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

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.6 by root, Sun Jul 10 22:19:48 2005 UTC vs.
Revision 1.91 by root, Mon Oct 30 23:30:29 2006 UTC

…		…
12	};	12	};
13		13
14	aio_unlink "/tmp/file", sub { };	14	aio_unlink "/tmp/file", sub { };
15		15
16	aio_read $fh, 30000, 1024, $buffer, 0, sub {	16	aio_read $fh, 30000, 1024, $buffer, 0, sub {
17	$_[0] >= 0 or die "read error: $!";	17	$_[0] > 0 or die "read error: $!";
18	};	18	};
19		19
20	# Event	20	# version 2+ has request and group objects
		21	use IO::AIO 2;
		22
		23	aioreq_pri 4; # give next request a very high priority
		24	my $req = aio_unlink "/tmp/file", sub { };
		25	$req->cancel; # cancel request if still in queue
		26
		27	my $grp = aio_group sub { print "all stats done\n" };
		28	add $grp aio_stat "..." for ...;
		29
		30	# AnyEvent integration
		31	open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!";
		32	my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb });
		33
		34	# Event integration
21	Event->io (fd => IO::AIO::poll_fileno,	35	Event->io (fd => IO::AIO::poll_fileno,
22	poll => 'r', async => 1,	36	poll => 'r',
23	cb => \&IO::AIO::poll_cb);	37	cb => \&IO::AIO::poll_cb);
24		38
25	# Glib/Gtk2	39	# Glib/Gtk2 integration
26	add_watch Glib::IO IO::AIO::poll_fileno,	40	add_watch Glib::IO IO::AIO::poll_fileno,
27	\&IO::AIO::poll_cb;	41	in => sub { IO::AIO::poll_cb; 1 };
28		42
29	# Tk	43	# Tk integration
30	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",	44	Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "",
31	readable => \&IO::AIO::poll_cb);	45	readable => \&IO::AIO::poll_cb);
32		46
		47	# Danga::Socket integration
		48	Danga::Socket->AddOtherFds (IO::AIO::poll_fileno =>
		49	\&IO::AIO::poll_cb);
		50
33	=head1 DESCRIPTION	51	=head1 DESCRIPTION
34		52
35	This module implements asynchronous I/O using whatever means your	53	This module implements asynchronous I/O using whatever means your
36	operating system supports.	54	operating system supports.
37		55
		56	Asynchronous means that operations that can normally block your program
		57	(e.g. reading from disk) will be done asynchronously: the operation
		58	will still block, but you can do something else in the meantime. This
		59	is extremely useful for programs that need to stay interactive even
		60	when doing heavy I/O (GUI programs, high performance network servers
		61	etc.), but can also be used to easily do operations in parallel that are
		62	normally done sequentially, e.g. stat'ing many files, which is much faster
		63	on a RAID volume or over NFS when you do a number of stat operations
		64	concurrently.
		65
		66	While most of this works on all types of file descriptors (for example
		67	sockets), using these functions on file descriptors that support
		68	nonblocking operation (again, sockets, pipes etc.) is very inefficient or
		69	might not work (aio_read fails on sockets/pipes/fifos). Use an event loop
		70	for that (such as the L<Event\|Event> module): IO::AIO will naturally fit
		71	into such an event loop itself.
		72
38	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
39	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
40	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
41	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
42	aio functions available on many operating systems. However, they are often	77	functions available on many operating systems. However, they are often
43	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
44	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
45	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.
46		82
47	Although the module will work with in the presence of other threads, it is	83	Although the module will work with in the presence of other (Perl-)
48	currently not reentrant, so use appropriate locking yourself.	84	threads, it is currently not reentrant in any way, so use appropriate
		85	locking yourself, always call C<poll_cb> from within the same thread, or
		86	never 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 = $_[0]
		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
49		183
50	=cut	184	=cut
51		185
52	package IO::AIO;	186	package IO::AIO;
53		187
		188	no warnings;
		189	use strict 'vars';
		190
54	use base 'Exporter';	191	use base 'Exporter';
55		192
56	use Fcntl ();
57
58	BEGIN {	193	BEGIN {
59	$VERSION = 0.2;	194	our $VERSION = '2.2';
60		195
61	@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
62	aio_fsync aio_fdatasync aio_readahead);	197	aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink
63	@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);
		200	our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice));
		201	our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush
		202	min_parallel max_parallel max_idle
		203	nreqs nready npending nthreads
		204	max_poll_time max_poll_reqs);
		205
		206	@IO::AIO::GRP::ISA = 'IO::AIO::REQ';
64		207
65	require XSLoader;	208	require XSLoader;
66	XSLoader::load IO::AIO, $VERSION;	209	XSLoader::load ("IO::AIO", $VERSION);
67	}	210	}
68		211
69	=head1 FUNCTIONS	212	=head1 FUNCTIONS
70		213
71	=head2 AIO FUNCTIONS	214	=head2 AIO REQUEST FUNCTIONS
72		215
73	All the C<aio_*> calls are more or less thin wrappers around the syscall	216	All the C<aio_*> calls are more or less thin wrappers around the syscall
74	with the same name (sans C<aio_>). The arguments are similar or identical,	217	with the same name (sans C<aio_>). The arguments are similar or identical,
75	and they all accept an additional C<$callback> argument which must be	218	and they all accept an additional (and optional) C<$callback> argument
76	a code reference. This code reference will get called with the syscall	219	which must be a code reference. This code reference will get called with
77	return code (e.g. most syscalls return C<-1> on error, unlike perl, which	220	the syscall return code (e.g. most syscalls return C<-1> on error, unlike
78	usually delivers "false") as it's sole argument when the given syscall has	221	perl, which usually delivers "false") as it's sole argument when the given
79	been executed asynchronously.	222	syscall has been executed asynchronously.
80		223
81	All functions that expect a filehandle will also accept a file descriptor.	224	All functions expecting a filehandle keep a copy of the filehandle
		225	internally until the request has finished.
82		226
		227	All functions return request objects of type L<IO::AIO::REQ> that allow
		228	further manipulation of those requests while they are in-flight.
		229
83	The filenames you pass to these routines I<must> be absolute. The reason	230	The pathnames you pass to these routines I<must> be absolute and
84	is that at the time the request is being executed, the current working	231	encoded as octets. The reason for the former is that at the time the
85	directory could have changed. Alternatively, you can make sure that you	232	request is being executed, the current working directory could have
86	never change the current working directory.	233	changed. Alternatively, you can make sure that you never change the
		234	current working directory anywhere in the program and then use relative
		235	paths.
		236
		237	To encode pathnames as octets, either make sure you either: a) always pass
		238	in filenames you got from outside (command line, readdir etc.) without
		239	tinkering, b) are ASCII or ISO 8859-1, c) use the Encode module and encode
		240	your pathnames to the locale (or other) encoding in effect in the user
		241	environment, d) use Glib::filename_from_unicode on unicode filenames or e)
		242	use something else to ensure your scalar has the correct contents.
		243
		244	This works, btw. independent of the internal UTF-8 bit, which IO::AIO
		245	handles correctly wether it is set or not.
87		246
88	=over 4	247	=over 4
89		248
		249	=item $prev_pri = aioreq_pri [$pri]
		250
		251	Returns the priority value that would be used for the next request and, if
		252	C<$pri> is given, sets the priority for the next aio request.
		253
		254	The default priority is C<0>, the minimum and maximum priorities are C<-4>
		255	and C<4>, respectively. Requests with higher priority will be serviced
		256	first.
		257
		258	The priority will be reset to C<0> after each call to one of the C<aio_*>
		259	functions.
		260
		261	Example: open a file with low priority, then read something from it with
		262	higher priority so the read request is serviced before other low priority
		263	open requests (potentially spamming the cache):
		264
		265	aioreq_pri -3;
		266	aio_open ..., sub {
		267	return unless $_[0];
		268
		269	aioreq_pri -2;
		270	aio_read $_[0], ..., sub {
		271	...
		272	};
		273	};
		274
		275	=item aioreq_nice $pri_adjust
		276
		277	Similar to C<aioreq_pri>, but subtracts the given value from the current
		278	priority, so the effect is cumulative.
		279
90	=item aio_open $pathname, $flags, $mode, $callback	280	=item aio_open $pathname, $flags, $mode, $callback->($fh)
91		281
92	Asynchronously open or create a file and call the callback with a newly	282	Asynchronously open or create a file and call the callback with a newly
93	created filehandle for the file.	283	created filehandle for the file.
94		284
95	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,	285	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
96	for an explanation.	286	for an explanation.
97		287
98	The C<$mode> argument is a bitmask. See the C<Fcntl> module for a	288	The C<$flags> argument is a bitmask. See the C<Fcntl> module for a
99	list. They are the same as used in C<sysopen>.	289	list. They are the same as used by C<sysopen>.
		290
		291	Likewise, C<$mode> specifies the mode of the newly created file, if it
		292	didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>,
		293	except that it is mandatory (i.e. use C<0> if you don't create new files,
		294	and C<0666> or C<0777> if you do).
100		295
101	Example:	296	Example:
102		297
103	aio_open "/etc/passwd", O_RDONLY, 0, sub {	298	aio_open "/etc/passwd", O_RDONLY, 0, sub {
104	if ($_[0]) {	299	if ($_[0]) {
…		…
107	} else {	302	} else {
108	die "open failed: $!\n";	303	die "open failed: $!\n";
109	}	304	}
110	};	305	};
111		306
112	=item aio_close $fh, $callback	307	=item aio_close $fh, $callback->($status)
113		308
114	Asynchronously close a file and call the callback with the result	309	Asynchronously close a file and call the callback with the result
115	code. I<WARNING:> although accepted, you should not pass in a perl	310	code. I<WARNING:> although accepted, you should not pass in a perl
116	filehandle here, as perl will likely close the file descriptor itself when	311	filehandle here, as perl will likely close the file descriptor another
117	the filehandle is destroyed. Normally, you can safely call perls C<close>	312	time when the filehandle is destroyed. Normally, you can safely call perls
118	or just let filehandles go out of scope.	313	C<close> or just let filehandles go out of scope.
119		314
		315	This is supposed to be a bug in the API, so that might change. It's
		316	therefore best to avoid this function.
		317
120	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback	318	=item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
121		319
122	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback	320	=item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval)
123		321
124	Reads or writes C<length> bytes from the specified C<fh> and C<offset>	322	Reads or writes C<length> bytes from the specified C<fh> and C<offset>
125	into the scalar given by C<data> and offset C<dataoffset> and calls the	323	into the scalar given by C<data> and offset C<dataoffset> and calls the
126	callback without the actual number of bytes read (or -1 on error, just	324	callback without the actual number of bytes read (or -1 on error, just
127	like the syscall).	325	like the syscall).
128		326
		327	The C<$data> scalar I<MUST NOT> be modified in any way while the request
		328	is outstanding. Modifying it can result in segfaults or WW3 (if the
		329	necessary/optional hardware is installed).
		330
129	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at	331	Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at
130	offset C<0> within the scalar:	332	offset C<0> within the scalar:
131		333
132	aio_read $fh, 7, 15, $buffer, 0, sub {	334	aio_read $fh, 7, 15, $buffer, 0, sub {
133	$_[0] >= 0 or die "read error: $!";	335	$_[0] > 0 or die "read error: $!";
134	print "read <$buffer>\n";	336	print "read $_[0] bytes: <$buffer>\n";
135	};	337	};
136		338
		339	=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval)
		340
		341	Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
		342	reading at byte offset C<$in_offset>, and starts writing at the current
		343	file offset of C<$out_fh>. Because of that, it is not safe to issue more
		344	than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
		345	other.
		346
		347	This call tries to make use of a native C<sendfile> syscall to provide
		348	zero-copy operation. For this to work, C<$out_fh> should refer to a
		349	socket, and C<$in_fh> should refer to mmap'able file.
		350
		351	If the native sendfile call fails or is not implemented, it will be
		352	emulated, so you can call C<aio_sendfile> on any type of filehandle
		353	regardless of the limitations of the operating system.
		354
		355	Please note, however, that C<aio_sendfile> can read more bytes from
		356	C<$in_fh> than are written, and there is no way to find out how many
		357	bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
		358	provides the number of bytes written to C<$out_fh>. Only if the result
		359	value equals C<$length> one can assume that C<$length> bytes have been
		360	read.
		361
137	=item aio_readahead $fh,$offset,$length, $callback	362	=item aio_readahead $fh,$offset,$length, $callback->($retval)
138		363
139	Asynchronously reads the specified byte range into the page cache, using
140	the C<readahead> syscall. If that syscall doesn't exist the status will be
141	C<-1> and C<$!> is set to ENOSYS.
142
143	readahead() populates the page cache with data from a file so that	364	C<aio_readahead> populates the page cache with data from a file so that
144	subsequent reads from that file will not block on disk I/O. The C<$offset>	365	subsequent reads from that file will not block on disk I/O. The C<$offset>
145	argument specifies the starting point from which data is to be read and	366	argument specifies the starting point from which data is to be read and
146	C<$length> specifies the number of bytes to be read. I/O is performed in	367	C<$length> specifies the number of bytes to be read. I/O is performed in
147	whole pages, so that offset is effectively rounded down to a page boundary	368	whole pages, so that offset is effectively rounded down to a page boundary
148	and bytes are read up to the next page boundary greater than or equal to	369	and bytes are read up to the next page boundary greater than or equal to
149	(off-set+length). aio_readahead() does not read beyond the end of the	370	(off-set+length). C<aio_readahead> does not read beyond the end of the
150	file. The current file offset of the file is left unchanged.	371	file. The current file offset of the file is left unchanged.
151		372
		373	If that syscall doesn't exist (likely if your OS isn't Linux) it will be
		374	emulated by simply reading the data, which would have a similar effect.
		375
152	=item aio_stat $fh_or_path, $callback	376	=item aio_stat $fh_or_path, $callback->($status)
153		377
154	=item aio_lstat $fh, $callback	378	=item aio_lstat $fh, $callback->($status)
155		379
156	Works like perl's C<stat> or C<lstat> in void context. The callback will	380	Works like perl's C<stat> or C<lstat> in void context. The callback will
157	be called after the stat and the results will be available using C<stat _>	381	be called after the stat and the results will be available using C<stat _>
158	or C<-s _> etc...	382	or C<-s _> etc...
159		383
…		…
169	aio_stat "/etc/passwd", sub {	393	aio_stat "/etc/passwd", sub {
170	$_[0] and die "stat failed: $!";	394	$_[0] and die "stat failed: $!";
171	print "size is ", -s _, "\n";	395	print "size is ", -s _, "\n";
172	};	396	};
173		397
174	=item aio_unlink $pathname, $callback	398	=item aio_unlink $pathname, $callback->($status)
175		399
176	Asynchronously unlink (delete) a file and call the callback with the	400	Asynchronously unlink (delete) a file and call the callback with the
177	result code.	401	result code.
178		402
		403	=item aio_mknod $path, $mode, $dev, $callback->($status)
		404
		405	[EXPERIMENTAL]
		406
		407	Asynchronously create a device node (or fifo). See mknod(2).
		408
		409	The only (POSIX-) portable way of calling this function is:
		410
		411	aio_mknod $path, IO::AIO::S_IFIFO \| $mode, 0, sub { ...
		412
		413	=item aio_link $srcpath, $dstpath, $callback->($status)
		414
		415	Asynchronously create a new link to the existing object at C<$srcpath> at
		416	the path C<$dstpath> and call the callback with the result code.
		417
		418	=item aio_symlink $srcpath, $dstpath, $callback->($status)
		419
		420	Asynchronously create a new symbolic link to the existing object at C<$srcpath> at
		421	the path C<$dstpath> and call the callback with the result code.
		422
		423	=item aio_readlink $path, $callback->($link)
		424
		425	Asynchronously read the symlink specified by C<$path> and pass it to
		426	the callback. If an error occurs, nothing or undef gets passed to the
		427	callback.
		428
		429	=item aio_rename $srcpath, $dstpath, $callback->($status)
		430
		431	Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as
		432	rename(2) and call the callback with the result code.
		433
		434	=item aio_rmdir $pathname, $callback->($status)
		435
		436	Asynchronously rmdir (delete) a directory and call the callback with the
		437	result code.
		438
		439	=item aio_readdir $pathname, $callback->($entries)
		440
		441	Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
		442	directory (i.e. opendir + readdir + closedir). The entries will not be
		443	sorted, and will B<NOT> include the C<.> and C<..> entries.
		444
		445	The callback a single argument which is either C<undef> or an array-ref
		446	with the filenames.
		447
		448	=item aio_copy $srcpath, $dstpath, $callback->($status)
		449
		450	Try to copy the I<file> (directories not supported as either source or
		451	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		452	the C<0> (error) or C<-1> ok.
		453
		454	This is a composite request that it creates the destination file with
		455	mode 0200 and copies the contents of the source file into it using
		456	C<aio_sendfile>, followed by restoring atime, mtime, access mode and
		457	uid/gid, in that order.
		458
		459	If an error occurs, the partial destination file will be unlinked, if
		460	possible, except when setting atime, mtime, access mode and uid/gid, where
		461	errors are being ignored.
		462
		463	=cut
		464
		465	sub aio_copy($$;$) {
		466	my ($src, $dst, $cb) = @_;
		467
		468	my $pri = aioreq_pri;
		469	my $grp = aio_group $cb;
		470
		471	aioreq_pri $pri;
		472	add $grp aio_open $src, O_RDONLY, 0, sub {
		473	if (my $src_fh = $_[0]) {
		474	my @stat = stat $src_fh;
		475
		476	aioreq_pri $pri;
		477	add $grp aio_open $dst, O_CREAT \| O_WRONLY \| O_TRUNC, 0200, sub {
		478	if (my $dst_fh = $_[0]) {
		479	aioreq_pri $pri;
		480	add $grp aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub {
		481	if ($_[0] == $stat[7]) {
		482	$grp->result (0);
		483	close $src_fh;
		484
		485	# those should not normally block. should. should.
		486	utime $stat[8], $stat[9], $dst;
		487	chmod $stat[2] & 07777, $dst_fh;
		488	chown $stat[4], $stat[5], $dst_fh;
		489	close $dst_fh;
		490	} else {
		491	$grp->result (-1);
		492	close $src_fh;
		493	close $dst_fh;
		494
		495	aioreq $pri;
		496	add $grp aio_unlink $dst;
		497	}
		498	};
		499	} else {
		500	$grp->result (-1);
		501	}
		502	},
		503
		504	} else {
		505	$grp->result (-1);
		506	}
		507	};
		508
		509	$grp
		510	}
		511
		512	=item aio_move $srcpath, $dstpath, $callback->($status)
		513
		514	Try to move the I<file> (directories not supported as either source or
		515	destination) from C<$srcpath> to C<$dstpath> and call the callback with
		516	the C<0> (error) or C<-1> ok.
		517
		518	This is a composite request that tries to rename(2) the file first. If
		519	rename files with C<EXDEV>, it copies the file with C<aio_copy> and, if
		520	that is successful, unlinking the C<$srcpath>.
		521
		522	=cut
		523
		524	sub aio_move($$;$) {
		525	my ($src, $dst, $cb) = @_;
		526
		527	my $pri = aioreq_pri;
		528	my $grp = aio_group $cb;
		529
		530	aioreq_pri $pri;
		531	add $grp aio_rename $src, $dst, sub {
		532	if ($_[0] && $! == EXDEV) {
		533	aioreq_pri $pri;
		534	add $grp aio_copy $src, $dst, sub {
		535	$grp->result ($_[0]);
		536
		537	if (!$_[0]) {
		538	aioreq_pri $pri;
		539	add $grp aio_unlink $src;
		540	}
		541	};
		542	} else {
		543	$grp->result ($_[0]);
		544	}
		545	};
		546
		547	$grp
		548	}
		549
		550	=item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs)
		551
		552	Scans a directory (similar to C<aio_readdir>) but additionally tries to
		553	efficiently separate the entries of directory C<$path> into two sets of
		554	names, directories you can recurse into (directories), and ones you cannot
		555	recurse into (everything else, including symlinks to directories).
		556
		557	C<aio_scandir> is a composite request that creates of many sub requests_
		558	C<$maxreq> specifies the maximum number of outstanding aio requests that
		559	this function generates. If it is C<< <= 0 >>, then a suitable default
		560	will be chosen (currently 4).
		561
		562	On error, the callback is called without arguments, otherwise it receives
		563	two array-refs with path-relative entry names.
		564
		565	Example:
		566
		567	aio_scandir $dir, 0, sub {
		568	my ($dirs, $nondirs) = @_;
		569	print "real directories: @$dirs\n";
		570	print "everything else: @$nondirs\n";
		571	};
		572
		573	Implementation notes.
		574
		575	The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can.
		576
		577	After reading the directory, the modification time, size etc. of the
		578	directory before and after the readdir is checked, and if they match (and
		579	isn't the current time), the link count will be used to decide how many
		580	entries are directories (if >= 2). Otherwise, no knowledge of the number
		581	of subdirectories will be assumed.
		582
		583	Then entries will be sorted into likely directories (everything without
		584	a non-initial dot currently) and likely non-directories (everything
		585	else). Then every entry plus an appended C</.> will be C<stat>'ed,
		586	likely directories first. If that succeeds, it assumes that the entry
		587	is a directory or a symlink to directory (which will be checked
		588	seperately). This is often faster than stat'ing the entry itself because
		589	filesystems might detect the type of the entry without reading the inode
		590	data (e.g. ext2fs filetype feature).
		591
		592	If the known number of directories (link count - 2) has been reached, the
		593	rest of the entries is assumed to be non-directories.
		594
		595	This only works with certainty on POSIX (= UNIX) filesystems, which
		596	fortunately are the vast majority of filesystems around.
		597
		598	It will also likely work on non-POSIX filesystems with reduced efficiency
		599	as those tend to return 0 or 1 as link counts, which disables the
		600	directory counting heuristic.
		601
		602	=cut
		603
		604	sub aio_scandir($$$) {
		605	my ($path, $maxreq, $cb) = @_;
		606
		607	my $pri = aioreq_pri;
		608
		609	my $grp = aio_group $cb;
		610
		611	$maxreq = 4 if $maxreq <= 0;
		612
		613	# stat once
		614	aioreq_pri $pri;
		615	add $grp aio_stat $path, sub {
		616	return $grp->result () if $_[0];
		617	my $now = time;
		618	my $hash1 = join ":", (stat _)[0,1,3,7,9];
		619
		620	# read the directory entries
		621	aioreq_pri $pri;
		622	add $grp aio_readdir $path, sub {
		623	my $entries = shift
		624	or return $grp->result ();
		625
		626	# stat the dir another time
		627	aioreq_pri $pri;
		628	add $grp aio_stat $path, sub {
		629	my $hash2 = join ":", (stat _)[0,1,3,7,9];
		630
		631	my $ndirs;
		632
		633	# take the slow route if anything looks fishy
		634	if ($hash1 ne $hash2 or (stat _)[9] == $now) {
		635	$ndirs = -1;
		636	} else {
		637	# if nlink == 2, we are finished
		638	# on non-posix-fs's, we rely on nlink < 2
		639	$ndirs = (stat _)[3] - 2
		640	or return $grp->result ([], $entries);
		641	}
		642
		643	# sort into likely dirs and likely nondirs
		644	# dirs == files without ".", short entries first
		645	$entries = [map $_->[0],
		646	sort { $b->[1] cmp $a->[1] }
		647	map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length],
		648	@$entries];
		649
		650	my (@dirs, @nondirs);
		651
		652	my $statgrp = add $grp aio_group sub {
		653	$grp->result (\@dirs, \@nondirs);
		654	};
		655
		656	limit $statgrp $maxreq;
		657	feed $statgrp sub {
		658	return unless @$entries;
		659	my $entry = pop @$entries;
		660
		661	aioreq_pri $pri;
		662	add $statgrp aio_stat "$path/$entry/.", sub {
		663	if ($_[0] < 0) {
		664	push @nondirs, $entry;
		665	} else {
		666	# need to check for real directory
		667	aioreq_pri $pri;
		668	add $statgrp aio_lstat "$path/$entry", sub {
		669	if (-d _) {
		670	push @dirs, $entry;
		671
		672	unless (--$ndirs) {
		673	push @nondirs, @$entries;
		674	feed $statgrp;
		675	}
		676	} else {
		677	push @nondirs, $entry;
		678	}
		679	}
		680	}
		681	};
		682	};
		683	};
		684	};
		685	};
		686
		687	$grp
		688	}
		689
179	=item aio_fsync $fh, $callback	690	=item aio_fsync $fh, $callback->($status)
180		691
181	Asynchronously call fsync on the given filehandle and call the callback	692	Asynchronously call fsync on the given filehandle and call the callback
182	with the fsync result code.	693	with the fsync result code.
183		694
184	=item aio_fdatasync $fh, $callback	695	=item aio_fdatasync $fh, $callback->($status)
185		696
186	Asynchronously call fdatasync on the given filehandle and call the	697	Asynchronously call fdatasync on the given filehandle and call the
187	callback with the fdatasync result code.	698	callback with the fdatasync result code.
188		699
		700	If this call isn't available because your OS lacks it or it couldn't be
		701	detected, it will be emulated by calling C<fsync> instead.
		702
		703	=item aio_group $callback->(...)
		704
		705	This is a very special aio request: Instead of doing something, it is a
		706	container for other aio requests, which is useful if you want to bundle
		707	many requests into a single, composite, request with a definite callback
		708	and the ability to cancel the whole request with its subrequests.
		709
		710	Returns an object of class L<IO::AIO::GRP>. See its documentation below
		711	for more info.
		712
		713	Example:
		714
		715	my $grp = aio_group sub {
		716	print "all stats done\n";
		717	};
		718
		719	add $grp
		720	(aio_stat ...),
		721	(aio_stat ...),
		722	...;
		723
		724	=item aio_nop $callback->()
		725
		726	This is a special request - it does nothing in itself and is only used for
		727	side effects, such as when you want to add a dummy request to a group so
		728	that finishing the requests in the group depends on executing the given
		729	code.
		730
		731	While this request does nothing, it still goes through the execution
		732	phase and still requires a worker thread. Thus, the callback will not
		733	be executed immediately but only after other requests in the queue have
		734	entered their execution phase. This can be used to measure request
		735	latency.
		736
		737	=item IO::AIO::aio_busy $fractional_seconds, $callback->() NOT EXPORTED
		738
		739	Mainly used for debugging and benchmarking, this aio request puts one of
		740	the request workers to sleep for the given time.
		741
		742	While it is theoretically handy to have simple I/O scheduling requests
		743	like sleep and file handle readable/writable, the overhead this creates is
		744	immense (it blocks a thread for a long time) so do not use this function
		745	except to put your application under artificial I/O pressure.
		746
189	=back	747	=back
190		748
		749	=head2 IO::AIO::REQ CLASS
		750
		751	All non-aggregate C<aio_*> functions return an object of this class when
		752	called in non-void context.
		753
		754	=over 4
		755
		756	=item cancel $req
		757
		758	Cancels the request, if possible. Has the effect of skipping execution
		759	when entering the B<execute> state and skipping calling the callback when
		760	entering the the B<result> state, but will leave the request otherwise
		761	untouched. That means that requests that currently execute will not be
		762	stopped and resources held by the request will not be freed prematurely.
		763
		764	=item cb $req $callback->(...)
		765
		766	Replace (or simply set) the callback registered to the request.
		767
		768	=back
		769
		770	=head2 IO::AIO::GRP CLASS
		771
		772	This class is a subclass of L<IO::AIO::REQ>, so all its methods apply to
		773	objects of this class, too.
		774
		775	A IO::AIO::GRP object is a special request that can contain multiple other
		776	aio requests.
		777
		778	You create one by calling the C<aio_group> constructing function with a
		779	callback that will be called when all contained requests have entered the
		780	C<done> state:
		781
		782	my $grp = aio_group sub {
		783	print "all requests are done\n";
		784	};
		785
		786	You add requests by calling the C<add> method with one or more
		787	C<IO::AIO::REQ> objects:
		788
		789	$grp->add (aio_unlink "...");
		790
		791	add $grp aio_stat "...", sub {
		792	$_[0] or return $grp->result ("error");
		793
		794	# add another request dynamically, if first succeeded
		795	add $grp aio_open "...", sub {
		796	$grp->result ("ok");
		797	};
		798	};
		799
		800	This makes it very easy to create composite requests (see the source of
		801	C<aio_move> for an application) that work and feel like simple requests.
		802
		803	=over 4
		804
		805	=item * The IO::AIO::GRP objects will be cleaned up during calls to
		806	C<IO::AIO::poll_cb>, just like any other request.
		807
		808	=item * They can be canceled like any other request. Canceling will cancel not
		809	only the request itself, but also all requests it contains.
		810
		811	=item * They can also can also be added to other IO::AIO::GRP objects.
		812
		813	=item * You must not add requests to a group from within the group callback (or
		814	any later time).
		815
		816	=back
		817
		818	Their lifetime, simplified, looks like this: when they are empty, they
		819	will finish very quickly. If they contain only requests that are in the
		820	C<done> state, they will also finish. Otherwise they will continue to
		821	exist.
		822
		823	That means after creating a group you have some time to add requests. And
		824	in the callbacks of those requests, you can add further requests to the
		825	group. And only when all those requests have finished will the the group
		826	itself finish.
		827
		828	=over 4
		829
		830	=item add $grp ...
		831
		832	=item $grp->add (...)
		833
		834	Add one or more requests to the group. Any type of L<IO::AIO::REQ> can
		835	be added, including other groups, as long as you do not create circular
		836	dependencies.
		837
		838	Returns all its arguments.
		839
		840	=item $grp->cancel_subs
		841
		842	Cancel all subrequests and clears any feeder, but not the group request
		843	itself. Useful when you queued a lot of events but got a result early.
		844
		845	=item $grp->result (...)
		846
		847	Set the result value(s) that will be passed to the group callback when all
		848	subrequests have finished and set thre groups errno to the current value
		849	of errno (just like calling C<errno> without an error number). By default,
		850	no argument will be passed and errno is zero.
		851
		852	=item $grp->errno ([$errno])
		853
		854	Sets the group errno value to C<$errno>, or the current value of errno
		855	when the argument is missing.
		856
		857	Every aio request has an associated errno value that is restored when
		858	the callback is invoked. This method lets you change this value from its
		859	default (0).
		860
		861	Calling C<result> will also set errno, so make sure you either set C<$!>
		862	before the call to C<result>, or call c<errno> after it.
		863
		864	=item feed $grp $callback->($grp)
		865
		866	Sets a feeder/generator on this group: every group can have an attached
		867	generator that generates requests if idle. The idea behind this is that,
		868	although you could just queue as many requests as you want in a group,
		869	this might starve other requests for a potentially long time. For
		870	example, C<aio_scandir> might generate hundreds of thousands C<aio_stat>
		871	requests, delaying any later requests for a long time.
		872
		873	To avoid this, and allow incremental generation of requests, you can
		874	instead a group and set a feeder on it that generates those requests. The
		875	feed callback will be called whenever there are few enough (see C<limit>,
		876	below) requests active in the group itself and is expected to queue more
		877	requests.
		878
		879	The feed callback can queue as many requests as it likes (i.e. C<add> does
		880	not impose any limits).
		881
		882	If the feed does not queue more requests when called, it will be
		883	automatically removed from the group.
		884
		885	If the feed limit is C<0>, it will be set to C<2> automatically.
		886
		887	Example:
		888
		889	# stat all files in @files, but only ever use four aio requests concurrently:
		890
		891	my $grp = aio_group sub { print "finished\n" };
		892	limit $grp 4;
		893	feed $grp sub {
		894	my $file = pop @files
		895	or return;
		896
		897	add $grp aio_stat $file, sub { ... };
		898	};
		899
		900	=item limit $grp $num
		901
		902	Sets the feeder limit for the group: The feeder will be called whenever
		903	the group contains less than this many requests.
		904
		905	Setting the limit to C<0> will pause the feeding process.
		906
		907	=back
		908
191	=head2 SUPPORT FUNCTIONS	909	=head2 SUPPORT FUNCTIONS
192		910
		911	=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
		912
193	=over 4	913	=over 4
194		914
195	=item $fileno = IO::AIO::poll_fileno	915	=item $fileno = IO::AIO::poll_fileno
196		916
197	Return the I<request result pipe filehandle>. This filehandle must be	917	Return the I<request result pipe file descriptor>. This filehandle must be
198	polled for reading by some mechanism outside this module (e.g. Event	918	polled for reading by some mechanism outside this module (e.g. Event or
199	or select, see below). If the pipe becomes readable you have to call	919	select, see below or the SYNOPSIS). If the pipe becomes readable you have
200	C<poll_cb> to check the results.	920	to call C<poll_cb> to check the results.
201		921
202	See C<poll_cb> for an example.	922	See C<poll_cb> for an example.
203		923
204	=item IO::AIO::poll_cb	924	=item IO::AIO::poll_cb
205		925
206	Process all outstanding events on the result pipe. You have to call this	926	Process some outstanding events on the result pipe. You have to call this
207	regularly. Returns the number of events processed. Returns immediately	927	regularly. Returns the number of events processed. Returns immediately
208	when no events are outstanding.	928	when no events are outstanding. The amount of events processed depends on
		929	the settings of C<IO::AIO::max_poll_req> and C<IO::AIO::max_poll_time>.
209		930
210	You can use Event to multiplex, e.g.:	931	If not all requests were processed for whatever reason, the filehandle
		932	will still be ready when C<poll_cb> returns.
		933
		934	Example: Install an Event watcher that automatically calls
		935	IO::AIO::poll_cb with high priority:
211		936
212	Event->io (fd => IO::AIO::poll_fileno,	937	Event->io (fd => IO::AIO::poll_fileno,
213	poll => 'r', async => 1,	938	poll => 'r', async => 1,
214	cb => \&IO::AIO::poll_cb);	939	cb => \&IO::AIO::poll_cb);
215		940
		941	=item IO::AIO::max_poll_reqs $nreqs
		942
		943	=item IO::AIO::max_poll_time $seconds
		944
		945	These set the maximum number of requests (default C<0>, meaning infinity)
		946	that are being processed by C<IO::AIO::poll_cb> in one call, respectively
		947	the maximum amount of time (default C<0>, meaning infinity) spent in
		948	C<IO::AIO::poll_cb> to process requests (more correctly the mininum amount
		949	of time C<poll_cb> is allowed to use).
		950
		951	Setting C<max_poll_time> to a non-zero value creates an overhead of one
		952	syscall per request processed, which is not normally a problem unless your
		953	callbacks are really really fast or your OS is really really slow (I am
		954	not mentioning Solaris here). Using C<max_poll_reqs> incurs no overhead.
		955
		956	Setting these is useful if you want to ensure some level of
		957	interactiveness when perl is not fast enough to process all requests in
		958	time.
		959
		960	For interactive programs, values such as C<0.01> to C<0.1> should be fine.
		961
		962	Example: Install an Event watcher that automatically calls
		963	IO::AIO::poll_cb with low priority, to ensure that other parts of the
		964	program get the CPU sometimes even under high AIO load.
		965
		966	# try not to spend much more than 0.1s in poll_cb
		967	IO::AIO::max_poll_time 0.1;
		968
		969	# use a low priority so other tasks have priority
		970	Event->io (fd => IO::AIO::poll_fileno,
		971	poll => 'r', nice => 1,
		972	cb => &IO::AIO::poll_cb);
		973
216	=item IO::AIO::poll_wait	974	=item IO::AIO::poll_wait
217		975
218	Wait till the result filehandle becomes ready for reading (simply does a	976	Wait till the result filehandle becomes ready for reading (simply does a
219	select on the filehandle. This is useful if you want to synchronously wait	977	C<select> on the filehandle. This is useful if you want to synchronously
220	for some requests to finish).	978	wait for some requests to finish).
221		979
222	See C<nreqs> for an example.	980	See C<nreqs> for an example.
223		981
		982	=item IO::AIO::poll
		983
		984	Waits until some requests have been handled.
		985
		986	Strictly equivalent to:
		987
		988	IO::AIO::poll_wait, IO::AIO::poll_cb
		989	if IO::AIO::nreqs;
		990
224	=item IO::AIO::nreqs	991	=item IO::AIO::flush
225		992
226	Returns the number of requests currently outstanding.	993	Wait till all outstanding AIO requests have been handled.
227		994
228	Example: wait till there are no outstanding requests anymore:	995	Strictly equivalent to:
229		996
230	IO::AIO::poll_wait, IO::AIO::poll_cb	997	IO::AIO::poll_wait, IO::AIO::poll_cb
231	while IO::AIO::nreqs;	998	while IO::AIO::nreqs;
232		999
		1000	=head3 CONTROLLING THE NUMBER OF THREADS
		1001
233	=item IO::AIO::min_parallel $nthreads	1002	=item IO::AIO::min_parallel $nthreads
234		1003
235	Set the minimum number of AIO threads to C<$nthreads>. The default is	1004	Set the minimum number of AIO threads to C<$nthreads>. The current
236	C<1>, which means a single asynchronous operation can be done at one time	1005	default is C<8>, which means eight asynchronous operations can execute
237	(the number of outstanding operations, however, is unlimited).	1006	concurrently at any one time (the number of outstanding requests,
		1007	however, is unlimited).
238		1008
		1009	IO::AIO starts threads only on demand, when an AIO request is queued and
		1010	no free thread exists. Please note that queueing up a hundred requests can
		1011	create demand for a hundred threads, even if it turns out that everything
		1012	is in the cache and could have been processed faster by a single thread.
		1013
239	It is recommended to keep the number of threads low, as some Linux	1014	It is recommended to keep the number of threads relatively low, as some
240	kernel versions will scale negatively with the number of threads (higher	1015	Linux kernel versions will scale negatively with the number of threads
241	parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32	1016	(higher parallelity => MUCH higher latency). With current Linux 2.6
242	threads should be fine.	1017	versions, 4-32 threads should be fine.
243		1018
244	Under normal circumstances you don't need to call this function, as this	1019	Under most circumstances you don't need to call this function, as the
245	module automatically starts some threads (the exact number might change,	1020	module selects a default that is suitable for low to moderate load.
246	and is currently 4).
247		1021
248	=item IO::AIO::max_parallel $nthreads	1022	=item IO::AIO::max_parallel $nthreads
249		1023
250	Sets the maximum number of AIO threads to C<$nthreads>. If more than	1024	Sets the maximum number of AIO threads to C<$nthreads>. If more than the
251	the specified number of threads are currently running, kill them. This	1025	specified number of threads are currently running, this function kills
252	function blocks until the limit is reached.	1026	them. This function blocks until the limit is reached.
		1027
		1028	While C<$nthreads> are zero, aio requests get queued but not executed
		1029	until the number of threads has been increased again.
253		1030
254	This module automatically runs C<max_parallel 0> at program end, to ensure	1031	This module automatically runs C<max_parallel 0> at program end, to ensure
255	that all threads are killed and that there are no outstanding requests.	1032	that all threads are killed and that there are no outstanding requests.
256		1033
257	Under normal circumstances you don't need to call this function.	1034	Under normal circumstances you don't need to call this function.
258		1035
		1036	=item IO::AIO::max_idle $nthreads
		1037
		1038	Limit the number of threads (default: 4) that are allowed to idle (i.e.,
		1039	threads that did not get a request to process within 10 seconds). That
		1040	means if a thread becomes idle while C<$nthreads> other threads are also
		1041	idle, it will free its resources and exit.
		1042
		1043	This is useful when you allow a large number of threads (e.g. 100 or 1000)
		1044	to allow for extremely high load situations, but want to free resources
		1045	under normal circumstances (1000 threads can easily consume 30MB of RAM).
		1046
		1047	The default is probably ok in most situations, especially if thread
		1048	creation is fast. If thread creation is very slow on your system you might
		1049	want to use larger values.
		1050
259	=item $oldnreqs = IO::AIO::max_outstanding $nreqs	1051	=item $oldmaxreqs = IO::AIO::max_outstanding $maxreqs
		1052
		1053	This is a very bad function to use in interactive programs because it
		1054	blocks, and a bad way to reduce concurrency because it is inexact: Better
		1055	use an C<aio_group> together with a feed callback.
260		1056
261	Sets the maximum number of outstanding requests to C<$nreqs>. If you	1057	Sets the maximum number of outstanding requests to C<$nreqs>. If you
262	try to queue up more than this number of requests, the caller will block until	1058	to queue up more than this number of requests, the next call to the
263	some requests have been handled.	1059	C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>)
		1060	function will block until the limit is no longer exceeded.
264		1061
265	The default is very large, so normally there is no practical limit. If you	1062	The default value is very large, so there is no practical limit on the
266	queue up many requests in a loop it it often improves speed if you set	1063	number of outstanding requests.
267	this to a relatively low number, such as C<100>.
268		1064
269	Under normal circumstances you don't need to call this function.	1065	You can still queue as many requests as you want. Therefore,
		1066	C<max_oustsanding> is mainly useful in simple scripts (with low values) or
		1067	as a stop gap to shield against fatal memory overflow (with large values).
		1068
		1069	=head3 STATISTICAL INFORMATION
		1070
		1071	=item IO::AIO::nreqs
		1072
		1073	Returns the number of requests currently in the ready, execute or pending
		1074	states (i.e. for which their callback has not been invoked yet).
		1075
		1076	Example: wait till there are no outstanding requests anymore:
		1077
		1078	IO::AIO::poll_wait, IO::AIO::poll_cb
		1079	while IO::AIO::nreqs;
		1080
		1081	=item IO::AIO::nready
		1082
		1083	Returns the number of requests currently in the ready state (not yet
		1084	executed).
		1085
		1086	=item IO::AIO::npending
		1087
		1088	Returns the number of requests currently in the pending state (executed,
		1089	but not yet processed by poll_cb).
270		1090
271	=back	1091	=back
272		1092
273	=cut	1093	=cut
274		1094
275	# support function to convert a fd into a perl filehandle	1095	# support function to convert a fd into a perl filehandle
276	sub _fd2fh {	1096	sub _fd2fh {
277	return undef if $_[0] < 0;	1097	return undef if $_[0] < 0;
278		1098
279	# try to be perl5.6-compatible	1099	# try to generate nice filehandles
280	local *AIO_FH;	1100	my $sym = "IO::AIO::fd#$_[0]";
281	open AIO_FH, "+<&=$_[0]"	1101	local *$sym;
		1102
		1103	open *$sym, "+<&=$_[0]" # usually works under any unix
		1104	or open *$sym, "<&=$_[0]" # cygwin needs this
		1105	or open *$sym, ">&=$_[0]" # or this
282	or return undef;	1106	or return undef;
283		1107
284	*AIO_FH	1108	*$sym
285	}	1109	}
286		1110
287	min_parallel 4;	1111	min_parallel 8;
288		1112
289	END {	1113	END {
290	max_parallel 0;	1114	min_parallel 1;
291	}	1115	flush;
		1116	};
292		1117
293	1;	1118	1;
294		1119
		1120	=head2 FORK BEHAVIOUR
		1121
		1122	This module should do "the right thing" when the process using it forks:
		1123
		1124	Before the fork, IO::AIO enters a quiescent state where no requests
		1125	can be added in other threads and no results will be processed. After
		1126	the fork the parent simply leaves the quiescent state and continues
		1127	request/result processing, while the child frees the request/result queue
		1128	(so that the requests started before the fork will only be handled in the
		1129	parent). Threads will be started on demand until the limit set in the
		1130	parent process has been reached again.
		1131
		1132	In short: the parent will, after a short pause, continue as if fork had
		1133	not been called, while the child will act as if IO::AIO has not been used
		1134	yet.
		1135
		1136	=head2 MEMORY USAGE
		1137
		1138	Per-request usage:
		1139
		1140	Each aio request uses - depending on your architecture - around 100-200
		1141	bytes of memory. In addition, stat requests need a stat buffer (possibly
		1142	a few hundred bytes), readdir requires a result buffer and so on. Perl
		1143	scalars and other data passed into aio requests will also be locked and
		1144	will consume memory till the request has entered the done state.
		1145
		1146	This is now awfully much, so queuing lots of requests is not usually a
		1147	problem.
		1148
		1149	Per-thread usage:
		1150
		1151	In the execution phase, some aio requests require more memory for
		1152	temporary buffers, and each thread requires a stack and other data
		1153	structures (usually around 16k-128k, depending on the OS).
		1154
		1155	=head1 KNOWN BUGS
		1156
		1157	Known bugs will be fixed in the next release.
		1158
295	=head1 SEE ALSO	1159	=head1 SEE ALSO
296		1160
297	L<Coro>, L<Linux::AIO>.	1161	L<Coro::AIO>.
298		1162
299	=head1 AUTHOR	1163	=head1 AUTHOR
300		1164
301	Marc Lehmann <schmorp@schmorp.de>	1165	Marc Lehmann <schmorp@schmorp.de>
302	http://home.schmorp.de/	1166	http://home.schmorp.de/

Diff Legend

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

Comparing IO-AIO/AIO.pm (file contents): Revision 1.6 by root, Sun Jul 10 22:19:48 2005 UTC vs. Revision 1.91 by root, Mon Oct 30 23:30:29 2006 UTC

Diff Legend

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.6 by root, Sun Jul 10 22:19:48 2005 UTC vs.
Revision 1.91 by root, Mon Oct 30 23:30:29 2006 UTC