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

Comparing BDB/BDB.pm (file contents):
Revision 1.1 by root, Mon Feb 5 18:40:55 2007 UTC vs.
Revision 1.30 by root, Tue Dec 25 14:23:21 2007 UTC

1=head1 NAME 1=head1 NAME
2 2
3BDB::AIO - Asynchronous Berkeley DB access 3BDB - Asynchronous Berkeley DB access
4 4
5=head1 SYNOPSIS 5=head1 SYNOPSIS
6 6
7 use BDB::AIO; 7 use BDB;
8
9 my $env = db_env_create;
10
11 mkdir "bdtest", 0700;
12 db_env_open
13 $env,
14 "bdtest",
15 BDB::INIT_LOCK | BDB::INIT_LOG | BDB::INIT_MPOOL
16 | BDB::INIT_TXN | BDB::RECOVER | BDB::USE_ENVIRON | BDB::CREATE,
17 0600;
18
19 $env->set_flags (BDB::AUTO_COMMIT | BDB::TXN_NOSYNC, 1);
20
21 my $db = db_create $env;
22 db_open $db, undef, "table", undef, BDB::BTREE, BDB::AUTO_COMMIT | BDB::CREATE
23 | BDB::READ_UNCOMMITTED, 0600;
24 db_put $db, undef, "key", "data", 0, sub {
25 db_del $db, undef, "key";
26 };
27 db_sync $db;
28
29 # when you also use Coro, management is easy:
30 use Coro::BDB;
31
32 # automatic result processing with AnyEvent:
33 our $FH; open $FH, "<&=" . BDB::poll_fileno;
34 our $WATCHER = AnyEvent->io (fh => $FH, poll => 'r', cb => \&BDB::poll_cb);
35
36 # automatic result processing with EV:
37 my $WATCHER = EV::io BDB::poll_fileno, EV::READ, \&BDB::poll_cb;
38
39 # with Glib:
40 add_watch Glib::IO BDB::poll_fileno,
41 in => sub { BDB::poll_cb; 1 };
42
43 # or simply flush manually
44 BDB::flush;
45
8 46
9=head1 DESCRIPTION 47=head1 DESCRIPTION
10 48
11=head2 EXAMPLE 49See the BerkeleyDB documentation (L<http://www.oracle.com/technology/documentation/berkeley-db/db/index.html>).
50The BDB API is very similar to the C API (the translation has been very faithful).
51
52See also the example sections in the document below and possibly the eg/
53subdirectory of the BDB distribution. Last not least see the IO::AIO
54documentation, as that module uses almost the same asynchronous request
55model as this module.
56
57I know this is woefully inadequate documentation. Send a patch!
58
12 59
13=head1 REQUEST ANATOMY AND LIFETIME 60=head1 REQUEST ANATOMY AND LIFETIME
14 61
15Every request method creates a request. which is a C data structure not 62Every request method creates a request. which is a C data structure not
16directly visible to Perl. 63directly visible to Perl.
55 102
56=back 103=back
57 104
58=cut 105=cut
59 106
60package BDB::AIO; 107package BDB;
61 108
62no warnings; 109no warnings;
63use strict 'vars'; 110use strict 'vars';
64 111
65use base 'Exporter'; 112use base 'Exporter';
66 113
67BEGIN { 114BEGIN {
68 our $VERSION = '0.1'; 115 our $VERSION = '1.42';
69 116
70 our @BDB_REQ = qw(); 117 our @BDB_REQ = qw(
118 db_env_open db_env_close db_env_txn_checkpoint db_env_lock_detect
119 db_env_memp_sync db_env_memp_trickle
120 db_open db_close db_compact db_sync db_upgrade
121 db_put db_get db_pget db_del db_key_range
122 db_txn_commit db_txn_abort db_txn_finish
123 db_c_close db_c_count db_c_put db_c_get db_c_pget db_c_del
124 db_sequence_open db_sequence_close
125 db_sequence_get db_sequence_remove
126 );
127 our @EXPORT = (@BDB_REQ, qw(dbreq_pri dbreq_nice db_env_create db_create));
128 our @EXPORT_OK = qw(
71 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush 129 poll_fileno poll_cb poll_wait flush
72 min_parallel max_parallel max_idle 130 min_parallel max_parallel max_idle
73 nreqs nready npending nthreads 131 nreqs nready npending nthreads
74 max_poll_time max_poll_reqs); 132 max_poll_time max_poll_reqs
133 );
75 134
76 require XSLoader; 135 require XSLoader;
77 XSLoader::load ("BDB::AIO", $VERSION); 136 XSLoader::load ("BDB", $VERSION);
78} 137}
79 138
139=head2 BERKELEYDB FUNCTIONS
140
141All of these are functions. The create functions simply return a new
142object and never block. All the remaining functions all take an optional
143callback as last argument. If it is missing, then the function will be
144executed synchronously. In both cases, C<$!> will reflect the return value
145of the function.
146
147BDB functions that cannot block (mostly functions that manipulate
148settings) are method calls on the relevant objects, so the rule of thumb
149is: if its a method, its not blocking, if its a function, it takes a
150callback as last argument.
151
152In the following, C<$int> signifies an integer return value,
153C<octetstring> is a "binary string" (i.e. a perl string with no character
154indices >255), C<U32> is an unsigned 32 bit integer, C<int> is some
155integer, C<NV> is a floating point value.
156
157The C<SV *> types are generic perl scalars (for input and output of data
158values), and the C<SV *callback> is the optional callback function to call
159when the request is completed.
160
161The various C<DB_ENV> etc. arguments are handles return by
162C<db_env_create>, C<db_create>, C<txn_begin> and so on. If they have an
163appended C<_ornull> this means they are optional and you can pass C<undef>
164for them, resulting a NULL pointer on the C level.
165
166=head3 BDB functions
167
168Functions in the BDB namespace, exported by default:
169
170 $env = db_env_create (U32 env_flags = 0)
171 flags: RPCCLIENT
172
173 db_env_open (DB_ENV *env, octetstring db_home, U32 open_flags, int mode, SV *callback = &PL_sv_undef)
174 open_flags: INIT_CDB INIT_LOCK INIT_LOG INIT_MPOOL INIT_REP INIT_TXN RECOVER RECOVER_FATAL USE_ENVIRON USE_ENVIRON_ROOT CREATE LOCKDOWN PRIVATE REGISTER SYSTEM_MEM
175 db_env_close (DB_ENV *env, U32 flags = 0, SV *callback = &PL_sv_undef)
176 db_env_txn_checkpoint (DB_ENV *env, U32 kbyte = 0, U32 min = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
177 flags: FORCE
178 db_env_lock_detect (DB_ENV *env, U32 flags = 0, U32 atype = DB_LOCK_DEFAULT, SV *dummy = 0, SV *callback = &PL_sv_undef)
179 atype: LOCK_DEFAULT LOCK_EXPIRE LOCK_MAXLOCKS LOCK_MAXWRITE LOCK_MINLOCKS LOCK_MINWRITE LOCK_OLDEST LOCK_RANDOM LOCK_YOUNGEST
180 db_env_memp_sync (DB_ENV *env, SV *dummy = 0, SV *callback = &PL_sv_undef)
181 db_env_memp_trickle (DB_ENV *env, int percent, SV *dummy = 0, SV *callback = &PL_sv_undef)
182
183 $db = db_create (DB_ENV *env = 0, U32 flags = 0)
184 flags: XA_CREATE
185
186 db_open (DB *db, DB_TXN_ornull *txnid, octetstring file, octetstring database, int type, U32 flags, int mode, SV *callback = &PL_sv_undef)
187 flags: AUTO_COMMIT CREATE EXCL MULTIVERSION NOMMAP RDONLY READ_UNCOMMITTED THREAD TRUNCATE
188 db_close (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
189 flags: DB_NOSYNC
190 db_upgrade (DB *db, octetstring file, U32 flags = 0, SV *callback = &PL_sv_undef)
191 db_compact (DB *db, DB_TXN_ornull *txn = 0, SV *start = 0, SV *stop = 0, SV *unused1 = 0, U32 flags = DB_FREE_SPACE, SV *unused2 = 0, SV *callback = &PL_sv_undef)
192 flags: FREELIST_ONLY FREE_SPACE
193 db_sync (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
194 db_key_range (DB *db, DB_TXN_ornull *txn, SV *key, SV *key_range, U32 flags = 0, SV *callback = &PL_sv_undef)
195 db_put (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
196 flags: APPEND NODUPDATA NOOVERWRITE
197 db_get (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
198 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
199 db_pget (DB *db, DB_TXN_ornull *txn, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
200 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
201 db_del (DB *db, DB_TXN_ornull *txn, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
202 db_txn_commit (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
203 flags: TXN_NOSYNC TXN_SYNC
204 db_txn_abort (DB_TXN *txn, SV *callback = &PL_sv_undef)
205
206 db_c_close (DBC *dbc, SV *callback = &PL_sv_undef)
207 db_c_count (DBC *dbc, SV *count, U32 flags = 0, SV *callback = &PL_sv_undef)
208 db_c_put (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
209 flags: AFTER BEFORE CURRENT KEYFIRST KEYLAST NODUPDATA
210 db_c_get (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
211 flags: CURRENT FIRST GET_BOTH GET_BOTH_RANGE GET_RECNO JOIN_ITEM LAST NEXT NEXT_DUP NEXT_NODUP PREV PREV_DUP PREV_NODUP SET SET_RANGE SET_RECNO READ_UNCOMMITTED MULTIPLE MULTIPLE_KEY RMW
212 db_c_pget (DBC *dbc, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
213 db_c_del (DBC *dbc, U32 flags = 0, SV *callback = &PL_sv_undef)
214
215 db_sequence_open (DB_SEQUENCE *seq, DB_TXN_ornull *txnid, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
216 flags: CREATE EXCL
217 db_sequence_close (DB_SEQUENCE *seq, U32 flags = 0, SV *callback = &PL_sv_undef)
218 db_sequence_get (DB_SEQUENCE *seq, DB_TXN_ornull *txnid, int delta, SV *seq_value, U32 flags = DB_TXN_NOSYNC, SV *callback = &PL_sv_undef)
219 flags: TXN_NOSYNC
220 db_sequence_remove (DB_SEQUENCE *seq, DB_TXN_ornull *txnid = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
221 flags: TXN_NOSYNC
222
223=head4 db_txn_finish (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
224
225This is not actually a Berkeley DB function but a BDB module
226extension. The background for this exytension is: It is very annoying to
227have to check every single BDB function for error returns and provide a
228codepath out of your transaction. While the BDB module still makes this
229possible, it contains the following extensions:
230
231When a transaction-protected function returns any operating system
232error (errno > 0), BDB will set the C<TXN_DEADLOCK> flag on the
233transaction. This flag is also set by Berkeley DB functions themselves
234when an operation fails with LOCK_DEADLOCK, and it causes all further
235operations on that transaction (including C<db_txn_commit>) to fail.
236
237The C<db_txn_finish> request will look at this flag, and, if it is set,
238will automatically call C<db_txn_abort> (setting errno to C<LOCK_DEADLOCK>
239if it isn't set to something else yet). If it isn't set, it will call
240C<db_txn_commit> and return the error normally.
241
242How to use this? Easy: just write your transaction normally:
243
244 my $txn = $db_env->txn_begin;
245 db_get $db, $txn, "key", my $data;
246 db_put $db, $txn, "key", $data + 1 unless $! == BDB::NOTFOUND;
247 db_txn_finish $txn;
248 die "transaction failed" if $!;
249
250That is, handle only the expected errors. If something unexpected happens
251(EIO, LOCK_NOTGRANTED or a deadlock in either db_get or db_put), then the remaining
252requests (db_put in this case) will simply be skipped (they will fail with
253LOCK_DEADLOCK) and the transaction will be aborted.
254
255You can use the C<< $txn->failed >> method to check wether a transaction
256has failed in this way and abort further processing (excluding
257C<db_txn_finish>).
258
259=head3 DB_ENV/database environment methods
260
261Methods available on DB_ENV/$env handles:
262
263 DESTROY (DB_ENV_ornull *env)
264 CODE:
265 if (env)
266 env->close (env, 0);
267
268 $int = $env->set_data_dir (const char *dir)
269 $int = $env->set_tmp_dir (const char *dir)
270 $int = $env->set_lg_dir (const char *dir)
271 $int = $env->set_shm_key (long shm_key)
272 $int = $env->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
273 $int = $env->set_flags (U32 flags, int onoff)
274 $env->set_errfile (FILE *errfile = 0)
275 $env->set_msgfile (FILE *msgfile = 0)
276 $int = $env->set_verbose (U32 which, int onoff = 1)
277 $int = $env->set_encrypt (const char *password, U32 flags = 0)
278 $int = $env->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
279 $int = $env->set_mp_max_openfd (int maxopenfd);
280 $int = $env->set_mp_max_write (int maxwrite, int maxwrite_sleep);
281 $int = $env->set_mp_mmapsize (int mmapsize_mb)
282 $int = $env->set_lk_detect (U32 detect = DB_LOCK_DEFAULT)
283 $int = $env->set_lk_max_lockers (U32 max)
284 $int = $env->set_lk_max_locks (U32 max)
285 $int = $env->set_lk_max_objects (U32 max)
286 $int = $env->set_lg_bsize (U32 max)
287 $int = $env->set_lg_max (U32 max)
288 $int = $env->mutex_set_increment (U32 increment)
289 $int = $env->mutex_set_tas_spins (U32 tas_spins)
290 $int = $env->mutex_set_max (U32 max)
291 $int = $env->mutex_set_align (U32 align)
292
293 $txn = $env->txn_begin (DB_TXN_ornull *parent = 0, U32 flags = 0)
294 flags: READ_COMMITTED READ_UNCOMMITTED TXN_NOSYNC TXN_NOWAIT TXN_SNAPSHOT TXN_SYNC TXN_WAIT TXN_WRITE_NOSYNC
295
296=head4 Example:
297
298 use AnyEvent;
299 use BDB;
300
301 our $FH; open $FH, "<&=" . BDB::poll_fileno;
302 our $WATCHER = AnyEvent->io (fh => $FH, poll => 'r', cb => \&BDB::poll_cb);
303
304 BDB::min_parallel 8;
305
306 my $env = db_env_create;
307
308 mkdir "bdtest", 0700;
309 db_env_open
310 $env,
311 "bdtest",
312 BDB::INIT_LOCK | BDB::INIT_LOG | BDB::INIT_MPOOL | BDB::INIT_TXN | BDB::RECOVER | BDB::USE_ENVIRON | BDB::CREATE,
313 0600;
314
315 $env->set_flags (BDB::AUTO_COMMIT | BDB::TXN_NOSYNC, 1);
316
317
318=head3 DB/database methods
319
320Methods available on DB/$db handles:
321
322 DESTROY (DB_ornull *db)
323 CODE:
324 if (db)
325 {
326 SV *env = (SV *)db->app_private;
327 db->close (db, 0);
328 SvREFCNT_dec (env);
329 }
330
331 $int = $db->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
332 $int = $db->set_flags (U32 flags)
333 flags: CHKSUM ENCRYPT TXN_NOT_DURABLE
334 Btree: DUP DUPSORT RECNUM REVSPLITOFF
335 Hash: DUP DUPSORT
336 Queue: INORDER
337 Recno: RENUMBER SNAPSHOT
338
339 $int = $db->set_encrypt (const char *password, U32 flags)
340 $int = $db->set_lorder (int lorder)
341 $int = $db->set_bt_minkey (U32 minkey)
342 $int = $db->set_re_delim (int delim)
343 $int = $db->set_re_pad (int re_pad)
344 $int = $db->set_re_source (char *source)
345 $int = $db->set_re_len (U32 re_len)
346 $int = $db->set_h_ffactor (U32 h_ffactor)
347 $int = $db->set_h_nelem (U32 h_nelem)
348 $int = $db->set_q_extentsize (U32 extentsize)
349
350 $dbc = $db->cursor (DB_TXN_ornull *txn = 0, U32 flags = 0)
351 flags: READ_COMMITTED READ_UNCOMMITTED WRITECURSOR TXN_SNAPSHOT
352 $seq = $db->sequence (U32 flags = 0)
353
354=head4 Example:
355
356 my $db = db_create $env;
357 db_open $db, undef, "table", undef, BDB::BTREE, BDB::AUTO_COMMIT | BDB::CREATE | BDB::READ_UNCOMMITTED, 0600;
358
359 for (1..1000) {
360 db_put $db, undef, "key $_", "data $_";
361
362 db_key_range $db, undef, "key $_", my $keyrange;
363 my ($lt, $eq, $gt) = @$keyrange;
364 }
365
366 db_del $db, undef, "key $_" for 1..1000;
367
368 db_sync $db;
369
370
371=head3 DB_TXN/transaction methods
372
373Methods available on DB_TXN/$txn handles:
374
375 DESTROY (DB_TXN_ornull *txn)
376 CODE:
377 if (txn)
378 txn->abort (txn);
379
380 $int = $txn->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
381 flags: SET_LOCK_TIMEOUT SET_TXN_TIMEOUT
382
383 $bool = $txn->failed
384 # see db_txn_finish documentation, above
385
386
387=head3 DBC/cursor methods
388
389Methods available on DBC/$dbc handles:
390
391 DESTROY (DBC_ornull *dbc)
392 CODE:
393 if (dbc)
394 dbc->c_close (dbc);
395
396 $int = $cursor->set_priority ($priority = PRIORITY_*)
397
398=head4 Example:
399
400 my $c = $db->cursor;
401
402 for (;;) {
403 db_c_get $c, my $key, my $data, BDB::NEXT;
404 warn "<$!,$key,$data>";
405 last if $!;
406 }
407
408 db_c_close $c;
409
410
411=head3 DB_SEQUENCE/sequence methods
412
413Methods available on DB_SEQUENCE/$seq handles:
414
415 DESTROY (DB_SEQUENCE_ornull *seq)
416 CODE:
417 if (seq)
418 seq->close (seq, 0);
419
420 $int = $seq->initial_value (db_seq_t value)
421 $int = $seq->set_cachesize (U32 size)
422 $int = $seq->set_flags (U32 flags)
423 flags: SEQ_DEC SEQ_INC SEQ_WRAP
424 $int = $seq->set_range (db_seq_t min, db_seq_t max)
425
426=head4 Example:
427
428 my $seq = $db->sequence;
429
430 db_sequence_open $seq, undef, "seq", BDB::CREATE;
431 db_sequence_get $seq, undef, 1, my $value;
432
433
80=head2 SUPPORT FUNCTIONS 434=head2 SUPPORT FUNCTIONS
81 435
82=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION 436=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
83 437
84=over 4 438=over 4
85 439
440=item $msg = BDB::strerror [$errno]
441
442Returns the string corresponding to the given errno value. If no argument
443is given, use C<$!>.
444
86=item $fileno = BDB::AIO::poll_fileno 445=item $fileno = BDB::poll_fileno
87 446
88Return the I<request result pipe file descriptor>. This filehandle must be 447Return the I<request result pipe file descriptor>. This filehandle must be
89polled for reading by some mechanism outside this module (e.g. Event or 448polled for reading by some mechanism outside this module (e.g. Event or
90select, see below or the SYNOPSIS). If the pipe becomes readable you have 449select, see below or the SYNOPSIS). If the pipe becomes readable you have
91to call C<poll_cb> to check the results. 450to call C<poll_cb> to check the results.
92 451
93See C<poll_cb> for an example. 452See C<poll_cb> for an example.
94 453
95=item BDB::AIO::poll_cb 454=item BDB::poll_cb
96 455
97Process some outstanding events on the result pipe. You have to call this 456Process some outstanding events on the result pipe. You have to call this
98regularly. Returns the number of events processed. Returns immediately 457regularly. Returns the number of events processed. Returns immediately
99when no events are outstanding. The amount of events processed depends on 458when no events are outstanding. The amount of events processed depends on
100the settings of C<BDB::AIO::max_poll_req> and C<BDB::AIO::max_poll_time>. 459the settings of C<BDB::max_poll_req> and C<BDB::max_poll_time>.
101 460
102If not all requests were processed for whatever reason, the filehandle 461If not all requests were processed for whatever reason, the filehandle
103will still be ready when C<poll_cb> returns. 462will still be ready when C<poll_cb> returns.
104 463
105Example: Install an Event watcher that automatically calls 464Example: Install an Event watcher that automatically calls
106BDB::AIO::poll_cb with high priority: 465BDB::poll_cb with high priority:
107 466
108 Event->io (fd => BDB::AIO::poll_fileno, 467 Event->io (fd => BDB::poll_fileno,
109 poll => 'r', async => 1, 468 poll => 'r', async => 1,
110 cb => \&BDB::AIO::poll_cb); 469 cb => \&BDB::poll_cb);
111 470
112=item BDB::AIO::max_poll_reqs $nreqs 471=item BDB::max_poll_reqs $nreqs
113 472
114=item BDB::AIO::max_poll_time $seconds 473=item BDB::max_poll_time $seconds
115 474
116These set the maximum number of requests (default C<0>, meaning infinity) 475These set the maximum number of requests (default C<0>, meaning infinity)
117that are being processed by C<BDB::AIO::poll_cb> in one call, respectively 476that are being processed by C<BDB::poll_cb> in one call, respectively
118the maximum amount of time (default C<0>, meaning infinity) spent in 477the maximum amount of time (default C<0>, meaning infinity) spent in
119C<BDB::AIO::poll_cb> to process requests (more correctly the mininum amount 478C<BDB::poll_cb> to process requests (more correctly the mininum amount
120of time C<poll_cb> is allowed to use). 479of time C<poll_cb> is allowed to use).
121 480
122Setting C<max_poll_time> to a non-zero value creates an overhead of one 481Setting C<max_poll_time> to a non-zero value creates an overhead of one
123syscall per request processed, which is not normally a problem unless your 482syscall per request processed, which is not normally a problem unless your
124callbacks are really really fast or your OS is really really slow (I am 483callbacks are really really fast or your OS is really really slow (I am
128interactiveness when perl is not fast enough to process all requests in 487interactiveness when perl is not fast enough to process all requests in
129time. 488time.
130 489
131For interactive programs, values such as C<0.01> to C<0.1> should be fine. 490For interactive programs, values such as C<0.01> to C<0.1> should be fine.
132 491
133Example: Install an Event watcher that automatically calls 492Example: Install an EV watcher that automatically calls
134BDB::AIO::poll_cb with low priority, to ensure that other parts of the 493BDB::poll_cb with low priority, to ensure that other parts of the
135program get the CPU sometimes even under high AIO load. 494program get the CPU sometimes even under high load.
136 495
137 # try not to spend much more than 0.1s in poll_cb 496 # try not to spend much more than 0.1s in poll_cb
138 BDB::AIO::max_poll_time 0.1; 497 BDB::max_poll_time 0.1;
139 498
140 # use a low priority so other tasks have priority 499 my $bdb_poll = EV::io BDB::poll_fileno, EV::READ, \&BDB::poll_cb);
141 Event->io (fd => BDB::AIO::poll_fileno,
142 poll => 'r', nice => 1,
143 cb => &BDB::AIO::poll_cb);
144 500
145=item BDB::AIO::poll_wait 501=item BDB::poll_wait
146 502
147If there are any outstanding requests and none of them in the result 503If there are any outstanding requests and none of them in the result
148phase, wait till the result filehandle becomes ready for reading (simply 504phase, wait till the result filehandle becomes ready for reading (simply
149does a C<select> on the filehandle. This is useful if you want to 505does a C<select> on the filehandle. This is useful if you want to
150synchronously wait for some requests to finish). 506synchronously wait for some requests to finish).
151 507
152See C<nreqs> for an example. 508See C<nreqs> for an example.
153 509
154=item BDB::AIO::poll 510=item BDB::poll
155 511
156Waits until some requests have been handled. 512Waits until some requests have been handled.
157 513
158Returns the number of requests processed, but is otherwise strictly 514Returns the number of requests processed, but is otherwise strictly
159equivalent to: 515equivalent to:
160 516
161 BDB::AIO::poll_wait, BDB::AIO::poll_cb 517 BDB::poll_wait, BDB::poll_cb
162 518
163=item BDB::AIO::flush 519=item BDB::flush
164 520
165Wait till all outstanding AIO requests have been handled. 521Wait till all outstanding BDB requests have been handled.
166 522
167Strictly equivalent to: 523Strictly equivalent to:
168 524
169 BDB::AIO::poll_wait, BDB::AIO::poll_cb 525 BDB::poll_wait, BDB::poll_cb
170 while BDB::AIO::nreqs; 526 while BDB::nreqs;
527
528=back
171 529
172=head3 CONTROLLING THE NUMBER OF THREADS 530=head3 CONTROLLING THE NUMBER OF THREADS
173 531
532=over 4
533
174=item BDB::AIO::min_parallel $nthreads 534=item BDB::min_parallel $nthreads
175 535
176Set the minimum number of AIO threads to C<$nthreads>. The current 536Set the minimum number of BDB threads to C<$nthreads>. The current
177default is C<8>, which means eight asynchronous operations can execute 537default is C<8>, which means eight asynchronous operations can execute
178concurrently at any one time (the number of outstanding requests, 538concurrently at any one time (the number of outstanding requests,
179however, is unlimited). 539however, is unlimited).
180 540
181BDB::AIO starts threads only on demand, when an AIO request is queued and 541BDB starts threads only on demand, when an BDB request is queued and
182no free thread exists. Please note that queueing up a hundred requests can 542no free thread exists. Please note that queueing up a hundred requests can
183create demand for a hundred threads, even if it turns out that everything 543create demand for a hundred threads, even if it turns out that everything
184is in the cache and could have been processed faster by a single thread. 544is in the cache and could have been processed faster by a single thread.
185 545
186It is recommended to keep the number of threads relatively low, as some 546It is recommended to keep the number of threads relatively low, as some
189versions, 4-32 threads should be fine. 549versions, 4-32 threads should be fine.
190 550
191Under most circumstances you don't need to call this function, as the 551Under most circumstances you don't need to call this function, as the
192module selects a default that is suitable for low to moderate load. 552module selects a default that is suitable for low to moderate load.
193 553
194=item BDB::AIO::max_parallel $nthreads 554=item BDB::max_parallel $nthreads
195 555
196Sets the maximum number of AIO threads to C<$nthreads>. If more than the 556Sets the maximum number of BDB threads to C<$nthreads>. If more than the
197specified number of threads are currently running, this function kills 557specified number of threads are currently running, this function kills
198them. This function blocks until the limit is reached. 558them. This function blocks until the limit is reached.
199 559
200While C<$nthreads> are zero, aio requests get queued but not executed 560While C<$nthreads> are zero, aio requests get queued but not executed
201until the number of threads has been increased again. 561until the number of threads has been increased again.
203This module automatically runs C<max_parallel 0> at program end, to ensure 563This module automatically runs C<max_parallel 0> at program end, to ensure
204that all threads are killed and that there are no outstanding requests. 564that all threads are killed and that there are no outstanding requests.
205 565
206Under normal circumstances you don't need to call this function. 566Under normal circumstances you don't need to call this function.
207 567
208=item BDB::AIO::max_idle $nthreads 568=item BDB::max_idle $nthreads
209 569
210Limit the number of threads (default: 4) that are allowed to idle (i.e., 570Limit the number of threads (default: 4) that are allowed to idle (i.e.,
211threads that did not get a request to process within 10 seconds). That 571threads that did not get a request to process within 10 seconds). That
212means if a thread becomes idle while C<$nthreads> other threads are also 572means if a thread becomes idle while C<$nthreads> other threads are also
213idle, it will free its resources and exit. 573idle, it will free its resources and exit.
218 578
219The default is probably ok in most situations, especially if thread 579The default is probably ok in most situations, especially if thread
220creation is fast. If thread creation is very slow on your system you might 580creation is fast. If thread creation is very slow on your system you might
221want to use larger values. 581want to use larger values.
222 582
223=item $oldmaxreqs = BDB::AIO::max_outstanding $maxreqs 583=item $oldmaxreqs = BDB::max_outstanding $maxreqs
224 584
225This is a very bad function to use in interactive programs because it 585This is a very bad function to use in interactive programs because it
226blocks, and a bad way to reduce concurrency because it is inexact: Better 586blocks, and a bad way to reduce concurrency because it is inexact: Better
227use an C<aio_group> together with a feed callback. 587use an C<aio_group> together with a feed callback.
228 588
236 596
237You can still queue as many requests as you want. Therefore, 597You can still queue as many requests as you want. Therefore,
238C<max_oustsanding> is mainly useful in simple scripts (with low values) or 598C<max_oustsanding> is mainly useful in simple scripts (with low values) or
239as a stop gap to shield against fatal memory overflow (with large values). 599as a stop gap to shield against fatal memory overflow (with large values).
240 600
601=item BDB::set_sync_prepare $cb
602
603Sets a callback that is called whenever a request is created without an
604explicit callback. It has to return two code references. The first is used
605as the request callback, and the second is called to wait until the first
606callback has been called. The default implementation works like this:
607
608 sub {
609 my $status;
610 (
611 sub { $status = $! },
612 sub { BDB::poll while !defined $status; $! = $status },
613 )
614 }
615
616=back
617
241=head3 STATISTICAL INFORMATION 618=head3 STATISTICAL INFORMATION
242 619
620=over 4
621
243=item BDB::AIO::nreqs 622=item BDB::nreqs
244 623
245Returns the number of requests currently in the ready, execute or pending 624Returns the number of requests currently in the ready, execute or pending
246states (i.e. for which their callback has not been invoked yet). 625states (i.e. for which their callback has not been invoked yet).
247 626
248Example: wait till there are no outstanding requests anymore: 627Example: wait till there are no outstanding requests anymore:
249 628
250 BDB::AIO::poll_wait, BDB::AIO::poll_cb 629 BDB::poll_wait, BDB::poll_cb
251 while BDB::AIO::nreqs; 630 while BDB::nreqs;
252 631
253=item BDB::AIO::nready 632=item BDB::nready
254 633
255Returns the number of requests currently in the ready state (not yet 634Returns the number of requests currently in the ready state (not yet
256executed). 635executed).
257 636
258=item BDB::AIO::npending 637=item BDB::npending
259 638
260Returns the number of requests currently in the pending state (executed, 639Returns the number of requests currently in the pending state (executed,
261but not yet processed by poll_cb). 640but not yet processed by poll_cb).
262 641
263=back 642=back
264 643
265=cut 644=cut
266 645
267# support function to convert a fd into a perl filehandle 646set_sync_prepare {
268sub _fd2fh { 647 my $status;
269 return undef if $_[0] < 0; 648 (
270 649 sub {
271 # try to generate nice filehandles 650 $status = $!;
272 my $sym = "BDB::AIO::fd#$_[0]"; 651 },
273 local *$sym; 652 sub {
274 653 BDB::poll while !defined $status;
275 open *$sym, "+<&=$_[0]" # usually works under any unix 654 $! = $status;
276 or open *$sym, "<&=$_[0]" # cygwin needs this 655 },
277 or open *$sym, ">&=$_[0]" # or this 656 )
278 or return undef; 657};
279
280 *$sym
281}
282 658
283min_parallel 8; 659min_parallel 8;
284 660
285END { flush } 661END { flush }
286 662
288 664
289=head2 FORK BEHAVIOUR 665=head2 FORK BEHAVIOUR
290 666
291This module should do "the right thing" when the process using it forks: 667This module should do "the right thing" when the process using it forks:
292 668
293Before the fork, IO::AIO enters a quiescent state where no requests 669Before the fork, BDB enters a quiescent state where no requests
294can be added in other threads and no results will be processed. After 670can be added in other threads and no results will be processed. After
295the fork the parent simply leaves the quiescent state and continues 671the fork the parent simply leaves the quiescent state and continues
296request/result processing, while the child frees the request/result queue 672request/result processing, while the child frees the request/result queue
297(so that the requests started before the fork will only be handled in the 673(so that the requests started before the fork will only be handled in the
298parent). Threads will be started on demand until the limit set in the 674parent). Threads will be started on demand until the limit set in the
299parent process has been reached again. 675parent process has been reached again.
300 676
301In short: the parent will, after a short pause, continue as if fork had 677In short: the parent will, after a short pause, continue as if fork had
302not been called, while the child will act as if IO::AIO has not been used 678not been called, while the child will act as if BDB has not been used
303yet. 679yet.
680
681Win32 note: there is no fork on win32, and perls emulation of it is too
682broken to be supported, so do not use BDB in a windows pseudo-fork, better
683yet, switch to a more capable platform.
304 684
305=head2 MEMORY USAGE 685=head2 MEMORY USAGE
306 686
307Per-request usage: 687Per-request usage:
308 688
310bytes of memory. In addition, stat requests need a stat buffer (possibly 690bytes of memory. In addition, stat requests need a stat buffer (possibly
311a few hundred bytes), readdir requires a result buffer and so on. Perl 691a few hundred bytes), readdir requires a result buffer and so on. Perl
312scalars and other data passed into aio requests will also be locked and 692scalars and other data passed into aio requests will also be locked and
313will consume memory till the request has entered the done state. 693will consume memory till the request has entered the done state.
314 694
315This is now awfully much, so queuing lots of requests is not usually a 695This is not awfully much, so queuing lots of requests is not usually a
316problem. 696problem.
317 697
318Per-thread usage: 698Per-thread usage:
319 699
320In the execution phase, some aio requests require more memory for 700In the execution phase, some aio requests require more memory for
321temporary buffers, and each thread requires a stack and other data 701temporary buffers, and each thread requires a stack and other data
322structures (usually around 16k-128k, depending on the OS). 702structures (usually around 16k-128k, depending on the OS).
323 703
324=head1 KNOWN BUGS 704=head1 KNOWN BUGS
325 705
326Known bugs will be fixed in the next release. 706Known bugs will be fixed in the next release, except:
707
708 If you use a transaction in any request, and the request returns
709 with an operating system error or DB_LOCK_NOTGRANTED, the internal
710 TXN_DEADLOCK flag will be set on the transaction. See C<db_txn_finish>,
711 above.
327 712
328=head1 SEE ALSO 713=head1 SEE ALSO
329 714
330L<Coro::AIO>. 715L<Coro::BDB>, L<IO::AIO>.
331 716
332=head1 AUTHOR 717=head1 AUTHOR
333 718
334 Marc Lehmann <schmorp@schmorp.de> 719 Marc Lehmann <schmorp@schmorp.de>
335 http://home.schmorp.de/ 720 http://home.schmorp.de/

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines