Linux-AIO/AIO.pm

=head1 NAME

Linux::AIO - linux-specific aio implemented using clone

=head1 SYNOPSIS

 use Linux::AIO;

=head1 DESCRIPTION

This module implements asynchronous I/O using the means available to Linux
- clone. It does not hook into the POSIX aio_* functions because Linux
does not yet support these in the kernel (even as of 2.6.12, only O_DIRECT
files are supported) and even if, it would only allow aio_read and write,
not open, stat and so on.

Instead, in this module a number of (non-posix) threads are started that
execute your read/writes and signal their completion. You don't need
thread support in your libc or perl, and the threads created by this
module will not be visible to the pthreads library.

NOTICE: the threads created by this module will automatically be killed
when the thread calling min_parallel exits. Make sure you only ever call
min_parallel from the same thread that loaded this module.

Although the module will work with in the presence of other threads, it is
not reentrant, so use appropriate locking yourself.

=head2 API NOTES

All the C<aio_*> calls are more or less thin wrappers around the syscall
with the same name (sans C<aio_>). The arguments are similar or identical,
and they all accept an additional C<$callback> argument which must be
a code reference. This code reference will get called with the syscall
return code (e.g. most syscalls return C<-1> on error, unlike perl, which
usually delivers "false") as it's sole argument when the given syscall has
been executed asynchronously.

All functions that expect a filehandle will also accept a file descriptor.

The filenames you pass to these routines I<must> be absolute. The reason
is that at the time the request is being executed, the current working
directory could have changed. Alternatively, you can make sure that you
never change the current working directory.

=over 4

=cut

package Linux::AIO;

use base 'Exporter';

BEGIN {
   $VERSION = 1.72;

   @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink
                aio_fsync aio_fdatasync aio_readahead);
   @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs);

   require XSLoader;
   XSLoader::load Linux::AIO, $VERSION;
}

=item Linux::AIO::min_parallel $nthreads

Set the minimum number of AIO threads to C<$nthreads>. The default is
C<1>, which means a single asynchronous operation can be done at one time
(the number of outstanding operations, however, is unlimited).

It is recommended to keep the number of threads low, as some linux
kernel versions will scale negatively with the number of threads (higher
parallelity => MUCH higher latency).

Under normal circumstances you don't need to call this function, as this
module automatically starts a single async thread.

=item Linux::AIO::max_parallel $nthreads

Sets the maximum number of AIO threads to C<$nthreads>. If more than
the specified number of threads are currently running, kill them. This
function blocks until the limit is reached.

This module automatically runs C<max_parallel 0> at program end, to ensure
that all threads are killed and that there are no outstanding requests.

Under normal circumstances you don't need to call this function.

=item $fileno = Linux::AIO::poll_fileno

Return the I<request result pipe filehandle>. This filehandle must be
polled for reading by some mechanism outside this module (e.g. Event
or select, see below). If the pipe becomes readable you have to call
C<poll_cb> to check the results.

See C<poll_cb> for an example.

=item Linux::AIO::poll_cb

Process all outstanding events on the result pipe. You have to call this
regularly. Returns the number of events processed. Returns immediately
when no events are outstanding.

You can use Event to multiplex, e.g.:

   Event->io (fd => Linux::AIO::poll_fileno,
              poll => 'r', async => 1,
              cb => \&Linux::AIO::poll_cb);

=item Linux::AIO::poll_wait

Wait till the result filehandle becomes ready for reading (simply does a
select on the filehandle. This is useful if you want to synchronously wait
for some requests to finish).

See C<nreqs> for an example.

=item Linux::AIO::nreqs

Returns the number of requests currently outstanding.

Example: wait till there are no outstanding requests anymore:

   Linux::AIO::poll_wait, Linux::AIO::poll_cb
      while Linux::AIO::nreqs;

=item aio_open $pathname, $flags, $mode, $callback

Asynchronously open or create a file and call the callback with the
filedescriptor (NOT a perl filehandle, sorry for that, but watch out, this
might change in the future).

The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
for an explanation.

The C<$mode> argument is a bitmask. See the C<Fcntl> module for a
list. They are the same as used in C<sysopen>.

Example:

   aio_open "/etc/passwd", O_RDONLY, 0, sub {
      if ($_[0] >= 0) {
         open my $fh, "<&=$_[0]";
         print "open successful, fh is $fh\n";
         ...
      } else {
         die "open failed: $!\n";
      }
   };

=item aio_close $fh, $callback

Asynchronously close a file and call the callback with the result code.

=item aio_read  $fh,$offset,$length, $data,$dataoffset,$callback

=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback

Reads or writes C<length> bytes from the specified C<fh> and C<offset>
into the scalar given by C<data> and offset C<dataoffset> and calls the
callback without the actual number of bytes read (or -1 on error, just
like the syscall).

Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at
offset C<0> within the scalar:

   aio_read $fh, 7, 15, $buffer, 0, sub {
      $_[0] >= 0 or die "read error: $!";
      print "read <$buffer>\n";
   };

=item aio_readahead $fh,$offset,$length, $callback

Asynchronously reads the specified byte range into the page cache, using
the C<readahead> syscall.

readahead() populates the page cache with data from a file so that
subsequent reads from that file will not block on disk I/O. The C<$offset>
argument specifies the starting point from which data is to be read and
C<$length> specifies the number of bytes to be read. I/O is performed in
whole pages, so that offset is effectively rounded down to a page boundary
and bytes are read up to the next page boundary greater than or equal to
(off-set+length). aio_readahead() does not read beyond the end of the
file. The current file offset of the file is left unchanged.

=item aio_stat  $fh_or_path, $callback

=item aio_lstat $fh, $callback

Works like perl's C<stat> or C<lstat> in void context. The callback will
be called after the stat and the results will be available using C<stat _>
or C<-s _> etc...

The pathname passed to C<aio_stat> must be absolute. See API NOTES, above,
for an explanation.

Currently, the stats are always 64-bit-stats, i.e. instead of returning an
error when stat'ing a large file, the results will be silently truncated
unless perl itself is compiled with large file support.

Example: Print the length of F</etc/passwd>:

   aio_stat "/etc/passwd", sub {
      $_[0] and die "stat failed: $!";
      print "size is ", -s _, "\n";
   };

=item aio_unlink $pathname, $callback

Asynchronously unlink (delete) a file and call the callback with the
result code.

=item aio_fsync $fh, $callback

Asynchronously call fsync on the given filehandle and call the callback
with the fsync result code.

=item aio_fdatasync $fh, $callback

Asynchronously call fdatasync on the given filehandle and call the
callback with the fdatasync result code.

=cut

min_parallel 1;

END {
   max_parallel 0;
}

1;

=back

=head1 BUGS

This module has been extensively tested in a large and very busy webserver
for many years now.

   - aio_open gives a fd, but all other functions expect a perl filehandle.

=head1 SEE ALSO

L<Coro>, L<IO::AIO>.

=head1 AUTHOR

 Marc Lehmann <schmorp@schmorp.de>
 http://home.schmorp.de/

=cut

Revision:	1.35
Committed:	Tue Jul 12 10:51:09 2005 UTC (18 years, 10 months ago) by root
Branch:	MAIN
Changes since 1.34:	+2 -3 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.1	=head1 NAME
2
3			Linux::AIO - linux-specific aio implemented using clone
4
5			=head1 SYNOPSIS
6
7			use Linux::AIO;
8
9			=head1 DESCRIPTION
10
11	root	1.27	This module implements asynchronous I/O using the means available to Linux
12			- clone. It does not hook into the POSIX aio_* functions because Linux
13			does not yet support these in the kernel (even as of 2.6.12, only O_DIRECT
14			files are supported) and even if, it would only allow aio_read and write,
15			not open, stat and so on.
16	root	1.14
17			Instead, in this module a number of (non-posix) threads are started that
18			execute your read/writes and signal their completion. You don't need
19			thread support in your libc or perl, and the threads created by this
20			module will not be visible to the pthreads library.
21	root	1.3
22	root	1.17	NOTICE: the threads created by this module will automatically be killed
23			when the thread calling min_parallel exits. Make sure you only ever call
24			min_parallel from the same thread that loaded this module.
25
26	root	1.27	Although the module will work with in the presence of other threads, it is
27			not reentrant, so use appropriate locking yourself.
28	root	1.16
29	root	1.30	=head2 API NOTES
30
31			All the C<aio_*> calls are more or less thin wrappers around the syscall
32			with the same name (sans C<aio_>). The arguments are similar or identical,
33			and they all accept an additional C<$callback> argument which must be
34			a code reference. This code reference will get called with the syscall
35			return code (e.g. most syscalls return C<-1> on error, unlike perl, which
36			usually delivers "false") as it's sole argument when the given syscall has
37			been executed asynchronously.
38
39			All functions that expect a filehandle will also accept a file descriptor.
40
41	root	1.32	The filenames you pass to these routines I<must> be absolute. The reason
42			is that at the time the request is being executed, the current working
43			directory could have changed. Alternatively, you can make sure that you
44	root	1.33	never change the current working directory.
45	root	1.32
46	root	1.1	=over 4
47
48			=cut
49
50			package Linux::AIO;
51
52	root	1.2	use base 'Exporter';
53
54	root	1.1	BEGIN {
55	root	1.35	$VERSION = 1.72;
56	root	1.1
57	root	1.29	@EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink
58			aio_fsync aio_fdatasync aio_readahead);
59	root	1.2	@EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs);
60
61	root	1.1	require XSLoader;
62			XSLoader::load Linux::AIO, $VERSION;
63			}
64
65	root	1.14	=item Linux::AIO::min_parallel $nthreads
66	root	1.1
67	root	1.16	Set the minimum number of AIO threads to C<$nthreads>. The default is
68			C<1>, which means a single asynchronous operation can be done at one time
69			(the number of outstanding operations, however, is unlimited).
70	root	1.1
71	root	1.24	It is recommended to keep the number of threads low, as some linux
72	root	1.14	kernel versions will scale negatively with the number of threads (higher
73			parallelity => MUCH higher latency).
74
75	root	1.30	Under normal circumstances you don't need to call this function, as this
76			module automatically starts a single async thread.
77
78	root	1.24	=item Linux::AIO::max_parallel $nthreads
79
80			Sets the maximum number of AIO threads to C<$nthreads>. If more than
81			the specified number of threads are currently running, kill them. This
82			function blocks until the limit is reached.
83
84			This module automatically runs C<max_parallel 0> at program end, to ensure
85			that all threads are killed and that there are no outstanding requests.
86
87	root	1.30	Under normal circumstances you don't need to call this function.
88
89	root	1.2	=item $fileno = Linux::AIO::poll_fileno
90
91	root	1.14	Return the I<request result pipe filehandle>. This filehandle must be
92			polled for reading by some mechanism outside this module (e.g. Event
93			or select, see below). If the pipe becomes readable you have to call
94			C<poll_cb> to check the results.
95	root	1.2
96	root	1.30	See C<poll_cb> for an example.
97
98	root	1.2	=item Linux::AIO::poll_cb
99
100			Process all outstanding events on the result pipe. You have to call this
101	root	1.14	regularly. Returns the number of events processed. Returns immediately
102			when no events are outstanding.
103	root	1.2
104	root	1.3	You can use Event to multiplex, e.g.:
105
106	root	1.14	Event->io (fd => Linux::AIO::poll_fileno,
107			poll => 'r', async => 1,
108	root	1.30	cb => \&Linux::AIO::poll_cb);
109	root	1.3
110	root	1.24	=item Linux::AIO::poll_wait
111
112			Wait till the result filehandle becomes ready for reading (simply does a
113			select on the filehandle. This is useful if you want to synchronously wait
114			for some requests to finish).
115
116	root	1.30	See C<nreqs> for an example.
117
118	root	1.2	=item Linux::AIO::nreqs
119
120			Returns the number of requests currently outstanding.
121	root	1.4
122	root	1.30	Example: wait till there are no outstanding requests anymore:
123
124	root	1.32	Linux::AIO::poll_wait, Linux::AIO::poll_cb
125			while Linux::AIO::nreqs;
126	root	1.30
127			=item aio_open $pathname, $flags, $mode, $callback
128	root	1.10
129			Asynchronously open or create a file and call the callback with the
130	root	1.14	filedescriptor (NOT a perl filehandle, sorry for that, but watch out, this
131			might change in the future).
132	root	1.10
133	root	1.34	The pathname passed to C<aio_open> must be absolute. See API NOTES, above,
134			for an explanation.
135
136	root	1.30	The C<$mode> argument is a bitmask. See the C<Fcntl> module for a
137			list. They are the same as used in C<sysopen>.
138
139			Example:
140
141			aio_open "/etc/passwd", O_RDONLY, 0, sub {
142			if ($_[0] >= 0) {
143	root	1.35	open my $fh, "<&=$_[0]";
144	root	1.30	print "open successful, fh is $fh\n";
145			...
146			} else {
147			die "open failed: $!\n";
148			}
149			};
150
151	root	1.14	=item aio_close $fh, $callback
152	root	1.10
153			Asynchronously close a file and call the callback with the result code.
154
155	root	1.14	=item aio_read $fh,$offset,$length, $data,$dataoffset,$callback
156	root	1.9
157	root	1.14	=item aio_write $fh,$offset,$length, $data,$dataoffset,$callback
158	root	1.4
159			Reads or writes C<length> bytes from the specified C<fh> and C<offset>
160			into the scalar given by C<data> and offset C<dataoffset> and calls the
161	root	1.30	callback without the actual number of bytes read (or -1 on error, just
162			like the syscall).
163
164			Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at
165			offset C<0> within the scalar:
166
167			aio_read $fh, 7, 15, $buffer, 0, sub {
168			$_[0] >= 0 or die "read error: $!";
169			print "read <$buffer>\n";
170			};
171
172			=item aio_readahead $fh,$offset,$length, $callback
173
174			Asynchronously reads the specified byte range into the page cache, using
175			the C<readahead> syscall.
176
177			readahead() populates the page cache with data from a file so that
178			subsequent reads from that file will not block on disk I/O. The C<$offset>
179			argument specifies the starting point from which data is to be read and
180			C<$length> specifies the number of bytes to be read. I/O is performed in
181			whole pages, so that offset is effectively rounded down to a page boundary
182			and bytes are read up to the next page boundary greater than or equal to
183			(off-set+length). aio_readahead() does not read beyond the end of the
184			file. The current file offset of the file is left unchanged.
185	root	1.4
186	root	1.14	=item aio_stat $fh_or_path, $callback
187	root	1.4
188	root	1.14	=item aio_lstat $fh, $callback
189	root	1.5
190	root	1.14	Works like perl's C<stat> or C<lstat> in void context. The callback will
191			be called after the stat and the results will be available using C<stat _>
192			or C<-s _> etc...
193	root	1.5
194	root	1.32	The pathname passed to C<aio_stat> must be absolute. See API NOTES, above,
195			for an explanation.
196
197	root	1.10	Currently, the stats are always 64-bit-stats, i.e. instead of returning an
198			error when stat'ing a large file, the results will be silently truncated
199			unless perl itself is compiled with large file support.
200	root	1.17
201	root	1.30	Example: Print the length of F</etc/passwd>:
202
203			aio_stat "/etc/passwd", sub {
204			$_[0] and die "stat failed: $!";
205			print "size is ", -s _, "\n";
206			};
207
208			=item aio_unlink $pathname, $callback
209
210			Asynchronously unlink (delete) a file and call the callback with the
211			result code.
212
213			=item aio_fsync $fh, $callback
214
215			Asynchronously call fsync on the given filehandle and call the callback
216			with the fsync result code.
217
218			=item aio_fdatasync $fh, $callback
219	root	1.17
220	root	1.30	Asynchronously call fdatasync on the given filehandle and call the
221			callback with the fdatasync result code.
222	root	1.2
223	root	1.1	=cut
224	root	1.16
225			min_parallel 1;
226	root	1.2
227			END {
228			max_parallel 0;
229			}
230	root	1.1
231			1;
232
233			=back
234
235			=head1 BUGS
236
237	root	1.14	This module has been extensively tested in a large and very busy webserver
238			for many years now.
239	root	1.1
240	root	1.14	- aio_open gives a fd, but all other functions expect a perl filehandle.
241	root	1.6
242	root	1.1	=head1 SEE ALSO
243
244	root	1.34	L<Coro>, L<IO::AIO>.
245	root	1.1
246			=head1 AUTHOR
247
248	root	1.22	Marc Lehmann <schmorp@schmorp.de>
249	root	1.20	http://home.schmorp.de/
250	root	1.1
251			=cut
252