… | |
… | |
3 | IO::AIO - Asynchronous Input/Output |
3 | IO::AIO - Asynchronous Input/Output |
4 | |
4 | |
5 | =head1 SYNOPSIS |
5 | =head1 SYNOPSIS |
6 | |
6 | |
7 | use IO::AIO; |
7 | use IO::AIO; |
|
|
8 | |
|
|
9 | aio_open "/etc/passwd", O_RDONLY, 0, sub { |
|
|
10 | my ($fh) = @_; |
|
|
11 | ... |
|
|
12 | }; |
|
|
13 | |
|
|
14 | aio_unlink "/tmp/file", sub { }; |
|
|
15 | |
|
|
16 | aio_read $fh, 30000, 1024, $buffer, 0, sub { |
|
|
17 | $_[0] > 0 or die "read error: $!"; |
|
|
18 | }; |
|
|
19 | |
|
|
20 | use IO::AIO 2; # version has aio objects |
|
|
21 | |
|
|
22 | my $req = aio_unlink "/tmp/file", sub { }; |
|
|
23 | $req->cancel; # cancel request if still in queue |
|
|
24 | |
|
|
25 | # AnyEvent |
|
|
26 | open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; |
|
|
27 | my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); |
|
|
28 | |
|
|
29 | # Event |
|
|
30 | Event->io (fd => IO::AIO::poll_fileno, |
|
|
31 | poll => 'r', |
|
|
32 | cb => \&IO::AIO::poll_cb); |
|
|
33 | |
|
|
34 | # Glib/Gtk2 |
|
|
35 | add_watch Glib::IO IO::AIO::poll_fileno, |
|
|
36 | in => sub { IO::AIO::poll_cb; 1 }; |
|
|
37 | |
|
|
38 | # Tk |
|
|
39 | Tk::Event::IO->fileevent (IO::AIO::poll_fileno, "", |
|
|
40 | readable => \&IO::AIO::poll_cb); |
|
|
41 | |
|
|
42 | # Danga::Socket |
|
|
43 | Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
|
|
44 | \&IO::AIO::poll_cb); |
|
|
45 | |
8 | |
46 | |
9 | =head1 DESCRIPTION |
47 | =head1 DESCRIPTION |
10 | |
48 | |
11 | This module implements asynchronous I/O using whatever means your |
49 | This module implements asynchronous I/O using whatever means your |
12 | operating system supports. |
50 | operating system supports. |
… | |
… | |
19 | not well-supported (Linux doesn't allow them on normal files currently, |
57 | not well-supported (Linux doesn't allow them on normal files currently, |
20 | for example), and they would only support aio_read and aio_write, so the |
58 | for example), and they would only support aio_read and aio_write, so the |
21 | remaining functionality would have to be implemented using threads anyway. |
59 | remaining functionality would have to be implemented using threads anyway. |
22 | |
60 | |
23 | Although the module will work with in the presence of other threads, it is |
61 | Although the module will work with in the presence of other threads, it is |
24 | currently not reentrant, so use appropriate locking yourself. |
62 | currently not reentrant, so use appropriate locking yourself, always call |
|
|
63 | C<poll_cb> from within the same thread, or never call C<poll_cb> (or other |
|
|
64 | C<aio_> functions) recursively. |
25 | |
65 | |
26 | =head2 API NOTES |
66 | =cut |
|
|
67 | |
|
|
68 | package IO::AIO; |
|
|
69 | |
|
|
70 | no warnings; |
|
|
71 | use strict 'vars'; |
|
|
72 | |
|
|
73 | use base 'Exporter'; |
|
|
74 | |
|
|
75 | BEGIN { |
|
|
76 | our $VERSION = '1.99'; |
|
|
77 | |
|
|
78 | our @EXPORT = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat |
|
|
79 | aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink |
|
|
80 | aio_fsync aio_fdatasync aio_readahead aio_rename aio_link aio_move); |
|
|
81 | our @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); |
|
|
82 | |
|
|
83 | require XSLoader; |
|
|
84 | XSLoader::load ("IO::AIO", $VERSION); |
|
|
85 | } |
|
|
86 | |
|
|
87 | =head1 FUNCTIONS |
|
|
88 | |
|
|
89 | =head2 AIO FUNCTIONS |
27 | |
90 | |
28 | All the C<aio_*> calls are more or less thin wrappers around the syscall |
91 | All the C<aio_*> calls are more or less thin wrappers around the syscall |
29 | with the same name (sans C<aio_>). The arguments are similar or identical, |
92 | with the same name (sans C<aio_>). The arguments are similar or identical, |
30 | and they all accept an additional C<$callback> argument which must be |
93 | and they all accept an additional (and optional) C<$callback> argument |
31 | a code reference. This code reference will get called with the syscall |
94 | which must be a code reference. This code reference will get called with |
32 | return code (e.g. most syscalls return C<-1> on error, unlike perl, which |
95 | the syscall return code (e.g. most syscalls return C<-1> on error, unlike |
33 | usually delivers "false") as it's sole argument when the given syscall has |
96 | perl, which usually delivers "false") as it's sole argument when the given |
34 | been executed asynchronously. |
97 | syscall has been executed asynchronously. |
35 | |
98 | |
36 | All functions that expect a filehandle will also accept a file descriptor. |
99 | All functions expecting a filehandle keep a copy of the filehandle |
|
|
100 | internally until the request has finished. |
37 | |
101 | |
|
|
102 | All non-composite requests (requests that are not broken down into |
|
|
103 | multiple requests) return objects of type L<IO::AIO::REQ> that allow |
|
|
104 | further manipulation of running requests. |
|
|
105 | |
38 | The filenames you pass to these routines I<must> be absolute. The reason |
106 | The pathnames you pass to these routines I<must> be absolute and |
39 | is that at the time the request is being executed, the current working |
107 | encoded in byte form. The reason for the former is that at the time the |
40 | directory could have changed. Alternatively, you can make sure that you |
108 | request is being executed, the current working directory could have |
|
|
109 | changed. Alternatively, you can make sure that you never change the |
41 | never change the current working directory. |
110 | current working directory. |
|
|
111 | |
|
|
112 | To encode pathnames to byte form, either make sure you either: a) |
|
|
113 | always pass in filenames you got from outside (command line, readdir |
|
|
114 | etc.), b) are ASCII or ISO 8859-1, c) use the Encode module and encode |
|
|
115 | your pathnames to the locale (or other) encoding in effect in the user |
|
|
116 | environment, d) use Glib::filename_from_unicode on unicode filenames or e) |
|
|
117 | use something else. |
42 | |
118 | |
43 | =over 4 |
119 | =over 4 |
44 | |
120 | |
45 | =cut |
|
|
46 | |
|
|
47 | package IO::AIO; |
|
|
48 | |
|
|
49 | use base 'Exporter'; |
|
|
50 | |
|
|
51 | use Fcntl (); |
|
|
52 | |
|
|
53 | BEGIN { |
|
|
54 | $VERSION = 0.1; |
|
|
55 | |
|
|
56 | @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink |
|
|
57 | aio_fsync aio_fdatasync aio_readahead); |
|
|
58 | @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel nreqs); |
|
|
59 | |
|
|
60 | require XSLoader; |
|
|
61 | XSLoader::load IO::AIO, $VERSION; |
|
|
62 | } |
|
|
63 | |
|
|
64 | =item IO::AIO::min_parallel $nthreads |
|
|
65 | |
|
|
66 | Set the minimum number of AIO threads to C<$nthreads>. The default is |
|
|
67 | C<1>, which means a single asynchronous operation can be done at one time |
|
|
68 | (the number of outstanding operations, however, is unlimited). |
|
|
69 | |
|
|
70 | It is recommended to keep the number of threads low, as some linux |
|
|
71 | kernel versions will scale negatively with the number of threads (higher |
|
|
72 | parallelity => MUCH higher latency). |
|
|
73 | |
|
|
74 | Under normal circumstances you don't need to call this function, as this |
|
|
75 | module automatically starts a single async thread. |
|
|
76 | |
|
|
77 | =item IO::AIO::max_parallel $nthreads |
|
|
78 | |
|
|
79 | Sets the maximum number of AIO threads to C<$nthreads>. If more than |
|
|
80 | the specified number of threads are currently running, kill them. This |
|
|
81 | function blocks until the limit is reached. |
|
|
82 | |
|
|
83 | This module automatically runs C<max_parallel 0> at program end, to ensure |
|
|
84 | that all threads are killed and that there are no outstanding requests. |
|
|
85 | |
|
|
86 | Under normal circumstances you don't need to call this function. |
|
|
87 | |
|
|
88 | =item $fileno = IO::AIO::poll_fileno |
|
|
89 | |
|
|
90 | Return the I<request result pipe filehandle>. This filehandle must be |
|
|
91 | polled for reading by some mechanism outside this module (e.g. Event |
|
|
92 | or select, see below). If the pipe becomes readable you have to call |
|
|
93 | C<poll_cb> to check the results. |
|
|
94 | |
|
|
95 | See C<poll_cb> for an example. |
|
|
96 | |
|
|
97 | =item IO::AIO::poll_cb |
|
|
98 | |
|
|
99 | Process all outstanding events on the result pipe. You have to call this |
|
|
100 | regularly. Returns the number of events processed. Returns immediately |
|
|
101 | when no events are outstanding. |
|
|
102 | |
|
|
103 | You can use Event to multiplex, e.g.: |
|
|
104 | |
|
|
105 | Event->io (fd => IO::AIO::poll_fileno, |
|
|
106 | poll => 'r', async => 1, |
|
|
107 | cb => \&IO::AIO::poll_cb); |
|
|
108 | |
|
|
109 | =item IO::AIO::poll_wait |
|
|
110 | |
|
|
111 | Wait till the result filehandle becomes ready for reading (simply does a |
|
|
112 | select on the filehandle. This is useful if you want to synchronously wait |
|
|
113 | for some requests to finish). |
|
|
114 | |
|
|
115 | See C<nreqs> for an example. |
|
|
116 | |
|
|
117 | =item IO::AIO::nreqs |
|
|
118 | |
|
|
119 | Returns the number of requests currently outstanding. |
|
|
120 | |
|
|
121 | Example: wait till there are no outstanding requests anymore: |
|
|
122 | |
|
|
123 | IO::AIO::poll_wait, IO::AIO::poll_cb |
|
|
124 | while IO::AIO::nreqs; |
|
|
125 | |
|
|
126 | =item aio_open $pathname, $flags, $mode, $callback |
121 | =item aio_open $pathname, $flags, $mode, $callback->($fh) |
127 | |
122 | |
128 | Asynchronously open or create a file and call the callback with a newly |
123 | Asynchronously open or create a file and call the callback with a newly |
129 | created filehandle for the file. |
124 | created filehandle for the file. |
130 | |
125 | |
131 | The pathname passed to C<aio_open> must be absolute. See API NOTES, above, |
126 | The pathname passed to C<aio_open> must be absolute. See API NOTES, above, |
132 | for an explanation. |
127 | for an explanation. |
133 | |
128 | |
134 | The C<$mode> argument is a bitmask. See the C<Fcntl> module for a |
129 | The C<$flags> argument is a bitmask. See the C<Fcntl> module for a |
135 | list. They are the same as used in C<sysopen>. |
130 | list. They are the same as used by C<sysopen>. |
|
|
131 | |
|
|
132 | Likewise, C<$mode> specifies the mode of the newly created file, if it |
|
|
133 | didn't exist and C<O_CREAT> has been given, just like perl's C<sysopen>, |
|
|
134 | except that it is mandatory (i.e. use C<0> if you don't create new files, |
|
|
135 | and C<0666> or C<0777> if you do). |
136 | |
136 | |
137 | Example: |
137 | Example: |
138 | |
138 | |
139 | aio_open "/etc/passwd", O_RDONLY, 0, sub { |
139 | aio_open "/etc/passwd", O_RDONLY, 0, sub { |
140 | if ($_[0]) { |
140 | if ($_[0]) { |
… | |
… | |
143 | } else { |
143 | } else { |
144 | die "open failed: $!\n"; |
144 | die "open failed: $!\n"; |
145 | } |
145 | } |
146 | }; |
146 | }; |
147 | |
147 | |
148 | =item aio_close $fh, $callback |
148 | =item aio_close $fh, $callback->($status) |
149 | |
149 | |
150 | Asynchronously close a file and call the callback with the result |
150 | Asynchronously close a file and call the callback with the result |
151 | code. I<WARNING:> although accepted, you should not pass in a perl |
151 | code. I<WARNING:> although accepted, you should not pass in a perl |
152 | filehandle here, as perl will likely close the file descriptor itself when |
152 | filehandle here, as perl will likely close the file descriptor another |
153 | the filehandle is destroyed. Normally, you can safely call perls C<close> |
153 | time when the filehandle is destroyed. Normally, you can safely call perls |
154 | or just let filehandles go out of scope. |
154 | C<close> or just let filehandles go out of scope. |
155 | |
155 | |
|
|
156 | This is supposed to be a bug in the API, so that might change. It's |
|
|
157 | therefore best to avoid this function. |
|
|
158 | |
156 | =item aio_read $fh,$offset,$length, $data,$dataoffset,$callback |
159 | =item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
157 | |
160 | |
158 | =item aio_write $fh,$offset,$length, $data,$dataoffset,$callback |
161 | =item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
159 | |
162 | |
160 | Reads or writes C<length> bytes from the specified C<fh> and C<offset> |
163 | 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 |
164 | into the scalar given by C<data> and offset C<dataoffset> and calls the |
162 | callback without the actual number of bytes read (or -1 on error, just |
165 | callback without the actual number of bytes read (or -1 on error, just |
163 | like the syscall). |
166 | like the syscall). |
164 | |
167 | |
|
|
168 | The C<$data> scalar I<MUST NOT> be modified in any way while the request |
|
|
169 | is outstanding. Modifying it can result in segfaults or WW3 (if the |
|
|
170 | necessary/optional hardware is installed). |
|
|
171 | |
165 | Example: Read 15 bytes at offset 7 into scalar C<$buffer>, strating at |
172 | Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at |
166 | offset C<0> within the scalar: |
173 | offset C<0> within the scalar: |
167 | |
174 | |
168 | aio_read $fh, 7, 15, $buffer, 0, sub { |
175 | aio_read $fh, 7, 15, $buffer, 0, sub { |
169 | $_[0] >= 0 or die "read error: $!"; |
176 | $_[0] > 0 or die "read error: $!"; |
170 | print "read <$buffer>\n"; |
177 | print "read $_[0] bytes: <$buffer>\n"; |
171 | }; |
178 | }; |
172 | |
179 | |
|
|
180 | =item aio_move $srcpath, $dstpath, $callback->($status) |
|
|
181 | |
|
|
182 | Try to move the I<file> (directories not supported as either source or |
|
|
183 | destination) from C<$srcpath> to C<$dstpath> and call the callback with |
|
|
184 | the C<0> (error) or C<-1> ok. |
|
|
185 | |
|
|
186 | This is a composite request that tries to rename(2) the file first. If |
|
|
187 | rename files with C<EXDEV>, it creates the destination file with mode 0200 |
|
|
188 | and copies the contents of the source file into it using C<aio_sendfile>, |
|
|
189 | followed by restoring atime, mtime, access mode and uid/gid, in that |
|
|
190 | order, and unlinking the C<$srcpath>. |
|
|
191 | |
|
|
192 | If an error occurs, the partial destination file will be unlinked, if |
|
|
193 | possible, except when setting atime, mtime, access mode and uid/gid, where |
|
|
194 | errors are being ignored. |
|
|
195 | |
|
|
196 | =cut |
|
|
197 | |
|
|
198 | sub aio_move($$$) { |
|
|
199 | my ($src, $dst, $cb) = @_; |
|
|
200 | |
|
|
201 | aio_rename $src, $dst, sub { |
|
|
202 | if ($_[0] && $! == EXDEV) { |
|
|
203 | aio_open $src, O_RDONLY, 0, sub { |
|
|
204 | if (my $src_fh = $_[0]) { |
|
|
205 | my @stat = stat $src_fh; |
|
|
206 | |
|
|
207 | aio_open $dst, O_WRONLY, 0200, sub { |
|
|
208 | if (my $dst_fh = $_[0]) { |
|
|
209 | aio_sendfile $dst_fh, $src_fh, 0, $stat[7], sub { |
|
|
210 | close $src_fh; |
|
|
211 | |
|
|
212 | if ($_[0] == $stat[7]) { |
|
|
213 | utime $stat[8], $stat[9], $dst; |
|
|
214 | chmod $stat[2] & 07777, $dst_fh; |
|
|
215 | chown $stat[4], $stat[5], $dst_fh; |
|
|
216 | close $dst_fh; |
|
|
217 | |
|
|
218 | aio_unlink $src, sub { |
|
|
219 | $cb->($_[0]); |
|
|
220 | }; |
|
|
221 | } else { |
|
|
222 | my $errno = $!; |
|
|
223 | aio_unlink $dst, sub { |
|
|
224 | $! = $errno; |
|
|
225 | $cb->(-1); |
|
|
226 | }; |
|
|
227 | } |
|
|
228 | }; |
|
|
229 | } else { |
|
|
230 | $cb->(-1); |
|
|
231 | } |
|
|
232 | }, |
|
|
233 | |
|
|
234 | } else { |
|
|
235 | $cb->(-1); |
|
|
236 | } |
|
|
237 | }; |
|
|
238 | } else { |
|
|
239 | $cb->($_[0]); |
|
|
240 | } |
|
|
241 | }; |
|
|
242 | } |
|
|
243 | |
|
|
244 | =item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
|
|
245 | |
|
|
246 | Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts |
|
|
247 | reading at byte offset C<$in_offset>, and starts writing at the current |
|
|
248 | file offset of C<$out_fh>. Because of that, it is not safe to issue more |
|
|
249 | than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each |
|
|
250 | other. |
|
|
251 | |
|
|
252 | This call tries to make use of a native C<sendfile> syscall to provide |
|
|
253 | zero-copy operation. For this to work, C<$out_fh> should refer to a |
|
|
254 | socket, and C<$in_fh> should refer to mmap'able file. |
|
|
255 | |
|
|
256 | If the native sendfile call fails or is not implemented, it will be |
|
|
257 | emulated, so you can call C<aio_sendfile> on any type of filehandle |
|
|
258 | regardless of the limitations of the operating system. |
|
|
259 | |
|
|
260 | Please note, however, that C<aio_sendfile> can read more bytes from |
|
|
261 | C<$in_fh> than are written, and there is no way to find out how many |
|
|
262 | bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only |
|
|
263 | provides the number of bytes written to C<$out_fh>. Only if the result |
|
|
264 | value equals C<$length> one can assume that C<$length> bytes have been |
|
|
265 | read. |
|
|
266 | |
173 | =item aio_readahead $fh,$offset,$length, $callback |
267 | =item aio_readahead $fh,$offset,$length, $callback->($retval) |
174 | |
268 | |
175 | Asynchronously reads the specified byte range into the page cache, using |
|
|
176 | the C<readahead> syscall. If that syscall doesn't exist the status will be |
|
|
177 | C<-1> and C<$!> is set to ENOSYS. |
|
|
178 | |
|
|
179 | readahead() populates the page cache with data from a file so that |
269 | C<aio_readahead> populates the page cache with data from a file so that |
180 | subsequent reads from that file will not block on disk I/O. The C<$offset> |
270 | subsequent reads from that file will not block on disk I/O. The C<$offset> |
181 | argument specifies the starting point from which data is to be read and |
271 | argument specifies the starting point from which data is to be read and |
182 | C<$length> specifies the number of bytes to be read. I/O is performed in |
272 | C<$length> specifies the number of bytes to be read. I/O is performed in |
183 | whole pages, so that offset is effectively rounded down to a page boundary |
273 | whole pages, so that offset is effectively rounded down to a page boundary |
184 | and bytes are read up to the next page boundary greater than or equal to |
274 | and bytes are read up to the next page boundary greater than or equal to |
185 | (off-set+length). aio_readahead() does not read beyond the end of the |
275 | (off-set+length). C<aio_readahead> does not read beyond the end of the |
186 | file. The current file offset of the file is left unchanged. |
276 | file. The current file offset of the file is left unchanged. |
187 | |
277 | |
|
|
278 | If that syscall doesn't exist (likely if your OS isn't Linux) it will be |
|
|
279 | emulated by simply reading the data, which would have a similar effect. |
|
|
280 | |
188 | =item aio_stat $fh_or_path, $callback |
281 | =item aio_stat $fh_or_path, $callback->($status) |
189 | |
282 | |
190 | =item aio_lstat $fh, $callback |
283 | =item aio_lstat $fh, $callback->($status) |
191 | |
284 | |
192 | Works like perl's C<stat> or C<lstat> in void context. The callback will |
285 | Works like perl's C<stat> or C<lstat> in void context. The callback will |
193 | be called after the stat and the results will be available using C<stat _> |
286 | be called after the stat and the results will be available using C<stat _> |
194 | or C<-s _> etc... |
287 | or C<-s _> etc... |
195 | |
288 | |
… | |
… | |
205 | aio_stat "/etc/passwd", sub { |
298 | aio_stat "/etc/passwd", sub { |
206 | $_[0] and die "stat failed: $!"; |
299 | $_[0] and die "stat failed: $!"; |
207 | print "size is ", -s _, "\n"; |
300 | print "size is ", -s _, "\n"; |
208 | }; |
301 | }; |
209 | |
302 | |
210 | =item aio_unlink $pathname, $callback |
303 | =item aio_unlink $pathname, $callback->($status) |
211 | |
304 | |
212 | Asynchronously unlink (delete) a file and call the callback with the |
305 | Asynchronously unlink (delete) a file and call the callback with the |
213 | result code. |
306 | result code. |
214 | |
307 | |
|
|
308 | =item aio_link $srcpath, $dstpath, $callback->($status) |
|
|
309 | |
|
|
310 | Asynchronously create a new link to the existing object at C<$srcpath> at |
|
|
311 | the path C<$dstpath> and call the callback with the result code. |
|
|
312 | |
|
|
313 | =item aio_symlink $srcpath, $dstpath, $callback->($status) |
|
|
314 | |
|
|
315 | Asynchronously create a new symbolic link to the existing object at C<$srcpath> at |
|
|
316 | the path C<$dstpath> and call the callback with the result code. |
|
|
317 | |
|
|
318 | =item aio_rename $srcpath, $dstpath, $callback->($status) |
|
|
319 | |
|
|
320 | Asynchronously rename the object at C<$srcpath> to C<$dstpath>, just as |
|
|
321 | rename(2) and call the callback with the result code. |
|
|
322 | |
|
|
323 | =item aio_rmdir $pathname, $callback->($status) |
|
|
324 | |
|
|
325 | Asynchronously rmdir (delete) a directory and call the callback with the |
|
|
326 | result code. |
|
|
327 | |
|
|
328 | =item aio_readdir $pathname, $callback->($entries) |
|
|
329 | |
|
|
330 | Unlike the POSIX call of the same name, C<aio_readdir> reads an entire |
|
|
331 | directory (i.e. opendir + readdir + closedir). The entries will not be |
|
|
332 | sorted, and will B<NOT> include the C<.> and C<..> entries. |
|
|
333 | |
|
|
334 | The callback a single argument which is either C<undef> or an array-ref |
|
|
335 | with the filenames. |
|
|
336 | |
|
|
337 | =item aio_scandir $path, $maxreq, $callback->($dirs, $nondirs) |
|
|
338 | |
|
|
339 | Scans a directory (similar to C<aio_readdir>) but additionally tries to |
|
|
340 | separate the entries of directory C<$path> into two sets of names, ones |
|
|
341 | you can recurse into (directories or links to them), and ones you cannot |
|
|
342 | recurse into (everything else). |
|
|
343 | |
|
|
344 | C<aio_scandir> is a composite request that consists of many sub |
|
|
345 | requests. C<$maxreq> specifies the maximum number of outstanding aio |
|
|
346 | requests that this function generates. If it is C<< <= 0 >>, then a |
|
|
347 | suitable default will be chosen (currently 8). |
|
|
348 | |
|
|
349 | On error, the callback is called without arguments, otherwise it receives |
|
|
350 | two array-refs with path-relative entry names. |
|
|
351 | |
|
|
352 | Example: |
|
|
353 | |
|
|
354 | aio_scandir $dir, 0, sub { |
|
|
355 | my ($dirs, $nondirs) = @_; |
|
|
356 | print "real directories: @$dirs\n"; |
|
|
357 | print "everything else: @$nondirs\n"; |
|
|
358 | }; |
|
|
359 | |
|
|
360 | Implementation notes. |
|
|
361 | |
|
|
362 | The C<aio_readdir> cannot be avoided, but C<stat()>'ing every entry can. |
|
|
363 | |
|
|
364 | After reading the directory, the modification time, size etc. of the |
|
|
365 | directory before and after the readdir is checked, and if they match (and |
|
|
366 | isn't the current time), the link count will be used to decide how many |
|
|
367 | entries are directories (if >= 2). Otherwise, no knowledge of the number |
|
|
368 | of subdirectories will be assumed. |
|
|
369 | |
|
|
370 | Then entries will be sorted into likely directories (everything without |
|
|
371 | a non-initial dot currently) and likely non-directories (everything |
|
|
372 | else). Then every entry plus an appended C</.> will be C<stat>'ed, |
|
|
373 | likely directories first. If that succeeds, it assumes that the entry |
|
|
374 | is a directory or a symlink to directory (which will be checked |
|
|
375 | seperately). This is often faster than stat'ing the entry itself because |
|
|
376 | filesystems might detect the type of the entry without reading the inode |
|
|
377 | data (e.g. ext2fs filetype feature). |
|
|
378 | |
|
|
379 | If the known number of directories (link count - 2) has been reached, the |
|
|
380 | rest of the entries is assumed to be non-directories. |
|
|
381 | |
|
|
382 | This only works with certainty on POSIX (= UNIX) filesystems, which |
|
|
383 | fortunately are the vast majority of filesystems around. |
|
|
384 | |
|
|
385 | It will also likely work on non-POSIX filesystems with reduced efficiency |
|
|
386 | as those tend to return 0 or 1 as link counts, which disables the |
|
|
387 | directory counting heuristic. |
|
|
388 | |
|
|
389 | =cut |
|
|
390 | |
|
|
391 | sub aio_scandir($$$) { |
|
|
392 | my ($path, $maxreq, $cb) = @_; |
|
|
393 | |
|
|
394 | $maxreq = 8 if $maxreq <= 0; |
|
|
395 | |
|
|
396 | # stat once |
|
|
397 | aio_stat $path, sub { |
|
|
398 | return $cb->() if $_[0]; |
|
|
399 | my $now = time; |
|
|
400 | my $hash1 = join ":", (stat _)[0,1,3,7,9]; |
|
|
401 | |
|
|
402 | # read the directory entries |
|
|
403 | aio_readdir $path, sub { |
|
|
404 | my $entries = shift |
|
|
405 | or return $cb->(); |
|
|
406 | |
|
|
407 | # stat the dir another time |
|
|
408 | aio_stat $path, sub { |
|
|
409 | my $hash2 = join ":", (stat _)[0,1,3,7,9]; |
|
|
410 | |
|
|
411 | my $ndirs; |
|
|
412 | |
|
|
413 | # take the slow route if anything looks fishy |
|
|
414 | if ($hash1 ne $hash2 or (stat _)[9] == $now) { |
|
|
415 | $ndirs = -1; |
|
|
416 | } else { |
|
|
417 | # if nlink == 2, we are finished |
|
|
418 | # on non-posix-fs's, we rely on nlink < 2 |
|
|
419 | $ndirs = (stat _)[3] - 2 |
|
|
420 | or return $cb->([], $entries); |
|
|
421 | } |
|
|
422 | |
|
|
423 | # sort into likely dirs and likely nondirs |
|
|
424 | # dirs == files without ".", short entries first |
|
|
425 | $entries = [map $_->[0], |
|
|
426 | sort { $b->[1] cmp $a->[1] } |
|
|
427 | map [$_, sprintf "%s%04d", (/.\./ ? "1" : "0"), length], |
|
|
428 | @$entries]; |
|
|
429 | |
|
|
430 | my (@dirs, @nondirs); |
|
|
431 | |
|
|
432 | my ($statcb, $schedcb); |
|
|
433 | my $nreq = 0; |
|
|
434 | |
|
|
435 | $schedcb = sub { |
|
|
436 | if (@$entries) { |
|
|
437 | if ($nreq < $maxreq) { |
|
|
438 | my $ent = pop @$entries; |
|
|
439 | $nreq++; |
|
|
440 | aio_stat "$path/$ent/.", sub { $statcb->($_[0], $ent) }; |
|
|
441 | } |
|
|
442 | } elsif (!$nreq) { |
|
|
443 | # finished |
|
|
444 | undef $statcb; |
|
|
445 | undef $schedcb; |
|
|
446 | $cb->(\@dirs, \@nondirs) if $cb; |
|
|
447 | undef $cb; |
|
|
448 | } |
|
|
449 | }; |
|
|
450 | $statcb = sub { |
|
|
451 | my ($status, $entry) = @_; |
|
|
452 | |
|
|
453 | if ($status < 0) { |
|
|
454 | $nreq--; |
|
|
455 | push @nondirs, $entry; |
|
|
456 | &$schedcb; |
|
|
457 | } else { |
|
|
458 | # need to check for real directory |
|
|
459 | aio_lstat "$path/$entry", sub { |
|
|
460 | $nreq--; |
|
|
461 | |
|
|
462 | if (-d _) { |
|
|
463 | push @dirs, $entry; |
|
|
464 | |
|
|
465 | if (!--$ndirs) { |
|
|
466 | push @nondirs, @$entries; |
|
|
467 | $entries = []; |
|
|
468 | } |
|
|
469 | } else { |
|
|
470 | push @nondirs, $entry; |
|
|
471 | } |
|
|
472 | |
|
|
473 | &$schedcb; |
|
|
474 | } |
|
|
475 | } |
|
|
476 | }; |
|
|
477 | |
|
|
478 | &$schedcb while @$entries && $nreq < $maxreq; |
|
|
479 | }; |
|
|
480 | }; |
|
|
481 | }; |
|
|
482 | } |
|
|
483 | |
215 | =item aio_fsync $fh, $callback |
484 | =item aio_fsync $fh, $callback->($status) |
216 | |
485 | |
217 | Asynchronously call fsync on the given filehandle and call the callback |
486 | Asynchronously call fsync on the given filehandle and call the callback |
218 | with the fsync result code. |
487 | with the fsync result code. |
219 | |
488 | |
220 | =item aio_fdatasync $fh, $callback |
489 | =item aio_fdatasync $fh, $callback->($status) |
221 | |
490 | |
222 | Asynchronously call fdatasync on the given filehandle and call the |
491 | Asynchronously call fdatasync on the given filehandle and call the |
223 | callback with the fdatasync result code. |
492 | callback with the fdatasync result code. |
|
|
493 | |
|
|
494 | If this call isn't available because your OS lacks it or it couldn't be |
|
|
495 | detected, it will be emulated by calling C<fsync> instead. |
|
|
496 | |
|
|
497 | =back |
|
|
498 | |
|
|
499 | =head2 IO::AIO::REQ CLASS |
|
|
500 | |
|
|
501 | All non-aggregate C<aio_*> functions return an object of this class when |
|
|
502 | called in non-void context. |
|
|
503 | |
|
|
504 | A request always moves through the following five states in its lifetime, |
|
|
505 | in order: B<ready> (request has been created, but has not been executed |
|
|
506 | yet), B<execute> (request is currently being executed), B<pending> |
|
|
507 | (request has been executed but callback has not been called yet), |
|
|
508 | B<result> (results are being processed synchronously, includes calling the |
|
|
509 | callback) and B<done> (request has reached the end of its lifetime and |
|
|
510 | holds no resources anymore). |
|
|
511 | |
|
|
512 | =over 4 |
|
|
513 | |
|
|
514 | =item $req->cancel |
|
|
515 | |
|
|
516 | Cancels the request, if possible. Has the effect of skipping execution |
|
|
517 | when entering the B<execute> state and skipping calling the callback when |
|
|
518 | entering the the B<result> state, but will leave the request otherwise |
|
|
519 | untouched. That means that requests that currently execute will not be |
|
|
520 | stopped and resources held by the request will not be freed prematurely. |
|
|
521 | |
|
|
522 | =back |
|
|
523 | |
|
|
524 | =head2 SUPPORT FUNCTIONS |
|
|
525 | |
|
|
526 | =over 4 |
|
|
527 | |
|
|
528 | =item $fileno = IO::AIO::poll_fileno |
|
|
529 | |
|
|
530 | Return the I<request result pipe file descriptor>. This filehandle must be |
|
|
531 | polled for reading by some mechanism outside this module (e.g. Event or |
|
|
532 | select, see below or the SYNOPSIS). If the pipe becomes readable you have |
|
|
533 | to call C<poll_cb> to check the results. |
|
|
534 | |
|
|
535 | See C<poll_cb> for an example. |
|
|
536 | |
|
|
537 | =item IO::AIO::poll_cb |
|
|
538 | |
|
|
539 | Process all outstanding events on the result pipe. You have to call this |
|
|
540 | regularly. Returns the number of events processed. Returns immediately |
|
|
541 | when no events are outstanding. |
|
|
542 | |
|
|
543 | Example: Install an Event watcher that automatically calls |
|
|
544 | IO::AIO::poll_cb with high priority: |
|
|
545 | |
|
|
546 | Event->io (fd => IO::AIO::poll_fileno, |
|
|
547 | poll => 'r', async => 1, |
|
|
548 | cb => \&IO::AIO::poll_cb); |
|
|
549 | |
|
|
550 | =item IO::AIO::poll_wait |
|
|
551 | |
|
|
552 | Wait till the result filehandle becomes ready for reading (simply does a |
|
|
553 | C<select> on the filehandle. This is useful if you want to synchronously wait |
|
|
554 | for some requests to finish). |
|
|
555 | |
|
|
556 | See C<nreqs> for an example. |
|
|
557 | |
|
|
558 | =item IO::AIO::nreqs |
|
|
559 | |
|
|
560 | Returns the number of requests currently outstanding (i.e. for which their |
|
|
561 | callback has not been invoked yet). |
|
|
562 | |
|
|
563 | Example: wait till there are no outstanding requests anymore: |
|
|
564 | |
|
|
565 | IO::AIO::poll_wait, IO::AIO::poll_cb |
|
|
566 | while IO::AIO::nreqs; |
|
|
567 | |
|
|
568 | =item IO::AIO::flush |
|
|
569 | |
|
|
570 | Wait till all outstanding AIO requests have been handled. |
|
|
571 | |
|
|
572 | Strictly equivalent to: |
|
|
573 | |
|
|
574 | IO::AIO::poll_wait, IO::AIO::poll_cb |
|
|
575 | while IO::AIO::nreqs; |
|
|
576 | |
|
|
577 | =item IO::AIO::poll |
|
|
578 | |
|
|
579 | Waits until some requests have been handled. |
|
|
580 | |
|
|
581 | Strictly equivalent to: |
|
|
582 | |
|
|
583 | IO::AIO::poll_wait, IO::AIO::poll_cb |
|
|
584 | if IO::AIO::nreqs; |
|
|
585 | |
|
|
586 | =item IO::AIO::min_parallel $nthreads |
|
|
587 | |
|
|
588 | Set the minimum number of AIO threads to C<$nthreads>. The current default |
|
|
589 | is C<4>, which means four asynchronous operations can be done at one time |
|
|
590 | (the number of outstanding operations, however, is unlimited). |
|
|
591 | |
|
|
592 | IO::AIO starts threads only on demand, when an AIO request is queued and |
|
|
593 | no free thread exists. |
|
|
594 | |
|
|
595 | It is recommended to keep the number of threads low, as some Linux |
|
|
596 | kernel versions will scale negatively with the number of threads (higher |
|
|
597 | parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 |
|
|
598 | threads should be fine. |
|
|
599 | |
|
|
600 | Under most circumstances you don't need to call this function, as the |
|
|
601 | module selects a default that is suitable for low to moderate load. |
|
|
602 | |
|
|
603 | =item IO::AIO::max_parallel $nthreads |
|
|
604 | |
|
|
605 | Sets the maximum number of AIO threads to C<$nthreads>. If more than the |
|
|
606 | specified number of threads are currently running, this function kills |
|
|
607 | them. This function blocks until the limit is reached. |
|
|
608 | |
|
|
609 | While C<$nthreads> are zero, aio requests get queued but not executed |
|
|
610 | until the number of threads has been increased again. |
|
|
611 | |
|
|
612 | This module automatically runs C<max_parallel 0> at program end, to ensure |
|
|
613 | that all threads are killed and that there are no outstanding requests. |
|
|
614 | |
|
|
615 | Under normal circumstances you don't need to call this function. |
|
|
616 | |
|
|
617 | =item $oldnreqs = IO::AIO::max_outstanding $nreqs |
|
|
618 | |
|
|
619 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
|
|
620 | try to queue up more than this number of requests, the caller will block until |
|
|
621 | some requests have been handled. |
|
|
622 | |
|
|
623 | The default is very large, so normally there is no practical limit. If you |
|
|
624 | queue up many requests in a loop it often improves speed if you set |
|
|
625 | this to a relatively low number, such as C<100>. |
|
|
626 | |
|
|
627 | Under normal circumstances you don't need to call this function. |
|
|
628 | |
|
|
629 | =back |
224 | |
630 | |
225 | =cut |
631 | =cut |
226 | |
632 | |
227 | # support function to convert a fd into a perl filehandle |
633 | # support function to convert a fd into a perl filehandle |
228 | sub _fd2fh { |
634 | sub _fd2fh { |
229 | return undef if $_[0] < 0; |
635 | return undef if $_[0] < 0; |
230 | |
636 | |
231 | # try to be perl5.6-compatible |
637 | # try to generate nice filehandles |
232 | local *AIO_FH; |
638 | my $sym = "IO::AIO::fd#$_[0]"; |
233 | open AIO_FH, "+<&=$_[0]" |
639 | local *$sym; |
|
|
640 | |
|
|
641 | open *$sym, "+<&=$_[0]" # usually works under any unix |
|
|
642 | or open *$sym, "<&=$_[0]" # cygwin needs this |
|
|
643 | or open *$sym, ">&=$_[0]" # or this |
234 | or return undef; |
644 | or return undef; |
235 | |
645 | |
236 | *AIO_FH |
646 | *$sym |
237 | } |
647 | } |
238 | |
648 | |
239 | min_parallel 4; |
649 | min_parallel 4; |
240 | |
650 | |
241 | END { |
651 | END { |
242 | max_parallel 0; |
652 | max_parallel 0; |
243 | } |
653 | } |
244 | |
654 | |
245 | 1; |
655 | 1; |
246 | |
656 | |
247 | =back |
657 | =head2 FORK BEHAVIOUR |
248 | |
658 | |
249 | =head1 BUGS |
659 | This module should do "the right thing" when the process using it forks: |
250 | |
660 | |
251 | - could be optimized to use more semaphores instead of filehandles. |
661 | Before the fork, IO::AIO enters a quiescent state where no requests |
|
|
662 | can be added in other threads and no results will be processed. After |
|
|
663 | the fork the parent simply leaves the quiescent state and continues |
|
|
664 | request/result processing, while the child clears the request/result |
|
|
665 | queue (so the requests started before the fork will only be handled in |
|
|
666 | the parent). Threads will be started on demand until the limit ste in the |
|
|
667 | parent process has been reached again. |
|
|
668 | |
|
|
669 | In short: the parent will, after a short pause, continue as if fork had |
|
|
670 | not been called, while the child will act as if IO::AIO has not been used |
|
|
671 | yet. |
252 | |
672 | |
253 | =head1 SEE ALSO |
673 | =head1 SEE ALSO |
254 | |
674 | |
255 | L<Coro>, L<Linux::AIO>. |
675 | L<Coro>, L<Linux::AIO> (obsolete). |
256 | |
676 | |
257 | =head1 AUTHOR |
677 | =head1 AUTHOR |
258 | |
678 | |
259 | Marc Lehmann <schmorp@schmorp.de> |
679 | Marc Lehmann <schmorp@schmorp.de> |
260 | http://home.schmorp.de/ |
680 | http://home.schmorp.de/ |