ViewVC Help

View File | Revision Log | Show Annotations | Download File

/cvs/IO-AIO/AIO.pm

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.18 by root, Mon Jul 11 03:29:39 2005 UTC vs.
Revision 1.116 by root, Wed Oct 3 21:27:51 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.	84	it is currently not reentrant in any way, so use appropriate locking
		85	yourself, always call C<poll_cb> from within the same thread, or never
		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
54		183
55	=cut	184	=cut
56		185
57	package IO::AIO;	186	package IO::AIO;
58		187
		188	no warnings;
		189	use strict 'vars';
		190
59	use base 'Exporter';	191	use base 'Exporter';
60		192
61	use Fcntl ();
62
63	BEGIN {	193	BEGIN {
64	$VERSION = 0.4;	194	our $VERSION = '2.5';
65		195
66	@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
67	aio_fsync aio_fdatasync aio_readahead);	197	aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink
68	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);	198	aio_readlink aio_fsync aio_fdatasync aio_readahead aio_rename aio_link
		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';
69		208
70	require XSLoader;	209	require XSLoader;
71	XSLoader::load IO::AIO, $VERSION;	210	XSLoader::load ("IO::AIO", $VERSION);
72	}	211	}
73		212
74	=head1 FUNCTIONS	213	=head1 FUNCTIONS
75		214
76	=head2 AIO FUNCTIONS	215	=head2 AIO REQUEST FUNCTIONS
77		216
78	All the C<aio_*> calls are more or less thin wrappers around the syscall	217	All the C<aio_*> calls are more or less thin wrappers around the syscall
79	with the same name (sans C<aio_>). The arguments are similar or identical,	218	with the same name (sans C<aio_>). The arguments are similar or identical,
80	and they all accept an additional (and optional) C<$callback> argument	219	and they all accept an additional (and optional) C<$callback> argument
81	which must be a code reference. This code reference will get called with	220	which must be a code reference. This code reference will get called with
82	the syscall return code (e.g. most syscalls return C<-1> on error, unlike	221	the syscall return code (e.g. most syscalls return C<-1> on error, unlike
83	perl, which usually delivers "false") as it's sole argument when the given	222	perl, which usually delivers "false") as it's sole argument when the given
84	syscall has been executed asynchronously.	223	syscall has been executed asynchronously.
85		224
86	All functions that expect a filehandle will also accept a file descriptor.	225	All functions expecting a filehandle keep a copy of the filehandle
		226	internally until the request has finished.
87		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
88	The filenames you pass to these routines I<must> be absolute. The reason	231	The pathnames you pass to these routines I<must> be absolute and
89	is that at the time the request is being executed, the current working	232	encoded as octets. The reason for the former is that at the time the
90	directory could have changed. Alternatively, you can make sure that you	233	request is being executed, the current working directory could have
91	never change the current working directory.	234	changed. Alternatively, you can make sure that you never change the
		235	current working directory anywhere in the program and then use relative
		236	paths.
		237
		238	To encode pathnames as octets, either make sure you either: a) always pass
		239	in filenames you got from outside (command line, readdir etc.) without
		240	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
		241	your pathnames to the locale (or other) encoding in effect in the user
		242	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
		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.
92		247
93	=over 4	248	=over 4
94		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
95	=item aio_open $pathname, $flags, $mode, $callback	283	=item aio_open $pathname, $flags, $mode, $callback->($fh)
96		284
97	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
98	created filehandle for the file.	286	created filehandle for the file.
99		287
100	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,
101	for an explanation.	289	for an explanation.
102		290
103	The C<$mode> argument is a bitmask. See the C<Fcntl> module for a	291	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
104	list. They are the same as used in C<sysopen>.	292	list. They are the same as used by C<sysopen>.
		293
		294	Likewise, C<$mode> specifies the mode of the newly created file, if it
		295	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
		296	except that it is mandatory (i.e. use C<0> if you don't create new files,
		297	and C<0666> or C<0777> if you do). Note that the C<$mode> will be modified
		298	by the umask in effect then the request is being executed, so better never
		299	change the umask.
105		300
106	Example:	301	Example:
107		302
108	aio_open "/etc/passwd", O_RDONLY, 0, sub {	303	aio_open "/etc/passwd", O_RDONLY, 0, sub {
109	if ($_[0]) {	304	if ($_[0]) {
…		…
112	} else {	307	} else {
113	die "open failed: $!\n";	308	die "open failed: $!\n";
114	}	309	}
115	};	310	};
116		311
		312
117	=item aio_close $fh, $callback	313	=item aio_close $fh, $callback->($status)
118		314
119	Asynchronously close a file and call the callback with the result	315	Asynchronously close a file and call the callback with the result
120	code. I<WARNING:> although accepted, you should not pass in a perl	316	code.
121	filehandle here, as perl will likely close the file descriptor itself when
122	the filehandle is destroyed. Normally, you can safely call perls C<close>
123	or just let filehandles go out of scope.
124		317
		318	Unlike the other functions operating on files, this function uses the
		319	PerlIO layer to close the filehandle. The reason is that the PerlIO API
		320	insists on closing the underlying fd itself, no matter what, and doesn't
		321	allow modifications to the fd. Unfortunately, it is not clear that you can
		322	call PerlIO from different threads (actually, its quite clear that this
		323	won't work in some cases), so while it likely works perfectly with simple
		324	file handles (such as the ones created by C<aio_open>) it might fail in
		325	interesting ways for others.
		326
		327	Having said that, aio_close tries to clean up the filehandle as much as
		328	possible before handing it to an io thread, and generally does work.
		329
		330
125	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	331	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
126		332
127	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	333	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
128		334
129	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	335	Reads or writes C<$length> bytes from the specified C<$fh> and C<$offset>
130	into the scalar given by C<data> and offset C<dataoffset> and calls the	336	into the scalar given by C<$data> and offset C<$dataoffset> and calls the
131	callback without the actual number of bytes read (or -1 on error, just	337	callback without the actual number of bytes read (or -1 on error, just
132	like the syscall).	338	like the syscall).
		339
		340	If C<$offset> is undefined, then the current file descriptor offset will
		341	be used (and updated), otherwise the file descriptor offset will not be
		342	changed by these calls.
		343
		344	If C<$length> is undefined in C<aio_write>, use the remaining length of C<$data>.
		345
		346	If C<$dataoffset> is less than zero, it will be counted from the end of
		347	C<$data>.
		348
		349	The C<$data> scalar I<MUST NOT> be modified in any way while the request
		350	is outstanding. Modifying it can result in segfaults or World War III (if
		351	the necessary/optional hardware is installed).
133		352
134	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at	353	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
135	offset C<0> within the scalar:	354	offset C<0> within the scalar:
136		355
137	aio_read $fh, 7, 15, $buffer, 0, sub {	356	aio_read $fh, 7, 15, $buffer, 0, sub {
138	$_[0] > 0 or die "read error: $!";	357	$_[0] > 0 or die "read error: $!";
139	print "read $_[0] bytes: <$buffer>\n";	358	print "read $_[0] bytes: <$buffer>\n";
140	};	359	};
141		360
		361
		362	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		363
		364	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
		365	reading at byte offset C<$in_offset>, and starts writing at the current
		366	file offset of C<$out_fh>. Because of that, it is not safe to issue more
		367	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
		368	other.
		369
		370	This call tries to make use of a native C<sendfile> syscall to provide
		371	zero-copy operation. For this to work, C<$out_fh> should refer to a
		372	socket, and C<$in_fh> should refer to mmap'able file.
		373
		374	If the native sendfile call fails or is not implemented, it will be
		375	emulated, so you can call C<aio_sendfile> on any type of filehandle
		376	regardless of the limitations of the operating system.
		377
		378	Please note, however, that C<aio_sendfile> can read more bytes from
		379	C<$in_fh> than are written, and there is no way to find out how many
		380	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
		381	provides the number of bytes written to C<$out_fh>. Only if the result
		382	value equals C<$length> one can assume that C<$length> bytes have been
		383	read.
		384
		385
142	=item aio_readahead $fh,$offset,$length, $callback	386	=item aio_readahead $fh,$offset,$length, $callback->($retval)
143		387
144	Asynchronously reads the specified byte range into the page cache, using
145	the C<readahead> syscall. If that syscall doesn't exist (likely if your OS
146	isn't Linux) the status will be C<-1> and C<$!> is set to ENOSYS.
147
148	readahead() populates the page cache with data from a file so that	388	C<aio_readahead> populates the page cache with data from a file so that
149	subsequent reads from that file will not block on disk I/O. The C<$offset>	389	subsequent reads from that file will not block on disk I/O. The C<$offset>
150	argument specifies the starting point from which data is to be read and	390	argument specifies the starting point from which data is to be read and
151	C<$length> specifies the number of bytes to be read. I/O is performed in	391	C<$length> specifies the number of bytes to be read. I/O is performed in
152	whole pages, so that offset is effectively rounded down to a page boundary	392	whole pages, so that offset is effectively rounded down to a page boundary
153	and bytes are read up to the next page boundary greater than or equal to	393	and bytes are read up to the next page boundary greater than or equal to
154	(off-set+length). aio_readahead() does not read beyond the end of the	394	(off-set+length). C<aio_readahead> does not read beyond the end of the
155	file. The current file offset of the file is left unchanged.	395	file. The current file offset of the file is left unchanged.
156		396
		397	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
		398	emulated by simply reading the data, which would have a similar effect.
		399
		400
157	=item aio_stat $fh_or_path, $callback	401	=item aio_stat $fh_or_path, $callback->($status)
158		402
159	=item aio_lstat $fh, $callback	403	=item aio_lstat $fh, $callback->($status)
160		404
161	Works like perl's C<stat> or C<lstat> in void context. The callback will	405	Works like perl's C<stat> or C<lstat> in void context. The callback will
162	be called after the stat and the results will be available using C<stat _>	406	be called after the stat and the results will be available using C<stat _>
163	or C<-s _> etc...	407	or C<-s _> etc...
164		408
…		…
174	aio_stat "/etc/passwd", sub {	418	aio_stat "/etc/passwd", sub {
175	$_[0] and die "stat failed: $!";	419	$_[0] and die "stat failed: $!";
176	print "size is ", -s _, "\n";	420	print "size is ", -s _, "\n";
177	};	421	};
178		422
		423
		424	=item aio_utime $fh_or_path, $atime, $mtime, $callback->($status)
		425
		426	Works like perl's C<utime> function (including the special case of $atime
		427	and $mtime being undef). Fractional times are supported if the underlying
		428	syscalls support them.
		429
		430	When called with a pathname, uses utimes(2) if available, otherwise
		431	utime(2). If called on a file descriptor, uses futimes(2) if available,
		432	otherwise returns ENOSYS, so this is not portable.
		433
		434	Examples:
		435
		436	# set atime and mtime to current time (basically touch(1)):
		437	aio_utime "path", undef, undef;
		438	# set atime to current time and mtime to beginning of the epoch:
		439	aio_utime "path", time, undef; # undef==0
		440
		441
		442	=item aio_chown $fh_or_path, $uid, $gid, $callback->($status)
		443
		444	Works like perl's C<chown> function, except that C<undef> for either $uid
		445	or $gid is being interpreted as "do not change" (but -1 can also be used).
		446
		447	Examples:
		448
		449	# same as "chown root path" in the shell:
		450	aio_chown "path", 0, -1;
		451	# same as above:
		452	aio_chown "path", 0, undef;
		453
		454
		455	=item aio_truncate $fh_or_path, $offset, $callback->($status)
		456
		457	Works like truncate(2) or ftruncate(2).
		458
		459
		460	=item aio_chmod $fh_or_path, $mode, $callback->($status)
		461
		462	Works like perl's C<chmod> function.
		463
		464
179	=item aio_unlink $pathname, $callback	465	=item aio_unlink $pathname, $callback->($status)
180		466
181	Asynchronously unlink (delete) a file and call the callback with the	467	Asynchronously unlink (delete) a file and call the callback with the
182	result code.	468	result code.
183		469
		470
		471	=item aio_mknod $path, $mode, $dev, $callback->($status)
		472
		473	[EXPERIMENTAL]
		474
		475	Asynchronously create a device node (or fifo). See mknod(2).
		476
		477	The only (POSIX-) portable way of calling this function is:
		478
		479	aio_mknod $path, IO::AIO::S_IFIFO | $mode, 0, sub { ...
		480
		481
		482	=item aio_link $srcpath, $dstpath, $callback->($status)
		483
		484	Asynchronously create a new link to the existing object at C<$srcpath> at
		485	the path C<$dstpath> and call the callback with the result code.
		486
		487
		488	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		489
		490	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		491	the path C<$dstpath> and call the callback with the result code.
		492
		493
		494	=item aio_readlink $path, $callback->($link)
		495
		496	Asynchronously read the symlink specified by C<$path> and pass it to
		497	the callback. If an error occurs, nothing or undef gets passed to the
		498	callback.
		499
		500
		501	=item aio_rename $srcpath, $dstpath, $callback->($status)
		502
		503	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		504	rename(2) and call the callback with the result code.
		505
		506
		507	=item aio_mkdir $pathname, $mode, $callback->($status)
		508
		509	Asynchronously mkdir (create) a directory and call the callback with
		510	the result code. C<$mode> will be modified by the umask at the time the
		511	request is executed, so do not change your umask.
		512
		513
		514	=item aio_rmdir $pathname, $callback->($status)
		515
		516	Asynchronously rmdir (delete) a directory and call the callback with the
		517	result code.
		518
		519
		520	=item aio_readdir $pathname, $callback->($entries)
		521
		522	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
		523	directory (i.e. opendir + readdir + closedir). The entries will not be
		524	sorted, and will B<NOT> include the C<.> and C<..> entries.
		525
		526	The callback a single argument which is either C<undef> or an array-ref
		527	with the filenames.
		528
		529
		530	=item aio_load $path, $data, $callback->($status)
		531
		532	This is a composite request that tries to fully load the given file into
		533	memory. Status is the same as with aio_read.
		534
		535	=cut
		536
		537	sub aio_load($$;$) {
		538	aio_block {
		539	my ($path, undef, $cb) = @_;
		540	my $data = \$_[1];
		541
		542	my $pri = aioreq_pri;
		543	my $grp = aio_group $cb;
		544
		545	aioreq_pri $pri;
		546	add $grp aio_open $path, O_RDONLY, 0, sub {
		547	my $fh = shift
		548	or return $grp->result (-1);
		549
		550	aioreq_pri $pri;
		551	add $grp aio_read $fh, 0, (-s $fh), $$data, 0, sub {
		552	$grp->result ($_[0]);
		553	};
		554	};
		555
		556	$grp
		557	}
		558	}
		559
		560	=item aio_copy $srcpath, $dstpath, $callback->($status)
		561
		562	Try to copy the I<file> (directories not supported as either source or
		563	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		564	the C<0> (error) or C<-1> ok.
		565
		566	This is a composite request that it creates the destination file with
		567	mode 0200 and copies the contents of the source file into it using
		568	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		569	uid/gid, in that order.
		570
		571	If an error occurs, the partial destination file will be unlinked, if
		572	possible, except when setting atime, mtime, access mode and uid/gid, where
		573	errors are being ignored.
		574
		575	=cut
		576
		577	sub aio_copy($$;$) {
		578	aio_block {
		579	my ($src, $dst, $cb) = @_;
		580
		581	my $pri = aioreq_pri;
		582	my $grp = aio_group $cb;
		583
		584	aioreq_pri $pri;
		585	add $grp aio_open $src, O_RDONLY, 0, sub {
		586	if (my $src_fh = $_[0]) {
		587	my @stat = stat $src_fh;
		588
		589	aioreq_pri $pri;
		590	add $grp aio_open $dst, O_CREAT | O_WRONLY | O_TRUNC, 0200, sub {
		591	if (my $dst_fh = $_[0]) {
		592	aioreq_pri $pri;
		593	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		594	if ($_[0] == $stat[7]) {
		595	$grp->result (0);
		596	close $src_fh;
		597
		598	# those should not normally block. should. should.
		599	utime $stat[8], $stat[9], $dst;
		600	chmod $stat[2] & 07777, $dst_fh;
		601	chown $stat[4], $stat[5], $dst_fh;
		602	close $dst_fh;
		603	} else {
		604	$grp->result (-1);
		605	close $src_fh;
		606	close $dst_fh;
		607
		608	aioreq $pri;
		609	add $grp aio_unlink $dst;
		610	}
		611	};
		612	} else {
		613	$grp->result (-1);
		614	}
		615	},
		616
		617	} else {
		618	$grp->result (-1);
		619	}
		620	};
		621
		622	$grp
		623	}
		624	}
		625
		626	=item aio_move $srcpath, $dstpath, $callback->($status)
		627
		628	Try to move the I<file> (directories not supported as either source or
		629	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		630	the C<0> (error) or C<-1> ok.
		631
		632	This is a composite request that tries to rename(2) the file first. If
		633	rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if
		634	that is successful, unlinking the C<$srcpath>.
		635
		636	=cut
		637
		638	sub aio_move($$;$) {
		639	aio_block {
		640	my ($src, $dst, $cb) = @_;
		641
		642	my $pri = aioreq_pri;
		643	my $grp = aio_group $cb;
		644
		645	aioreq_pri $pri;
		646	add $grp aio_rename $src, $dst, sub {
		647	if ($_[0] && $! == EXDEV) {
		648	aioreq_pri $pri;
		649	add $grp aio_copy $src, $dst, sub {
		650	$grp->result ($_[0]);
		651
		652	if (!$_[0]) {
		653	aioreq_pri $pri;
		654	add $grp aio_unlink $src;
		655	}
		656	};
		657	} else {
		658	$grp->result ($_[0]);
		659	}
		660	};
		661
		662	$grp
		663	}
		664	}
		665
		666	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
		667
		668	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		669	efficiently separate the entries of directory C<$path> into two sets of
		670	names, directories you can recurse into (directories), and ones you cannot
		671	recurse into (everything else, including symlinks to directories).
		672
		673	C<aio_scandir> is a composite request that creates of many sub requests_
		674	C<$maxreq> specifies the maximum number of outstanding aio requests that
		675	this function generates. If it is C<< <= 0 >>, then a suitable default
		676	will be chosen (currently 4).
		677
		678	On error, the callback is called without arguments, otherwise it receives
		679	two array-refs with path-relative entry names.
		680
		681	Example:
		682
		683	aio_scandir $dir, 0, sub {
		684	my ($dirs, $nondirs) = @_;
		685	print "real directories: @$dirs\n";
		686	print "everything else: @$nondirs\n";
		687	};
		688
		689	Implementation notes.
		690
		691	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		692
		693	After reading the directory, the modification time, size etc. of the
		694	directory before and after the readdir is checked, and if they match (and
		695	isn't the current time), the link count will be used to decide how many
		696	entries are directories (if >= 2). Otherwise, no knowledge of the number
		697	of subdirectories will be assumed.
		698
		699	Then entries will be sorted into likely directories (everything without
		700	a non-initial dot currently) and likely non-directories (everything
		701	else). Then every entry plus an appended C</.> will be C<stat>'ed,
		702	likely directories first. If that succeeds, it assumes that the entry
		703	is a directory or a symlink to directory (which will be checked
		704	seperately). This is often faster than stat'ing the entry itself because
		705	filesystems might detect the type of the entry without reading the inode
		706	data (e.g. ext2fs filetype feature).
		707
		708	If the known number of directories (link count - 2) has been reached, the
		709	rest of the entries is assumed to be non-directories.
		710
		711	This only works with certainty on POSIX (= UNIX) filesystems, which
		712	fortunately are the vast majority of filesystems around.
		713
		714	It will also likely work on non-POSIX filesystems with reduced efficiency
		715	as those tend to return 0 or 1 as link counts, which disables the
		716	directory counting heuristic.
		717
		718	=cut
		719
		720	sub aio_scandir($$;$) {
		721	aio_block {
		722	my ($path, $maxreq, $cb) = @_;
		723
		724	my $pri = aioreq_pri;
		725
		726	my $grp = aio_group $cb;
		727
		728	$maxreq = 4 if $maxreq <= 0;
		729
		730	# stat once
		731	aioreq_pri $pri;
		732	add $grp aio_stat $path, sub {
		733	return $grp->result () if $_[0];
		734	my $now = time;
		735	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		736
		737	# read the directory entries
		738	aioreq_pri $pri;
		739	add $grp aio_readdir $path, sub {
		740	my $entries = shift
		741	or return $grp->result ();
		742
		743	# stat the dir another time
		744	aioreq_pri $pri;
		745	add $grp aio_stat $path, sub {
		746	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		747
		748	my $ndirs;
		749
		750	# take the slow route if anything looks fishy
		751	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		752	$ndirs = -1;
		753	} else {
		754	# if nlink == 2, we are finished
		755	# on non-posix-fs's, we rely on nlink < 2
		756	$ndirs = (stat _)[3] - 2
		757	or return $grp->result ([], $entries);
		758	}
		759
		760	# sort into likely dirs and likely nondirs
		761	# dirs == files without ".", short entries first
		762	$entries = [map $_->[0],
		763	sort { $b->[1] cmp $a->[1] }
		764	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
		765	@$entries];
		766
		767	my (@dirs, @nondirs);
		768
		769	my $statgrp = add $grp aio_group sub {
		770	$grp->result (\@dirs, \@nondirs);
		771	};
		772
		773	limit $statgrp $maxreq;
		774	feed $statgrp sub {
		775	return unless @$entries;
		776	my $entry = pop @$entries;
		777
		778	aioreq_pri $pri;
		779	add $statgrp aio_stat "$path/$entry/.", sub {
		780	if ($_[0] < 0) {
		781	push @nondirs, $entry;
		782	} else {
		783	# need to check for real directory
		784	aioreq_pri $pri;
		785	add $statgrp aio_lstat "$path/$entry", sub {
		786	if (-d _) {
		787	push @dirs, $entry;
		788
		789	unless (--$ndirs) {
		790	push @nondirs, @$entries;
		791	feed $statgrp;
		792	}
		793	} else {
		794	push @nondirs, $entry;
		795	}
		796	}
		797	}
		798	};
		799	};
		800	};
		801	};
		802	};
		803
		804	$grp
		805	}
		806	}
		807
		808	=item aio_rmtree $path, $callback->($status)
		809
		810	Delete a directory tree starting (and including) C<$path>, return the
		811	status of the final C<rmdir> only. This is a composite request that
		812	uses C<aio_scandir> to recurse into and rmdir directories, and unlink
		813	everything else.
		814
		815	=cut
		816
		817	sub aio_rmtree;
		818	sub aio_rmtree($;$) {
		819	aio_block {
		820	my ($path, $cb) = @_;
		821
		822	my $pri = aioreq_pri;
		823	my $grp = aio_group $cb;
		824
		825	aioreq_pri $pri;
		826	add $grp aio_scandir $path, 0, sub {
		827	my ($dirs, $nondirs) = @_;
		828
		829	my $dirgrp = aio_group sub {
		830	add $grp aio_rmdir $path, sub {
		831	$grp->result ($_[0]);
		832	};
		833	};
		834
		835	(aioreq_pri $pri), add $dirgrp aio_rmtree "$path/$_" for @$dirs;
		836	(aioreq_pri $pri), add $dirgrp aio_unlink "$path/$_" for @$nondirs;
		837
		838	add $grp $dirgrp;
		839	};
		840
		841	$grp
		842	}
		843	}
		844
184	=item aio_fsync $fh, $callback	845	=item aio_fsync $fh, $callback->($status)
185		846
186	Asynchronously call fsync on the given filehandle and call the callback	847	Asynchronously call fsync on the given filehandle and call the callback
187	with the fsync result code.	848	with the fsync result code.
188		849
189	=item aio_fdatasync $fh, $callback	850	=item aio_fdatasync $fh, $callback->($status)
190		851
191	Asynchronously call fdatasync on the given filehandle and call the	852	Asynchronously call fdatasync on the given filehandle and call the
192	callback with the fdatasync result code.	853	callback with the fdatasync result code.
193		854
		855	If this call isn't available because your OS lacks it or it couldn't be
		856	detected, it will be emulated by calling C<fsync> instead.
		857
		858	=item aio_group $callback->(...)
		859
		860	This is a very special aio request: Instead of doing something, it is a
		861	container for other aio requests, which is useful if you want to bundle
		862	many requests into a single, composite, request with a definite callback
		863	and the ability to cancel the whole request with its subrequests.
		864
		865	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		866	for more info.
		867
		868	Example:
		869
		870	my $grp = aio_group sub {
		871	print "all stats done\n";
		872	};
		873
		874	add $grp
		875	(aio_stat ...),
		876	(aio_stat ...),
		877	...;
		878
		879	=item aio_nop $callback->()
		880
		881	This is a special request - it does nothing in itself and is only used for
		882	side effects, such as when you want to add a dummy request to a group so
		883	that finishing the requests in the group depends on executing the given
		884	code.
		885
		886	While this request does nothing, it still goes through the execution
		887	phase and still requires a worker thread. Thus, the callback will not
		888	be executed immediately but only after other requests in the queue have
		889	entered their execution phase. This can be used to measure request
		890	latency.
		891
		892	=item IO::AIO::aio_busy $fractional_seconds, $callback->() *NOT EXPORTED*
		893
		894	Mainly used for debugging and benchmarking, this aio request puts one of
		895	the request workers to sleep for the given time.
		896
		897	While it is theoretically handy to have simple I/O scheduling requests
		898	like sleep and file handle readable/writable, the overhead this creates is
		899	immense (it blocks a thread for a long time) so do not use this function
		900	except to put your application under artificial I/O pressure.
		901
194	=back	902	=back
195		903
		904	=head2 IO::AIO::REQ CLASS
		905
		906	All non-aggregate C<aio_*> functions return an object of this class when
		907	called in non-void context.
		908
		909	=over 4
		910
		911	=item cancel $req
		912
		913	Cancels the request, if possible. Has the effect of skipping execution
		914	when entering the B<execute> state and skipping calling the callback when
		915	entering the the B<result> state, but will leave the request otherwise
		916	untouched. That means that requests that currently execute will not be
		917	stopped and resources held by the request will not be freed prematurely.
		918
		919	=item cb $req $callback->(...)
		920
		921	Replace (or simply set) the callback registered to the request.
		922
		923	=back
		924
		925	=head2 IO::AIO::GRP CLASS
		926
		927	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		928	objects of this class, too.
		929
		930	A IO::AIO::GRP object is a special request that can contain multiple other
		931	aio requests.
		932
		933	You create one by calling the C<aio_group> constructing function with a
		934	callback that will be called when all contained requests have entered the
		935	C<done> state:
		936
		937	my $grp = aio_group sub {
		938	print "all requests are done\n";
		939	};
		940
		941	You add requests by calling the C<add> method with one or more
		942	C<IO::AIO::REQ> objects:
		943
		944	$grp->add (aio_unlink "...");
		945
		946	add $grp aio_stat "...", sub {
		947	$_[0] or return $grp->result ("error");
		948
		949	# add another request dynamically, if first succeeded
		950	add $grp aio_open "...", sub {
		951	$grp->result ("ok");
		952	};
		953	};
		954
		955	This makes it very easy to create composite requests (see the source of
		956	C<aio_move> for an application) that work and feel like simple requests.
		957
		958	=over 4
		959
		960	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		961	C<IO::AIO::poll_cb>, just like any other request.
		962
		963	=item * They can be canceled like any other request. Canceling will cancel not
		964	only the request itself, but also all requests it contains.
		965
		966	=item * They can also can also be added to other IO::AIO::GRP objects.
		967
		968	=item * You must not add requests to a group from within the group callback (or
		969	any later time).
		970
		971	=back
		972
		973	Their lifetime, simplified, looks like this: when they are empty, they
		974	will finish very quickly. If they contain only requests that are in the
		975	C<done> state, they will also finish. Otherwise they will continue to
		976	exist.
		977
		978	That means after creating a group you have some time to add requests. And
		979	in the callbacks of those requests, you can add further requests to the
		980	group. And only when all those requests have finished will the the group
		981	itself finish.
		982
		983	=over 4
		984
		985	=item add $grp ...
		986
		987	=item $grp->add (...)
		988
		989	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		990	be added, including other groups, as long as you do not create circular
		991	dependencies.
		992
		993	Returns all its arguments.
		994
		995	=item $grp->cancel_subs
		996
		997	Cancel all subrequests and clears any feeder, but not the group request
		998	itself. Useful when you queued a lot of events but got a result early.
		999
		1000	=item $grp->result (...)
		1001
		1002	Set the result value(s) that will be passed to the group callback when all
		1003	subrequests have finished and set thre groups errno to the current value
		1004	of errno (just like calling C<errno> without an error number). By default,
		1005	no argument will be passed and errno is zero.
		1006
		1007	=item $grp->errno ([$errno])
		1008
		1009	Sets the group errno value to C<$errno>, or the current value of errno
		1010	when the argument is missing.
		1011
		1012	Every aio request has an associated errno value that is restored when
		1013	the callback is invoked. This method lets you change this value from its
		1014	default (0).
		1015
		1016	Calling C<result> will also set errno, so make sure you either set C<$!>
		1017	before the call to C<result>, or call c<errno> after it.
		1018
		1019	=item feed $grp $callback->($grp)
		1020
		1021	Sets a feeder/generator on this group: every group can have an attached
		1022	generator that generates requests if idle. The idea behind this is that,
		1023	although you could just queue as many requests as you want in a group,
		1024	this might starve other requests for a potentially long time. For
		1025	example, C<aio_scandir> might generate hundreds of thousands C<aio_stat>
		1026	requests, delaying any later requests for a long time.
		1027
		1028	To avoid this, and allow incremental generation of requests, you can
		1029	instead a group and set a feeder on it that generates those requests. The
		1030	feed callback will be called whenever there are few enough (see C<limit>,
		1031	below) requests active in the group itself and is expected to queue more
		1032	requests.
		1033
		1034	The feed callback can queue as many requests as it likes (i.e. C<add> does
		1035	not impose any limits).
		1036
		1037	If the feed does not queue more requests when called, it will be
		1038	automatically removed from the group.
		1039
		1040	If the feed limit is C<0>, it will be set to C<2> automatically.
		1041
		1042	Example:
		1043
		1044	# stat all files in @files, but only ever use four aio requests concurrently:
		1045
		1046	my $grp = aio_group sub { print "finished\n" };
		1047	limit $grp 4;
		1048	feed $grp sub {
		1049	my $file = pop @files
		1050	or return;
		1051
		1052	add $grp aio_stat $file, sub { ... };
		1053	};
		1054
		1055	=item limit $grp $num
		1056
		1057	Sets the feeder limit for the group: The feeder will be called whenever
		1058	the group contains less than this many requests.
		1059
		1060	Setting the limit to C<0> will pause the feeding process.
		1061
		1062	=back
		1063
196	=head2 SUPPORT FUNCTIONS	1064	=head2 SUPPORT FUNCTIONS
197		1065
		1066	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		1067
198	=over 4	1068	=over 4
199		1069
200	=item $fileno = IO::AIO::poll_fileno	1070	=item $fileno = IO::AIO::poll_fileno
201		1071
202	Return the I<request result pipe filehandle>. This filehandle must be	1072	Return the I<request result pipe file descriptor>. This filehandle must be
203	polled for reading by some mechanism outside this module (e.g. Event	1073	polled for reading by some mechanism outside this module (e.g. Event or
204	or select, see below). If the pipe becomes readable you have to call	1074	select, see below or the SYNOPSIS). If the pipe becomes readable you have
205	C<poll_cb> to check the results.	1075	to call C<poll_cb> to check the results.
206		1076
207	See C<poll_cb> for an example.	1077	See C<poll_cb> for an example.
208		1078
209	=item IO::AIO::poll_cb	1079	=item IO::AIO::poll_cb
210		1080
211	Process all outstanding events on the result pipe. You have to call this	1081	Process some outstanding events on the result pipe. You have to call this
212	regularly. Returns the number of events processed. Returns immediately	1082	regularly. Returns the number of events processed. Returns immediately
213	when no events are outstanding.	1083	when no events are outstanding. The amount of events processed depends on
		1084	the settings of C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
214		1085
215	You can use Event to multiplex, e.g.:	1086	If not all requests were processed for whatever reason, the filehandle
		1087	will still be ready when C<poll_cb> returns.
		1088
		1089	Example: Install an Event watcher that automatically calls
		1090	IO::AIO::poll_cb with high priority:
216		1091
217	Event->io (fd => IO::AIO::poll_fileno,	1092	Event->io (fd => IO::AIO::poll_fileno,
218	poll => 'r', async => 1,	1093	poll => 'r', async => 1,
219	cb => \&IO::AIO::poll_cb);	1094	cb => \&IO::AIO::poll_cb);
220		1095
		1096	=item IO::AIO::max_poll_reqs $nreqs
		1097
		1098	=item IO::AIO::max_poll_time $seconds
		1099
		1100	These set the maximum number of requests (default C<0>, meaning infinity)
		1101	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		1102	the maximum amount of time (default C<0>, meaning infinity) spent in
		1103	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		1104	of time C<poll_cb> is allowed to use).
		1105
		1106	Setting C<max_poll_time> to a non-zero value creates an overhead of one
		1107	syscall per request processed, which is not normally a problem unless your
		1108	callbacks are really really fast or your OS is really really slow (I am
		1109	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
		1110
		1111	Setting these is useful if you want to ensure some level of
		1112	interactiveness when perl is not fast enough to process all requests in
		1113	time.
		1114
		1115	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
		1116
		1117	Example: Install an Event watcher that automatically calls
		1118	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		1119	program get the CPU sometimes even under high AIO load.
		1120
		1121	# try not to spend much more than 0.1s in poll_cb
		1122	IO::AIO::max_poll_time 0.1;
		1123
		1124	# use a low priority so other tasks have priority
		1125	Event->io (fd => IO::AIO::poll_fileno,
		1126	poll => 'r', nice => 1,
		1127	cb => &IO::AIO::poll_cb);
		1128
221	=item IO::AIO::poll_wait	1129	=item IO::AIO::poll_wait
222		1130
		1131	If there are any outstanding requests and none of them in the result
223	Wait till the result filehandle becomes ready for reading (simply does a	1132	phase, wait till the result filehandle becomes ready for reading (simply
224	select on the filehandle. This is useful if you want to synchronously wait	1133	does a C<select> on the filehandle. This is useful if you want to
225	for some requests to finish).	1134	synchronously wait for some requests to finish).
226		1135
227	See C<nreqs> for an example.	1136	See C<nreqs> for an example.
228		1137
		1138	=item IO::AIO::poll
		1139
		1140	Waits until some requests have been handled.
		1141
		1142	Returns the number of requests processed, but is otherwise strictly
		1143	equivalent to:
		1144
		1145	IO::AIO::poll_wait, IO::AIO::poll_cb
		1146
229	=item IO::AIO::nreqs	1147	=item IO::AIO::flush
230		1148
231	Returns the number of requests currently outstanding.	1149	Wait till all outstanding AIO requests have been handled.
232		1150
233	Example: wait till there are no outstanding requests anymore:	1151	Strictly equivalent to:
234		1152
235	IO::AIO::poll_wait, IO::AIO::poll_cb	1153	IO::AIO::poll_wait, IO::AIO::poll_cb
236	while IO::AIO::nreqs;	1154	while IO::AIO::nreqs;
237		1155
		1156	=back
		1157
		1158	=head3 CONTROLLING THE NUMBER OF THREADS
		1159
		1160	=over
		1161
		1162	=item IO::AIO::min_parallel $nthreads
		1163
		1164	Set the minimum number of AIO threads to C<$nthreads>. The current
		1165	default is C<8>, which means eight asynchronous operations can execute
		1166	concurrently at any one time (the number of outstanding requests,
		1167	however, is unlimited).
		1168
		1169	IO::AIO starts threads only on demand, when an AIO request is queued and
		1170	no free thread exists. Please note that queueing up a hundred requests can
		1171	create demand for a hundred threads, even if it turns out that everything
		1172	is in the cache and could have been processed faster by a single thread.
		1173
		1174	It is recommended to keep the number of threads relatively low, as some
		1175	Linux kernel versions will scale negatively with the number of threads
		1176	(higher parallelity => MUCH higher latency). With current Linux 2.6
		1177	versions, 4-32 threads should be fine.
		1178
		1179	Under most circumstances you don't need to call this function, as the
		1180	module selects a default that is suitable for low to moderate load.
		1181
		1182	=item IO::AIO::max_parallel $nthreads
		1183
		1184	Sets the maximum number of AIO threads to C<$nthreads>. If more than the
		1185	specified number of threads are currently running, this function kills
		1186	them. This function blocks until the limit is reached.
		1187
		1188	While C<$nthreads> are zero, aio requests get queued but not executed
		1189	until the number of threads has been increased again.
		1190
		1191	This module automatically runs C<max_parallel 0> at program end, to ensure
		1192	that all threads are killed and that there are no outstanding requests.
		1193
		1194	Under normal circumstances you don't need to call this function.
		1195
		1196	=item IO::AIO::max_idle $nthreads
		1197
		1198	Limit the number of threads (default: 4) that are allowed to idle (i.e.,
		1199	threads that did not get a request to process within 10 seconds). That
		1200	means if a thread becomes idle while C<$nthreads> other threads are also
		1201	idle, it will free its resources and exit.
		1202
		1203	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1204	to allow for extremely high load situations, but want to free resources
		1205	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1206
		1207	The default is probably ok in most situations, especially if thread
		1208	creation is fast. If thread creation is very slow on your system you might
		1209	want to use larger values.
		1210
		1211	=item $oldmaxreqs = IO::AIO::max_outstanding $maxreqs
		1212
		1213	This is a very bad function to use in interactive programs because it
		1214	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1215	use an C<aio_group> together with a feed callback.
		1216
		1217	Sets the maximum number of outstanding requests to C<$nreqs>. If you
		1218	do queue up more than this number of requests, the next call to the
		1219	C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
		1220	function will block until the limit is no longer exceeded.
		1221
		1222	The default value is very large, so there is no practical limit on the
		1223	number of outstanding requests.
		1224
		1225	You can still queue as many requests as you want. Therefore,
		1226	C<max_oustsanding> is mainly useful in simple scripts (with low values) or
		1227	as a stop gap to shield against fatal memory overflow (with large values).
		1228
		1229	=back
		1230
		1231	=head3 STATISTICAL INFORMATION
		1232
		1233	=over
		1234
238	=item IO::AIO::flush	1235	=item IO::AIO::nreqs
239		1236
240	Wait till all outstanding AIO requests have been handled.	1237	Returns the number of requests currently in the ready, execute or pending
		1238	states (i.e. for which their callback has not been invoked yet).
241		1239
242	Strictly equivalent to:	1240	Example: wait till there are no outstanding requests anymore:
243		1241
244	IO::AIO::poll_wait, IO::AIO::poll_cb	1242	IO::AIO::poll_wait, IO::AIO::poll_cb
245	while IO::AIO::nreqs;	1243	while IO::AIO::nreqs;
246		1244
		1245	=item IO::AIO::nready
		1246
		1247	Returns the number of requests currently in the ready state (not yet
		1248	executed).
		1249
247	=item IO::AIO::poll	1250	=item IO::AIO::npending
248		1251
249	Waits until some requests have been handled.	1252	Returns the number of requests currently in the pending state (executed,
250		1253	but not yet processed by poll_cb).
251	Strictly equivalent to:
252
253	IO::AIO::poll_wait, IO::AIO::poll_cb
254	if IO::AIO::nreqs;
255
256	=item IO::AIO::min_parallel $nthreads
257
258	Set the minimum number of AIO threads to C<$nthreads>. The default is
259	C<1>, which means a single asynchronous operation can be done at one time
260	(the number of outstanding operations, however, is unlimited).
261
262	It is recommended to keep the number of threads low, as some Linux
263	kernel versions will scale negatively with the number of threads (higher
264	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32
265	threads should be fine.
266
267	Under normal circumstances you don't need to call this function, as this
268	module automatically starts some threads (the exact number might change,
269	and is currently 4).
270
271	=item IO::AIO::max_parallel $nthreads
272
273	Sets the maximum number of AIO threads to C<$nthreads>. If more than
274	the specified number of threads are currently running, kill them. This
275	function blocks until the limit is reached.
276
277	This module automatically runs C<max_parallel 0> at program end, to ensure
278	that all threads are killed and that there are no outstanding requests.
279
280	Under normal circumstances you don't need to call this function.
281
282	=item $oldnreqs = IO::AIO::max_outstanding $nreqs
283
284	Sets the maximum number of outstanding requests to C<$nreqs>. If you
285	try to queue up more than this number of requests, the caller will block until
286	some requests have been handled.
287
288	The default is very large, so normally there is no practical limit. If you
289	queue up many requests in a loop it it often improves speed if you set
290	this to a relatively low number, such as C<100>.
291
292	Under normal circumstances you don't need to call this function.
293		1254
294	=back	1255	=back
295		1256
296	=cut	1257	=cut
297		1258
298	# support function to convert a fd into a perl filehandle
299	sub _fd2fh {
300	return undef if $_[0] < 0;
301
302	# try to be perl5.6-compatible
303	local *AIO_FH;
304	open AIO_FH, "+<&=$_[0]"
305	or return undef;
306
307	*AIO_FH
308	}
309
310	min_parallel 4;	1259	min_parallel 8;
311		1260
312	END {	1261	END { flush }
313	max_parallel 0;
314	}
315		1262
316	1;	1263	1;
317		1264
		1265	=head2 FORK BEHAVIOUR
		1266
		1267	This module should do "the right thing" when the process using it forks:
		1268
		1269	Before the fork, IO::AIO enters a quiescent state where no requests
		1270	can be added in other threads and no results will be processed. After
		1271	the fork the parent simply leaves the quiescent state and continues
		1272	request/result processing, while the child frees the request/result queue
		1273	(so that the requests started before the fork will only be handled in the
		1274	parent). Threads will be started on demand until the limit set in the
		1275	parent process has been reached again.
		1276
		1277	In short: the parent will, after a short pause, continue as if fork had
		1278	not been called, while the child will act as if IO::AIO has not been used
		1279	yet.
		1280
		1281	=head2 MEMORY USAGE
		1282
		1283	Per-request usage:
		1284
		1285	Each aio request uses - depending on your architecture - around 100-200
		1286	bytes of memory. In addition, stat requests need a stat buffer (possibly
		1287	a few hundred bytes), readdir requires a result buffer and so on. Perl
		1288	scalars and other data passed into aio requests will also be locked and
		1289	will consume memory till the request has entered the done state.
		1290
		1291	This is not awfully much, so queuing lots of requests is not usually a
		1292	problem.
		1293
		1294	Per-thread usage:
		1295
		1296	In the execution phase, some aio requests require more memory for
		1297	temporary buffers, and each thread requires a stack and other data
		1298	structures (usually around 16k-128k, depending on the OS).
		1299
		1300	=head1 KNOWN BUGS
		1301
		1302	Known bugs will be fixed in the next release.
		1303
318	=head1 SEE ALSO	1304	=head1 SEE ALSO
319		1305
320	L<Coro>, L<Linux::AIO>.	1306	L<Coro::AIO>.
321		1307
322	=head1 AUTHOR	1308	=head1 AUTHOR
323		1309
324	Marc Lehmann <schmorp@schmorp.de>	1310	Marc Lehmann <schmorp@schmorp.de>
325	http://home.schmorp.de/	1311	http://home.schmorp.de/

Diff Legend

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