1 |
root |
1.1 |
NAME |
2 |
|
|
IO::AIO - Asynchronous Input/Output |
3 |
|
|
|
4 |
|
|
SYNOPSIS |
5 |
|
|
use IO::AIO; |
6 |
|
|
|
7 |
|
|
DESCRIPTION |
8 |
|
|
This module implements asynchronous I/O using whatever means your |
9 |
root |
1.2 |
operating system supports. |
10 |
root |
1.1 |
|
11 |
root |
1.2 |
Currently, a number of threads are started that execute your read/writes |
12 |
|
|
and signal their completion. You don't need thread support in your libc |
13 |
|
|
or perl, and the threads created by this module will not be visible to |
14 |
|
|
the pthreads library. In the future, this module might make use of the |
15 |
|
|
native aio functions available on many operating systems. However, they |
16 |
|
|
are often not well-supported (Linux doesn't allow them on normal files |
17 |
|
|
currently, for example), and they would only support aio_read and |
18 |
|
|
aio_write, so the remaining functionality would have to be implemented |
19 |
|
|
using threads anyway. |
20 |
root |
1.1 |
|
21 |
|
|
Although the module will work with in the presence of other threads, it |
22 |
root |
1.2 |
is currently not reentrant, so use appropriate locking yourself. |
23 |
root |
1.1 |
|
24 |
|
|
API NOTES |
25 |
|
|
All the "aio_*" calls are more or less thin wrappers around the syscall |
26 |
|
|
with the same name (sans "aio_"). The arguments are similar or |
27 |
|
|
identical, and they all accept an additional $callback argument which |
28 |
|
|
must be a code reference. This code reference will get called with the |
29 |
|
|
syscall return code (e.g. most syscalls return -1 on error, unlike perl, |
30 |
|
|
which usually delivers "false") as it's sole argument when the given |
31 |
|
|
syscall has been executed asynchronously. |
32 |
|
|
|
33 |
|
|
All functions that expect a filehandle will also accept a file |
34 |
|
|
descriptor. |
35 |
|
|
|
36 |
|
|
The filenames you pass to these routines *must* be absolute. The reason |
37 |
|
|
is that at the time the request is being executed, the current working |
38 |
|
|
directory could have changed. Alternatively, you can make sure that you |
39 |
|
|
never change the current working directory. |
40 |
|
|
|
41 |
|
|
IO::AIO::min_parallel $nthreads |
42 |
|
|
Set the minimum number of AIO threads to $nthreads. The default is |
43 |
|
|
1, which means a single asynchronous operation can be done at one |
44 |
|
|
time (the number of outstanding operations, however, is unlimited). |
45 |
|
|
|
46 |
root |
1.3 |
It is recommended to keep the number of threads low, as some Linux |
47 |
root |
1.1 |
kernel versions will scale negatively with the number of threads |
48 |
root |
1.3 |
(higher parallelity => MUCH higher latency). With current Linux 2.6 |
49 |
|
|
versions, 4-32 threads should be fine. |
50 |
root |
1.1 |
|
51 |
|
|
Under normal circumstances you don't need to call this function, as |
52 |
root |
1.3 |
this module automatically starts some threads (the exact number |
53 |
|
|
might change, and is currently 4). |
54 |
root |
1.1 |
|
55 |
|
|
IO::AIO::max_parallel $nthreads |
56 |
|
|
Sets the maximum number of AIO threads to $nthreads. If more than |
57 |
|
|
the specified number of threads are currently running, kill them. |
58 |
|
|
This function blocks until the limit is reached. |
59 |
|
|
|
60 |
|
|
This module automatically runs "max_parallel 0" at program end, to |
61 |
|
|
ensure that all threads are killed and that there are no outstanding |
62 |
|
|
requests. |
63 |
|
|
|
64 |
|
|
Under normal circumstances you don't need to call this function. |
65 |
|
|
|
66 |
root |
1.3 |
$oldnreqs = IO::AIO::max_outstanding $nreqs |
67 |
|
|
Sets the maximum number of outstanding requests to $nreqs. If you |
68 |
|
|
try to queue up more than this number of requests, the caller will |
69 |
|
|
block until some requests have been handled. |
70 |
|
|
|
71 |
|
|
The default is very large, so normally there is no practical limit. |
72 |
|
|
If you queue up many requests in a loop it it often improves speed |
73 |
|
|
if you set this to a relatively low number, such as 100. |
74 |
|
|
|
75 |
|
|
Under normal circumstances you don't need to call this function. |
76 |
|
|
|
77 |
root |
1.1 |
$fileno = IO::AIO::poll_fileno |
78 |
|
|
Return the *request result pipe filehandle*. This filehandle must be |
79 |
|
|
polled for reading by some mechanism outside this module (e.g. Event |
80 |
|
|
or select, see below). If the pipe becomes readable you have to call |
81 |
|
|
"poll_cb" to check the results. |
82 |
|
|
|
83 |
|
|
See "poll_cb" for an example. |
84 |
|
|
|
85 |
|
|
IO::AIO::poll_cb |
86 |
|
|
Process all outstanding events on the result pipe. You have to call |
87 |
|
|
this regularly. Returns the number of events processed. Returns |
88 |
|
|
immediately when no events are outstanding. |
89 |
|
|
|
90 |
|
|
You can use Event to multiplex, e.g.: |
91 |
|
|
|
92 |
|
|
Event->io (fd => IO::AIO::poll_fileno, |
93 |
|
|
poll => 'r', async => 1, |
94 |
|
|
cb => \&IO::AIO::poll_cb); |
95 |
|
|
|
96 |
|
|
IO::AIO::poll_wait |
97 |
|
|
Wait till the result filehandle becomes ready for reading (simply |
98 |
|
|
does a select on the filehandle. This is useful if you want to |
99 |
|
|
synchronously wait for some requests to finish). |
100 |
|
|
|
101 |
|
|
See "nreqs" for an example. |
102 |
|
|
|
103 |
|
|
IO::AIO::nreqs |
104 |
|
|
Returns the number of requests currently outstanding. |
105 |
|
|
|
106 |
|
|
Example: wait till there are no outstanding requests anymore: |
107 |
|
|
|
108 |
|
|
IO::AIO::poll_wait, IO::AIO::poll_cb |
109 |
|
|
while IO::AIO::nreqs; |
110 |
|
|
|
111 |
|
|
aio_open $pathname, $flags, $mode, $callback |
112 |
root |
1.2 |
Asynchronously open or create a file and call the callback with a |
113 |
|
|
newly created filehandle for the file. |
114 |
|
|
|
115 |
|
|
The pathname passed to "aio_open" must be absolute. See API NOTES, |
116 |
|
|
above, for an explanation. |
117 |
root |
1.1 |
|
118 |
|
|
The $mode argument is a bitmask. See the "Fcntl" module for a list. |
119 |
|
|
They are the same as used in "sysopen". |
120 |
|
|
|
121 |
|
|
Example: |
122 |
|
|
|
123 |
|
|
aio_open "/etc/passwd", O_RDONLY, 0, sub { |
124 |
root |
1.2 |
if ($_[0]) { |
125 |
|
|
print "open successful, fh is $_[0]\n"; |
126 |
root |
1.1 |
... |
127 |
|
|
} else { |
128 |
|
|
die "open failed: $!\n"; |
129 |
|
|
} |
130 |
|
|
}; |
131 |
|
|
|
132 |
|
|
aio_close $fh, $callback |
133 |
|
|
Asynchronously close a file and call the callback with the result |
134 |
root |
1.2 |
code. *WARNING:* although accepted, you should not pass in a perl |
135 |
|
|
filehandle here, as perl will likely close the file descriptor |
136 |
|
|
itself when the filehandle is destroyed. Normally, you can safely |
137 |
|
|
call perls "close" or just let filehandles go out of scope. |
138 |
root |
1.1 |
|
139 |
|
|
aio_read $fh,$offset,$length, $data,$dataoffset,$callback |
140 |
|
|
aio_write $fh,$offset,$length, $data,$dataoffset,$callback |
141 |
|
|
Reads or writes "length" bytes from the specified "fh" and "offset" |
142 |
|
|
into the scalar given by "data" and offset "dataoffset" and calls |
143 |
|
|
the callback without the actual number of bytes read (or -1 on |
144 |
|
|
error, just like the syscall). |
145 |
|
|
|
146 |
|
|
Example: Read 15 bytes at offset 7 into scalar $buffer, strating at |
147 |
|
|
offset 0 within the scalar: |
148 |
|
|
|
149 |
|
|
aio_read $fh, 7, 15, $buffer, 0, sub { |
150 |
|
|
$_[0] >= 0 or die "read error: $!"; |
151 |
|
|
print "read <$buffer>\n"; |
152 |
|
|
}; |
153 |
|
|
|
154 |
|
|
aio_readahead $fh,$offset,$length, $callback |
155 |
|
|
Asynchronously reads the specified byte range into the page cache, |
156 |
root |
1.2 |
using the "readahead" syscall. If that syscall doesn't exist the |
157 |
|
|
status will be -1 and $! is set to ENOSYS. |
158 |
root |
1.1 |
|
159 |
|
|
readahead() populates the page cache with data from a file so that |
160 |
|
|
subsequent reads from that file will not block on disk I/O. The |
161 |
|
|
$offset argument specifies the starting point from which data is to |
162 |
|
|
be read and $length specifies the number of bytes to be read. I/O is |
163 |
|
|
performed in whole pages, so that offset is effectively rounded down |
164 |
|
|
to a page boundary and bytes are read up to the next page boundary |
165 |
|
|
greater than or equal to (off-set+length). aio_readahead() does not |
166 |
|
|
read beyond the end of the file. The current file offset of the file |
167 |
|
|
is left unchanged. |
168 |
|
|
|
169 |
|
|
aio_stat $fh_or_path, $callback |
170 |
|
|
aio_lstat $fh, $callback |
171 |
|
|
Works like perl's "stat" or "lstat" in void context. The callback |
172 |
|
|
will be called after the stat and the results will be available |
173 |
|
|
using "stat _" or "-s _" etc... |
174 |
|
|
|
175 |
|
|
The pathname passed to "aio_stat" must be absolute. See API NOTES, |
176 |
|
|
above, for an explanation. |
177 |
|
|
|
178 |
|
|
Currently, the stats are always 64-bit-stats, i.e. instead of |
179 |
|
|
returning an error when stat'ing a large file, the results will be |
180 |
|
|
silently truncated unless perl itself is compiled with large file |
181 |
|
|
support. |
182 |
|
|
|
183 |
|
|
Example: Print the length of /etc/passwd: |
184 |
|
|
|
185 |
|
|
aio_stat "/etc/passwd", sub { |
186 |
|
|
$_[0] and die "stat failed: $!"; |
187 |
|
|
print "size is ", -s _, "\n"; |
188 |
|
|
}; |
189 |
|
|
|
190 |
|
|
aio_unlink $pathname, $callback |
191 |
|
|
Asynchronously unlink (delete) a file and call the callback with the |
192 |
|
|
result code. |
193 |
|
|
|
194 |
|
|
aio_fsync $fh, $callback |
195 |
|
|
Asynchronously call fsync on the given filehandle and call the |
196 |
|
|
callback with the fsync result code. |
197 |
|
|
|
198 |
|
|
aio_fdatasync $fh, $callback |
199 |
|
|
Asynchronously call fdatasync on the given filehandle and call the |
200 |
|
|
callback with the fdatasync result code. |
201 |
|
|
|
202 |
|
|
BUGS |
203 |
root |
1.2 |
- could be optimized to use more semaphores instead of filehandles. |
204 |
root |
1.1 |
|
205 |
|
|
SEE ALSO |
206 |
root |
1.2 |
Coro, Linux::AIO. |
207 |
root |
1.1 |
|
208 |
|
|
AUTHOR |
209 |
|
|
Marc Lehmann <schmorp@schmorp.de> |
210 |
|
|
http://home.schmorp.de/ |
211 |
|
|
|