… | |
… | |
306 | } else { |
306 | } else { |
307 | die "open failed: $!\n"; |
307 | die "open failed: $!\n"; |
308 | } |
308 | } |
309 | }; |
309 | }; |
310 | |
310 | |
|
|
311 | In addition to all the common open modes/flags ("O_RDONLY", |
|
|
312 | "O_WRONLY", "O_RDWR", "O_CREAT", "O_TRUNC", "O_EXCL" and |
|
|
313 | "O_APPEND"), the following POSIX and non-POSIX constants are |
|
|
314 | available (missing ones on your system are, as usual, 0): |
|
|
315 | |
|
|
316 | "O_ASYNC", "O_DIRECT", "O_NOATIME", "O_CLOEXEC", "O_NOCTTY", |
|
|
317 | "O_NOFOLLOW", "O_NONBLOCK", "O_EXEC", "O_SEARCH", "O_DIRECTORY", |
|
|
318 | "O_DSYNC", "O_RSYNC", "O_SYNC" and "O_TTY_INIT". |
|
|
319 | |
311 | aio_close $fh, $callback->($status) |
320 | aio_close $fh, $callback->($status) |
312 | Asynchronously close a file and call the callback with the result |
321 | Asynchronously close a file and call the callback with the result |
313 | code. |
322 | code. |
314 | |
323 | |
315 | Unfortunately, you can't do this to perl. Perl *insists* very |
324 | Unfortunately, you can't do this to perl. Perl *insists* very |
… | |
… | |
358 | aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
367 | aio_sendfile $out_fh, $in_fh, $in_offset, $length, $callback->($retval) |
359 | Tries to copy $length bytes from $in_fh to $out_fh. It starts |
368 | Tries to copy $length bytes from $in_fh to $out_fh. It starts |
360 | reading at byte offset $in_offset, and starts writing at the current |
369 | reading at byte offset $in_offset, and starts writing at the current |
361 | file offset of $out_fh. Because of that, it is not safe to issue |
370 | file offset of $out_fh. Because of that, it is not safe to issue |
362 | more than one "aio_sendfile" per $out_fh, as they will interfere |
371 | more than one "aio_sendfile" per $out_fh, as they will interfere |
363 | with each other. |
372 | with each other. The same $in_fh works fine though, as this function |
|
|
373 | does not move or use the file offset of $in_fh. |
364 | |
374 | |
365 | Please note that "aio_sendfile" can read more bytes from $in_fh than |
375 | Please note that "aio_sendfile" can read more bytes from $in_fh than |
366 | are written, and there is no way to find out how many bytes have |
376 | are written, and there is no way to find out how many more bytes |
367 | been read from "aio_sendfile" alone, as "aio_sendfile" only provides |
377 | have been read from "aio_sendfile" alone, as "aio_sendfile" only |
368 | the number of bytes written to $out_fh. Only if the result value |
378 | provides the number of bytes written to $out_fh. Only if the result |
369 | equals $length one can assume that $length bytes have been read. |
379 | value equals $length one can assume that $length bytes have been |
|
|
380 | read. |
370 | |
381 | |
371 | Unlike with other "aio_" functions, it makes a lot of sense to use |
382 | Unlike with other "aio_" functions, it makes a lot of sense to use |
372 | "aio_sendfile" on non-blocking sockets, as long as one end |
383 | "aio_sendfile" on non-blocking sockets, as long as one end |
373 | (typically the $in_fh) is a file - the file I/O will then be |
384 | (typically the $in_fh) is a file - the file I/O will then be |
374 | asynchronous, while the socket I/O will be non-blocking. Note, |
385 | asynchronous, while the socket I/O will be non-blocking. Note, |
375 | however, that you can run into a trap where "aio_sendfile" reads |
386 | however, that you can run into a trap where "aio_sendfile" reads |
376 | some data with readahead, then fails to write all data, and when the |
387 | some data with readahead, then fails to write all data, and when the |
377 | socket is ready the next time, the data in the cache is already |
388 | socket is ready the next time, the data in the cache is already |
378 | lost, forcing "aio_sendfile" to again hit the disk. Explicit |
389 | lost, forcing "aio_sendfile" to again hit the disk. Explicit |
379 | "aio_read" + "aio_write" let's you control resource usage much |
390 | "aio_read" + "aio_write" let's you better control resource usage. |
380 | better. |
|
|
381 | |
391 | |
382 | This call tries to make use of a native "sendfile" syscall to |
392 | This call tries to make use of a native "sendfile"-like syscall to |
383 | provide zero-copy operation. For this to work, $out_fh should refer |
393 | provide zero-copy operation. For this to work, $out_fh should refer |
384 | to a socket, and $in_fh should refer to an mmap'able file. |
394 | to a socket, and $in_fh should refer to an mmap'able file. |
385 | |
395 | |
386 | If a native sendfile cannot be found or it fails with "ENOSYS", |
396 | If a native sendfile cannot be found or it fails with "ENOSYS", |
387 | "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or "ENOTSOCK", |
397 | "EINVAL", "ENOTSUP", "EOPNOTSUPP", "EAFNOSUPPORT", "EPROTOTYPE" or |
388 | it will be emulated, so you can call "aio_sendfile" on any type of |
398 | "ENOTSOCK", it will be emulated, so you can call "aio_sendfile" on |
389 | filehandle regardless of the limitations of the operating system. |
399 | any type of filehandle regardless of the limitations of the |
|
|
400 | operating system. |
|
|
401 | |
|
|
402 | As native sendfile syscalls (as practically any non-POSIX interface |
|
|
403 | hacked together in a hurry to improve benchmark numbers) tend to be |
|
|
404 | rather buggy on many systems, this implementation tries to work |
|
|
405 | around some known bugs in Linux and FreeBSD kernels (probably |
|
|
406 | others, too), but that might fail, so you really really should check |
|
|
407 | the return value of "aio_sendfile" - fewre bytes than expected might |
|
|
408 | have been transferred. |
390 | |
409 | |
391 | aio_readahead $fh,$offset,$length, $callback->($retval) |
410 | aio_readahead $fh,$offset,$length, $callback->($retval) |
392 | "aio_readahead" populates the page cache with data from a file so |
411 | "aio_readahead" populates the page cache with data from a file so |
393 | that subsequent reads from that file will not block on disk I/O. The |
412 | that subsequent reads from that file will not block on disk I/O. The |
394 | $offset argument specifies the starting point from which data is to |
413 | $offset argument specifies the starting point from which data is to |
… | |
… | |
569 | The flags are a combination of the following constants, ORed |
588 | The flags are a combination of the following constants, ORed |
570 | together (the flags will also be passed to the callback, possibly |
589 | together (the flags will also be passed to the callback, possibly |
571 | modified): |
590 | modified): |
572 | |
591 | |
573 | IO::AIO::READDIR_DENTS |
592 | IO::AIO::READDIR_DENTS |
574 | When this flag is off, then the callback gets an arrayref with |
593 | When this flag is off, then the callback gets an arrayref |
575 | of names only (as with "aio_readdir"), otherwise it gets an |
594 | consisting of names only (as with "aio_readdir"), otherwise it |
576 | arrayref with "[$name, $type, $inode]" arrayrefs, each |
595 | gets an arrayref with "[$name, $type, $inode]" arrayrefs, each |
577 | describing a single directory entry in more detail. |
596 | describing a single directory entry in more detail. |
578 | |
597 | |
579 | $name is the name of the entry. |
598 | $name is the name of the entry. |
580 | |
599 | |
581 | $type is one of the "IO::AIO::DT_xxx" constants: |
600 | $type is one of the "IO::AIO::DT_xxx" constants: |
… | |
… | |
594 | unspecified content on systems that do not deliver the inode |
613 | unspecified content on systems that do not deliver the inode |
595 | information. |
614 | information. |
596 | |
615 | |
597 | IO::AIO::READDIR_DIRS_FIRST |
616 | IO::AIO::READDIR_DIRS_FIRST |
598 | When this flag is set, then the names will be returned in an |
617 | When this flag is set, then the names will be returned in an |
599 | order where likely directories come first. This is useful when |
618 | order where likely directories come first, in optimal stat |
600 | you need to quickly find directories, or you want to find all |
619 | order. This is useful when you need to quickly find directories, |
601 | directories while avoiding to stat() each entry. |
620 | or you want to find all directories while avoiding to stat() |
|
|
621 | each entry. |
602 | |
622 | |
603 | If the system returns type information in readdir, then this is |
623 | If the system returns type information in readdir, then this is |
604 | used to find directories directly. Otherwise, likely directories |
624 | used to find directories directly. Otherwise, likely directories |
605 | are files beginning with ".", or otherwise files with no dots, |
625 | are names beginning with ".", or otherwise names with no dots, |
606 | of which files with short names are tried first. |
626 | of which names with short names are tried first. |
607 | |
627 | |
608 | IO::AIO::READDIR_STAT_ORDER |
628 | IO::AIO::READDIR_STAT_ORDER |
609 | When this flag is set, then the names will be returned in an |
629 | When this flag is set, then the names will be returned in an |
610 | order suitable for stat()'ing each one. That is, when you plan |
630 | order suitable for stat()'ing each one. That is, when you plan |
611 | to stat() all files in the given directory, then the returned |
631 | to stat() all files in the given directory, then the returned |
… | |
… | |
1022 | |
1042 | |
1023 | See "poll_cb" for an example. |
1043 | See "poll_cb" for an example. |
1024 | |
1044 | |
1025 | IO::AIO::poll_cb |
1045 | IO::AIO::poll_cb |
1026 | Process some outstanding events on the result pipe. You have to call |
1046 | Process some outstanding events on the result pipe. You have to call |
1027 | this regularly. Returns 0 if all events could be processed, or -1 if |
1047 | this regularly. Returns 0 if all events could be processed (or there |
1028 | it returned earlier for whatever reason. Returns immediately when no |
1048 | were no events to process), or -1 if it returned earlier for |
1029 | events are outstanding. The amount of events processed depends on |
1049 | whatever reason. Returns immediately when no events are outstanding. |
1030 | the settings of "IO::AIO::max_poll_req" and |
1050 | The amount of events processed depends on the settings of |
1031 | "IO::AIO::max_poll_time". |
1051 | "IO::AIO::max_poll_req" and "IO::AIO::max_poll_time". |
1032 | |
1052 | |
1033 | If not all requests were processed for whatever reason, the |
1053 | If not all requests were processed for whatever reason, the |
1034 | filehandle will still be ready when "poll_cb" returns, so normally |
1054 | filehandle will still be ready when "poll_cb" returns, so normally |
1035 | you don't have to do anything special to have it called later. |
1055 | you don't have to do anything special to have it called later. |
|
|
1056 | |
|
|
1057 | Apart from calling "IO::AIO::poll_cb" when the event filehandle |
|
|
1058 | becomes ready, it can be beneficial to call this function from loops |
|
|
1059 | which submit a lot of requests, to make sure the results get |
|
|
1060 | processed when they become available and not just when the loop is |
|
|
1061 | finished and the event loop takes over again. This function returns |
|
|
1062 | very fast when there are no outstanding requests. |
1036 | |
1063 | |
1037 | Example: Install an Event watcher that automatically calls |
1064 | Example: Install an Event watcher that automatically calls |
1038 | IO::AIO::poll_cb with high priority (more examples can be found in |
1065 | IO::AIO::poll_cb with high priority (more examples can be found in |
1039 | the SYNOPSIS section, at the top of this document): |
1066 | the SYNOPSIS section, at the top of this document): |
1040 | |
1067 | |
… | |
… | |
1153 | IO::AIO::idle_timeout $seconds |
1180 | IO::AIO::idle_timeout $seconds |
1154 | Sets the minimum idle timeout (default 10) after which worker |
1181 | Sets the minimum idle timeout (default 10) after which worker |
1155 | threads are allowed to exit. SEe "IO::AIO::max_idle". |
1182 | threads are allowed to exit. SEe "IO::AIO::max_idle". |
1156 | |
1183 | |
1157 | IO::AIO::max_outstanding $maxreqs |
1184 | IO::AIO::max_outstanding $maxreqs |
|
|
1185 | Sets the maximum number of outstanding requests to $nreqs. If you do |
|
|
1186 | queue up more than this number of requests, the next call to |
|
|
1187 | "IO::AIO::poll_cb" (and other functions calling "poll_cb", such as |
|
|
1188 | "IO::AIO::flush" or "IO::AIO::poll") will block until the limit is |
|
|
1189 | no longer exceeded. |
|
|
1190 | |
|
|
1191 | In other words, this setting does not enforce a queue limit, but can |
|
|
1192 | be used to make poll functions block if the limit is exceeded. |
|
|
1193 | |
1158 | This is a very bad function to use in interactive programs because |
1194 | This is a very bad function to use in interactive programs because |
1159 | it blocks, and a bad way to reduce concurrency because it is |
1195 | it blocks, and a bad way to reduce concurrency because it is |
1160 | inexact: Better use an "aio_group" together with a feed callback. |
1196 | inexact: Better use an "aio_group" together with a feed callback. |
1161 | |
1197 | |
1162 | Sets the maximum number of outstanding requests to $nreqs. If you do |
1198 | It's main use is in scripts without an event loop - when you want to |
1163 | queue up more than this number of requests, the next call to the |
1199 | stat a lot of files, you can write somehting like this: |
1164 | "poll_cb" (and "poll_some" and other functions calling "poll_cb") |
|
|
1165 | function will block until the limit is no longer exceeded. |
|
|
1166 | |
1200 | |
1167 | The default value is very large, so there is no practical limit on |
1201 | IO::AIO::max_outstanding 32; |
|
|
1202 | |
|
|
1203 | for my $path (...) { |
|
|
1204 | aio_stat $path , ...; |
|
|
1205 | IO::AIO::poll_cb; |
|
|
1206 | } |
|
|
1207 | |
|
|
1208 | IO::AIO::flush; |
|
|
1209 | |
|
|
1210 | The call to "poll_cb" inside the loop will normally return |
|
|
1211 | instantly, but as soon as more thna 32 reqeusts are in-flight, it |
|
|
1212 | will block until some requests have been handled. This keeps the |
|
|
1213 | loop from pushing a large number of "aio_stat" requests onto the |
|
|
1214 | queue. |
|
|
1215 | |
|
|
1216 | The default value for "max_outstanding" is very large, so there is |
1168 | the number of outstanding requests. |
1217 | no practical limit on the number of outstanding requests. |
1169 | |
|
|
1170 | You can still queue as many requests as you want. Therefore, |
|
|
1171 | "max_outstanding" is mainly useful in simple scripts (with low |
|
|
1172 | values) or as a stop gap to shield against fatal memory overflow |
|
|
1173 | (with large values). |
|
|
1174 | |
1218 | |
1175 | STATISTICAL INFORMATION |
1219 | STATISTICAL INFORMATION |
1176 | IO::AIO::nreqs |
1220 | IO::AIO::nreqs |
1177 | Returns the number of requests currently in the ready, execute or |
1221 | Returns the number of requests currently in the ready, execute or |
1178 | pending states (i.e. for which their callback has not been invoked |
1222 | pending states (i.e. for which their callback has not been invoked |
… | |
… | |
1324 | # Danga::Socket integration |
1368 | # Danga::Socket integration |
1325 | Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
1369 | Danga::Socket->AddOtherFds (IO::AIO::poll_fileno => |
1326 | \&IO::AIO::poll_cb); |
1370 | \&IO::AIO::poll_cb); |
1327 | |
1371 | |
1328 | FORK BEHAVIOUR |
1372 | FORK BEHAVIOUR |
1329 | This module should do "the right thing" when the process using it forks: |
1373 | Usage of pthreads in a program changes the semantics of fork |
|
|
1374 | considerably. Specifically, only async-safe functions can be called |
|
|
1375 | after fork. Perl doesn't know about this, so in general, you cannot call |
|
|
1376 | fork with defined behaviour in perl. IO::AIO uses pthreads, so this |
|
|
1377 | applies, but many other extensions and (for inexplicable reasons) perl |
|
|
1378 | itself often is linked against pthreads, so this limitation applies. |
1330 | |
1379 | |
1331 | Before the fork, IO::AIO enters a quiescent state where no requests can |
1380 | Some operating systems have extensions that allow safe use of fork, and |
1332 | be added in other threads and no results will be processed. After the |
1381 | this module should do "the right thing" on those, and tries on others. |
1333 | fork the parent simply leaves the quiescent state and continues |
1382 | At the time of this writing (2011) only GNU/Linux supports these |
1334 | request/result processing, while the child frees the request/result |
1383 | extensions to POSIX. |
1335 | queue (so that the requests started before the fork will only be handled |
|
|
1336 | in the parent). Threads will be started on demand until the limit set in |
|
|
1337 | the parent process has been reached again. |
|
|
1338 | |
|
|
1339 | In short: the parent will, after a short pause, continue as if fork had |
|
|
1340 | not been called, while the child will act as if IO::AIO has not been |
|
|
1341 | used yet. |
|
|
1342 | |
1384 | |
1343 | MEMORY USAGE |
1385 | MEMORY USAGE |
1344 | Per-request usage: |
1386 | Per-request usage: |
1345 | |
1387 | |
1346 | Each aio request uses - depending on your architecture - around 100-200 |
1388 | Each aio request uses - depending on your architecture - around 100-200 |