… | |
… | |
26 | $req->cancel; # cancel request if still in queue |
26 | $req->cancel; # cancel request if still in queue |
27 | |
27 | |
28 | my $grp = aio_group sub { print "all stats done\n" }; |
28 | my $grp = aio_group sub { print "all stats done\n" }; |
29 | add $grp aio_stat "..." for ...; |
29 | add $grp aio_stat "..." for ...; |
30 | |
30 | |
31 | # AnyEvent integration |
31 | # AnyEvent integration (EV, Event, Glib, Tk, urxvt, pureperl...) |
32 | open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; |
32 | open my $fh, "<&=" . IO::AIO::poll_fileno or die "$!"; |
33 | my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); |
33 | my $w = AnyEvent->io (fh => $fh, poll => 'r', cb => sub { IO::AIO::poll_cb }); |
|
|
34 | |
|
|
35 | # EV integration |
|
|
36 | my $w = EV::io IO::AIO::poll_fileno, EV::READ, \&IO::AIO::poll_cb; |
34 | |
37 | |
35 | # Event integration |
38 | # Event integration |
36 | Event->io (fd => IO::AIO::poll_fileno, |
39 | Event->io (fd => IO::AIO::poll_fileno, |
37 | poll => 'r', |
40 | poll => 'r', |
38 | cb => \&IO::AIO::poll_cb); |
41 | cb => \&IO::AIO::poll_cb); |
… | |
… | |
62 | etc.), but can also be used to easily do operations in parallel that are |
65 | etc.), but can also be used to easily do operations in parallel that are |
63 | normally done sequentially, e.g. stat'ing many files, which is much faster |
66 | normally done sequentially, e.g. stat'ing many files, which is much faster |
64 | on a RAID volume or over NFS when you do a number of stat operations |
67 | on a RAID volume or over NFS when you do a number of stat operations |
65 | concurrently. |
68 | concurrently. |
66 | |
69 | |
67 | While most of this works on all types of file descriptors (for example |
70 | While most of this works on all types of file descriptors (for |
68 | sockets), using these functions on file descriptors that support |
71 | example sockets), using these functions on file descriptors that |
69 | nonblocking operation (again, sockets, pipes etc.) is very inefficient or |
72 | support nonblocking operation (again, sockets, pipes etc.) is very |
70 | might not work (aio_read fails on sockets/pipes/fifos). Use an event loop |
73 | inefficient. Use an event loop for that (such as the L<Event|Event> |
71 | for that (such as the L<Event|Event> module): IO::AIO will naturally fit |
74 | module): IO::AIO will naturally fit into such an event loop itself. |
72 | into such an event loop itself. |
|
|
73 | |
75 | |
74 | In this version, a number of threads are started that execute your |
76 | In this version, a number of threads are started that execute your |
75 | requests and signal their completion. You don't need thread support |
77 | requests and signal their completion. You don't need thread support |
76 | in perl, and the threads created by this module will not be visible |
78 | in perl, and the threads created by this module will not be visible |
77 | to perl. In the future, this module might make use of the native aio |
79 | to perl. In the future, this module might make use of the native aio |
… | |
… | |
79 | not well-supported or restricted (GNU/Linux doesn't allow them on normal |
81 | not well-supported or restricted (GNU/Linux doesn't allow them on normal |
80 | files currently, for example), and they would only support aio_read and |
82 | files currently, for example), and they would only support aio_read and |
81 | aio_write, so the remaining functionality would have to be implemented |
83 | aio_write, so the remaining functionality would have to be implemented |
82 | using threads anyway. |
84 | using threads anyway. |
83 | |
85 | |
84 | Although the module will work with in the presence of other (Perl-) |
86 | Although the module will work in the presence of other (Perl-) threads, |
85 | threads, it is currently not reentrant in any way, so use appropriate |
87 | it is currently not reentrant in any way, so use appropriate locking |
86 | locking yourself, always call C<poll_cb> from within the same thread, or |
88 | yourself, always call C<poll_cb> from within the same thread, or never |
87 | never call C<poll_cb> (or other C<aio_> functions) recursively. |
89 | call C<poll_cb> (or other C<aio_> functions) recursively. |
88 | |
90 | |
89 | =head2 EXAMPLE |
91 | =head2 EXAMPLE |
90 | |
92 | |
91 | This is a simple example that uses the Event module and loads |
93 | This is a simple example that uses the Event module and loads |
92 | F</etc/passwd> asynchronously: |
94 | F</etc/passwd> asynchronously: |
… | |
… | |
184 | |
186 | |
185 | =cut |
187 | =cut |
186 | |
188 | |
187 | package IO::AIO; |
189 | package IO::AIO; |
188 | |
190 | |
|
|
191 | use Carp (); |
|
|
192 | |
189 | no warnings; |
193 | no warnings; |
190 | use strict 'vars'; |
194 | use strict 'vars'; |
191 | |
195 | |
192 | use base 'Exporter'; |
196 | use base 'Exporter'; |
193 | |
197 | |
194 | BEGIN { |
198 | BEGIN { |
195 | our $VERSION = '2.4'; |
199 | our $VERSION = '2.6'; |
196 | |
200 | |
197 | our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat |
201 | our @AIO_REQ = qw(aio_sendfile aio_read aio_write aio_open aio_close aio_stat |
198 | aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink |
202 | aio_lstat aio_unlink aio_rmdir aio_readdir aio_scandir aio_symlink |
199 | aio_readlink aio_fsync aio_fdatasync aio_readahead aio_rename aio_link |
203 | aio_readlink aio_sync aio_fsync aio_fdatasync aio_readahead aio_rename aio_link |
200 | aio_move aio_copy aio_group aio_nop aio_mknod aio_load aio_rmtree aio_mkdir |
204 | aio_move aio_copy aio_group aio_nop aio_mknod aio_load aio_rmtree aio_mkdir |
201 | aio_chown aio_chmod aio_utime); |
205 | aio_chown aio_chmod aio_utime aio_truncate); |
202 | our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice aio_block)); |
206 | our @EXPORT = (@AIO_REQ, qw(aioreq_pri aioreq_nice aio_block)); |
203 | our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush |
207 | our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush |
204 | min_parallel max_parallel max_idle |
208 | min_parallel max_parallel max_idle |
205 | nreqs nready npending nthreads |
209 | nreqs nready npending nthreads |
206 | max_poll_time max_poll_reqs); |
210 | max_poll_time max_poll_reqs); |
… | |
… | |
312 | |
316 | |
313 | |
317 | |
314 | =item aio_close $fh, $callback->($status) |
318 | =item aio_close $fh, $callback->($status) |
315 | |
319 | |
316 | Asynchronously close a file and call the callback with the result |
320 | Asynchronously close a file and call the callback with the result |
317 | code. I<WARNING:> although accepted, you should not pass in a perl |
321 | code. |
318 | filehandle here, as perl will likely close the file descriptor another |
|
|
319 | time when the filehandle is destroyed. Normally, you can safely call perls |
|
|
320 | C<close> or just let filehandles go out of scope. |
|
|
321 | |
322 | |
322 | This is supposed to be a bug in the API, so that might change. It's |
323 | Unfortunately, you can't do this to perl. Perl I<insists> very strongly on |
323 | therefore best to avoid this function. |
324 | closing the file descriptor associated with the filehandle itself. Here is |
|
|
325 | what aio_close will try: |
|
|
326 | |
|
|
327 | 1. dup()licate the fd |
|
|
328 | 2. asynchronously close() the duplicated fd |
|
|
329 | 3. dup()licate the fd once more |
|
|
330 | 4. let perl close() the filehandle |
|
|
331 | 5. asynchronously close the duplicated fd |
|
|
332 | |
|
|
333 | The idea is that the first close() flushes stuff to disk that closing an |
|
|
334 | fd will flush, so when perl closes the fd, nothing much will need to be |
|
|
335 | flushed. The second async. close() will then flush stuff to disk that |
|
|
336 | closing the last fd to the file will flush. |
|
|
337 | |
|
|
338 | Just FYI, SuSv3 has this to say on close: |
|
|
339 | |
|
|
340 | All outstanding record locks owned by the process on the file |
|
|
341 | associated with the file descriptor shall be removed. |
|
|
342 | |
|
|
343 | If fildes refers to a socket, close() shall cause the socket to be |
|
|
344 | destroyed. ... close() shall block for up to the current linger |
|
|
345 | interval until all data is transmitted. |
|
|
346 | [this actually sounds like a specification bug, but who knows] |
|
|
347 | |
|
|
348 | And at least Linux additionally actually flushes stuff on every close, |
|
|
349 | even when the file itself is still open. |
|
|
350 | |
|
|
351 | Sounds enourmously inefficient and complicated? Yes... please show me how |
|
|
352 | to nuke perl's fd out of existence... |
|
|
353 | |
|
|
354 | =cut |
|
|
355 | |
|
|
356 | sub aio_close($;$) { |
|
|
357 | aio_block { |
|
|
358 | my ($fh, $cb) = @_; |
|
|
359 | |
|
|
360 | my $pri = aioreq_pri; |
|
|
361 | my $grp = aio_group $cb; |
|
|
362 | |
|
|
363 | my $fd = fileno $fh; |
|
|
364 | |
|
|
365 | defined $fd or Carp::croak "aio_close called with fd-less filehandle"; |
|
|
366 | |
|
|
367 | # if the dups fail we will simply get EBADF |
|
|
368 | my $fd2 = _dup $fd; |
|
|
369 | aioreq_pri $pri; |
|
|
370 | add $grp _aio_close $fd2, sub { |
|
|
371 | my $fd2 = _dup $fd; |
|
|
372 | close $fh; |
|
|
373 | aioreq_pri $pri; |
|
|
374 | add $grp _aio_close $fd2, sub { |
|
|
375 | $grp->result ($_[0]); |
|
|
376 | }; |
|
|
377 | }; |
|
|
378 | |
|
|
379 | $grp |
|
|
380 | } |
|
|
381 | } |
324 | |
382 | |
325 | |
383 | |
326 | =item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
384 | =item aio_read $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
327 | |
385 | |
328 | =item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
386 | =item aio_write $fh,$offset,$length, $data,$dataoffset, $callback->($retval) |
329 | |
387 | |
330 | Reads or writes C<length> bytes from the specified C<fh> and C<offset> |
388 | Reads or writes C<$length> bytes from the specified C<$fh> and C<$offset> |
331 | into the scalar given by C<data> and offset C<dataoffset> and calls the |
389 | into the scalar given by C<$data> and offset C<$dataoffset> and calls the |
332 | callback without the actual number of bytes read (or -1 on error, just |
390 | callback without the actual number of bytes read (or -1 on error, just |
333 | like the syscall). |
391 | like the syscall). |
334 | |
392 | |
|
|
393 | If C<$offset> is undefined, then the current file descriptor offset will |
|
|
394 | be used (and updated), otherwise the file descriptor offset will not be |
|
|
395 | changed by these calls. |
|
|
396 | |
|
|
397 | If C<$length> is undefined in C<aio_write>, use the remaining length of C<$data>. |
|
|
398 | |
|
|
399 | If C<$dataoffset> is less than zero, it will be counted from the end of |
|
|
400 | C<$data>. |
|
|
401 | |
335 | The C<$data> scalar I<MUST NOT> be modified in any way while the request |
402 | The C<$data> scalar I<MUST NOT> be modified in any way while the request |
336 | is outstanding. Modifying it can result in segfaults or WW3 (if the |
403 | is outstanding. Modifying it can result in segfaults or World War III (if |
337 | necessary/optional hardware is installed). |
404 | the necessary/optional hardware is installed). |
338 | |
405 | |
339 | Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at |
406 | Example: Read 15 bytes at offset 7 into scalar C<$buffer>, starting at |
340 | offset C<0> within the scalar: |
407 | offset C<0> within the scalar: |
341 | |
408 | |
342 | aio_read $fh, 7, 15, $buffer, 0, sub { |
409 | aio_read $fh, 7, 15, $buffer, 0, sub { |
… | |
… | |
417 | utime(2). If called on a file descriptor, uses futimes(2) if available, |
484 | utime(2). If called on a file descriptor, uses futimes(2) if available, |
418 | otherwise returns ENOSYS, so this is not portable. |
485 | otherwise returns ENOSYS, so this is not portable. |
419 | |
486 | |
420 | Examples: |
487 | Examples: |
421 | |
488 | |
422 | # set atime and mtime to current time: |
489 | # set atime and mtime to current time (basically touch(1)): |
423 | aio_utime "path", undef, undef; |
490 | aio_utime "path", undef, undef; |
424 | # set atime to current time and mtime to beginning of the epoch: |
491 | # set atime to current time and mtime to beginning of the epoch: |
425 | aio_utime "path", time, undef; # undef==0 |
492 | aio_utime "path", time, undef; # undef==0 |
426 | |
493 | |
427 | |
494 | |
… | |
… | |
434 | |
501 | |
435 | # same as "chown root path" in the shell: |
502 | # same as "chown root path" in the shell: |
436 | aio_chown "path", 0, -1; |
503 | aio_chown "path", 0, -1; |
437 | # same as above: |
504 | # same as above: |
438 | aio_chown "path", 0, undef; |
505 | aio_chown "path", 0, undef; |
|
|
506 | |
|
|
507 | |
|
|
508 | =item aio_truncate $fh_or_path, $offset, $callback->($status) |
|
|
509 | |
|
|
510 | Works like truncate(2) or ftruncate(2). |
439 | |
511 | |
440 | |
512 | |
441 | =item aio_chmod $fh_or_path, $mode, $callback->($status) |
513 | =item aio_chmod $fh_or_path, $mode, $callback->($status) |
442 | |
514 | |
443 | Works like perl's C<chmod> function. |
515 | Works like perl's C<chmod> function. |
… | |
… | |
821 | |
893 | |
822 | $grp |
894 | $grp |
823 | } |
895 | } |
824 | } |
896 | } |
825 | |
897 | |
|
|
898 | =item aio_sync $callback->($status) |
|
|
899 | |
|
|
900 | Asynchronously call sync and call the callback when finished. |
|
|
901 | |
826 | =item aio_fsync $fh, $callback->($status) |
902 | =item aio_fsync $fh, $callback->($status) |
827 | |
903 | |
828 | Asynchronously call fsync on the given filehandle and call the callback |
904 | Asynchronously call fsync on the given filehandle and call the callback |
829 | with the fsync result code. |
905 | with the fsync result code. |
830 | |
906 | |
… | |
… | |
1194 | This is a very bad function to use in interactive programs because it |
1270 | This is a very bad function to use in interactive programs because it |
1195 | blocks, and a bad way to reduce concurrency because it is inexact: Better |
1271 | blocks, and a bad way to reduce concurrency because it is inexact: Better |
1196 | use an C<aio_group> together with a feed callback. |
1272 | use an C<aio_group> together with a feed callback. |
1197 | |
1273 | |
1198 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
1274 | Sets the maximum number of outstanding requests to C<$nreqs>. If you |
1199 | to queue up more than this number of requests, the next call to the |
1275 | do queue up more than this number of requests, the next call to the |
1200 | C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>) |
1276 | C<poll_cb> (and C<poll_some> and other functions calling C<poll_cb>) |
1201 | function will block until the limit is no longer exceeded. |
1277 | function will block until the limit is no longer exceeded. |
1202 | |
1278 | |
1203 | The default value is very large, so there is no practical limit on the |
1279 | The default value is very large, so there is no practical limit on the |
1204 | number of outstanding requests. |
1280 | number of outstanding requests. |
… | |
… | |
1234 | but not yet processed by poll_cb). |
1310 | but not yet processed by poll_cb). |
1235 | |
1311 | |
1236 | =back |
1312 | =back |
1237 | |
1313 | |
1238 | =cut |
1314 | =cut |
1239 | |
|
|
1240 | # support function to convert a fd into a perl filehandle |
|
|
1241 | sub _fd2fh { |
|
|
1242 | return undef if $_[0] < 0; |
|
|
1243 | |
|
|
1244 | # try to generate nice filehandles |
|
|
1245 | my $sym = "IO::AIO::fd#$_[0]"; |
|
|
1246 | local *$sym; |
|
|
1247 | |
|
|
1248 | open *$sym, "+<&=$_[0]" # usually works under any unix |
|
|
1249 | or open *$sym, "<&=$_[0]" # cygwin needs this |
|
|
1250 | or open *$sym, ">&=$_[0]" # or this |
|
|
1251 | or return undef; |
|
|
1252 | |
|
|
1253 | *$sym |
|
|
1254 | } |
|
|
1255 | |
1315 | |
1256 | min_parallel 8; |
1316 | min_parallel 8; |
1257 | |
1317 | |
1258 | END { flush } |
1318 | END { flush } |
1259 | |
1319 | |
… | |
… | |
1283 | bytes of memory. In addition, stat requests need a stat buffer (possibly |
1343 | bytes of memory. In addition, stat requests need a stat buffer (possibly |
1284 | a few hundred bytes), readdir requires a result buffer and so on. Perl |
1344 | a few hundred bytes), readdir requires a result buffer and so on. Perl |
1285 | scalars and other data passed into aio requests will also be locked and |
1345 | scalars and other data passed into aio requests will also be locked and |
1286 | will consume memory till the request has entered the done state. |
1346 | will consume memory till the request has entered the done state. |
1287 | |
1347 | |
1288 | This is now awfully much, so queuing lots of requests is not usually a |
1348 | This is not awfully much, so queuing lots of requests is not usually a |
1289 | problem. |
1349 | problem. |
1290 | |
1350 | |
1291 | Per-thread usage: |
1351 | Per-thread usage: |
1292 | |
1352 | |
1293 | In the execution phase, some aio requests require more memory for |
1353 | In the execution phase, some aio requests require more memory for |