… | |
… | |
6 | |
6 | |
7 | use Linux::AIO; |
7 | use Linux::AIO; |
8 | |
8 | |
9 | =head1 DESCRIPTION |
9 | =head1 DESCRIPTION |
10 | |
10 | |
11 | This module implements asynchroneous i/o using the means available to |
11 | This module implements asynchronous I/O using the means available to Linux |
12 | linux - clone. It does not hook into the POSIX aio_* functions because |
12 | - clone. It does not hook into the POSIX aio_* functions because Linux |
13 | linux does not yet support these in the kernel. Instead, a number of |
13 | does not yet support these in the kernel (even as of 2.6.12, only O_DIRECT |
14 | threads are started that execute your read/writes and signal their |
14 | files are supported) and even if, it would only allow aio_read and write, |
15 | completion. |
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 | 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 | never change the current working directory. |
16 | |
45 | |
17 | =over 4 |
46 | =over 4 |
18 | |
47 | |
19 | =cut |
48 | =cut |
20 | |
49 | |
21 | package Linux::AIO; |
50 | package Linux::AIO; |
22 | |
51 | |
23 | use base 'Exporter'; |
52 | use base 'Exporter'; |
24 | |
53 | |
25 | BEGIN { |
54 | BEGIN { |
26 | $VERSION = 0.1; |
55 | $VERSION = 1.71; |
27 | |
56 | |
28 | @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat); |
57 | @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink |
|
|
58 | aio_fsync aio_fdatasync aio_readahead); |
29 | @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs); |
59 | @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs); |
30 | |
60 | |
31 | require XSLoader; |
61 | require XSLoader; |
32 | XSLoader::load Linux::AIO, $VERSION; |
62 | XSLoader::load Linux::AIO, $VERSION; |
33 | } |
63 | } |
34 | |
64 | |
35 | =item Linux::AIO::min_parallel($nthreads) |
65 | =item Linux::AIO::min_parallel $nthreads |
36 | |
66 | |
37 | Set the minimum number of AIO threads to $nthreads. You I<have> to call |
67 | Set the minimum number of AIO threads to C<$nthreads>. The default is |
38 | this function with a positive number at leats once, otherwise no threads |
68 | C<1>, which means a single asynchronous operation can be done at one time |
39 | will be started and you aio-operations will seem to hang. |
69 | (the number of outstanding operations, however, is unlimited). |
|
|
70 | |
|
|
71 | It is recommended to keep the number of threads low, as some linux |
|
|
72 | kernel versions will scale negatively with the number of threads (higher |
|
|
73 | parallelity => MUCH higher latency). |
|
|
74 | |
|
|
75 | Under normal circumstances you don't need to call this function, as this |
|
|
76 | module automatically starts a single async thread. |
|
|
77 | |
|
|
78 | =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 | Under normal circumstances you don't need to call this function. |
40 | |
88 | |
41 | =item $fileno = Linux::AIO::poll_fileno |
89 | =item $fileno = Linux::AIO::poll_fileno |
42 | |
90 | |
43 | Return the request result pipe filehandle. This filehandle must be polled |
91 | Return the I<request result pipe filehandle>. This filehandle must be |
44 | for reading. If the pipe becomes readable you have to call C<poll_cb>. |
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 | |
|
|
96 | See C<poll_cb> for an example. |
45 | |
97 | |
46 | =item Linux::AIO::poll_cb |
98 | =item Linux::AIO::poll_cb |
47 | |
99 | |
48 | Process all outstanding events on the result pipe. You have to call this |
100 | Process all outstanding events on the result pipe. You have to call this |
49 | regularly. Returns the number of events processed. |
101 | regularly. Returns the number of events processed. Returns immediately |
|
|
102 | when no events are outstanding. |
50 | |
103 | |
51 | You can use Event to multiplex, e.g.: |
104 | You can use Event to multiplex, e.g.: |
52 | |
105 | |
53 | Event->io(fd => Linux::AIO::poll_fileno, |
106 | Event->io (fd => Linux::AIO::poll_fileno, |
54 | poll => 'r', async => 1, |
107 | poll => 'r', async => 1, |
55 | cb => \&Linux::AIO::poll_cb ); |
108 | cb => \&Linux::AIO::poll_cb); |
56 | |
109 | |
|
|
110 | =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 | See C<nreqs> for an example. |
57 | |
117 | |
58 | =item Linux::AIO::nreqs |
118 | =item Linux::AIO::nreqs |
59 | |
119 | |
60 | Returns the number of requests currently outstanding. |
120 | Returns the number of requests currently outstanding. |
61 | |
121 | |
|
|
122 | Example: wait till there are no outstanding requests anymore: |
|
|
123 | |
|
|
124 | Linux::AIO::poll_wait, Linux::AIO::poll_cb |
|
|
125 | while Linux::AIO::nreqs; |
|
|
126 | |
62 | =item aio_open($pathname, $flags, $mode, $callback) |
127 | =item aio_open $pathname, $flags, $mode, $callback |
63 | |
128 | |
64 | Asynchronously open or create a file and call the callback with the |
129 | Asynchronously open or create a file and call the callback with the |
65 | filedescriptor. |
130 | filedescriptor (NOT a perl filehandle, sorry for that, but watch out, this |
|
|
131 | might change in the future). |
66 | |
132 | |
|
|
133 | The pathname passed to C<aio_open> must be absolute. See API NOTES, above, |
|
|
134 | for an explanation. |
|
|
135 | |
|
|
136 | The C<$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: $!\n"; |
|
|
149 | } |
|
|
150 | }; |
|
|
151 | |
67 | =item aio_close($fh, $callback) |
152 | =item aio_close $fh, $callback |
68 | |
153 | |
69 | Asynchronously close a file and call the callback with the result code. |
154 | Asynchronously close a file and call the callback with the result code. |
70 | |
155 | |
71 | =item aio_read($fh,$offset,$length, $data,$dataoffset,$callback) |
156 | =item aio_read $fh,$offset,$length, $data,$dataoffset,$callback |
72 | |
157 | |
73 | =item aio_write($fh,$offset,$length, $data,$dataoffset,$callback) |
158 | =item aio_write $fh,$offset,$length, $data,$dataoffset,$callback |
74 | |
159 | |
75 | Reads or writes C<length> bytes from the specified C<fh> and C<offset> |
160 | Reads or writes C<length> bytes from the specified C<fh> and C<offset> |
76 | into the scalar given by C<data> and offset C<dataoffset> and calls the |
161 | into the scalar given by C<data> and offset C<dataoffset> and calls the |
77 | callback without the actual number of bytes read (or undef on error). |
162 | callback without the actual number of bytes read (or -1 on error, just |
|
|
163 | like the syscall). |
78 | |
164 | |
|
|
165 | Example: Read 15 bytes at offset 7 into scalar C<$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<$length> specifies the number of bytes to be read. I/O is performed in |
|
|
182 | whole pages, so that offset is effectively rounded down to a page boundary |
|
|
183 | and bytes are read up to the next page boundary greater than or equal to |
|
|
184 | (off-set+length). aio_readahead() does not read beyond the end of the |
|
|
185 | file. The current file offset of the file is left unchanged. |
|
|
186 | |
79 | =item aio_stat($fh_or_path,$callback) |
187 | =item aio_stat $fh_or_path, $callback |
80 | |
188 | |
81 | =item aio_lstat($fh,$callback) |
189 | =item aio_lstat $fh, $callback |
82 | |
190 | |
83 | Works like perl's C<stat> or C<lstat> in void context, i.e. the callback |
191 | Works like perl's C<stat> or C<lstat> in void context. The callback will |
84 | will be called after the stat and the results will be available using |
192 | be called after the stat and the results will be available using C<stat _> |
85 | C<stat _> or C<-s _> etc... |
193 | or C<-s _> etc... |
|
|
194 | |
|
|
195 | The pathname passed to C<aio_stat> must be absolute. See API NOTES, above, |
|
|
196 | for an explanation. |
86 | |
197 | |
87 | Currently, the stats are always 64-bit-stats, i.e. instead of returning an |
198 | Currently, the stats are always 64-bit-stats, i.e. instead of returning an |
88 | error when stat'ing a large file, the results will be silently truncated |
199 | error when stat'ing a large file, the results will be silently truncated |
89 | unless perl itself is compiled with large file support. |
200 | unless perl itself is compiled with large file support. |
90 | |
201 | |
|
|
202 | 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 | |
|
|
221 | Asynchronously call fdatasync on the given filehandle and call the |
|
|
222 | callback with the fdatasync result code. |
|
|
223 | |
91 | =cut |
224 | =cut |
|
|
225 | |
|
|
226 | min_parallel 1; |
92 | |
227 | |
93 | END { |
228 | END { |
94 | max_parallel 0; |
229 | max_parallel 0; |
95 | } |
230 | } |
96 | |
231 | |
… | |
… | |
98 | |
233 | |
99 | =back |
234 | =back |
100 | |
235 | |
101 | =head1 BUGS |
236 | =head1 BUGS |
102 | |
237 | |
103 | This module has not yet been extensively tested. Watch out! |
238 | This module has been extensively tested in a large and very busy webserver |
|
|
239 | for many years now. |
104 | |
240 | |
105 | - perl-threads/fork interaction poorly tested. |
|
|
106 | - aio_open gives a fd, but all other functions expect a filehandle. |
241 | - aio_open gives a fd, but all other functions expect a perl filehandle. |
107 | |
242 | |
108 | =head1 SEE ALSO |
243 | =head1 SEE ALSO |
109 | |
244 | |
110 | L<Coro>. |
245 | L<Coro>, L<IO::AIO>. |
111 | |
246 | |
112 | =head1 AUTHOR |
247 | =head1 AUTHOR |
113 | |
248 | |
114 | Marc Lehmann <pcg@goof.com> |
249 | Marc Lehmann <schmorp@schmorp.de> |
115 | http://www.goof.com/pcg/marc/ |
250 | http://home.schmorp.de/ |
116 | |
251 | |
117 | =cut |
252 | =cut |
118 | |
253 | |