perl's C<stat> or C<lstat> in void context. The callback will
…		…
296	for an explanation.	484	for an explanation.
297		485
298	Currently, the stats are always 64-bit-stats, i.e. instead of returning an	486	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
299	error when stat'ing a large file, the results will be silently truncated	487	error when stat'ing a large file, the results will be silently truncated
300	unless perl itself is compiled with large file support.	488	unless perl itself is compiled with large file support.
		489
		490	To help interpret the mode and dev/rdev stat values, IO::AIO offers the
		491	following constants and functions (if not implemented, the constants will
		492	be C<0> and the functions will either C<croak> or fall back on traditional
		493	behaviour).
		494
		495	C<S_IFMT>, C<S_IFIFO>, C<S_IFCHR>, C<S_IFBLK>, C<S_IFLNK>, C<S_IFREG>,
		496	C<S_IFDIR>, C<S_IFWHT>, C<S_IFSOCK>, C<IO::AIO::major $dev_t>,
		497	C<IO::AIO::minor $dev_t>, C<IO::AIO::makedev $major, $minor>.
301		498
302	Example: Print the length of F</etc/passwd>:	499	Example: Print the length of F</etc/passwd>:
303		500
304	aio_stat "/etc/passwd", sub {	501	aio_stat "/etc/passwd", sub {
305	$_[0] and die "stat failed: $!";	502	$_[0] and die "stat failed: $!";
306	print "size is ", -s _, "\n";	503	print "size is ", -s _, "\n";
307	};	504	};
308		505
		506
		507	=item aio_statvfs $fh_or_path, $callback->($statvfs)
		508
		509	Works like the POSIX C<statvfs> or C<fstatvfs> syscalls, depending on
		510	whether a file handle or path was passed.
		511
		512	On success, the callback is passed a hash reference with the following
		513	members: C<bsize>, C<frsize>, C<blocks>, C<bfree>, C<bavail>, C<files>,
		514	C<ffree>, C<favail>, C<fsid>, C<flag> and C<namemax>. On failure, C<undef>
		515	is passed.
		516
		517	The following POSIX IO::AIO::ST_* constants are defined: C<ST_RDONLY> and
		518	C<ST_NOSUID>.
		519
		520	The following non-POSIX IO::AIO::ST_* flag masks are defined to
		521	their correct value when available, or to C<0> on systems that do
		522	not support them: C<ST_NODEV>, C<ST_NOEXEC>, C<ST_SYNCHRONOUS>,
		523	C<ST_MANDLOCK>, C<ST_WRITE>, C<ST_APPEND>, C<ST_IMMUTABLE>, C<ST_NOATIME>,
		524	C<ST_NODIRATIME> and C<ST_RELATIME>.
		525
		526	Example: stat C</wd> and dump out the data if successful.
		527
		528	aio_statvfs "/wd", sub {
		529	my $f = $_[0]
		530	or die "statvfs: $!";
		531
		532	use Data::Dumper;
		533	say Dumper $f;
		534	};
		535
		536	# result:
		537	{
		538	bsize => 1024,
		539	bfree => 4333064312,
		540	blocks => 10253828096,
		541	files => 2050765568,
		542	flag => 4096,
		543	favail => 2042092649,
		544	bavail => 4333064312,
		545	ffree => 2042092649,
		546	namemax => 255,
		547	frsize => 1024,
		548	fsid => 1810
		549	}
		550
		551
		552	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		553
		554	Works like perl's C<utime> function (including the special case of $atime
		555	and $mtime being undef). Fractional times are supported if the underlying
		556	syscalls support them.
		557
		558	When called with a pathname, uses utimes(2) if available, otherwise
		559	utime(2). If called on a file descriptor, uses futimes(2) if available,
		560	otherwise returns ENOSYS, so this is not portable.
		561
		562	Examples:
		563
		564	# set atime and mtime to current time (basically touch(1)):
		565	aio_utime "path", undef, undef;
		566	# set atime to current time and mtime to beginning of the epoch:
		567	aio_utime "path", time, undef; # undef==0
		568
		569
		570	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		571
		572	Works like perl's C<chown> function, except that C<undef> for either $uid
		573	or $gid is being interpreted as "do not change" (but -1 can also be used).
		574
		575	Examples:
		576
		577	# same as "chown root path" in the shell:
		578	aio_chown "path", 0, -1;
		579	# same as above:
		580	aio_chown "path", 0, undef;
		581
		582
		583	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		584
		585	Works like truncate(2) or ftruncate(2).
		586
		587
		588	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		589
		590	Works like perl's C<chmod> function.
		591
		592
309	=item aio_unlink $pathname, $callback->($status)	593	=item aio_unlink $pathname, $callback->($status)
310		594
311	Asynchronously unlink (delete) a file and call the callback with the	595	Asynchronously unlink (delete) a file and call the callback with the
312	result code.	596	result code.
313		597
		598
		599	=item aio_mknod $path, $mode, $dev, $callback->($status)
		600
		601	[EXPERIMENTAL]
		602
		603	Asynchronously create a device node (or fifo). See mknod(2).
		604
		605	The only (POSIX-) portable way of calling this function is:
		606
		607	aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		608
		609	See C<aio_stat> for info about some potentially helpful extra constants
		610	and functions.
		611
314	=item aio_link $srcpath, $dstpath, $callback->($status)	612	=item aio_link $srcpath, $dstpath, $callback->($status)
315		613
316	Asynchronously create a new link to the existing object at C<$srcpath> at	614	Asynchronously create a new link to the existing object at C<$srcpath> at
317	the path C<$dstpath> and call the callback with the result code.	615	the path C<$dstpath> and call the callback with the result code.
318		616
		617
319	=item aio_symlink $srcpath, $dstpath, $callback->($status)	618	=item aio_symlink $srcpath, $dstpath, $callback->($status)
320		619
321	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at	620	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
322	the path C<$dstpath> and call the callback with the result code.	621	the path C<$dstpath> and call the callback with the result code.
323		622
		623
		624	=item aio_readlink $path, $callback->($link)
		625
		626	Asynchronously read the symlink specified by C<$path> and pass it to
		627	the callback. If an error occurs, nothing or undef gets passed to the
		628	callback.
		629
		630
324	=item aio_rename $srcpath, $dstpath, $callback->($status)	631	=item aio_rename $srcpath, $dstpath, $callback->($status)
325		632
326	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as	633	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
327	rename(2) and call the callback with the result code.	634	rename(2) and call the callback with the result code.
328		635
		636
		637	=item aio_mkdir $pathname, $mode, $callback->($status)
		638
		639	Asynchronously mkdir (create) a directory and call the callback with
		640	the result code. C<$mode> will be modified by the umask at the time the
		641	request is executed, so do not change your umask.
		642
		643
329	=item aio_rmdir $pathname, $callback->($status)	644	=item aio_rmdir $pathname, $callback->($status)
330		645
331	Asynchronously rmdir (delete) a directory and call the callback with the	646	Asynchronously rmdir (delete) a directory and call the callback with the
332	result code.	647	result code.
		648
333		649
334	=item aio_readdir $pathname, $callback->($entries)	650	=item aio_readdir $pathname, $callback->($entries)
335		651
336	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire	652	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
337	directory (i.e. opendir + readdir + closedir). The entries will not be	653	directory (i.e. opendir + readdir + closedir). The entries will not be
338	sorted, and will B<NOT> include the C<.> and C<..> entries.	654	sorted, and will B<NOT> include the C<.> and C<..> entries.
339		655
340	The callback a single argument which is either C<undef> or an array-ref	656	The callback is passed a single argument which is either C<undef> or an
341	with the filenames.	657	array-ref with the filenames.
		658
		659
		660	=item aio_readdirx $pathname, $flags, $callback->($entries, $flags)
		661
		662	Quite similar to C<aio_readdir>, but the C<$flags> argument allows to tune
		663	behaviour and output format. In case of an error, C<$entries> will be
		664	C<undef>.
		665
		666	The flags are a combination of the following constants, ORed together (the
		667	flags will also be passed to the callback, possibly modified):
		668
		669	=over 4
		670
		671	=item IO::AIO::READDIR_DENTS
		672
		673	When this flag is off, then the callback gets an arrayref consisting of
		674	names only (as with C<aio_readdir>), otherwise it gets an arrayref with
		675	C<[$name, $type, $inode]> arrayrefs, each describing a single directory
		676	entry in more detail.
		677
		678	C<$name> is the name of the entry.
		679
		680	C<$type> is one of the C<IO::AIO::DT_xxx> constants:
		681
		682	C<IO::AIO::DT_UNKNOWN>, C<IO::AIO::DT_FIFO>, C<IO::AIO::DT_CHR>, C<IO::AIO::DT_DIR>,
		683	C<IO::AIO::DT_BLK>, C<IO::AIO::DT_REG>, C<IO::AIO::DT_LNK>, C<IO::AIO::DT_SOCK>,
		684	C<IO::AIO::DT_WHT>.
		685
		686	C<IO::AIO::DT_UNKNOWN> means just that: readdir does not know. If you need to
		687	know, you have to run stat yourself. Also, for speed reasons, the C<$type>
		688	scalars are read-only: you can not modify them.
		689
		690	C<$inode> is the inode number (which might not be exact on systems with 64
		691	bit inode numbers and 32 bit perls). This field has unspecified content on
		692	systems that do not deliver the inode information.
		693
		694	=item IO::AIO::READDIR_DIRS_FIRST
		695
		696	When this flag is set, then the names will be returned in an order where
		697	likely directories come first, in optimal stat order. This is useful when
		698	you need to quickly find directories, or you want to find all directories
		699	while avoiding to stat() each entry.
		700
		701	If the system returns type information in readdir, then this is used
		702	to find directories directly. Otherwise, likely directories are names
		703	beginning with ".", or otherwise names with no dots, of which names with
		704	short names are tried first.
		705
		706	=item IO::AIO::READDIR_STAT_ORDER
		707
		708	When this flag is set, then the names will be returned in an order
		709	suitable for stat()'ing each one. That is, when you plan to stat()
		710	all files in the given directory, then the returned order will likely
		711	be fastest.
		712
		713	If both this flag and C<IO::AIO::READDIR_DIRS_FIRST> are specified, then
		714	the likely dirs come first, resulting in a less optimal stat order.
		715
		716	=item IO::AIO::READDIR_FOUND_UNKNOWN
		717
		718	This flag should not be set when calling C<aio_readdirx>. Instead, it
		719	is being set by C<aio_readdirx>, when any of the C<$type>'s found were
		720	C<IO::AIO::DT_UNKNOWN>. The absense of this flag therefore indicates that all
		721	C<$type>'s are known, which can be used to speed up some algorithms.
		722
		723	=back
		724
		725
		726	=item aio_load $path, $data, $callback->($status)
		727
		728	This is a composite request that tries to fully load the given file into
		729	memory. Status is the same as with aio_read.
		730
		731	=cut
		732
		733	sub aio_load($$;$) {
		734	my ($path, undef, $cb) = @_;
		735	my $data = \$_[1];
		736
		737	my $pri = aioreq_pri;
		738	my $grp = aio_group $cb;
		739
		740	aioreq_pri $pri;
		741	add $grp aio_open $path, O_RDONLY, 0, sub {
		742	my $fh = shift
		743	or return $grp->result (-1);
		744
		745	aioreq_pri $pri;
		746	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		747	$grp->result ($_[0]);
		748	};
		749	};
		750
		751	$grp
		752	}
		753
		754	=item aio_copy $srcpath, $dstpath, $callback->($status)
		755
		756	Try to copy the I<file> (directories not supported as either source or
		757	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		758	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		759
		760	This is a composite request that creates the destination file with
		761	mode 0200 and copies the contents of the source file into it using
		762	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		763	uid/gid, in that order.
		764
		765	If an error occurs, the partial destination file will be unlinked, if
		766	possible, except when setting atime, mtime, access mode and uid/gid, where
		767	errors are being ignored.
		768
		769	=cut
		770
		771	sub aio_copy($$;$) {
		772	my ($src, $dst, $cb) = @_;
		773
		774	my $pri = aioreq_pri;
		775	my $grp = aio_group $cb;
		776
		777	aioreq_pri $pri;
		778	add $grp aio_open $src, O_RDONLY, 0, sub {
		779	if (my $src_fh = $_[0]) {
		780	my @stat = stat $src_fh; # hmm, might block over nfs?
		781
		782	aioreq_pri $pri;
		783	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		784	if (my $dst_fh = $_[0]) {
		785	aioreq_pri $pri;
		786	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		787	if ($_[0] == $stat[7]) {
		788	$grp->result (0);
		789	close $src_fh;
		790
		791	my $ch = sub {
		792	aioreq_pri $pri;
		793	add $grp aio_chmod $dst_fh, $stat[2] & 07777, sub {
		794	aioreq_pri $pri;
		795	add $grp aio_chown $dst_fh, $stat[4], $stat[5], sub {
		796	aioreq_pri $pri;
		797	add $grp aio_close $dst_fh;
		798	}
		799	};
		800	};
		801
		802	aioreq_pri $pri;
		803	add $grp aio_utime $dst_fh, $stat[8], $stat[9], sub {
		804	if ($_[0] < 0 && $! == ENOSYS) {
		805	aioreq_pri $pri;
		806	add $grp aio_utime $dst, $stat[8], $stat[9], $ch;
		807	} else {
		808	$ch->();
		809	}
		810	};
		811	} else {
		812	$grp->result (-1);
		813	close $src_fh;
		814	close $dst_fh;
		815
		816	aioreq $pri;
		817	add $grp aio_unlink $dst;
		818	}
		819	};
		820	} else {
		821	$grp->result (-1);
		822	}
		823	},
		824
		825	} else {
		826	$grp->result (-1);
		827	}
		828	};
		829
		830	$grp
		831	}
		832
		833	=item aio_move $srcpath, $dstpath, $callback->($status)
		834
		835	Try to move the I<file> (directories not supported as either source or
		836	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		837	a status of C<0> (ok) or C<-1> (error, see C<$!>).
		838
		839	This is a composite request that tries to rename(2) the file first; if
		840	rename fails with C<EXDEV>, it copies the file with C<aio_copy> and, if
		841	that is successful, unlinks the C<$srcpath>.
		842
		843	=cut
		844
		845	sub aio_move($$;$) {
		846	my ($src, $dst, $cb) = @_;
		847
		848	my $pri = aioreq_pri;
		849	my $grp = aio_group $cb;
		850
		851	aioreq_pri $pri;
		852	add $grp aio_rename $src, $dst, sub {
		853	if ($_[0] && $! == EXDEV) {
		854	aioreq_pri $pri;
		855	add $grp aio_copy $src, $dst, sub {
		856	$grp->result ($_[0]);
		857
		858	if (!$_[0]) {
		859	aioreq_pri $pri;
		860	add $grp aio_unlink $src;
		861	}
		862	};
		863	} else {
		864	$grp->result ($_[0]);
		865	}
		866	};
		867
		868	$grp
		869	}
342		870
343	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)	871	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
344		872
345	Scans a directory (similar to C<aio_readdir>) but additionally tries to	873	Scans a directory (similar to C<aio_readdir>) but additionally tries to
346	separate the entries of directory C<$path> into two sets of names, ones	874	efficiently separate the entries of directory C<$path> into two sets of
347	you can recurse into (directories or links to them), and ones you cannot	875	names, directories you can recurse into (directories), and ones you cannot
348	recurse into (everything else).	876	recurse into (everything else, including symlinks to directories).
349		877
350	C<aio_scandir> is a composite request that consists of many sub	878	C<aio_scandir> is a composite request that creates of many sub requests_
351	requests. C<$maxreq> specifies the maximum number of outstanding aio	879	C<$maxreq> specifies the maximum number of outstanding aio requests that
352	requests that this function generates. If it is C<< <= 0 >>, then a	880	this function generates. If it is C<< <= 0 >>, then a suitable default
353	suitable default will be chosen (currently 8).	881	will be chosen (currently 4).
354		882
355	On error, the callback is called without arguments, otherwise it receives	883	On error, the callback is called without arguments, otherwise it receives
356	two array-refs with path-relative entry names.	884	two array-refs with path-relative entry names.
357		885
358	Example:	886	Example:
…		…
365		893
366	Implementation notes.	894	Implementation notes.
367		895
368	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.	896	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
369		897
		898	If readdir returns file type information, then this is used directly to
		899	find directories.
		900
370	After reading the directory, the modification time, size etc. of the	901	Otherwise, after reading the directory, the modification time, size etc.
371	directory before and after the readdir is checked, and if they match (and	902	of the directory before and after the readdir is checked, and if they
372	isn't the current time), the link count will be used to decide how many	903	match (and isn't the current time), the link count will be used to decide
373	entries are directories (if >= 2). Otherwise, no knowledge of the number	904	how many entries are directories (if >= 2). Otherwise, no knowledge of the
374	of subdirectories will be assumed.	905	number of subdirectories will be assumed.
375		906
376	Then entries will be sorted into likely directories (everything without	907	Then entries will be sorted into likely directories a non-initial dot
377	a non-initial dot currently) and likely non-directories (everything	908	currently) and likely non-directories (see C<aio_readdirx>). Then every
378	else). Then every entry plus an appended C</.> will be C<stat>'ed,	909	entry plus an appended C</.> will be C<stat>'ed, likely directories first,
379	likely directories first. If that succeeds, it assumes that the entry	910	in order of their inode numbers. If that succeeds, it assumes that the
380	is a directory or a symlink to directory (which will be checked	911	entry is a directory or a symlink to directory (which will be checked
381	seperately). This is often faster than stat'ing the entry itself because	912	seperately). This is often faster than stat'ing the entry itself because
382	filesystems might detect the type of the entry without reading the inode	913	filesystems might detect the type of the entry without reading the inode
383	data (e.g. ext2fs filetype feature).	914	data (e.g. ext2fs filetype feature), even on systems that cannot return
		915	the filetype information on readdir.
384		916
385	If the known number of directories (link count - 2) has been reached, the	917	If the known number of directories (link count - 2) has been reached, the
386	rest of the entries is assumed to be non-directories.	918	rest of the entries is assumed to be non-directories.
387		919
388	This only works with certainty on POSIX (= UNIX) filesystems, which	920	This only works with certainty on POSIX (= UNIX) filesystems, which
…		…
392	as those tend to return 0 or 1 as link counts, which disables the	924	as those tend to return 0 or 1 as link counts, which disables the
393	directory counting heuristic.	925	directory counting heuristic.
394		926
395	=cut	927	=cut
396		928
397	sub aio_scandir($$$) {	929	sub aio_scandir($$;$) {
398	my ($path, $maxreq, $cb) = @_;	930	my ($path, $maxreq, $cb) = @_;
399		931
		932	my $pri = aioreq_pri;
		933
400	my $grp = aio_group;	934	my $grp = aio_group $cb;
401		935
402	$maxreq = 8 if $maxreq <= 0;	936	$maxreq = 4 if $maxreq <= 0;
403		937
404	# stat once	938	# stat once
		939	aioreq_pri $pri;
405	add $grp aio_stat $path, sub {	940	add $grp aio_stat $path, sub {
406	return $cb->() if $_[0];	941	return $grp->result () if $_[0];
407	my $now = time;	942	my $now = time;
408	my $hash1 = join ":", (stat _)[0,1,3,7,9];	943	my $hash1 = join ":", (stat _)[0,1,3,7,9];
409		944
410	# read the directory entries	945	# read the directory entries
		946	aioreq_pri $pri;
411	add $grp aio_readdir $path, sub {	947	add $grp aio_readdirx $path, READDIR_DIRS_FIRST, sub {
412	my $entries = shift	948	my $entries = shift
413	or return $cb->();	949	or return $grp->result ();
414		950
415	# stat the dir another time	951	# stat the dir another time
		952	aioreq_pri $pri;
416	add $grp aio_stat $path, sub {	953	add $grp aio_stat $path, sub {
417	my $hash2 = join ":", (stat _)[0,1,3,7,9];	954	my $hash2 = join ":", (stat _)[0,1,3,7,9];
418		955
419	my $ndirs;	956	my $ndirs;
420		957
421	# take the slow route if anything looks fishy	958	# take the slow route if anything looks fishy
422	if ($hash1 ne $hash2 or (stat _)[9] == $now) {	959	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
423	$ndirs = -1;	960	$ndirs = -1;
424	} else {	961	} else {
425	# if nlink == 2, we are finished	962	# if nlink == 2, we are finished
426	# on non-posix-fs's, we rely on nlink < 2	963	# for non-posix-fs's, we rely on nlink < 2
427	$ndirs = (stat _)[3] - 2	964	$ndirs = (stat _)[3] - 2
428	or return $cb->([], $entries);	965	or return $grp->result ([], $entries);
429	}	966	}
430		967
431	# sort into likely dirs and likely nondirs
432	# dirs == files without ".", short entries first
433	$entries = [map $_->[0],
434	sort { $b->[1] cmp $a->[1] }
435	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
436	@$entries];
437
438	my (@dirs, @nondirs);	968	my (@dirs, @nondirs);
439		969
440	my ($statcb, $schedcb);	970	my $statgrp = add $grp aio_group sub {
441	my $nreq = 0;	971	$grp->result (\@dirs, \@nondirs);
		972	};
442		973
443	$schedcb = sub {	974	limit $statgrp $maxreq;
444	if (@$entries) {	975	feed $statgrp sub {
445	if ($nreq < $maxreq) {	976	return unless @$entries;
446	my $ent = pop @$entries;	977	my $entry = shift @$entries;
		978
		979	aioreq_pri $pri;
		980	add $statgrp aio_stat "$path/$entry/.", sub {
		981	if ($_[0] < 0) {
		982	push @nondirs, $entry;
		983	} else {
		984	# need to check for real directory
		985	aioreq_pri $pri;
		986	add $statgrp aio_lstat "$path/$entry", sub {
		987	if (-d _) {
		988	push @dirs, $entry;
		989
		990	unless (--$ndirs) {
		991	push @nondirs, @$entries;
		992	feed $statgrp;
		993	}
		994	} else {
		995	push @nondirs, $entry;
		996	}
447	$nreq++;	997	}
448	add $grp aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) };
449	}	998	}
450	} elsif (!$nreq) {
451	# finished
452	undef $statcb;
453	undef $schedcb;
454	$cb->(\@dirs, \@nondirs) if $cb;
455	undef $cb;
456	}	999	};
457	};	1000	};
458	$statcb = sub {
459	my ($status, $entry) = @_;
460
461	if ($status < 0) {
462	$nreq--;
463	push @nondirs, $entry;
464	&$schedcb;
465	} else {
466	# need to check for real directory
467	add $grp aio_lstat "$path/$entry", sub {
468	$nreq--;
469
470	if (-d _) {
471	push @dirs, $entry;
472
473	if (!--$ndirs) {
474	push @nondirs, @$entries;
475	$entries = [];
476	}
477	} else {
478	push @nondirs, $entry;
479	}
480
481	&$schedcb;
482	}
483	}
484	};
485
486	&$schedcb while @$entries && $nreq < $maxreq;
487	};	1001	};
488	};	1002	};
489	};	1003	};
490		1004
491	$grp	1005	$grp
492	}	1006	}
493		1007
		1008	=item aio_rmtree $path, $callback->($status)
		1009
		1010	Delete a directory tree starting (and including) C<$path>, return the
		1011	status of the final C<rmdir> only. This is a composite request that
		1012	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		1013	everything else.
		1014
		1015	=cut
		1016
		1017	sub aio_rmtree;
		1018	sub aio_rmtree($;$) {
		1019	my ($path, $cb) = @_;
		1020
		1021	my $pri = aioreq_pri;
		1022	my $grp = aio_group $cb;
		1023
		1024	aioreq_pri $pri;
		1025	add $grp aio_scandir $path, 0, sub {
		1026	my ($dirs, $nondirs) = @_;
		1027
		1028	my $dirgrp = aio_group sub {
		1029	add $grp aio_rmdir $path, sub {
		1030	$grp->result ($_[0]);
		1031	};
		1032	};
		1033
		1034	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		1035	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		1036
		1037	add $grp $dirgrp;
		1038	};
		1039
		1040	$grp
		1041	}
		1042
		1043	=item aio_sync $callback->($status)
		1044
		1045	Asynchronously call sync and call the callback when finished.
		1046
494	=item aio_fsync $fh, $callback->($status)	1047	=item aio_fsync $fh, $callback->($status)
495		1048
496	Asynchronously call fsync on the given filehandle and call the callback	1049	Asynchronously call fsync on the given filehandle and call the callback
497	with the fsync result code.	1050	with the fsync result code.
498		1051
…		…
502	callback with the fdatasync result code.	1055	callback with the fdatasync result code.
503		1056
504	If this call isn't available because your OS lacks it or it couldn't be	1057	If this call isn't available because your OS lacks it or it couldn't be
505	detected, it will be emulated by calling C<fsync> instead.	1058	detected, it will be emulated by calling C<fsync> instead.
506		1059
		1060	=item aio_sync_file_range $fh, $offset, $nbytes, $flags, $callback->($status)
		1061
		1062	Sync the data portion of the file specified by C<$offset> and C<$length>
		1063	to disk (but NOT the metadata), by calling the Linux-specific
		1064	sync_file_range call. If sync_file_range is not available or it returns
		1065	ENOSYS, then fdatasync or fsync is being substituted.
		1066
		1067	C<$flags> can be a combination of C<IO::AIO::SYNC_FILE_RANGE_WAIT_BEFORE>,
		1068	C<IO::AIO::SYNC_FILE_RANGE_WRITE> and
		1069	C<IO::AIO::SYNC_FILE_RANGE_WAIT_AFTER>: refer to the sync_file_range
		1070	manpage for details.
		1071
		1072	=item aio_pathsync $path, $callback->($status)
		1073
		1074	This request tries to open, fsync and close the given path. This is a
		1075	composite request intended to sync directories after directory operations
		1076	(E.g. rename). This might not work on all operating systems or have any
		1077	specific effect, but usually it makes sure that directory changes get
		1078	written to disc. It works for anything that can be opened for read-only,
		1079	not just directories.
		1080
		1081	Future versions of this function might fall back to other methods when
		1082	C<fsync> on the directory fails (such as calling C<sync>).
		1083
		1084	Passes C<0> when everything went ok, and C<-1> on error.
		1085
		1086	=cut
		1087
		1088	sub aio_pathsync($;$) {
		1089	my ($path, $cb) = @_;
		1090
		1091	my $pri = aioreq_pri;
		1092	my $grp = aio_group $cb;
		1093
		1094	aioreq_pri $pri;
		1095	add $grp aio_open $path, O_RDONLY, 0, sub {
		1096	my ($fh) = @_;
		1097	if ($fh) {
		1098	aioreq_pri $pri;
		1099	add $grp aio_fsync $fh, sub {
		1100	$grp->result ($_[0]);
		1101
		1102	aioreq_pri $pri;
		1103	add $grp aio_close $fh;
		1104	};
		1105	} else {
		1106	$grp->result (-1);
		1107	}
		1108	};
		1109
		1110	$grp
		1111	}
		1112
		1113	=item aio_msync $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1114
		1115	This is a rather advanced IO::AIO call, which only works on mmap(2)ed
		1116	scalars (see the C<IO::AIO::mmap> function, although it also works on data
		1117	scalars managed by the L<Sys::Mmap> or L<Mmap> modules, note that the
		1118	scalar must only be modified in-place while an aio operation is pending on
		1119	it).
		1120
		1121	It calls the C<msync> function of your OS, if available, with the memory
		1122	area starting at C<$offset> in the string and ending C<$length> bytes
		1123	later. If C<$length> is negative, counts from the end, and if C<$length>
		1124	is C<undef>, then it goes till the end of the string. The flags can be
		1125	a combination of C<IO::AIO::MS_ASYNC>, C<IO::AIO::MS_INVALIDATE> and
		1126	C<IO::AIO::MS_SYNC>.
		1127
		1128	=item aio_mtouch $scalar, $offset = 0, $length = undef, flags = 0, $callback->($status)
		1129
		1130	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1131	scalars.
		1132
		1133	It touches (reads or writes) all memory pages in the specified
		1134	range inside the scalar. All caveats and parameters are the same
		1135	as for C<aio_msync>, above, except for flags, which must be either
		1136	C<0> (which reads all pages and ensures they are instantiated) or
		1137	C<IO::AIO::MT_MODIFY>, which modifies the memory page s(by reading and
		1138	writing an octet from it, which dirties the page).
		1139
		1140	=item aio_mlock $scalar, $offset = 0, $length = undef, $callback->($status)
		1141
		1142	This is a rather advanced IO::AIO call, which works best on mmap(2)ed
		1143	scalars.
		1144
		1145	It reads in all the pages of the underlying storage into memory (if any)
		1146	and locks them, so they are not getting swapped/paged out or removed.
		1147
		1148	If C<$length> is undefined, then the scalar will be locked till the end.
		1149
		1150	On systems that do not implement C<mlock>, this function returns C<-1>
		1151	and sets errno to C<ENOSYS>.
		1152
		1153	Note that the corresponding C<munlock> is synchronous and is
		1154	documented under L<MISCELLANEOUS FUNCTIONS>.
		1155
		1156	Example: open a file, mmap and mlock it - both will be undone when
		1157	C<$data> gets destroyed.
		1158
		1159	open my $fh, "<", $path or die "$path: $!";
		1160	my $data;
		1161	IO::AIO::mmap $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh;
		1162	aio_mlock $data; # mlock in background
		1163
		1164	=item aio_mlockall $flags, $callback->($status)
		1165
		1166	Calls the C<mlockall> function with the given C<$flags> (a combination of
		1167	C<IO::AIO::MCL_CURRENT> and C<IO::AIO::MCL_FUTURE>).
		1168
		1169	On systems that do not implement C<mlockall>, this function returns C<-1>
		1170	and sets errno to C<ENOSYS>.
		1171
		1172	Note that the corresponding C<munlockall> is synchronous and is
		1173	documented under L<MISCELLANEOUS FUNCTIONS>.
		1174
		1175	Example: asynchronously lock all current and future pages into memory.
		1176
		1177	aio_mlockall IO::AIO::MCL_FUTURE;
		1178
507	=item aio_group $callback->()	1179	=item aio_group $callback->(...)
508
509	[EXPERIMENTAL]
510		1180
511	This is a very special aio request: Instead of doing something, it is a	1181	This is a very special aio request: Instead of doing something, it is a
512	container for other aio requests, which is useful if you want to bundle	1182	container for other aio requests, which is useful if you want to bundle
513	many requests into a single, composite, request.	1183	many requests into a single, composite, request with a definite callback
		1184	and the ability to cancel the whole request with its subrequests.
514		1185
515	Returns an object of class L<IO::AIO::GRP>. See its documentation below	1186	Returns an object of class L<IO::AIO::GRP>. See its documentation below
516	for more info.	1187	for more info.
517		1188
518	Example:	1189	Example:
…		…
524	add $grp	1195	add $grp
525	(aio_stat ...),	1196	(aio_stat ...),
526	(aio_stat ...),	1197	(aio_stat ...),
527	...;	1198	...;
528		1199
		1200	=item aio_nop $callback->()
		1201
		1202	This is a special request - it does nothing in itself and is only used for
		1203	side effects, such as when you want to add a dummy request to a group so
		1204	that finishing the requests in the group depends on executing the given
		1205	code.
		1206
		1207	While this request does nothing, it still goes through the execution
		1208	phase and still requires a worker thread. Thus, the callback will not
		1209	be executed immediately but only after other requests in the queue have
		1210	entered their execution phase. This can be used to measure request
		1211	latency.
		1212
529	=item aio_sleep $fractional_seconds, $callback->() *NOT EXPORTED*	1213	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
530		1214
531	Mainly used for debugging and benchmarking, this aio request puts one of	1215	Mainly used for debugging and benchmarking, this aio request puts one of
532	the request workers to sleep for the given time.	1216	the request workers to sleep for the given time.
533		1217
		1218	While it is theoretically handy to have simple I/O scheduling requests
		1219	like sleep and file handle readable/writable, the overhead this creates is
		1220	immense (it blocks a thread for a long time) so do not use this function
		1221	except to put your application under artificial I/O pressure.
		1222
534	=back	1223	=back
535		1224
536	=head2 IO::AIO::REQ CLASS	1225	=head2 IO::AIO::REQ CLASS
537		1226
538	All non-aggregate C<aio_*> functions return an object of this class when	1227	All non-aggregate C<aio_*> functions return an object of this class when
539	called in non-void context.	1228	called in non-void context.
540		1229
541	A request always moves through the following five states in its lifetime,
542	in order: B<ready> (request has been created, but has not been executed
543	yet), B<execute> (request is currently being executed), B<pending>
544	(request has been executed but callback has not been called yet),
545	B<result> (results are being processed synchronously, includes calling the
546	callback) and B<done> (request has reached the end of its lifetime and
547	holds no resources anymore).
548
549	=over 4	1230	=over 4
550		1231
551	=item $req->cancel	1232	=item cancel $req
552		1233
553	Cancels the request, if possible. Has the effect of skipping execution	1234	Cancels the request, if possible. Has the effect of skipping execution
554	when entering the B<execute> state and skipping calling the callback when	1235	when entering the B<execute> state and skipping calling the callback when
555	entering the the B<result> state, but will leave the request otherwise	1236	entering the the B<result> state, but will leave the request otherwise
556	untouched. That means that requests that currently execute will not be	1237	untouched (with the exception of readdir). That means that requests that
557	stopped and resources held by the request will not be freed prematurely.	1238	currently execute will not be stopped and resources held by the request
		1239	will not be freed prematurely.
		1240
		1241	=item cb $req $callback->(...)
		1242
		1243	Replace (or simply set) the callback registered to the request.
558		1244
559	=back	1245	=back
560		1246
561	=head2 IO::AIO::GRP CLASS	1247	=head2 IO::AIO::GRP CLASS
562		1248
…		…
577	You add requests by calling the C<add> method with one or more	1263	You add requests by calling the C<add> method with one or more
578	C<IO::AIO::REQ> objects:	1264	C<IO::AIO::REQ> objects:
579		1265
580	$grp->add (aio_unlink "...");	1266	$grp->add (aio_unlink "...");
581		1267
582	add $grp aio_stat "...", sub { ... };	1268	add $grp aio_stat "...", sub {
		1269	$_[0] or return $grp->result ("error");
		1270
		1271	# add another request dynamically, if first succeeded
		1272	add $grp aio_open "...", sub {
		1273	$grp->result ("ok");
		1274	};
		1275	};
583		1276
584	This makes it very easy to create composite requests (see the source of	1277	This makes it very easy to create composite requests (see the source of
585	C<aio_move> for an application) that work and feel like simple requests.	1278	C<aio_move> for an application) that work and feel like simple requests.
586		1279
		1280	=over 4
		1281
587	The IO::AIO::GRP objects will be cleaned up during calls to	1282	=item * The IO::AIO::GRP objects will be cleaned up during calls to
588	C<IO::AIO::poll_cb>, just like any other request.	1283	C<IO::AIO::poll_cb>, just like any other request.
589		1284
590	They can be canceled like any other request. Canceling will cancel not	1285	=item * They can be canceled like any other request. Canceling will cancel not
591	just the request itself, but also all requests it contains.	1286	only the request itself, but also all requests it contains.
592		1287
593	They can also can also be added to other IO::AIO::GRP objects.	1288	=item * They can also can also be added to other IO::AIO::GRP objects.
		1289
		1290	=item * You must not add requests to a group from within the group callback (or
		1291	any later time).
		1292
		1293	=back
594		1294
595	Their lifetime, simplified, looks like this: when they are empty, they	1295	Their lifetime, simplified, looks like this: when they are empty, they
596	will finish very quickly. If they contain only requests that are in the	1296	will finish very quickly. If they contain only requests that are in the
597	C<done> state, they will also finish. Otherwise they will continue to	1297	C<done> state, they will also finish. Otherwise they will continue to
598	exist.	1298	exist.
599		1299
		1300	That means after creating a group you have some time to add requests
		1301	(precisely before the callback has been invoked, which is only done within
		1302	the C<poll_cb>). And in the callbacks of those requests, you can add
		1303	further requests to the group. And only when all those requests have
		1304	finished will the the group itself finish.
		1305
600	=over 4	1306	=over 4
601		1307
		1308	=item add $grp ...
		1309
602	=item $grp->add (...)	1310	=item $grp->add (...)
603		1311
604	=item add $grp ...	1312	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		1313	be added, including other groups, as long as you do not create circular
		1314	dependencies.
605		1315
606	Add one or more	1316	Returns all its arguments.
607	Cancels the request, if possible. Has the effect of skipping execution	1317
608	when entering the B<execute> state and skipping calling the callback when	1318	=item $grp->cancel_subs
609	entering the the B<result> state, but will leave the request otherwise	1319
610	untouched. That means that requests that currently execute will not be	1320	Cancel all subrequests and clears any feeder, but not the group request
611	stopped and resources held by the request will not be freed prematurely.	1321	itself. Useful when you queued a lot of events but got a result early.
		1322
		1323	The group request will finish normally (you cannot add requests to the
		1324	group).
		1325
		1326	=item $grp->result (...)
		1327
		1328	Set the result value(s) that will be passed to the group callback when all
		1329	subrequests have finished and set the groups errno to the current value
		1330	of errno (just like calling C<errno> without an error number). By default,
		1331	no argument will be passed and errno is zero.
		1332
		1333	=item $grp->errno ([$errno])
		1334
		1335	Sets the group errno value to C<$errno>, or the current value of errno
		1336	when the argument is missing.
		1337
		1338	Every aio request has an associated errno value that is restored when
		1339	the callback is invoked. This method lets you change this value from its
		1340	default (0).
		1341
		1342	Calling C<result> will also set errno, so make sure you either set C<$!>
		1343	before the call to C<result>, or call c<errno> after it.
		1344
		1345	=item feed $grp $callback->($grp)
		1346
		1347	Sets a feeder/generator on this group: every group can have an attached
		1348	generator that generates requests if idle. The idea behind this is that,
		1349	although you could just queue as many requests as you want in a group,
		1350	this might starve other requests for a potentially long time. For example,
		1351	C<aio_scandir> might generate hundreds of thousands C<aio_stat> requests,
		1352	delaying any later requests for a long time.
		1353
		1354	To avoid this, and allow incremental generation of requests, you can
		1355	instead a group and set a feeder on it that generates those requests. The
		1356	feed callback will be called whenever there are few enough (see C<limit>,
		1357	below) requests active in the group itself and is expected to queue more
		1358	requests.
		1359
		1360	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1361	not impose any limits).
		1362
		1363	If the feed does not queue more requests when called, it will be
		1364	automatically removed from the group.
		1365
		1366	If the feed limit is C<0> when this method is called, it will be set to
		1367	C<2> automatically.
		1368
		1369	Example:
		1370
		1371	# stat all files in @files, but only ever use four aio requests concurrently:
		1372
		1373	my $grp = aio_group sub { print "finished\n" };
		1374	limit $grp 4;
		1375	feed $grp sub {
		1376	my $file = pop @files
		1377	or return;
		1378
		1379	add $grp aio_stat $file, sub { ... };
		1380	};
		1381
		1382	=item limit $grp $num
		1383
		1384	Sets the feeder limit for the group: The feeder will be called whenever
		1385	the group contains less than this many requests.
		1386
		1387	Setting the limit to C<0> will pause the feeding process.
		1388
		1389	The default value for the limit is C<0>, but note that setting a feeder
		1390	automatically bumps it up to C<2>.
612		1391
613	=back	1392	=back
614		1393
615	=head2 SUPPORT FUNCTIONS	1394	=head2 SUPPORT FUNCTIONS
616		1395
		1396	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1397
617	=over 4	1398	=over 4
618		1399
619	=item $fileno = IO::AIO::poll_fileno	1400	=item $fileno = IO::AIO::poll_fileno
620		1401
621	Return the I<request result pipe file descriptor>. This filehandle must be	1402	Return the I<request result pipe file descriptor>. This filehandle must be
622	polled for reading by some mechanism outside this module (e.g. Event or	1403	polled for reading by some mechanism outside this module (e.g. EV, Glib,
623	select, see below or the SYNOPSIS). If the pipe becomes readable you have	1404	select and so on, see below or the SYNOPSIS). If the pipe becomes readable
624	to call C<poll_cb> to check the results.	1405	you have to call C<poll_cb> to check the results.
625		1406
626	See C<poll_cb> for an example.	1407	See C<poll_cb> for an example.
627		1408
628	=item IO::AIO::poll_cb	1409	=item IO::AIO::poll_cb
629		1410
630	Process all outstanding events on the result pipe. You have to call this	1411	Process some outstanding events on the result pipe. You have to call
631	regularly. Returns the number of events processed. Returns immediately	1412	this regularly. Returns C<0> if all events could be processed (or there
632	when no events are outstanding.	1413	were no events to process), or C<-1> if it returned earlier for whatever
		1414	reason. Returns immediately when no events are outstanding. The amount of
		1415	events processed depends on the settings of C<IO::AIO::max_poll_req> and
		1416	C<IO::AIO::max_poll_time>.
		1417
		1418	If not all requests were processed for whatever reason, the filehandle
		1419	will still be ready when C<poll_cb> returns, so normally you don't have to
		1420	do anything special to have it called later.
		1421
		1422	Apart from calling C<IO::AIO::poll_cb> when the event filehandle becomes
		1423	ready, it can be beneficial to call this function from loops which submit
		1424	a lot of requests, to make sure the results get processed when they become
		1425	available and not just when the loop is finished and the event loop takes
		1426	over again. This function returns very fast when there are no outstanding
		1427	requests.
633		1428
634	Example: Install an Event watcher that automatically calls	1429	Example: Install an Event watcher that automatically calls
635	IO::AIO::poll_cb with high priority:	1430	IO::AIO::poll_cb with high priority (more examples can be found in the
		1431	SYNOPSIS section, at the top of this document):
636		1432
637	Event->io (fd => IO::AIO::poll_fileno,	1433	Event->io (fd => IO::AIO::poll_fileno,
638	poll => 'r', async => 1,	1434	poll => 'r', async => 1,
639	cb => \&IO::AIO::poll_cb);	1435	cb => \&IO::AIO::poll_cb);
640		1436
641	=item IO::AIO::poll_wait	1437	=item IO::AIO::poll_wait
642		1438
		1439	If there are any outstanding requests and none of them in the result
643	Wait till the result filehandle becomes ready for reading (simply does a	1440	phase, wait till the result filehandle becomes ready for reading (simply
644	C<select> on the filehandle. This is useful if you want to synchronously wait	1441	does a C<select> on the filehandle. This is useful if you want to
645	for some requests to finish).	1442	synchronously wait for some requests to finish).
646		1443
647	See C<nreqs> for an example.	1444	See C<nreqs> for an example.
648		1445
		1446	=item IO::AIO::poll
		1447
		1448	Waits until some requests have been handled.
		1449
		1450	Returns the number of requests processed, but is otherwise strictly
		1451	equivalent to:
		1452
		1453	IO::AIO::poll_wait, IO::AIO::poll_cb
		1454
649	=item IO::AIO::nreqs	1455	=item IO::AIO::flush
650		1456
651	Returns the number of requests currently outstanding (i.e. for which their	1457	Wait till all outstanding AIO requests have been handled.
652	callback has not been invoked yet).
653		1458
654	Example: wait till there are no outstanding requests anymore:	1459	Strictly equivalent to:
655		1460
656	IO::AIO::poll_wait, IO::AIO::poll_cb	1461	IO::AIO::poll_wait, IO::AIO::poll_cb
657	while IO::AIO::nreqs;	1462	while IO::AIO::nreqs;
658		1463
659	=item IO::AIO::flush	1464	=item IO::AIO::max_poll_reqs $nreqs
660		1465
661	Wait till all outstanding AIO requests have been handled.	1466	=item IO::AIO::max_poll_time $seconds
662		1467
663	Strictly equivalent to:	1468	These set the maximum number of requests (default C<0>, meaning infinity)
		1469	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1470	the maximum amount of time (default C<0>, meaning infinity) spent in
		1471	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1472	of time C<poll_cb> is allowed to use).
664		1473
665	IO::AIO::poll_wait, IO::AIO::poll_cb	1474	Setting C<max_poll_time> to a non-zero value creates an overhead of one
666	while IO::AIO::nreqs;	1475	syscall per request processed, which is not normally a problem unless your
		1476	callbacks are really really fast or your OS is really really slow (I am
		1477	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
667		1478
668	=item IO::AIO::poll	1479	Setting these is useful if you want to ensure some level of
		1480	interactiveness when perl is not fast enough to process all requests in
		1481	time.
669		1482
670	Waits until some requests have been handled.	1483	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
671		1484
672	Strictly equivalent to:	1485	Example: Install an Event watcher that automatically calls
		1486	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1487	program get the CPU sometimes even under high AIO load.
673		1488
674	IO::AIO::poll_wait, IO::AIO::poll_cb	1489	# try not to spend much more than 0.1s in poll_cb
675	if IO::AIO::nreqs;	1490	IO::AIO::max_poll_time 0.1;
		1491
		1492	# use a low priority so other tasks have priority
		1493	Event->io (fd => IO::AIO::poll_fileno,
		1494	poll => 'r', nice => 1,
		1495	cb => &IO::AIO::poll_cb);
		1496
		1497	=back
		1498
		1499	=head3 CONTROLLING THE NUMBER OF THREADS
		1500
		1501	=over
676		1502
677	=item IO::AIO::min_parallel $nthreads	1503	=item IO::AIO::min_parallel $nthreads
678		1504
679	Set the minimum number of AIO threads to C<$nthreads>. The current default	1505	Set the minimum number of AIO threads to C<$nthreads>. The current
680	is C<4>, which means four asynchronous operations can be done at one time	1506	default is C<8>, which means eight asynchronous operations can execute
681	(the number of outstanding operations, however, is unlimited).	1507	concurrently at any one time (the number of outstanding requests,
		1508	however, is unlimited).
682		1509
683	IO::AIO starts threads only on demand, when an AIO request is queued and	1510	IO::AIO starts threads only on demand, when an AIO request is queued and
684	no free thread exists.	1511	no free thread exists. Please note that queueing up a hundred requests can
		1512	create demand for a hundred threads, even if it turns out that everything
		1513	is in the cache and could have been processed faster by a single thread.
685		1514
686	It is recommended to keep the number of threads low, as some Linux	1515	It is recommended to keep the number of threads relatively low, as some
687	kernel versions will scale negatively with the number of threads (higher	1516	Linux kernel versions will scale negatively with the number of threads
688	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1517	(higher parallelity => MUCH higher latency). With current Linux 2.6
689	threads should be fine.	1518	versions, 4-32 threads should be fine.
690		1519
691	Under most circumstances you don't need to call this function, as the	1520	Under most circumstances you don't need to call this function, as the
692	module selects a default that is suitable for low to moderate load.	1521	module selects a default that is suitable for low to moderate load.
693		1522
694	=item IO::AIO::max_parallel $nthreads	1523	=item IO::AIO::max_parallel $nthreads
…		…
703	This module automatically runs C<max_parallel 0> at program end, to ensure	1532	This module automatically runs C<max_parallel 0> at program end, to ensure
704	that all threads are killed and that there are no outstanding requests.	1533	that all threads are killed and that there are no outstanding requests.
705		1534
706	Under normal circumstances you don't need to call this function.	1535	Under normal circumstances you don't need to call this function.
707		1536
		1537	=item IO::AIO::max_idle $nthreads
		1538
		1539	Limit the number of threads (default: 4) that are allowed to idle
		1540	(i.e., threads that did not get a request to process within the idle
		1541	timeout (default: 10 seconds). That means if a thread becomes idle while
		1542	C<$nthreads> other threads are also idle, it will free its resources and
		1543	exit.
		1544
		1545	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1546	to allow for extremely high load situations, but want to free resources
		1547	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1548
		1549	The default is probably ok in most situations, especially if thread
		1550	creation is fast. If thread creation is very slow on your system you might
		1551	want to use larger values.
		1552
		1553	=item IO::AIO::idle_timeout $seconds
		1554
		1555	Sets the minimum idle timeout (default 10) after which worker threads are
		1556	allowed to exit. SEe C<IO::AIO::max_idle>.
		1557
708	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1558	=item IO::AIO::max_outstanding $maxreqs
		1559
		1560	This is a very bad function to use in interactive programs because it
		1561	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1562	use an C<aio_group> together with a feed callback.
709		1563
710	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1564	Sets the maximum number of outstanding requests to C<$nreqs>. If you
711	try to queue up more than this number of requests, the caller will block until	1565	do queue up more than this number of requests, the next call to the
712	some requests have been handled.	1566	C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
		1567	function will block until the limit is no longer exceeded.
713		1568
714	The default is very large, so normally there is no practical limit. If you	1569	The default value is very large, so there is no practical limit on the
715	queue up many requests in a loop it often improves speed if you set	1570	number of outstanding requests.
716	this to a relatively low number, such as C<100>.
717		1571
718	Under normal circumstances you don't need to call this function.	1572	You can still queue as many requests as you want. Therefore,
		1573	C<max_outstanding> is mainly useful in simple scripts (with low values) or
		1574	as a stop gap to shield against fatal memory overflow (with large values).
719		1575
720	=back	1576	=back
721		1577
		1578	=head3 STATISTICAL INFORMATION
		1579
		1580	=over
		1581
		1582	=item IO::AIO::nreqs
		1583
		1584	Returns the number of requests currently in the ready, execute or pending
		1585	states (i.e. for which their callback has not been invoked yet).
		1586
		1587	Example: wait till there are no outstanding requests anymore:
		1588
		1589	IO::AIO::poll_wait, IO::AIO::poll_cb
		1590	while IO::AIO::nreqs;
		1591
		1592	=item IO::AIO::nready
		1593
		1594	Returns the number of requests currently in the ready state (not yet
		1595	executed).
		1596
		1597	=item IO::AIO::npending
		1598
		1599	Returns the number of requests currently in the pending state (executed,
		1600	but not yet processed by poll_cb).
		1601
		1602	=back
		1603
		1604	=head3 MISCELLANEOUS FUNCTIONS
		1605
		1606	IO::AIO implements some functions that might be useful, but are not
		1607	asynchronous.
		1608
		1609	=over 4
		1610
		1611	=item IO::AIO::sendfile $ofh, $ifh, $offset, $count
		1612
		1613	Calls the C<eio_sendfile_sync> function, which is like C<aio_sendfile>,
		1614	but is blocking (this makes most sense if you know the input data is
		1615	likely cached already and the output filehandle is set to non-blocking
		1616	operations).
		1617
		1618	Returns the number of bytes copied, or C<-1> on error.
		1619
		1620	=item IO::AIO::fadvise $fh, $offset, $len, $advice
		1621
		1622	Simply calls the C<posix_fadvise> function (see its
		1623	manpage for details). The following advice constants are
		1624	avaiable: C<IO::AIO::FADV_NORMAL>, C<IO::AIO::FADV_SEQUENTIAL>,
		1625	C<IO::AIO::FADV_RANDOM>, C<IO::AIO::FADV_NOREUSE>,
		1626	C<IO::AIO::FADV_WILLNEED>, C<IO::AIO::FADV_DONTNEED>.
		1627
		1628	On systems that do not implement C<posix_fadvise>, this function returns
		1629	ENOSYS, otherwise the return value of C<posix_fadvise>.
		1630
		1631	=item IO::AIO::madvise $scalar, $offset, $len, $advice
		1632
		1633	Simply calls the C<posix_madvise> function (see its
		1634	manpage for details). The following advice constants are
		1635	avaiable: C<IO::AIO::MADV_NORMAL>, C<IO::AIO::MADV_SEQUENTIAL>,
		1636	C<IO::AIO::MADV_RANDOM>, C<IO::AIO::MADV_WILLNEED>, C<IO::AIO::MADV_DONTNEED>.
		1637
		1638	On systems that do not implement C<posix_madvise>, this function returns
		1639	ENOSYS, otherwise the return value of C<posix_madvise>.
		1640
		1641	=item IO::AIO::mprotect $scalar, $offset, $len, $protect
		1642
		1643	Simply calls the C<mprotect> function on the preferably AIO::mmap'ed
		1644	$scalar (see its manpage for details). The following protect
		1645	constants are avaiable: C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_READ>,
		1646	C<IO::AIO::PROT_WRITE>, C<IO::AIO::PROT_EXEC>.
		1647
		1648	On systems that do not implement C<mprotect>, this function returns
		1649	ENOSYS, otherwise the return value of C<mprotect>.
		1650
		1651	=item IO::AIO::mmap $scalar, $length, $prot, $flags, $fh[, $offset]
		1652
		1653	Memory-maps a file (or anonymous memory range) and attaches it to the
		1654	given C<$scalar>, which will act like a string scalar.
		1655
		1656	The only operations allowed on the scalar are C<substr>/C<vec> that don't
		1657	change the string length, and most read-only operations such as copying it
		1658	or searching it with regexes and so on.
		1659
		1660	Anything else is unsafe and will, at best, result in memory leaks.
		1661
		1662	The memory map associated with the C<$scalar> is automatically removed
		1663	when the C<$scalar> is destroyed, or when the C<IO::AIO::mmap> or
		1664	C<IO::AIO::munmap> functions are called.
		1665
		1666	This calls the C<mmap>(2) function internally. See your system's manual
		1667	page for details on the C<$length>, C<$prot> and C<$flags> parameters.
		1668
		1669	The C<$length> must be larger than zero and smaller than the actual
		1670	filesize.
		1671
		1672	C<$prot> is a combination of C<IO::AIO::PROT_NONE>, C<IO::AIO::PROT_EXEC>,
		1673	C<IO::AIO::PROT_READ> and/or C<IO::AIO::PROT_WRITE>,
		1674
		1675	C<$flags> can be a combination of C<IO::AIO::MAP_SHARED> or
		1676	C<IO::AIO::MAP_PRIVATE>, or a number of system-specific flags (when
		1677	not available, the are defined as 0): C<IO::AIO::MAP_ANONYMOUS>
		1678	(which is set to C<MAP_ANON> if your system only provides this
		1679	constant), C<IO::AIO::MAP_HUGETLB>, C<IO::AIO::MAP_LOCKED>,
		1680	C<IO::AIO::MAP_NORESERVE>, C<IO::AIO::MAP_POPULATE> or
		1681	C<IO::AIO::MAP_NONBLOCK>
		1682
		1683	If C<$fh> is C<undef>, then a file descriptor of C<-1> is passed.
		1684
		1685	C<$offset> is the offset from the start of the file - it generally must be
		1686	a multiple of C<IO::AIO::PAGESIZE> and defaults to C<0>.
		1687
		1688	Example:
		1689
		1690	use Digest::MD5;
		1691	use IO::AIO;
		1692
		1693	open my $fh, "<verybigfile"
		1694	or die "$!";
		1695
		1696	IO::AIO::mmap my $data, -s $fh, IO::AIO::PROT_READ, IO::AIO::MAP_SHARED, $fh
		1697	or die "verybigfile: $!";
		1698
		1699	my $fast_md5 = md5 $data;
		1700
		1701	=item IO::AIO::munmap $scalar
		1702
		1703	Removes a previous mmap and undefines the C<$scalar>.
		1704
		1705	=item IO::AIO::munlock $scalar, $offset = 0, $length = undef
		1706
		1707	Calls the C<munlock> function, undoing the effects of a previous
		1708	C<aio_mlock> call (see its description for details).
		1709
		1710	=item IO::AIO::munlockall
		1711
		1712	Calls the C<munlockall> function.
		1713
		1714	On systems that do not implement C<munlockall>, this function returns
		1715	ENOSYS, otherwise the return value of C<munlockall>.
		1716
		1717	=back
		1718
722	=cut	1719	=cut
723		1720
724	# support function to convert a fd into a perl filehandle
725	sub _fd2fh {
726	return undef if $_[0] < 0;
727
728	# try to generate nice filehandles
729	my $sym = "IO::AIO::fd#$_[0]";
730	local *$sym;
731
732	open *$sym, "+<&=$_[0]" # usually works under any unix
733	or open *$sym, "<&=$_[0]" # cygwin needs this
734	or open *$sym, ">&=$_[0]" # or this
735	or return undef;
736
737	*$sym
738	}
739
740	min_parallel 4;	1721	min_parallel 8;
741		1722
742	END {	1723	END { flush }
743	max_parallel 0;
744	}
745		1724
746	1;	1725	1;
		1726
		1727	=head1 EVENT LOOP INTEGRATION
		1728
		1729	It is recommended to use L<AnyEvent::AIO> to integrate IO::AIO
		1730	automatically into many event loops:
		1731
		1732	# AnyEvent integration (EV, Event, Glib, Tk, POE, urxvt, pureperl...)
		1733	use AnyEvent::AIO;
		1734
		1735	You can also integrate IO::AIO manually into many event loops, here are
		1736	some examples of how to do this:
		1737
		1738	# EV integration
		1739	my $aio_w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb;
		1740
		1741	# Event integration
		1742	Event->io (fd => IO::AIO::poll_fileno,
		1743	poll => 'r',
		1744	cb => \&IO::AIO::poll_cb);
		1745
		1746	# Glib/Gtk2 integration
		1747	add_watch Glib::IO IO::AIO::poll_fileno,
		1748	in => sub { IO::AIO::poll_cb; 1 };
		1749
		1750	# Tk integration
		1751	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
		1752	readable => \&IO::AIO::poll_cb);
		1753
		1754	# Danga::Socket integration
		1755	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		1756	\&IO::AIO::poll_cb);
747		1757
748	=head2 FORK BEHAVIOUR	1758	=head2 FORK BEHAVIOUR
749		1759
750	This module should do "the right thing" when the process using it forks:	1760	This module should do "the right thing" when the process using it forks:
751		1761
752	Before the fork, IO::AIO enters a quiescent state where no requests	1762	Before the fork, IO::AIO enters a quiescent state where no requests
753	can be added in other threads and no results will be processed. After	1763	can be added in other threads and no results will be processed. After
754	the fork the parent simply leaves the quiescent state and continues	1764	the fork the parent simply leaves the quiescent state and continues
755	request/result processing, while the child clears the request/result	1765	request/result processing, while the child frees the request/result queue
756	queue (so the requests started before the fork will only be handled in	1766	(so that the requests started before the fork will only be handled in the
757	the parent). Threads will be started on demand until the limit ste in the	1767	parent). Threads will be started on demand until the limit set in the
758	parent process has been reached again.	1768	parent process has been reached again.
759		1769
760	In short: the parent will, after a short pause, continue as if fork had	1770	In short: the parent will, after a short pause, continue as if fork had
761	not been called, while the child will act as if IO::AIO has not been used	1771	not been called, while the child will act as if IO::AIO has not been used
762	yet.	1772	yet.
763		1773
		1774	=head2 MEMORY USAGE
		1775
		1776	Per-request usage:
		1777
		1778	Each aio request uses - depending on your architecture - around 100-200
		1779	bytes of memory. In addition, stat requests need a stat buffer (possibly
		1780	a few hundred bytes), readdir requires a result buffer and so on. Perl
		1781	scalars and other data passed into aio requests will also be locked and
		1782	will consume memory till the request has entered the done state.
		1783
		1784	This is not awfully much, so queuing lots of requests is not usually a
		1785	problem.
		1786
		1787	Per-thread usage:
		1788
		1789	In the execution phase, some aio requests require more memory for
		1790	temporary buffers, and each thread requires a stack and other data
		1791	structures (usually around 16k-128k, depending on the OS).
		1792
		1793	=head1 KNOWN BUGS
		1794
		1795	Known bugs will be fixed in the next release.
		1796
764	=head1 SEE ALSO	1797	=head1 SEE ALSO
765		1798
766	L<Coro>, L<Linux::AIO> (obsolete).	1799	L<AnyEvent::AIO> for easy integration into event loops, L<Coro::AIO> for a
		1800	more natural syntax.
767		1801
768	=head1 AUTHOR	1802	=head1 AUTHOR
769		1803
770	Marc Lehmann <schmorp@schmorp.de>	1804	Marc Lehmann <schmorp@schmorp.de>
771	http://home.schmorp.de/	1805	http://home.schmorp.de/

Diff Legend

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