ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/Linux-AIO/AIO.pm
Revision: 1.34
Committed: Sun Jul 10 17:07:34 2005 UTC (18 years, 10 months ago) by root
Branch: MAIN
CVS Tags: rel-1_71
Changes since 1.33: +4 -1 lines
Log Message:
*** empty log message ***

File Contents

# 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 1.31 $VERSION = 1.71;
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     open my $fh, "<&$_[0]"; # create a copy for perl
144     aio_close $_[0], sub { }; # close the aio handle
145     print "open successful, fh is $fh\n";
146     ...
147     } else {
148     die "open failed: $!\n";
149     }
150     };
151    
152 root 1.14 =item aio_close $fh, $callback
153 root 1.10
154     Asynchronously close a file and call the callback with the result code.
155    
156 root 1.14 =item aio_read $fh,$offset,$length, $data,$dataoffset,$callback
157 root 1.9
158 root 1.14 =item aio_write $fh,$offset,$length, $data,$dataoffset,$callback
159 root 1.4
160     Reads or writes C<length> bytes from the specified C<fh> and C<offset>
161     into the scalar given by C<data> and offset C<dataoffset> and calls the
162 root 1.30 callback without the actual number of bytes read (or -1 on error, just
163     like the syscall).
164    
165     Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at
166     offset C<0> within the scalar:
167    
168     aio_read $fh, 7, 15, $buffer, 0, sub {
169     $_[0] >= 0 or die "read error: $!";
170     print "read <$buffer>\n";
171     };
172    
173     =item aio_readahead $fh,$offset,$length, $callback
174    
175     Asynchronously reads the specified byte range into the page cache, using
176     the C<readahead> syscall.
177    
178     readahead() populates the page cache with data from a file so that
179     subsequent reads from that file will not block on disk I/O. The C<$offset>
180     argument specifies the starting point from which data is to be read and
181     C<$length> specifies the number of bytes to be read. I/O is performed in
182     whole pages, so that offset is effectively rounded down to a page boundary
183     and bytes are read up to the next page boundary greater than or equal to
184     (off-set+length). aio_readahead() does not read beyond the end of the
185     file. The current file offset of the file is left unchanged.
186 root 1.4
187 root 1.14 =item aio_stat $fh_or_path, $callback
188 root 1.4
189 root 1.14 =item aio_lstat $fh, $callback
190 root 1.5
191 root 1.14 Works like perl's C<stat> or C<lstat> in void context. The callback will
192     be called after the stat and the results will be available using C<stat _>
193     or C<-s _> etc...
194 root 1.5
195 root 1.32 The pathname passed to C<aio_stat> must be absolute. See API NOTES, above,
196     for an explanation.
197    
198 root 1.10 Currently, the stats are always 64-bit-stats, i.e. instead of returning an
199     error when stat'ing a large file, the results will be silently truncated
200     unless perl itself is compiled with large file support.
201 root 1.17
202 root 1.30 Example: Print the length of F</etc/passwd>:
203    
204     aio_stat "/etc/passwd", sub {
205     $_[0] and die "stat failed: $!";
206     print "size is ", -s _, "\n";
207     };
208    
209     =item aio_unlink $pathname, $callback
210    
211     Asynchronously unlink (delete) a file and call the callback with the
212     result code.
213    
214     =item aio_fsync $fh, $callback
215    
216     Asynchronously call fsync on the given filehandle and call the callback
217     with the fsync result code.
218    
219     =item aio_fdatasync $fh, $callback
220 root 1.17
221 root 1.30 Asynchronously call fdatasync on the given filehandle and call the
222     callback with the fdatasync result code.
223 root 1.2
224 root 1.1 =cut
225 root 1.16
226     min_parallel 1;
227 root 1.2
228     END {
229     max_parallel 0;
230     }
231 root 1.1
232     1;
233    
234     =back
235    
236     =head1 BUGS
237    
238 root 1.14 This module has been extensively tested in a large and very busy webserver
239     for many years now.
240 root 1.1
241 root 1.14 - aio_open gives a fd, but all other functions expect a perl filehandle.
242 root 1.6
243 root 1.1 =head1 SEE ALSO
244    
245 root 1.34 L<Coro>, L<IO::AIO>.
246 root 1.1
247     =head1 AUTHOR
248    
249 root 1.22 Marc Lehmann <schmorp@schmorp.de>
250 root 1.20 http://home.schmorp.de/
251 root 1.1
252     =cut
253