ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/IO-AIO/AIO.pm
(Generate patch)

Comparing IO-AIO/AIO.pm (file contents):
Revision 1.32 by root, Wed Aug 17 05:26:20 2005 UTC vs.
Revision 1.37 by root, Tue Aug 23 12:37:19 2005 UTC

… …
63use base 'Exporter'; 63use base 'Exporter';
64 64
65use Fcntl (); 65use Fcntl ();
66 66
67BEGIN { 67BEGIN {
68 $VERSION = 1.3; 68 $VERSION = 1.6;
69
70 69
71 @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink 70 @EXPORT = qw(aio_read aio_write aio_open aio_close aio_stat aio_lstat aio_unlink
72 aio_rmdir aio_symlink aio_fsync aio_fdatasync aio_readahead); 71 aio_rmdir aio_readdir aio_symlink aio_fsync aio_fdatasync aio_readahead);
73 @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs); 72 @EXPORT_OK = qw(poll_fileno poll_cb min_parallel max_parallel max_outstanding nreqs);
74 73
75 require XSLoader; 74 require XSLoader;
76 XSLoader::load IO::AIO, $VERSION; 75 XSLoader::load IO::AIO, $VERSION;
77} 76}
… …
163 aio_read $fh, 7, 15, $buffer, 0, sub { 162 aio_read $fh, 7, 15, $buffer, 0, sub {
164 $_[0] > 0 or die "read error: $!"; 163 $_[0] > 0 or die "read error: $!";
165 print "read $_[0] bytes: <$buffer>\n"; 164 print "read $_[0] bytes: <$buffer>\n";
166 }; 165 };
167 166
167=item aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback
168
169Tries to copy C<$length> bytes from C<$in_fh> to C<$out_fh>. It starts
170reading at byte offset C<$in_offset>, and starts writing at the current
171file offset of C<$out_fh>. Because of that, it is not safe to issue more
172than one C<aio_sendfile> per C<$out_fh>, as they will interfere with each
173other.
174
175This call tries to make use of a native C<sendfile> syscall to provide
176zero-copy operation. For this to work, C<$out_fh> should refer to a
177socket, and C<$in_fh> should refer to mmap'able file.
178
179If the native sendfile call fails or is not implemented, it will be
180emulated, so you can call C<aio_sendfile> on any type of filehandle
181regardless of the limitations of the operating system.
182
183Please note, however, that C<aio_sendfile> can read more bytes from
184C<$in_fh> than are written, and there is no way to find out how many
185bytes have been read from C<aio_sendfile> alone, as C<aio_sendfile> only
186provides the number of bytes written to C<$out_fh>. Only if the result
187value equals C<$length> one can assume that C<$length> bytes have been
188read.
189
168=item aio_readahead $fh,$offset,$length, $callback 190=item aio_readahead $fh,$offset,$length, $callback
169 191
170C<aio_readahead> populates the page cache with data from a file so that 192C<aio_readahead> populates the page cache with data from a file so that
171subsequent reads from that file will not block on disk I/O. The C<$offset> 193subsequent reads from that file will not block on disk I/O. The C<$offset>
172argument specifies the starting point from which data is to be read and 194argument specifies the starting point from which data is to be read and
… …
209=item aio_rmdir $pathname, $callback 231=item aio_rmdir $pathname, $callback
210 232
211Asynchronously rmdir (delete) a directory and call the callback with the 233Asynchronously rmdir (delete) a directory and call the callback with the
212result code. 234result code.
213 235
236=item aio_readdir $pathname $callback
237
238Unlike the POSIX call of the same name, C<aio_readdir> reads an entire
239directory (i.e. opendir + readdir + closedir). The entries will not be
240sorted, and will B<NOT> include the C<.> and C<..> entries.
241
242The callback a single argument which is either C<undef> or an array-ref
243with the filenames.
244
214=item aio_fsync $fh, $callback 245=item aio_fsync $fh, $callback
215 246
216Asynchronously call fsync on the given filehandle and call the callback 247Asynchronously call fsync on the given filehandle and call the callback
217with the fsync result code. 248with the fsync result code.
218 249
… …
288 IO::AIO::poll_wait, IO::AIO::poll_cb 319 IO::AIO::poll_wait, IO::AIO::poll_cb
289 if IO::AIO::nreqs; 320 if IO::AIO::nreqs;
290 321
291=item IO::AIO::min_parallel $nthreads 322=item IO::AIO::min_parallel $nthreads
292 323
293Set the minimum number of AIO threads to C<$nthreads>. The default is 324Set the minimum number of AIO threads to C<$nthreads>. The current default
294C<1>, which means a single asynchronous operation can be done at one time 325is C<4>, which means four asynchronous operations can be done at one time
295(the number of outstanding operations, however, is unlimited). 326(the number of outstanding operations, however, is unlimited).
327
328IO::AIO starts threads only on demand, when an AIO request is queued and
329no free thread exists.
296 330
297It is recommended to keep the number of threads low, as some Linux 331It is recommended to keep the number of threads low, as some Linux
298kernel versions will scale negatively with the number of threads (higher 332kernel versions will scale negatively with the number of threads (higher
299parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32 333parallelity => MUCH higher latency). With current Linux 2.6 versions, 4-32
300threads should be fine. 334threads should be fine.
301 335
302Under normal circumstances you don't need to call this function, as this 336Under most circumstances you don't need to call this function, as the
303module automatically starts some threads (the exact number might change, 337module selects a default that is suitable for low to moderate load.
304and is currently 4).
305 338
306=item IO::AIO::max_parallel $nthreads 339=item IO::AIO::max_parallel $nthreads
307 340
308Sets the maximum number of AIO threads to C<$nthreads>. If more than 341Sets the maximum number of AIO threads to C<$nthreads>. If more than the
309the specified number of threads are currently running, kill them. This 342specified number of threads are currently running, this function kills
310function blocks until the limit is reached. 343them. This function blocks until the limit is reached.
344
345While C<$nthreads> are zero, aio requests get queued but not executed
346until the number of threads has been increased again.
311 347
312This module automatically runs C<max_parallel 0> at program end, to ensure 348This module automatically runs C<max_parallel 0> at program end, to ensure
313that all threads are killed and that there are no outstanding requests. 349that all threads are killed and that there are no outstanding requests.
314 350
315Under normal circumstances you don't need to call this function. 351Under normal circumstances you don't need to call this function.
… …
319Sets the maximum number of outstanding requests to C<$nreqs>. If you 355Sets the maximum number of outstanding requests to C<$nreqs>. If you
320try to queue up more than this number of requests, the caller will block until 356try to queue up more than this number of requests, the caller will block until
321some requests have been handled. 357some requests have been handled.
322 358
323The default is very large, so normally there is no practical limit. If you 359The default is very large, so normally there is no practical limit. If you
324queue up many requests in a loop it it often improves speed if you set 360queue up many requests in a loop it often improves speed if you set
325this to a relatively low number, such as C<100>. 361this to a relatively low number, such as C<100>.
326 362
327Under normal circumstances you don't need to call this function. 363Under normal circumstances you don't need to call this function.
328 364
329=back 365=back
… …
354 390
3551; 3911;
356 392
357=head2 FORK BEHAVIOUR 393=head2 FORK BEHAVIOUR
358 394
359Before the fork IO::AIO enters a quiescent state where no requests can be 395Before the fork, IO::AIO enters a quiescent state where no requests
360added in other threads and no results will be processed. After the fork 396can be added in other threads and no results will be processed. After
361the parent simply leaves the quiescent state and continues request/result 397the fork the parent simply leaves the quiescent state and continues
362processing, while the child clears the request/result queue and starts the 398request/result processing, while the child clears the request/result
363same number of threads as were in use by the parent. 399queue (so the requests started before the fork will only be handled in
400the parent). Threats will be started on demand until the limit ste in the
401parent process has been reached again.
364 402
365=head1 SEE ALSO 403=head1 SEE ALSO
366 404
367L<Coro>, L<Linux::AIO>. 405L<Coro>, L<Linux::AIO>.
368 406

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines