ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Linux-AIO/AIO.pm
Revision: 1.31
Committed: Sun Jul 10 01:03:06 2005 UTC (18 years, 10 months ago) by root
Branch: MAIN
Changes since 1.30: +1 -1 lines
Log Message:
*** empty log message ***

File Contents

# Content
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 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
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
22 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 Although the module will work with in the presence of other threads, it is
27 not reentrant, so use appropriate locking yourself.
28
29 =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 =over 4
42
43 =cut
44
45 package Linux::AIO;
46
47 use base 'Exporter';
48
49 BEGIN {
50 $VERSION = 1.71;
51
52 @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink
53 aio_fsync aio_fdatasync aio_readahead);
54 @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs);
55
56 require XSLoader;
57 XSLoader::load Linux::AIO, $VERSION;
58 }
59
60 =item Linux::AIO::min_parallel $nthreads
61
62 Set the minimum number of AIO threads to C<$nthreads>. The default is
63 C<1>, which means a single asynchronous operation can be done at one time
64 (the number of outstanding operations, however, is unlimited).
65
66 It is recommended to keep the number of threads low, as some linux
67 kernel versions will scale negatively with the number of threads (higher
68 parallelity => MUCH higher latency).
69
70 Under normal circumstances you don't need to call this function, as this
71 module automatically starts a single async thread.
72
73 =item Linux::AIO::max_parallel $nthreads
74
75 Sets the maximum number of AIO threads to C<$nthreads>. If more than
76 the specified number of threads are currently running, kill them. This
77 function blocks until the limit is reached.
78
79 This module automatically runs C<max_parallel 0> at program end, to ensure
80 that all threads are killed and that there are no outstanding requests.
81
82 Under normal circumstances you don't need to call this function.
83
84 =item $fileno = Linux::AIO::poll_fileno
85
86 Return the I<request result pipe filehandle>. This filehandle must be
87 polled for reading by some mechanism outside this module (e.g. Event
88 or select, see below). If the pipe becomes readable you have to call
89 C<poll_cb> to check the results.
90
91 See C<poll_cb> for an example.
92
93 =item Linux::AIO::poll_cb
94
95 Process all outstanding events on the result pipe. You have to call this
96 regularly. Returns the number of events processed. Returns immediately
97 when no events are outstanding.
98
99 You can use Event to multiplex, e.g.:
100
101 Event->io (fd => Linux::AIO::poll_fileno,
102 poll => 'r', async => 1,
103 cb => \&Linux::AIO::poll_cb);
104
105 =item Linux::AIO::poll_wait
106
107 Wait till the result filehandle becomes ready for reading (simply does a
108 select on the filehandle. This is useful if you want to synchronously wait
109 for some requests to finish).
110
111 See C<nreqs> for an example.
112
113 =item Linux::AIO::nreqs
114
115 Returns the number of requests currently outstanding.
116
117 Example: wait till there are no outstanding requests anymore:
118
119 Linux::AIO::poll_wait while Linux::AIO::nreqs;
120
121 =item aio_open $pathname, $flags, $mode, $callback
122
123 Asynchronously open or create a file and call the callback with the
124 filedescriptor (NOT a perl filehandle, sorry for that, but watch out, this
125 might change in the future).
126
127 The C<$mode> argument is a bitmask. See the C<Fcntl> module for a
128 list. They are the same as used in C<sysopen>.
129
130 Example:
131
132 aio_open "/etc/passwd", O_RDONLY, 0, sub {
133 if ($_[0] >= 0) {
134 open my $fh, "<&$_[0]"; # create a copy for perl
135 aio_close $_[0], sub { }; # close the aio handle
136 print "open successful, fh is $fh\n";
137 ...
138 } else {
139 die "open failed: $!\n";
140 }
141 };
142
143 =item aio_close $fh, $callback
144
145 Asynchronously close a file and call the callback with the result code.
146
147 =item aio_read $fh,$offset,$length, $data,$dataoffset,$callback
148
149 =item aio_write $fh,$offset,$length, $data,$dataoffset,$callback
150
151 Reads or writes C<length> bytes from the specified C<fh> and C<offset>
152 into the scalar given by C<data> and offset C<dataoffset> and calls the
153 callback without the actual number of bytes read (or -1 on error, just
154 like the syscall).
155
156 Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at
157 offset C<0> within the scalar:
158
159 aio_read $fh, 7, 15, $buffer, 0, sub {
160 $_[0] >= 0 or die "read error: $!";
161 print "read <$buffer>\n";
162 };
163
164 =item aio_readahead $fh,$offset,$length, $callback
165
166 Asynchronously reads the specified byte range into the page cache, using
167 the C<readahead> syscall.
168
169 readahead() populates the page cache with data from a file so that
170 subsequent reads from that file will not block on disk I/O. The C<$offset>
171 argument specifies the starting point from which data is to be read and
172 C<$length> specifies the number of bytes to be read. I/O is performed in
173 whole pages, so that offset is effectively rounded down to a page boundary
174 and bytes are read up to the next page boundary greater than or equal to
175 (off-set+length). aio_readahead() does not read beyond the end of the
176 file. The current file offset of the file is left unchanged.
177
178 =item aio_stat $fh_or_path, $callback
179
180 =item aio_lstat $fh, $callback
181
182 Works like perl's C<stat> or C<lstat> in void context. The callback will
183 be called after the stat and the results will be available using C<stat _>
184 or C<-s _> etc...
185
186 Currently, the stats are always 64-bit-stats, i.e. instead of returning an
187 error when stat'ing a large file, the results will be silently truncated
188 unless perl itself is compiled with large file support.
189
190 Example: Print the length of F</etc/passwd>:
191
192 aio_stat "/etc/passwd", sub {
193 $_[0] and die "stat failed: $!";
194 print "size is ", -s _, "\n";
195 };
196
197 =item aio_unlink $pathname, $callback
198
199 Asynchronously unlink (delete) a file and call the callback with the
200 result code.
201
202 =item aio_fsync $fh, $callback
203
204 Asynchronously call fsync on the given filehandle and call the callback
205 with the fsync result code.
206
207 =item aio_fdatasync $fh, $callback
208
209 Asynchronously call fdatasync on the given filehandle and call the
210 callback with the fdatasync result code.
211
212 =cut
213
214 min_parallel 1;
215
216 END {
217 max_parallel 0;
218 }
219
220 1;
221
222 =back
223
224 =head1 BUGS
225
226 This module has been extensively tested in a large and very busy webserver
227 for many years now.
228
229 - aio_open gives a fd, but all other functions expect a perl filehandle.
230
231 =head1 SEE ALSO
232
233 L<Coro>.
234
235 =head1 AUTHOR
236
237 Marc Lehmann <schmorp@schmorp.de>
238 http://home.schmorp.de/
239
240 =cut
241