[ViewVC] Diff of: cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.30 by root, Wed Aug 17 04:47:38 2005 UTC vs.
Revision 1.111 by root, Mon Aug 13 12:08:13 2007 UTC

…		…
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", O_RDONLY, 0, sub {
10	my ($fh) = @_;	10	my $fh = shift
		11	or die "/etc/passwd: $!";
11	...	12	...
12	};	13	};
13		14
14	aio_unlink "/tmp/file", sub { };	15	aio_unlink "/tmp/file", sub { };
15		16
16	aio_read $fh, 30000, 1024, $buffer, 0, sub {	17	aio_read $fh, 30000, 1024, $buffer, 0, sub {
17	$_[0] > 0 or die "read error: $!";	18	$_[0] > 0 or die "read error: $!";
18	};	19	};
19		20
20	# Event	21	# version 2+ has request and group objects
		22	use IO::AIO 2;
		23
		24	aioreq_pri 4; # give next request a very high priority
		25	my $req = aio_unlink "/tmp/file", sub { };
		26	$req->cancel; # cancel request if still in queue
		27
		28	my $grp = aio_group sub { print "all stats done\n" };
		29	add $grp aio_stat "..." for ...;
		30
		31	# AnyEvent integration
		32	open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
		33	my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
		34
		35	# Event integration
21	Event->io (fd => IO::AIO::poll_fileno,	36	Event->io (fd => IO::AIO::poll_fileno,
22	poll => 'r',	37	poll => 'r',
23	cb => \&IO::AIO::poll_cb);	38	cb => \&IO::AIO::poll_cb);
24		39
25	# Glib/Gtk2	40	# Glib/Gtk2 integration
26	add_watch Glib::IO IO::AIO::poll_fileno,	41	add_watch Glib::IO IO::AIO::poll_fileno,
27	in => sub { IO::AIO::poll_cb; 1 };	42	in => sub { IO::AIO::poll_cb; 1 };
28		43
29	# Tk	44	# Tk integration
30	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",	45	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
31	readable => \&IO::AIO::poll_cb);	46	readable => \&IO::AIO::poll_cb);
32		47
33	# Danga::Socket	48	# Danga::Socket integration
34	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>	49	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
35	\&IO::AIO::poll_cb);	50	\&IO::AIO::poll_cb);
36		51
37
38	=head1 DESCRIPTION	52	=head1 DESCRIPTION
39		53
40	This module implements asynchronous I/O using whatever means your	54	This module implements asynchronous I/O using whatever means your
41	operating system supports.	55	operating system supports.
42		56
		57	Asynchronous means that operations that can normally block your program
		58	(e.g. reading from disk) will be done asynchronously: the operation
		59	will still block, but you can do something else in the meantime. This
		60	is extremely useful for programs that need to stay interactive even
		61	when doing heavy I/O (GUI programs, high performance network servers
		62	etc.), but can also be used to easily do operations in parallel that are
		63	normally done sequentially, e.g. stat'ing many files, which is much faster
		64	on a RAID volume or over NFS when you do a number of stat operations
		65	concurrently.
		66
		67	While most of this works on all types of file descriptors (for
		68	example sockets), using these functions on file descriptors that
		69	support nonblocking operation (again, sockets, pipes etc.) is very
		70	inefficient. Use an event loop for that (such as the L<Event\|Event>
		71	module): IO::AIO will naturally fit into such an event loop itself.
		72
43	Currently, a number of threads are started that execute your read/writes	73	In this version, a number of threads are started that execute your
44	and signal their completion. You don't need thread support in your libc or	74	requests and signal their completion. You don't need thread support
45	perl, and the threads created by this module will not be visible to the	75	in perl, and the threads created by this module will not be visible
46	pthreads library. In the future, this module might make use of the native	76	to perl. In the future, this module might make use of the native aio
47	aio functions available on many operating systems. However, they are often	77	functions available on many operating systems. However, they are often
48	not well-supported (Linux doesn't allow them on normal files currently,	78	not well-supported or restricted (GNU/Linux doesn't allow them on normal
49	for example), and they would only support aio_read and aio_write, so the	79	files currently, for example), and they would only support aio_read and
50	remaining functionality would have to be implemented using threads anyway.	80	aio_write, so the remaining functionality would have to be implemented
		81	using threads anyway.
51		82
52	Although the module will work with in the presence of other threads, it is	83	Although the module will work in the presence of other (Perl-) threads,
53	currently not reentrant, so use appropriate locking yourself, always call	84	it is currently not reentrant in any way, so use appropriate locking
54	C<poll_cb> from within the same thread, or never call C<poll_cb> (or other	85	yourself, always call C<poll_cb> from within the same thread, or never
55	C<aio_> functions) recursively.	86	call C<poll_cb> (or other C<aio_> functions) recursively.
		87
		88	=head2 EXAMPLE
		89
		90	This is a simple example that uses the Event module and loads
		91	F</etc/passwd> asynchronously:
		92
		93	use Fcntl;
		94	use Event;
		95	use IO::AIO;
		96
		97	# register the IO::AIO callback with Event
		98	Event->io (fd => IO::AIO::poll_fileno,
		99	poll => 'r',
		100	cb => \&IO::AIO::poll_cb);
		101
		102	# queue the request to open /etc/passwd
		103	aio_open "/etc/passwd", O_RDONLY, 0, sub {
		104	my $fh = shift
		105	or die "error while opening: $!";
		106
		107	# stat'ing filehandles is generally non-blocking
		108	my $size = -s $fh;
		109
		110	# queue a request to read the file
		111	my $contents;
		112	aio_read $fh, 0, $size, $contents, 0, sub {
		113	$_[0] == $size
		114	or die "short read: $!";
		115
		116	close $fh;
		117
		118	# file contents now in $contents
		119	print $contents;
		120
		121	# exit event loop and program
		122	Event::unloop;
		123	};
		124	};
		125
		126	# possibly queue up other requests, or open GUI windows,
		127	# check for sockets etc. etc.
		128
		129	# process events as long as there are some:
		130	Event::loop;
		131
		132	=head1 REQUEST ANATOMY AND LIFETIME
		133
		134	Every C<aio_*> function creates a request. which is a C data structure not
		135	directly visible to Perl.
		136
		137	If called in non-void context, every request function returns a Perl
		138	object representing the request. In void context, nothing is returned,
		139	which saves a bit of memory.
		140
		141	The perl object is a fairly standard ref-to-hash object. The hash contents
		142	are not used by IO::AIO so you are free to store anything you like in it.
		143
		144	During their existance, aio requests travel through the following states,
		145	in order:
		146
		147	=over 4
		148
		149	=item ready
		150
		151	Immediately after a request is created it is put into the ready state,
		152	waiting for a thread to execute it.
		153
		154	=item execute
		155
		156	A thread has accepted the request for processing and is currently
		157	executing it (e.g. blocking in read).
		158
		159	=item pending
		160
		161	The request has been executed and is waiting for result processing.
		162
		163	While request submission and execution is fully asynchronous, result
		164	processing is not and relies on the perl interpreter calling C<poll_cb>
		165	(or another function with the same effect).
		166
		167	=item result
		168
		169	The request results are processed synchronously by C<poll_cb>.
		170
		171	The C<poll_cb> function will process all outstanding aio requests by
		172	calling their callbacks, freeing memory associated with them and managing
		173	any groups they are contained in.
		174
		175	=item done
		176
		177	Request has reached the end of its lifetime and holds no resources anymore
		178	(except possibly for the Perl object, but its connection to the actual
		179	aio request is severed and calling its methods will either do nothing or
		180	result in a runtime error).
		181
		182	=back
56		183
57	=cut	184	=cut
58		185
59	package IO::AIO;	186	package IO::AIO;
60		187
61	no warnings;	188	no warnings;
		189	use strict 'vars';
62		190
63	use base 'Exporter';	191	use base 'Exporter';
64		192
65	use Fcntl ();
66
67	BEGIN {	193	BEGIN {
68	$VERSION = 1.2;	194	our $VERSION = '2.4';
69		195
70	@EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink	196	our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat
		197	aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink
71	aio_rmdir aio_symlink aio_fsync aio_fdatasync aio_readahead);	198	aio_readlink aio_fsync aio_fdatasync aio_readahead aio_rename aio_link
72	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);	199	aio_move aio_copy aio_group aio_nop aio_mknod aio_load aio_rmtree aio_mkdir
		200	aio_chown aio_chmod aio_utime aio_truncate);
		201	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice aio_block));
		202	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		203	min_parallel max_parallel max_idle
		204	nreqs nready npending nthreads
		205	max_poll_time max_poll_reqs);
		206
		207	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
73		208
74	require XSLoader;	209	require XSLoader;
75	XSLoader::load IO::AIO, $VERSION;	210	XSLoader::load ("IO::AIO", $VERSION);
76	}	211	}
77		212
78	=head1 FUNCTIONS	213	=head1 FUNCTIONS
79		214
80	=head2 AIO FUNCTIONS	215	=head2 AIO REQUEST FUNCTIONS
81		216
82	All the C<aio_*> calls are more or less thin wrappers around the syscall	217	All the C<aio_*> calls are more or less thin wrappers around the syscall
83	with the same name (sans C<aio_>). The arguments are similar or identical,	218	with the same name (sans C<aio_>). The arguments are similar or identical,
84	and they all accept an additional (and optional) C<$callback> argument	219	and they all accept an additional (and optional) C<$callback> argument
85	which must be a code reference. This code reference will get called with	220	which must be a code reference. This code reference will get called with
…		…
88	syscall has been executed asynchronously.	223	syscall has been executed asynchronously.
89		224
90	All functions expecting a filehandle keep a copy of the filehandle	225	All functions expecting a filehandle keep a copy of the filehandle
91	internally until the request has finished.	226	internally until the request has finished.
92		227
		228	All functions return request objects of type L<IO::AIO::REQ> that allow
		229	further manipulation of those requests while they are in-flight.
		230
93	The pathnames you pass to these routines I<must> be absolute and	231	The pathnames you pass to these routines I<must> be absolute and
94	encoded in byte form. The reason for the former is that at the time the	232	encoded as octets. The reason for the former is that at the time the
95	request is being executed, the current working directory could have	233	request is being executed, the current working directory could have
96	changed. Alternatively, you can make sure that you never change the	234	changed. Alternatively, you can make sure that you never change the
97	current working directory.	235	current working directory anywhere in the program and then use relative
		236	paths.
98		237
99	To encode pathnames to byte form, either make sure you either: a)	238	To encode pathnames as octets, either make sure you either: a) always pass
100	always pass in filenames you got from outside (command line, readdir	239	in filenames you got from outside (command line, readdir etc.) without
101	etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode	240	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
102	your pathnames to the locale (or other) encoding in effect in the user	241	your pathnames to the locale (or other) encoding in effect in the user
103	environment, d) use Glib::filename_from_unicode on unicode filenames or e)	242	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
104	use something else.	243	use something else to ensure your scalar has the correct contents.
		244
		245	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		246	handles correctly wether it is set or not.
105		247
106	=over 4	248	=over 4
107		249
		250	=item $prev_pri = aioreq_pri [$pri]
		251
		252	Returns the priority value that would be used for the next request and, if
		253	C<$pri> is given, sets the priority for the next aio request.
		254
		255	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		256	and C<4>, respectively. Requests with higher priority will be serviced
		257	first.
		258
		259	The priority will be reset to C<0> after each call to one of the C<aio_*>
		260	functions.
		261
		262	Example: open a file with low priority, then read something from it with
		263	higher priority so the read request is serviced before other low priority
		264	open requests (potentially spamming the cache):
		265
		266	aioreq_pri -3;
		267	aio_open ..., sub {
		268	return unless $_[0];
		269
		270	aioreq_pri -2;
		271	aio_read $_[0], ..., sub {
		272	...
		273	};
		274	};
		275
		276
		277	=item aioreq_nice $pri_adjust
		278
		279	Similar to C<aioreq_pri>, but subtracts the given value from the current
		280	priority, so the effect is cumulative.
		281
		282
108	=item aio_open $pathname, $flags, $mode, $callback	283	=item aio_open $pathname, $flags, $mode, $callback->($fh)
109		284
110	Asynchronously open or create a file and call the callback with a newly	285	Asynchronously open or create a file and call the callback with a newly
111	created filehandle for the file.	286	created filehandle for the file.
112		287
113	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	288	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
…		…
117	list. They are the same as used by C<sysopen>.	292	list. They are the same as used by C<sysopen>.
118		293
119	Likewise, C<$mode> specifies the mode of the newly created file, if it	294	Likewise, C<$mode> specifies the mode of the newly created file, if it
120	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,	295	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
121	except that it is mandatory (i.e. use C<0> if you don't create new files,	296	except that it is mandatory (i.e. use C<0> if you don't create new files,
122	and C<0666> or C<0777> if you do).	297	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		298	by the umask in effect then the request is being executed, so better never
		299	change the umask.
123		300
124	Example:	301	Example:
125		302
126	aio_open "/etc/passwd", O_RDONLY, 0, sub {	303	aio_open "/etc/passwd", O_RDONLY, 0, sub {
127	if ($_[0]) {	304	if ($_[0]) {
…		…
130	} else {	307	} else {
131	die "open failed: $!\n";	308	die "open failed: $!\n";
132	}	309	}
133	};	310	};
134		311
		312
135	=item aio_close $fh, $callback	313	=item aio_close $fh, $callback->($status)
136		314
137	Asynchronously close a file and call the callback with the result	315	Asynchronously close a file and call the callback with the result
138	code. I<WARNING:> although accepted, you should not pass in a perl	316	code. I<WARNING:> although accepted, you should not pass in a perl
139	filehandle here, as perl will likely close the file descriptor another	317	filehandle here, as perl will likely close the file descriptor another
140	time when the filehandle is destroyed. Normally, you can safely call perls	318	time when the filehandle is destroyed. Normally, you can safely call perls
141	C<close> or just let filehandles go out of scope.	319	C<close> or just let filehandles go out of scope.
142		320
143	This is supposed to be a bug in the API, so that might change. It's	321	This is supposed to be a bug in the API, so that might change. It's
144	therefore best to avoid this function.	322	therefore best to avoid this function.
145		323
		324
146	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	325	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
147		326
148	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	327	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
149		328
150	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	329	Reads or writes C<$length> bytes from the specified C<$fh> and C<$offset>
151	into the scalar given by C<data> and offset C<dataoffset> and calls the	330	into the scalar given by C<$data> and offset C<$dataoffset> and calls the
152	callback without the actual number of bytes read (or -1 on error, just	331	callback without the actual number of bytes read (or -1 on error, just
153	like the syscall).	332	like the syscall).
		333
		334	If C<$offset> is undefined, then the current file offset will be used (and
		335	updated), otherwise the file offset will not be changed by these calls.
		336
		337	If C<$length> is undefined in C<aio_write>, use the remaining length of C<$data>.
		338
		339	If C<$dataoffset> is less than zero, it will be counted from the end of
		340	C<$data>.
		341
		342	The C<$data> scalar I<MUST NOT> be modified in any way while the request
		343	is outstanding. Modifying it can result in segfaults or World War III (if
		344	the necessary/optional hardware is installed).
154		345
155	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	346	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
156	offset C<0> within the scalar:	347	offset C<0> within the scalar:
157		348
158	aio_read $fh, 7, 15, $buffer, 0, sub {	349	aio_read $fh, 7, 15, $buffer, 0, sub {
159	$_[0] > 0 or die "read error: $!";	350	$_[0] > 0 or die "read error: $!";
160	print "read $_[0] bytes: <$buffer>\n";	351	print "read $_[0] bytes: <$buffer>\n";
161	};	352	};
162		353
		354
		355	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		356
		357	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
		358	reading at byte offset C<$in_offset>, and starts writing at the current
		359	file offset of C<$out_fh>. Because of that, it is not safe to issue more
		360	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
		361	other.
		362
		363	This call tries to make use of a native C<sendfile> syscall to provide
		364	zero-copy operation. For this to work, C<$out_fh> should refer to a
		365	socket, and C<$in_fh> should refer to mmap'able file.
		366
		367	If the native sendfile call fails or is not implemented, it will be
		368	emulated, so you can call C<aio_sendfile> on any type of filehandle
		369	regardless of the limitations of the operating system.
		370
		371	Please note, however, that C<aio_sendfile> can read more bytes from
		372	C<$in_fh> than are written, and there is no way to find out how many
		373	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
		374	provides the number of bytes written to C<$out_fh>. Only if the result
		375	value equals C<$length> one can assume that C<$length> bytes have been
		376	read.
		377
		378
163	=item aio_readahead $fh,$offset,$length, $callback	379	=item aio_readahead $fh,$offset,$length, $callback->($retval)
164		380
165	C<aio_readahead> populates the page cache with data from a file so that	381	C<aio_readahead> populates the page cache with data from a file so that
166	subsequent reads from that file will not block on disk I/O. The C<$offset>	382	subsequent reads from that file will not block on disk I/O. The C<$offset>
167	argument specifies the starting point from which data is to be read and	383	argument specifies the starting point from which data is to be read and
168	C<$length> specifies the number of bytes to be read. I/O is performed in	384	C<$length> specifies the number of bytes to be read. I/O is performed in
…		…
172	file. The current file offset of the file is left unchanged.	388	file. The current file offset of the file is left unchanged.
173		389
174	If that syscall doesn't exist (likely if your OS isn't Linux) it will be	390	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
175	emulated by simply reading the data, which would have a similar effect.	391	emulated by simply reading the data, which would have a similar effect.
176		392
		393
177	=item aio_stat $fh_or_path, $callback	394	=item aio_stat $fh_or_path, $callback->($status)
178		395
179	=item aio_lstat $fh, $callback	396	=item aio_lstat $fh, $callback->($status)
180		397
181	Works like perl's C<stat> or C<lstat> in void context. The callback will	398	Works like perl's C<stat> or C<lstat> in void context. The callback will
182	be called after the stat and the results will be available using C<stat _>	399	be called after the stat and the results will be available using C<stat _>
183	or C<-s _> etc...	400	or C<-s _> etc...
184		401
…		…
194	aio_stat "/etc/passwd", sub {	411	aio_stat "/etc/passwd", sub {
195	$_[0] and die "stat failed: $!";	412	$_[0] and die "stat failed: $!";
196	print "size is ", -s _, "\n";	413	print "size is ", -s _, "\n";
197	};	414	};
198		415
		416
		417	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		418
		419	Works like perl's C<utime> function (including the special case of $atime
		420	and $mtime being undef). Fractional times are supported if the underlying
		421	syscalls support them.
		422
		423	When called with a pathname, uses utimes(2) if available, otherwise
		424	utime(2). If called on a file descriptor, uses futimes(2) if available,
		425	otherwise returns ENOSYS, so this is not portable.
		426
		427	Examples:
		428
		429	# set atime and mtime to current time (basically touch(1)):
		430	aio_utime "path", undef, undef;
		431	# set atime to current time and mtime to beginning of the epoch:
		432	aio_utime "path", time, undef; # undef==0
		433
		434
		435	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		436
		437	Works like perl's C<chown> function, except that C<undef> for either $uid
		438	or $gid is being interpreted as "do not change" (but -1 can also be used).
		439
		440	Examples:
		441
		442	# same as "chown root path" in the shell:
		443	aio_chown "path", 0, -1;
		444	# same as above:
		445	aio_chown "path", 0, undef;
		446
		447
		448	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		449
		450	Works like truncate(2) or ftruncate(2).
		451
		452
		453	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		454
		455	Works like perl's C<chmod> function.
		456
		457
199	=item aio_unlink $pathname, $callback	458	=item aio_unlink $pathname, $callback->($status)
200		459
201	Asynchronously unlink (delete) a file and call the callback with the	460	Asynchronously unlink (delete) a file and call the callback with the
202	result code.	461	result code.
203		462
		463
		464	=item aio_mknod $path, $mode, $dev, $callback->($status)
		465
		466	[EXPERIMENTAL]
		467
		468	Asynchronously create a device node (or fifo). See mknod(2).
		469
		470	The only (POSIX-) portable way of calling this function is:
		471
		472	aio_mknod $path, IO::AIO::S_IFIFO \| $mode, 0, sub { ...
		473
		474
		475	=item aio_link $srcpath, $dstpath, $callback->($status)
		476
		477	Asynchronously create a new link to the existing object at C<$srcpath> at
		478	the path C<$dstpath> and call the callback with the result code.
		479
		480
		481	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		482
		483	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		484	the path C<$dstpath> and call the callback with the result code.
		485
		486
		487	=item aio_readlink $path, $callback->($link)
		488
		489	Asynchronously read the symlink specified by C<$path> and pass it to
		490	the callback. If an error occurs, nothing or undef gets passed to the
		491	callback.
		492
		493
		494	=item aio_rename $srcpath, $dstpath, $callback->($status)
		495
		496	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		497	rename(2) and call the callback with the result code.
		498
		499
		500	=item aio_mkdir $pathname, $mode, $callback->($status)
		501
		502	Asynchronously mkdir (create) a directory and call the callback with
		503	the result code. C<$mode> will be modified by the umask at the time the
		504	request is executed, so do not change your umask.
		505
		506
204	=item aio_rmdir $pathname, $callback	507	=item aio_rmdir $pathname, $callback->($status)
205		508
206	Asynchronously rmdir (delete) a directory and call the callback with the	509	Asynchronously rmdir (delete) a directory and call the callback with the
207	result code.	510	result code.
208		511
		512
		513	=item aio_readdir $pathname, $callback->($entries)
		514
		515	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
		516	directory (i.e. opendir + readdir + closedir). The entries will not be
		517	sorted, and will B<NOT> include the C<.> and C<..> entries.
		518
		519	The callback a single argument which is either C<undef> or an array-ref
		520	with the filenames.
		521
		522
		523	=item aio_load $path, $data, $callback->($status)
		524
		525	This is a composite request that tries to fully load the given file into
		526	memory. Status is the same as with aio_read.
		527
		528	=cut
		529
		530	sub aio_load($$;$) {
		531	aio_block {
		532	my ($path, undef, $cb) = @_;
		533	my $data = \$_[1];
		534
		535	my $pri = aioreq_pri;
		536	my $grp = aio_group $cb;
		537
		538	aioreq_pri $pri;
		539	add $grp aio_open $path, O_RDONLY, 0, sub {
		540	my $fh = shift
		541	or return $grp->result (-1);
		542
		543	aioreq_pri $pri;
		544	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		545	$grp->result ($_[0]);
		546	};
		547	};
		548
		549	$grp
		550	}
		551	}
		552
		553	=item aio_copy $srcpath, $dstpath, $callback->($status)
		554
		555	Try to copy the I<file> (directories not supported as either source or
		556	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		557	the C<0> (error) or C<-1> ok.
		558
		559	This is a composite request that it creates the destination file with
		560	mode 0200 and copies the contents of the source file into it using
		561	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		562	uid/gid, in that order.
		563
		564	If an error occurs, the partial destination file will be unlinked, if
		565	possible, except when setting atime, mtime, access mode and uid/gid, where
		566	errors are being ignored.
		567
		568	=cut
		569
		570	sub aio_copy($$;$) {
		571	aio_block {
		572	my ($src, $dst, $cb) = @_;
		573
		574	my $pri = aioreq_pri;
		575	my $grp = aio_group $cb;
		576
		577	aioreq_pri $pri;
		578	add $grp aio_open $src, O_RDONLY, 0, sub {
		579	if (my $src_fh = $_[0]) {
		580	my @stat = stat $src_fh;
		581
		582	aioreq_pri $pri;
		583	add $grp aio_open $dst, O_CREAT \| O_WRONLY \| O_TRUNC, 0200, sub {
		584	if (my $dst_fh = $_[0]) {
		585	aioreq_pri $pri;
		586	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		587	if ($_[0] == $stat[7]) {
		588	$grp->result (0);
		589	close $src_fh;
		590
		591	# those should not normally block. should. should.
		592	utime $stat[8], $stat[9], $dst;
		593	chmod $stat[2] & 07777, $dst_fh;
		594	chown $stat[4], $stat[5], $dst_fh;
		595	close $dst_fh;
		596	} else {
		597	$grp->result (-1);
		598	close $src_fh;
		599	close $dst_fh;
		600
		601	aioreq $pri;
		602	add $grp aio_unlink $dst;
		603	}
		604	};
		605	} else {
		606	$grp->result (-1);
		607	}
		608	},
		609
		610	} else {
		611	$grp->result (-1);
		612	}
		613	};
		614
		615	$grp
		616	}
		617	}
		618
		619	=item aio_move $srcpath, $dstpath, $callback->($status)
		620
		621	Try to move the I<file> (directories not supported as either source or
		622	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		623	the C<0> (error) or C<-1> ok.
		624
		625	This is a composite request that tries to rename(2) the file first. If
		626	rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if
		627	that is successful, unlinking the C<$srcpath>.
		628
		629	=cut
		630
		631	sub aio_move($$;$) {
		632	aio_block {
		633	my ($src, $dst, $cb) = @_;
		634
		635	my $pri = aioreq_pri;
		636	my $grp = aio_group $cb;
		637
		638	aioreq_pri $pri;
		639	add $grp aio_rename $src, $dst, sub {
		640	if ($_[0] && $! == EXDEV) {
		641	aioreq_pri $pri;
		642	add $grp aio_copy $src, $dst, sub {
		643	$grp->result ($_[0]);
		644
		645	if (!$_[0]) {
		646	aioreq_pri $pri;
		647	add $grp aio_unlink $src;
		648	}
		649	};
		650	} else {
		651	$grp->result ($_[0]);
		652	}
		653	};
		654
		655	$grp
		656	}
		657	}
		658
		659	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
		660
		661	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		662	efficiently separate the entries of directory C<$path> into two sets of
		663	names, directories you can recurse into (directories), and ones you cannot
		664	recurse into (everything else, including symlinks to directories).
		665
		666	C<aio_scandir> is a composite request that creates of many sub requests_
		667	C<$maxreq> specifies the maximum number of outstanding aio requests that
		668	this function generates. If it is C<< <= 0 >>, then a suitable default
		669	will be chosen (currently 4).
		670
		671	On error, the callback is called without arguments, otherwise it receives
		672	two array-refs with path-relative entry names.
		673
		674	Example:
		675
		676	aio_scandir $dir, 0, sub {
		677	my ($dirs, $nondirs) = @_;
		678	print "real directories: @$dirs\n";
		679	print "everything else: @$nondirs\n";
		680	};
		681
		682	Implementation notes.
		683
		684	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		685
		686	After reading the directory, the modification time, size etc. of the
		687	directory before and after the readdir is checked, and if they match (and
		688	isn't the current time), the link count will be used to decide how many
		689	entries are directories (if >= 2). Otherwise, no knowledge of the number
		690	of subdirectories will be assumed.
		691
		692	Then entries will be sorted into likely directories (everything without
		693	a non-initial dot currently) and likely non-directories (everything
		694	else). Then every entry plus an appended C</.> will be C<stat>'ed,
		695	likely directories first. If that succeeds, it assumes that the entry
		696	is a directory or a symlink to directory (which will be checked
		697	seperately). This is often faster than stat'ing the entry itself because
		698	filesystems might detect the type of the entry without reading the inode
		699	data (e.g. ext2fs filetype feature).
		700
		701	If the known number of directories (link count - 2) has been reached, the
		702	rest of the entries is assumed to be non-directories.
		703
		704	This only works with certainty on POSIX (= UNIX) filesystems, which
		705	fortunately are the vast majority of filesystems around.
		706
		707	It will also likely work on non-POSIX filesystems with reduced efficiency
		708	as those tend to return 0 or 1 as link counts, which disables the
		709	directory counting heuristic.
		710
		711	=cut
		712
		713	sub aio_scandir($$;$) {
		714	aio_block {
		715	my ($path, $maxreq, $cb) = @_;
		716
		717	my $pri = aioreq_pri;
		718
		719	my $grp = aio_group $cb;
		720
		721	$maxreq = 4 if $maxreq <= 0;
		722
		723	# stat once
		724	aioreq_pri $pri;
		725	add $grp aio_stat $path, sub {
		726	return $grp->result () if $_[0];
		727	my $now = time;
		728	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		729
		730	# read the directory entries
		731	aioreq_pri $pri;
		732	add $grp aio_readdir $path, sub {
		733	my $entries = shift
		734	or return $grp->result ();
		735
		736	# stat the dir another time
		737	aioreq_pri $pri;
		738	add $grp aio_stat $path, sub {
		739	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		740
		741	my $ndirs;
		742
		743	# take the slow route if anything looks fishy
		744	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		745	$ndirs = -1;
		746	} else {
		747	# if nlink == 2, we are finished
		748	# on non-posix-fs's, we rely on nlink < 2
		749	$ndirs = (stat _)[3] - 2
		750	or return $grp->result ([], $entries);
		751	}
		752
		753	# sort into likely dirs and likely nondirs
		754	# dirs == files without ".", short entries first
		755	$entries = [map $_->[0],
		756	sort { $b->[1] cmp $a->[1] }
		757	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
		758	@$entries];
		759
		760	my (@dirs, @nondirs);
		761
		762	my $statgrp = add $grp aio_group sub {
		763	$grp->result (\@dirs, \@nondirs);
		764	};
		765
		766	limit $statgrp $maxreq;
		767	feed $statgrp sub {
		768	return unless @$entries;
		769	my $entry = pop @$entries;
		770
		771	aioreq_pri $pri;
		772	add $statgrp aio_stat "$path/$entry/.", sub {
		773	if ($_[0] < 0) {
		774	push @nondirs, $entry;
		775	} else {
		776	# need to check for real directory
		777	aioreq_pri $pri;
		778	add $statgrp aio_lstat "$path/$entry", sub {
		779	if (-d _) {
		780	push @dirs, $entry;
		781
		782	unless (--$ndirs) {
		783	push @nondirs, @$entries;
		784	feed $statgrp;
		785	}
		786	} else {
		787	push @nondirs, $entry;
		788	}
		789	}
		790	}
		791	};
		792	};
		793	};
		794	};
		795	};
		796
		797	$grp
		798	}
		799	}
		800
		801	=item aio_rmtree $path, $callback->($status)
		802
		803	Delete a directory tree starting (and including) C<$path>, return the
		804	status of the final C<rmdir> only. This is a composite request that
		805	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		806	everything else.
		807
		808	=cut
		809
		810	sub aio_rmtree;
		811	sub aio_rmtree($;$) {
		812	aio_block {
		813	my ($path, $cb) = @_;
		814
		815	my $pri = aioreq_pri;
		816	my $grp = aio_group $cb;
		817
		818	aioreq_pri $pri;
		819	add $grp aio_scandir $path, 0, sub {
		820	my ($dirs, $nondirs) = @_;
		821
		822	my $dirgrp = aio_group sub {
		823	add $grp aio_rmdir $path, sub {
		824	$grp->result ($_[0]);
		825	};
		826	};
		827
		828	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		829	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		830
		831	add $grp $dirgrp;
		832	};
		833
		834	$grp
		835	}
		836	}
		837
209	=item aio_fsync $fh, $callback	838	=item aio_fsync $fh, $callback->($status)
210		839
211	Asynchronously call fsync on the given filehandle and call the callback	840	Asynchronously call fsync on the given filehandle and call the callback
212	with the fsync result code.	841	with the fsync result code.
213		842
214	=item aio_fdatasync $fh, $callback	843	=item aio_fdatasync $fh, $callback->($status)
215		844
216	Asynchronously call fdatasync on the given filehandle and call the	845	Asynchronously call fdatasync on the given filehandle and call the
217	callback with the fdatasync result code.	846	callback with the fdatasync result code.
218		847
219	If this call isn't available because your OS lacks it or it couldn't be	848	If this call isn't available because your OS lacks it or it couldn't be
220	detected, it will be emulated by calling C<fsync> instead.	849	detected, it will be emulated by calling C<fsync> instead.
221		850
		851	=item aio_group $callback->(...)
		852
		853	This is a very special aio request: Instead of doing something, it is a
		854	container for other aio requests, which is useful if you want to bundle
		855	many requests into a single, composite, request with a definite callback
		856	and the ability to cancel the whole request with its subrequests.
		857
		858	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		859	for more info.
		860
		861	Example:
		862
		863	my $grp = aio_group sub {
		864	print "all stats done\n";
		865	};
		866
		867	add $grp
		868	(aio_stat ...),
		869	(aio_stat ...),
		870	...;
		871
		872	=item aio_nop $callback->()
		873
		874	This is a special request - it does nothing in itself and is only used for
		875	side effects, such as when you want to add a dummy request to a group so
		876	that finishing the requests in the group depends on executing the given
		877	code.
		878
		879	While this request does nothing, it still goes through the execution
		880	phase and still requires a worker thread. Thus, the callback will not
		881	be executed immediately but only after other requests in the queue have
		882	entered their execution phase. This can be used to measure request
		883	latency.
		884
		885	=item IO::AIO::aio_busy $fractional_seconds, $callback->() NOT EXPORTED
		886
		887	Mainly used for debugging and benchmarking, this aio request puts one of
		888	the request workers to sleep for the given time.
		889
		890	While it is theoretically handy to have simple I/O scheduling requests
		891	like sleep and file handle readable/writable, the overhead this creates is
		892	immense (it blocks a thread for a long time) so do not use this function
		893	except to put your application under artificial I/O pressure.
		894
222	=back	895	=back
223		896
		897	=head2 IO::AIO::REQ CLASS
		898
		899	All non-aggregate C<aio_*> functions return an object of this class when
		900	called in non-void context.
		901
		902	=over 4
		903
		904	=item cancel $req
		905
		906	Cancels the request, if possible. Has the effect of skipping execution
		907	when entering the B<execute> state and skipping calling the callback when
		908	entering the the B<result> state, but will leave the request otherwise
		909	untouched. That means that requests that currently execute will not be
		910	stopped and resources held by the request will not be freed prematurely.
		911
		912	=item cb $req $callback->(...)
		913
		914	Replace (or simply set) the callback registered to the request.
		915
		916	=back
		917
		918	=head2 IO::AIO::GRP CLASS
		919
		920	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		921	objects of this class, too.
		922
		923	A IO::AIO::GRP object is a special request that can contain multiple other
		924	aio requests.
		925
		926	You create one by calling the C<aio_group> constructing function with a
		927	callback that will be called when all contained requests have entered the
		928	C<done> state:
		929
		930	my $grp = aio_group sub {
		931	print "all requests are done\n";
		932	};
		933
		934	You add requests by calling the C<add> method with one or more
		935	C<IO::AIO::REQ> objects:
		936
		937	$grp->add (aio_unlink "...");
		938
		939	add $grp aio_stat "...", sub {
		940	$_[0] or return $grp->result ("error");
		941
		942	# add another request dynamically, if first succeeded
		943	add $grp aio_open "...", sub {
		944	$grp->result ("ok");
		945	};
		946	};
		947
		948	This makes it very easy to create composite requests (see the source of
		949	C<aio_move> for an application) that work and feel like simple requests.
		950
		951	=over 4
		952
		953	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		954	C<IO::AIO::poll_cb>, just like any other request.
		955
		956	=item * They can be canceled like any other request. Canceling will cancel not
		957	only the request itself, but also all requests it contains.
		958
		959	=item * They can also can also be added to other IO::AIO::GRP objects.
		960
		961	=item * You must not add requests to a group from within the group callback (or
		962	any later time).
		963
		964	=back
		965
		966	Their lifetime, simplified, looks like this: when they are empty, they
		967	will finish very quickly. If they contain only requests that are in the
		968	C<done> state, they will also finish. Otherwise they will continue to
		969	exist.
		970
		971	That means after creating a group you have some time to add requests. And
		972	in the callbacks of those requests, you can add further requests to the
		973	group. And only when all those requests have finished will the the group
		974	itself finish.
		975
		976	=over 4
		977
		978	=item add $grp ...
		979
		980	=item $grp->add (...)
		981
		982	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		983	be added, including other groups, as long as you do not create circular
		984	dependencies.
		985
		986	Returns all its arguments.
		987
		988	=item $grp->cancel_subs
		989
		990	Cancel all subrequests and clears any feeder, but not the group request
		991	itself. Useful when you queued a lot of events but got a result early.
		992
		993	=item $grp->result (...)
		994
		995	Set the result value(s) that will be passed to the group callback when all
		996	subrequests have finished and set thre groups errno to the current value
		997	of errno (just like calling C<errno> without an error number). By default,
		998	no argument will be passed and errno is zero.
		999
		1000	=item $grp->errno ([$errno])
		1001
		1002	Sets the group errno value to C<$errno>, or the current value of errno
		1003	when the argument is missing.
		1004
		1005	Every aio request has an associated errno value that is restored when
		1006	the callback is invoked. This method lets you change this value from its
		1007	default (0).
		1008
		1009	Calling C<result> will also set errno, so make sure you either set C<$!>
		1010	before the call to C<result>, or call c<errno> after it.
		1011
		1012	=item feed $grp $callback->($grp)
		1013
		1014	Sets a feeder/generator on this group: every group can have an attached
		1015	generator that generates requests if idle. The idea behind this is that,
		1016	although you could just queue as many requests as you want in a group,
		1017	this might starve other requests for a potentially long time. For
		1018	example, C<aio_scandir> might generate hundreds of thousands C<aio_stat>
		1019	requests, delaying any later requests for a long time.
		1020
		1021	To avoid this, and allow incremental generation of requests, you can
		1022	instead a group and set a feeder on it that generates those requests. The
		1023	feed callback will be called whenever there are few enough (see C<limit>,
		1024	below) requests active in the group itself and is expected to queue more
		1025	requests.
		1026
		1027	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1028	not impose any limits).
		1029
		1030	If the feed does not queue more requests when called, it will be
		1031	automatically removed from the group.
		1032
		1033	If the feed limit is C<0>, it will be set to C<2> automatically.
		1034
		1035	Example:
		1036
		1037	# stat all files in @files, but only ever use four aio requests concurrently:
		1038
		1039	my $grp = aio_group sub { print "finished\n" };
		1040	limit $grp 4;
		1041	feed $grp sub {
		1042	my $file = pop @files
		1043	or return;
		1044
		1045	add $grp aio_stat $file, sub { ... };
		1046	};
		1047
		1048	=item limit $grp $num
		1049
		1050	Sets the feeder limit for the group: The feeder will be called whenever
		1051	the group contains less than this many requests.
		1052
		1053	Setting the limit to C<0> will pause the feeding process.
		1054
		1055	=back
		1056
224	=head2 SUPPORT FUNCTIONS	1057	=head2 SUPPORT FUNCTIONS
		1058
		1059	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
225		1060
226	=over 4	1061	=over 4
227		1062
228	=item $fileno = IO::AIO::poll_fileno	1063	=item $fileno = IO::AIO::poll_fileno
229		1064
…		…
234		1069
235	See C<poll_cb> for an example.	1070	See C<poll_cb> for an example.
236		1071
237	=item IO::AIO::poll_cb	1072	=item IO::AIO::poll_cb
238		1073
239	Process all outstanding events on the result pipe. You have to call this	1074	Process some outstanding events on the result pipe. You have to call this
240	regularly. Returns the number of events processed. Returns immediately	1075	regularly. Returns the number of events processed. Returns immediately
241	when no events are outstanding.	1076	when no events are outstanding. The amount of events processed depends on
		1077	the settings of C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
		1078
		1079	If not all requests were processed for whatever reason, the filehandle
		1080	will still be ready when C<poll_cb> returns.
242		1081
243	Example: Install an Event watcher that automatically calls	1082	Example: Install an Event watcher that automatically calls
244	IO::AIO::poll_cb with high priority:	1083	IO::AIO::poll_cb with high priority:
245		1084
246	Event->io (fd => IO::AIO::poll_fileno,	1085	Event->io (fd => IO::AIO::poll_fileno,
247	poll => 'r', async => 1,	1086	poll => 'r', async => 1,
248	cb => \&IO::AIO::poll_cb);	1087	cb => \&IO::AIO::poll_cb);
249		1088
		1089	=item IO::AIO::max_poll_reqs $nreqs
		1090
		1091	=item IO::AIO::max_poll_time $seconds
		1092
		1093	These set the maximum number of requests (default C<0>, meaning infinity)
		1094	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1095	the maximum amount of time (default C<0>, meaning infinity) spent in
		1096	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1097	of time C<poll_cb> is allowed to use).
		1098
		1099	Setting C<max_poll_time> to a non-zero value creates an overhead of one
		1100	syscall per request processed, which is not normally a problem unless your
		1101	callbacks are really really fast or your OS is really really slow (I am
		1102	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
		1103
		1104	Setting these is useful if you want to ensure some level of
		1105	interactiveness when perl is not fast enough to process all requests in
		1106	time.
		1107
		1108	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
		1109
		1110	Example: Install an Event watcher that automatically calls
		1111	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1112	program get the CPU sometimes even under high AIO load.
		1113
		1114	# try not to spend much more than 0.1s in poll_cb
		1115	IO::AIO::max_poll_time 0.1;
		1116
		1117	# use a low priority so other tasks have priority
		1118	Event->io (fd => IO::AIO::poll_fileno,
		1119	poll => 'r', nice => 1,
		1120	cb => &IO::AIO::poll_cb);
		1121
250	=item IO::AIO::poll_wait	1122	=item IO::AIO::poll_wait
251		1123
		1124	If there are any outstanding requests and none of them in the result
252	Wait till the result filehandle becomes ready for reading (simply does a	1125	phase, wait till the result filehandle becomes ready for reading (simply
253	C<select> on the filehandle. This is useful if you want to synchronously wait	1126	does a C<select> on the filehandle. This is useful if you want to
254	for some requests to finish).	1127	synchronously wait for some requests to finish).
255		1128
256	See C<nreqs> for an example.	1129	See C<nreqs> for an example.
257		1130
		1131	=item IO::AIO::poll
		1132
		1133	Waits until some requests have been handled.
		1134
		1135	Returns the number of requests processed, but is otherwise strictly
		1136	equivalent to:
		1137
		1138	IO::AIO::poll_wait, IO::AIO::poll_cb
		1139
258	=item IO::AIO::nreqs	1140	=item IO::AIO::flush
259		1141
260	Returns the number of requests currently outstanding (i.e. for which their	1142	Wait till all outstanding AIO requests have been handled.
261	callback has not been invoked yet).
262		1143
263	Example: wait till there are no outstanding requests anymore:	1144	Strictly equivalent to:
264		1145
265	IO::AIO::poll_wait, IO::AIO::poll_cb	1146	IO::AIO::poll_wait, IO::AIO::poll_cb
266	while IO::AIO::nreqs;	1147	while IO::AIO::nreqs;
267		1148
		1149	=back
		1150
		1151	=head3 CONTROLLING THE NUMBER OF THREADS
		1152
		1153	=over
		1154
		1155	=item IO::AIO::min_parallel $nthreads
		1156
		1157	Set the minimum number of AIO threads to C<$nthreads>. The current
		1158	default is C<8>, which means eight asynchronous operations can execute
		1159	concurrently at any one time (the number of outstanding requests,
		1160	however, is unlimited).
		1161
		1162	IO::AIO starts threads only on demand, when an AIO request is queued and
		1163	no free thread exists. Please note that queueing up a hundred requests can
		1164	create demand for a hundred threads, even if it turns out that everything
		1165	is in the cache and could have been processed faster by a single thread.
		1166
		1167	It is recommended to keep the number of threads relatively low, as some
		1168	Linux kernel versions will scale negatively with the number of threads
		1169	(higher parallelity => MUCH higher latency). With current Linux 2.6
		1170	versions, 4-32 threads should be fine.
		1171
		1172	Under most circumstances you don't need to call this function, as the
		1173	module selects a default that is suitable for low to moderate load.
		1174
		1175	=item IO::AIO::max_parallel $nthreads
		1176
		1177	Sets the maximum number of AIO threads to C<$nthreads>. If more than the
		1178	specified number of threads are currently running, this function kills
		1179	them. This function blocks until the limit is reached.
		1180
		1181	While C<$nthreads> are zero, aio requests get queued but not executed
		1182	until the number of threads has been increased again.
		1183
		1184	This module automatically runs C<max_parallel 0> at program end, to ensure
		1185	that all threads are killed and that there are no outstanding requests.
		1186
		1187	Under normal circumstances you don't need to call this function.
		1188
		1189	=item IO::AIO::max_idle $nthreads
		1190
		1191	Limit the number of threads (default: 4) that are allowed to idle (i.e.,
		1192	threads that did not get a request to process within 10 seconds). That
		1193	means if a thread becomes idle while C<$nthreads> other threads are also
		1194	idle, it will free its resources and exit.
		1195
		1196	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1197	to allow for extremely high load situations, but want to free resources
		1198	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1199
		1200	The default is probably ok in most situations, especially if thread
		1201	creation is fast. If thread creation is very slow on your system you might
		1202	want to use larger values.
		1203
		1204	=item $oldmaxreqs = IO::AIO::max_outstanding $maxreqs
		1205
		1206	This is a very bad function to use in interactive programs because it
		1207	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1208	use an C<aio_group> together with a feed callback.
		1209
		1210	Sets the maximum number of outstanding requests to C<$nreqs>. If you
		1211	to queue up more than this number of requests, the next call to the
		1212	C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
		1213	function will block until the limit is no longer exceeded.
		1214
		1215	The default value is very large, so there is no practical limit on the
		1216	number of outstanding requests.
		1217
		1218	You can still queue as many requests as you want. Therefore,
		1219	C<max_oustsanding> is mainly useful in simple scripts (with low values) or
		1220	as a stop gap to shield against fatal memory overflow (with large values).
		1221
		1222	=back
		1223
		1224	=head3 STATISTICAL INFORMATION
		1225
		1226	=over
		1227
268	=item IO::AIO::flush	1228	=item IO::AIO::nreqs
269		1229
270	Wait till all outstanding AIO requests have been handled.	1230	Returns the number of requests currently in the ready, execute or pending
		1231	states (i.e. for which their callback has not been invoked yet).
271		1232
272	Strictly equivalent to:	1233	Example: wait till there are no outstanding requests anymore:
273		1234
274	IO::AIO::poll_wait, IO::AIO::poll_cb	1235	IO::AIO::poll_wait, IO::AIO::poll_cb
275	while IO::AIO::nreqs;	1236	while IO::AIO::nreqs;
276		1237
		1238	=item IO::AIO::nready
		1239
		1240	Returns the number of requests currently in the ready state (not yet
		1241	executed).
		1242
277	=item IO::AIO::poll	1243	=item IO::AIO::npending
278		1244
279	Waits until some requests have been handled.	1245	Returns the number of requests currently in the pending state (executed,
280		1246	but not yet processed by poll_cb).
281	Strictly equivalent to:
282
283	IO::AIO::poll_wait, IO::AIO::poll_cb
284	if IO::AIO::nreqs;
285
286	=item IO::AIO::min_parallel $nthreads
287
288	Set the minimum number of AIO threads to C<$nthreads>. The default is
289	C<1>, which means a single asynchronous operation can be done at one time
290	(the number of outstanding operations, however, is unlimited).
291
292	It is recommended to keep the number of threads low, as some Linux
293	kernel versions will scale negatively with the number of threads (higher
294	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32
295	threads should be fine.
296
297	Under normal circumstances you don't need to call this function, as this
298	module automatically starts some threads (the exact number might change,
299	and is currently 4).
300
301	=item IO::AIO::max_parallel $nthreads
302
303	Sets the maximum number of AIO threads to C<$nthreads>. If more than
304	the specified number of threads are currently running, kill them. This
305	function blocks until the limit is reached.
306
307	This module automatically runs C<max_parallel 0> at program end, to ensure
308	that all threads are killed and that there are no outstanding requests.
309
310	Under normal circumstances you don't need to call this function.
311
312	=item $oldnreqs = IO::AIO::max_outstanding $nreqs
313
314	Sets the maximum number of outstanding requests to C<$nreqs>. If you
315	try to queue up more than this number of requests, the caller will block until
316	some requests have been handled.
317
318	The default is very large, so normally there is no practical limit. If you
319	queue up many requests in a loop it it often improves speed if you set
320	this to a relatively low number, such as C<100>.
321
322	Under normal circumstances you don't need to call this function.
323		1247
324	=back	1248	=back
325		1249
326	=cut	1250	=cut
327		1251
…		…
339	or return undef;	1263	or return undef;
340		1264
341	*$sym	1265	*$sym
342	}	1266	}
343		1267
344	min_parallel 4;	1268	min_parallel 8;
345		1269
346	END {	1270	END { flush }
347	max_parallel 0;
348	}
349		1271
350	1;	1272	1;
351		1273
352	=head2 FORK BEHAVIOUR	1274	=head2 FORK BEHAVIOUR
353		1275
354	Before the fork IO::AIO first handles all outstanding requests - if other	1276	This module should do "the right thing" when the process using it forks:
355	threads add requests during this period, this time is prolonged. It then	1277
356	enters a quiescent state where no requests can be added in other threads	1278	Before the fork, IO::AIO enters a quiescent state where no requests
357	and no results will be processed. After the fork the parent simply leaves	1279	can be added in other threads and no results will be processed. After
358	the quiescent state and continues request processing, while the child	1280	the fork the parent simply leaves the quiescent state and continues
359	starts the same number of threads as were in use by the parent.	1281	request/result processing, while the child frees the request/result queue
		1282	(so that the requests started before the fork will only be handled in the
		1283	parent). Threads will be started on demand until the limit set in the
		1284	parent process has been reached again.
		1285
		1286	In short: the parent will, after a short pause, continue as if fork had
		1287	not been called, while the child will act as if IO::AIO has not been used
		1288	yet.
		1289
		1290	=head2 MEMORY USAGE
		1291
		1292	Per-request usage:
		1293
		1294	Each aio request uses - depending on your architecture - around 100-200
		1295	bytes of memory. In addition, stat requests need a stat buffer (possibly
		1296	a few hundred bytes), readdir requires a result buffer and so on. Perl
		1297	scalars and other data passed into aio requests will also be locked and
		1298	will consume memory till the request has entered the done state.
		1299
		1300	This is not awfully much, so queuing lots of requests is not usually a
		1301	problem.
		1302
		1303	Per-thread usage:
		1304
		1305	In the execution phase, some aio requests require more memory for
		1306	temporary buffers, and each thread requires a stack and other data
		1307	structures (usually around 16k-128k, depending on the OS).
		1308
		1309	=head1 KNOWN BUGS
		1310
		1311	Known bugs will be fixed in the next release.
360		1312
361	=head1 SEE ALSO	1313	=head1 SEE ALSO
362		1314
363	L<Coro>, L<Linux::AIO>.	1315	L<Coro::AIO>.
364		1316
365	=head1 AUTHOR	1317	=head1 AUTHOR
366		1318
367	Marc Lehmann <schmorp@schmorp.de>	1319	Marc Lehmann <schmorp@schmorp.de>
368	http://home.schmorp.de/	1320	http://home.schmorp.de/

Diff Legend

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

Comparing IO-AIO/AIO.pm (file contents): Revision 1.30 by root, Wed Aug 17 04:47:38 2005 UTC vs. Revision 1.111 by root, Mon Aug 13 12:08:13 2007 UTC

Diff Legend

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.30 by root, Wed Aug 17 04:47:38 2005 UTC vs.
Revision 1.111 by root, Mon Aug 13 12:08:13 2007 UTC