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 1.15 by root, Thu Sep 13 21:34:00 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 8
9=head1 DESCRIPTION 9=head1 DESCRIPTION
10 10
11=head2 EXAMPLE 11See the BerkeleyDB documentation (L<http://www.oracle.com/technology/documentation/berkeley-db/db/index.html>).
12The BDB API is very similar to the C API (the translation has been very faithful).
13
14See also the example sections in the document below and possibly the eg/
15subdirectory of the BDB distribution. Last not least see the IO::AIO
16documentation, as that module uses almost the same asynchronous request
17model as this module.
18
19I know this is woefully inadequate documentation. Send a patch!
20
12 21
13=head1 REQUEST ANATOMY AND LIFETIME 22=head1 REQUEST ANATOMY AND LIFETIME
14 23
15Every request method creates a request. which is a C data structure not 24Every request method creates a request. which is a C data structure not
16directly visible to Perl. 25directly visible to Perl.
55 64
56=back 65=back
57 66
58=cut 67=cut
59 68
60package BDB::AIO; 69package BDB;
61 70
62no warnings; 71no warnings;
63use strict 'vars'; 72use strict 'vars';
64 73
65use base 'Exporter'; 74use base 'Exporter';
66 75
67BEGIN { 76BEGIN {
68 our $VERSION = '0.1'; 77 our $VERSION = '1.1';
69 78
70 our @BDB_REQ = qw(); 79 our @BDB_REQ = qw(
80 db_env_open db_env_close db_env_txn_checkpoint db_env_lock_detect
81 db_env_memp_sync db_env_memp_trickle
82 db_open db_close db_compact db_sync db_put db_get db_pget db_del db_key_range
83 db_txn_commit db_txn_abort db_txn_finish
84 db_c_close db_c_count db_c_put db_c_get db_c_pget db_c_del
85 db_sequence_open db_sequence_close
86 db_sequence_get db_sequence_remove
87 );
88 our @EXPORT = (@BDB_REQ, qw(dbreq_pri dbreq_nice db_env_create db_create));
89 our @EXPORT_OK = qw(
71 our @EXPORT_OK = qw(poll_fileno poll_cb poll_wait flush 90 poll_fileno poll_cb poll_wait flush
72 min_parallel max_parallel max_idle 91 min_parallel max_parallel max_idle
73 nreqs nready npending nthreads 92 nreqs nready npending nthreads
74 max_poll_time max_poll_reqs); 93 max_poll_time max_poll_reqs
94 );
75 95
76 require XSLoader; 96 require XSLoader;
77 XSLoader::load ("BDB::AIO", $VERSION); 97 XSLoader::load ("BDB", $VERSION);
78} 98}
79 99
100=head2 BERKELEYDB FUNCTIONS
101
102All of these are functions. The create functions simply return a new
103object and never block. All the remaining functions all take an optional
104callback as last argument. If it is missing, then the fucntion will be
105executed synchronously.
106
107BDB functions that cannot block (mostly functions that manipulate
108settings) are method calls on the relevant objects, so the rule of thumb
109is: if its a method, its not blocking, if its a function, it takes a
110callback as last argument.
111
112In the following, C<$int> signifies an integer return value,
113C<octetstring> is a "binary string" (i.e. a perl string with no character
114indices >255), C<U32> is an unsigned 32 bit integer, C<int> is some
115integer, C<NV> is a floating point value.
116
117The C<SV *> types are generic perl scalars (for input and output of data
118values), and the C<SV *callback> is the optional callback function to call
119when the request is completed.
120
121The various C<DB_ENV> etc. arguments are handles return by
122C<db_env_create>, C<db_create>, C<txn_begin> and so on. If they have an
123appended C<_ornull> this means they are optional and you can pass C<undef>
124for them, resulting a NULL pointer on the C level.
125
126=head3 BDB functions
127
128Functions in the BDB namespace, exported by default:
129
130 $env = db_env_create (U32 env_flags = 0)
131 flags: RPCCLIENT
132
133 db_env_open (DB_ENV *env, octetstring db_home, U32 open_flags, int mode, SV *callback = &PL_sv_undef)
134 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
135 db_env_close (DB_ENV *env, U32 flags = 0, SV *callback = &PL_sv_undef)
136 db_env_txn_checkpoint (DB_ENV *env, U32 kbyte = 0, U32 min = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
137 flags: FORCE
138 db_env_lock_detect (DB_ENV *env, U32 flags = 0, U32 atype = DB_LOCK_DEFAULT, SV *dummy = 0, SV *callback = &PL_sv_undef)
139 atype: LOCK_DEFAULT LOCK_EXPIRE LOCK_MAXLOCKS LOCK_MAXWRITE LOCK_MINLOCKS LOCK_MINWRITE LOCK_OLDEST LOCK_RANDOM LOCK_YOUNGEST
140 db_env_memp_sync (DB_ENV *env, SV *dummy = 0, SV *callback = &PL_sv_undef)
141 db_env_memp_trickle (DB_ENV *env, int percent, SV *dummy = 0, SV *callback = &PL_sv_undef)
142
143 $db = db_create (DB_ENV *env = 0, U32 flags = 0)
144 flags: XA_CREATE
145
146 db_open (DB *db, DB_TXN_ornull *txnid, octetstring file, octetstring database, int type, U32 flags, int mode, SV *callback = &PL_sv_undef)
147 flags: AUTO_COMMIT CREATE EXCL MULTIVERSION NOMMAP RDONLY READ_UNCOMMITTED THREAD TRUNCATE
148 db_close (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
149 flags: DB_NOSYNC
150 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)
151 flags: FREELIST_ONLY FREE_SPACE
152 db_sync (DB *db, U32 flags = 0, SV *callback = &PL_sv_undef)
153 db_key_range (DB *db, DB_TXN_ornull *txn, SV *key, SV *key_range, U32 flags = 0, SV *callback = &PL_sv_undef)
154 db_put (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
155 flags: APPEND NODUPDATA NOOVERWRITE
156 db_get (DB *db, DB_TXN_ornull *txn, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
157 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
158 db_pget (DB *db, DB_TXN_ornull *txn, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
159 flags: CONSUME CONSUME_WAIT GET_BOTH SET_RECNO MULTIPLE READ_COMMITTED READ_UNCOMMITTED RMW
160 db_del (DB *db, DB_TXN_ornull *txn, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
161 db_txn_commit (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
162 flags: TXN_NOSYNC TXN_SYNC
163 db_txn_abort (DB_TXN *txn, SV *callback = &PL_sv_undef)
164
165 db_c_close (DBC *dbc, SV *callback = &PL_sv_undef)
166 db_c_count (DBC *dbc, SV *count, U32 flags = 0, SV *callback = &PL_sv_undef)
167 db_c_put (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
168 flags: AFTER BEFORE CURRENT KEYFIRST KEYLAST NODUPDATA
169 db_c_get (DBC *dbc, SV *key, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
170 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
171 db_c_pget (DBC *dbc, SV *key, SV *pkey, SV *data, U32 flags = 0, SV *callback = &PL_sv_undef)
172 db_c_del (DBC *dbc, U32 flags = 0, SV *callback = &PL_sv_undef)
173
174 db_sequence_open (DB_SEQUENCE *seq, DB_TXN_ornull *txnid, SV *key, U32 flags = 0, SV *callback = &PL_sv_undef)
175 flags: CREATE EXCL
176 db_sequence_close (DB_SEQUENCE *seq, U32 flags = 0, SV *callback = &PL_sv_undef)
177 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)
178 flags: TXN_NOSYNC
179 db_sequence_remove (DB_SEQUENCE *seq, DB_TXN_ornull *txnid = 0, U32 flags = 0, SV *callback = &PL_sv_undef)
180 flags: TXN_NOSYNC
181
182=head4 db_txn_finish (DB_TXN *txn, U32 flags = 0, SV *callback = &PL_sv_undef)
183
184This is not a Berkeley DB function but a BDB module extension. It is very
185annoying to have to check every single BDB function for error returns and
186provide a codepath out of your transaction. While the BDB module still
187makes this possible, it contains the following extensions:
188
189When a transaction-protected function returns any operating system
190error (errno > 0), BDB will set the C<TXN_DEADLOCK> flag on the
191transaction. This flag is also set by Berkeley DB functions externally
192when an operation fails with LOCK_DEADLOCK, and it causes all further
193operations on that transaction (including C<db_txn_commit>) to fail.
194
195The C<db_txn_finish> request will look at this flag, and, if it is set,
196will automatically call C<db_txn_abort> (setting errno to C<LOCK_DEADLOCK>
197if it isn't set). If it isn't set, it will call C<db_txn_commit> and
198return the error normally.
199
200How to use this? Easy: just write your transaction normally:
201
202 my $txn = $db_env->txn_begin;
203 db_get $db, $txn, "key", my $data;
204 db_put $db, $txn, "key", $data + 1 unless $! == BDB::NOTFOUND;
205 db_txn_finish $txn;
206 die "transaction failed" if $!;
207
208That is, handle only the expected errors. If something unexpected happens
209(EIO, LOCK_NOTGRANTED or a deadlock in either db_get or db_put), then the remaining
210requests (db_put in this case) will simply be skipped (they will fail with
211LOCK_DEADLOCK) and the transaction will be aborted.
212
213You cna use the C<< $txn->failed >> method to check wether a transaction
214has failed in this way and abort further processing (excluding
215C<db_txn_finish>).
216
217=head3 DB_ENV/database environment methods
218
219Methods available on DB_ENV/$env handles:
220
221 DESTROY (DB_ENV_ornull *env)
222 CODE:
223 if (env)
224 env->close (env, 0);
225
226 $int = $env->set_data_dir (const char *dir)
227 $int = $env->set_tmp_dir (const char *dir)
228 $int = $env->set_lg_dir (const char *dir)
229 $int = $env->set_shm_key (long shm_key)
230 $int = $env->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
231 $int = $env->set_flags (U32 flags, int onoff)
232 $env->set_errfile (FILE *errfile = 0)
233 $env->set_msgfile (FILE *msgfile = 0)
234 $int = $env->set_verbose (U32 which, int onoff = 1)
235 $int = $env->set_encrypt (const char *password, U32 flags = 0)
236 $int = $env->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
237 $int = $env->set_mp_max_openfd (int maxopenfd);
238 $int = $env->set_mp_max_write (int maxwrite, int maxwrite_sleep);
239 $int = $env->set_mp_mmapsize (int mmapsize_mb)
240 $int = $env->set_lk_detect (U32 detect = DB_LOCK_DEFAULT)
241 $int = $env->set_lk_max_lockers (U32 max)
242 $int = $env->set_lk_max_locks (U32 max)
243 $int = $env->set_lk_max_objects (U32 max)
244 $int = $env->set_lg_bsize (U32 max)
245 $int = $env->set_lg_max (U32 max)
246
247 $txn = $env->txn_begin (DB_TXN_ornull *parent = 0, U32 flags = 0)
248 flags: READ_COMMITTED READ_UNCOMMITTED TXN_NOSYNC TXN_NOWAIT TXN_SNAPSHOT TXN_SYNC TXN_WAIT TXN_WRITE_NOSYNC
249
250=head4 Example:
251
252 use AnyEvent;
253 use BDB;
254
255 our $FH; open $FH, "<&=" . BDB::poll_fileno;
256 our $WATCHER = AnyEvent->io (fh => $FH, poll => 'r', cb => \&BDB::poll_cb);
257
258 BDB::min_parallel 8;
259
260 my $env = db_env_create;
261
262 mkdir "bdtest", 0700;
263 db_env_open
264 $env,
265 "bdtest",
266 BDB::INIT_LOCK | BDB::INIT_LOG | BDB::INIT_MPOOL | BDB::INIT_TXN | BDB::RECOVER | BDB::USE_ENVIRON | BDB::CREATE,
267 0600;
268
269 $env->set_flags (BDB::AUTO_COMMIT | BDB::TXN_NOSYNC, 1);
270
271
272=head3 DB/database methods
273
274Methods available on DB/$db handles:
275
276 DESTROY (DB_ornull *db)
277 CODE:
278 if (db)
279 {
280 SV *env = (SV *)db->app_private;
281 db->close (db, 0);
282 SvREFCNT_dec (env);
283 }
284
285 $int = $db->set_cachesize (U32 gbytes, U32 bytes, int ncache = 0)
286 $int = $db->set_flags (U32 flags)
287 flags: CHKSUM ENCRYPT TXN_NOT_DURABLE
288 Btree: DUP DUPSORT RECNUM REVSPLITOFF
289 Hash: DUP DUPSORT
290 Queue: INORDER
291 Recno: RENUMBER SNAPSHOT
292
293 $int = $db->set_encrypt (const char *password, U32 flags)
294 $int = $db->set_lorder (int lorder)
295 $int = $db->set_bt_minkey (U32 minkey)
296 $int = $db->set_re_delim (int delim)
297 $int = $db->set_re_pad (int re_pad)
298 $int = $db->set_re_source (char *source)
299 $int = $db->set_re_len (U32 re_len)
300 $int = $db->set_h_ffactor (U32 h_ffactor)
301 $int = $db->set_h_nelem (U32 h_nelem)
302 $int = $db->set_q_extentsize (U32 extentsize)
303
304 $dbc = $db->cursor (DB_TXN_ornull *txn = 0, U32 flags = 0)
305 flags: READ_COMMITTED READ_UNCOMMITTED WRITECURSOR TXN_SNAPSHOT
306 $seq = $db->sequence (U32 flags = 0)
307
308=head4 Example:
309
310 my $db = db_create $env;
311 db_open $db, undef, "table", undef, BDB::BTREE, BDB::AUTO_COMMIT | BDB::CREATE | BDB::READ_UNCOMMITTED, 0600;
312
313 for (1..1000) {
314 db_put $db, undef, "key $_", "data $_";
315
316 db_key_range $db, undef, "key $_", my $keyrange;
317 my ($lt, $eq, $gt) = @$keyrange;
318 }
319
320 db_del $db, undef, "key $_" for 1..1000;
321
322 db_sync $db;
323
324
325=head3 DB_TXN/transaction methods
326
327Methods available on DB_TXN/$txn handles:
328
329 DESTROY (DB_TXN_ornull *txn)
330 CODE:
331 if (txn)
332 txn->abort (txn);
333
334 $int = $txn->set_timeout (NV timeout_seconds, U32 flags = SET_TXN_TIMEOUT)
335 flags: SET_LOCK_TIMEOUT SET_TXN_TIMEOUT
336
337 $bool = $txn->failed
338 # see db_txn_finish documentation, above
339
340
341=head3 DBC/cursor methods
342
343Methods available on DBC/$dbc handles:
344
345 DESTROY (DBC_ornull *dbc)
346 CODE:
347 if (dbc)
348 dbc->c_close (dbc);
349
350=head4 Example:
351
352 my $c = $db->cursor;
353
354 for (;;) {
355 db_c_get $c, my $key, my $data, BDB::NEXT;
356 warn "<$!,$key,$data>";
357 last if $!;
358 }
359
360 db_c_close $c;
361
362
363=head3 DB_SEQUENCE/sequence methods
364
365Methods available on DB_SEQUENCE/$seq handles:
366
367 DESTROY (DB_SEQUENCE_ornull *seq)
368 CODE:
369 if (seq)
370 seq->close (seq, 0);
371
372 $int = $seq->initial_value (db_seq_t value)
373 $int = $seq->set_cachesize (U32 size)
374 $int = $seq->set_flags (U32 flags)
375 flags: SEQ_DEC SEQ_INC SEQ_WRAP
376 $int = $seq->set_range (db_seq_t min, db_seq_t max)
377
378=head4 Example:
379
380 my $seq = $db->sequence;
381
382 db_sequence_open $seq, undef, "seq", BDB::CREATE;
383 db_sequence_get $seq, undef, 1, my $value;
384
385
80=head2 SUPPORT FUNCTIONS 386=head2 SUPPORT FUNCTIONS
81 387
82=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION 388=head3 EVENT PROCESSING AND EVENT LOOP INTEGRATION
83 389
84=over 4 390=over 4
85 391
86=item $fileno = BDB::AIO::poll_fileno 392=item $fileno = BDB::poll_fileno
87 393
88Return the I<request result pipe file descriptor>. This filehandle must be 394Return the I<request result pipe file descriptor>. This filehandle must be
89polled for reading by some mechanism outside this module (e.g. Event or 395polled 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 396select, see below or the SYNOPSIS). If the pipe becomes readable you have
91to call C<poll_cb> to check the results. 397to call C<poll_cb> to check the results.
92 398
93See C<poll_cb> for an example. 399See C<poll_cb> for an example.
94 400
95=item BDB::AIO::poll_cb 401=item BDB::poll_cb
96 402
97Process some outstanding events on the result pipe. You have to call this 403Process some outstanding events on the result pipe. You have to call this
98regularly. Returns the number of events processed. Returns immediately 404regularly. Returns the number of events processed. Returns immediately
99when no events are outstanding. The amount of events processed depends on 405when 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>. 406the settings of C<BDB::max_poll_req> and C<BDB::max_poll_time>.
101 407
102If not all requests were processed for whatever reason, the filehandle 408If not all requests were processed for whatever reason, the filehandle
103will still be ready when C<poll_cb> returns. 409will still be ready when C<poll_cb> returns.
104 410
105Example: Install an Event watcher that automatically calls 411Example: Install an Event watcher that automatically calls
106BDB::AIO::poll_cb with high priority: 412BDB::poll_cb with high priority:
107 413
108 Event->io (fd => BDB::AIO::poll_fileno, 414 Event->io (fd => BDB::poll_fileno,
109 poll => 'r', async => 1, 415 poll => 'r', async => 1,
110 cb => \&BDB::AIO::poll_cb); 416 cb => \&BDB::poll_cb);
111 417
112=item BDB::AIO::max_poll_reqs $nreqs 418=item BDB::max_poll_reqs $nreqs
113 419
114=item BDB::AIO::max_poll_time $seconds 420=item BDB::max_poll_time $seconds
115 421
116These set the maximum number of requests (default C<0>, meaning infinity) 422These 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 423that are being processed by C<BDB::poll_cb> in one call, respectively
118the maximum amount of time (default C<0>, meaning infinity) spent in 424the maximum amount of time (default C<0>, meaning infinity) spent in
119C<BDB::AIO::poll_cb> to process requests (more correctly the mininum amount 425C<BDB::poll_cb> to process requests (more correctly the mininum amount
120of time C<poll_cb> is allowed to use). 426of time C<poll_cb> is allowed to use).
121 427
122Setting C<max_poll_time> to a non-zero value creates an overhead of one 428Setting 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 429syscall 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 430callbacks are really really fast or your OS is really really slow (I am
129time. 435time.
130 436
131For interactive programs, values such as C<0.01> to C<0.1> should be fine. 437For interactive programs, values such as C<0.01> to C<0.1> should be fine.
132 438
133Example: Install an Event watcher that automatically calls 439Example: Install an Event watcher that automatically calls
134BDB::AIO::poll_cb with low priority, to ensure that other parts of the 440BDB::poll_cb with low priority, to ensure that other parts of the
135program get the CPU sometimes even under high AIO load. 441program get the CPU sometimes even under high AIO load.
136 442
137 # try not to spend much more than 0.1s in poll_cb 443 # try not to spend much more than 0.1s in poll_cb
138 BDB::AIO::max_poll_time 0.1; 444 BDB::max_poll_time 0.1;
139 445
140 # use a low priority so other tasks have priority 446 # use a low priority so other tasks have priority
141 Event->io (fd => BDB::AIO::poll_fileno, 447 Event->io (fd => BDB::poll_fileno,
142 poll => 'r', nice => 1, 448 poll => 'r', nice => 1,
143 cb => &BDB::AIO::poll_cb); 449 cb => &BDB::poll_cb);
144 450
145=item BDB::AIO::poll_wait 451=item BDB::poll_wait
146 452
147If there are any outstanding requests and none of them in the result 453If there are any outstanding requests and none of them in the result
148phase, wait till the result filehandle becomes ready for reading (simply 454phase, wait till the result filehandle becomes ready for reading (simply
149does a C<select> on the filehandle. This is useful if you want to 455does a C<select> on the filehandle. This is useful if you want to
150synchronously wait for some requests to finish). 456synchronously wait for some requests to finish).
151 457
152See C<nreqs> for an example. 458See C<nreqs> for an example.
153 459
154=item BDB::AIO::poll 460=item BDB::poll
155 461
156Waits until some requests have been handled. 462Waits until some requests have been handled.
157 463
158Returns the number of requests processed, but is otherwise strictly 464Returns the number of requests processed, but is otherwise strictly
159equivalent to: 465equivalent to:
160 466
161 BDB::AIO::poll_wait, BDB::AIO::poll_cb 467 BDB::poll_wait, BDB::poll_cb
162 468
163=item BDB::AIO::flush 469=item BDB::flush
164 470
165Wait till all outstanding AIO requests have been handled. 471Wait till all outstanding AIO requests have been handled.
166 472
167Strictly equivalent to: 473Strictly equivalent to:
168 474
169 BDB::AIO::poll_wait, BDB::AIO::poll_cb 475 BDB::poll_wait, BDB::poll_cb
170 while BDB::AIO::nreqs; 476 while BDB::nreqs;
477
478=back
171 479
172=head3 CONTROLLING THE NUMBER OF THREADS 480=head3 CONTROLLING THE NUMBER OF THREADS
173 481
482=over 4
483
174=item BDB::AIO::min_parallel $nthreads 484=item BDB::min_parallel $nthreads
175 485
176Set the minimum number of AIO threads to C<$nthreads>. The current 486Set the minimum number of AIO threads to C<$nthreads>. The current
177default is C<8>, which means eight asynchronous operations can execute 487default is C<8>, which means eight asynchronous operations can execute
178concurrently at any one time (the number of outstanding requests, 488concurrently at any one time (the number of outstanding requests,
179however, is unlimited). 489however, is unlimited).
180 490
181BDB::AIO starts threads only on demand, when an AIO request is queued and 491BDB starts threads only on demand, when an AIO request is queued and
182no free thread exists. Please note that queueing up a hundred requests can 492no 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 493create 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. 494is in the cache and could have been processed faster by a single thread.
185 495
186It is recommended to keep the number of threads relatively low, as some 496It is recommended to keep the number of threads relatively low, as some
189versions, 4-32 threads should be fine. 499versions, 4-32 threads should be fine.
190 500
191Under most circumstances you don't need to call this function, as the 501Under most circumstances you don't need to call this function, as the
192module selects a default that is suitable for low to moderate load. 502module selects a default that is suitable for low to moderate load.
193 503
194=item BDB::AIO::max_parallel $nthreads 504=item BDB::max_parallel $nthreads
195 505
196Sets the maximum number of AIO threads to C<$nthreads>. If more than the 506Sets the maximum number of AIO threads to C<$nthreads>. If more than the
197specified number of threads are currently running, this function kills 507specified number of threads are currently running, this function kills
198them. This function blocks until the limit is reached. 508them. This function blocks until the limit is reached.
199 509
203This module automatically runs C<max_parallel 0> at program end, to ensure 513This module automatically runs C<max_parallel 0> at program end, to ensure
204that all threads are killed and that there are no outstanding requests. 514that all threads are killed and that there are no outstanding requests.
205 515
206Under normal circumstances you don't need to call this function. 516Under normal circumstances you don't need to call this function.
207 517
208=item BDB::AIO::max_idle $nthreads 518=item BDB::max_idle $nthreads
209 519
210Limit the number of threads (default: 4) that are allowed to idle (i.e., 520Limit the number of threads (default: 4) that are allowed to idle (i.e.,
211threads that did not get a request to process within 10 seconds). That 521threads that did not get a request to process within 10 seconds). That
212means if a thread becomes idle while C<$nthreads> other threads are also 522means if a thread becomes idle while C<$nthreads> other threads are also
213idle, it will free its resources and exit. 523idle, it will free its resources and exit.
218 528
219The default is probably ok in most situations, especially if thread 529The default is probably ok in most situations, especially if thread
220creation is fast. If thread creation is very slow on your system you might 530creation is fast. If thread creation is very slow on your system you might
221want to use larger values. 531want to use larger values.
222 532
223=item $oldmaxreqs = BDB::AIO::max_outstanding $maxreqs 533=item $oldmaxreqs = BDB::max_outstanding $maxreqs
224 534
225This is a very bad function to use in interactive programs because it 535This 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 536blocks, and a bad way to reduce concurrency because it is inexact: Better
227use an C<aio_group> together with a feed callback. 537use an C<aio_group> together with a feed callback.
228 538
236 546
237You can still queue as many requests as you want. Therefore, 547You can still queue as many requests as you want. Therefore,
238C<max_oustsanding> is mainly useful in simple scripts (with low values) or 548C<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). 549as a stop gap to shield against fatal memory overflow (with large values).
240 550
551=item BDB::set_sync_prepare $cb
552
553Sets a callback that is called whenever a request is created without an
554explicit callback. It has to return two code references. The first is used
555as the request callback, and the second is called to wait until the first
556callback has been called. The default implementation works like this:
557
558 sub {
559 my $status;
560 (
561 sub { $status = $! },
562 sub { BDB::poll while !defined $status; $! = $status },
563 )
564 }
565
566=back
567
241=head3 STATISTICAL INFORMATION 568=head3 STATISTICAL INFORMATION
242 569
570=over 4
571
243=item BDB::AIO::nreqs 572=item BDB::nreqs
244 573
245Returns the number of requests currently in the ready, execute or pending 574Returns the number of requests currently in the ready, execute or pending
246states (i.e. for which their callback has not been invoked yet). 575states (i.e. for which their callback has not been invoked yet).
247 576
248Example: wait till there are no outstanding requests anymore: 577Example: wait till there are no outstanding requests anymore:
249 578
250 BDB::AIO::poll_wait, BDB::AIO::poll_cb 579 BDB::poll_wait, BDB::poll_cb
251 while BDB::AIO::nreqs; 580 while BDB::nreqs;
252 581
253=item BDB::AIO::nready 582=item BDB::nready
254 583
255Returns the number of requests currently in the ready state (not yet 584Returns the number of requests currently in the ready state (not yet
256executed). 585executed).
257 586
258=item BDB::AIO::npending 587=item BDB::npending
259 588
260Returns the number of requests currently in the pending state (executed, 589Returns the number of requests currently in the pending state (executed,
261but not yet processed by poll_cb). 590but not yet processed by poll_cb).
262 591
263=back 592=back
264 593
265=cut 594=cut
266 595
267# support function to convert a fd into a perl filehandle 596set_sync_prepare {
268sub _fd2fh { 597 my $status;
269 return undef if $_[0] < 0; 598 (
270 599 sub {
271 # try to generate nice filehandles 600 $status = $!;
272 my $sym = "BDB::AIO::fd#$_[0]"; 601 },
273 local *$sym; 602 sub {
274 603 BDB::poll while !defined $status;
275 open *$sym, "+<&=$_[0]" # usually works under any unix 604 $! = $status;
276 or open *$sym, "<&=$_[0]" # cygwin needs this 605 },
277 or open *$sym, ">&=$_[0]" # or this 606 )
278 or return undef; 607};
279
280 *$sym
281}
282 608
283min_parallel 8; 609min_parallel 8;
284 610
285END { flush } 611END { flush }
286 612
310bytes of memory. In addition, stat requests need a stat buffer (possibly 636bytes of memory. In addition, stat requests need a stat buffer (possibly
311a few hundred bytes), readdir requires a result buffer and so on. Perl 637a 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 638scalars and other data passed into aio requests will also be locked and
313will consume memory till the request has entered the done state. 639will consume memory till the request has entered the done state.
314 640
315This is now awfully much, so queuing lots of requests is not usually a 641This is not awfully much, so queuing lots of requests is not usually a
316problem. 642problem.
317 643
318Per-thread usage: 644Per-thread usage:
319 645
320In the execution phase, some aio requests require more memory for 646In the execution phase, some aio requests require more memory for
321temporary buffers, and each thread requires a stack and other data 647temporary buffers, and each thread requires a stack and other data
322structures (usually around 16k-128k, depending on the OS). 648structures (usually around 16k-128k, depending on the OS).
323 649
324=head1 KNOWN BUGS 650=head1 KNOWN BUGS
325 651
326Known bugs will be fixed in the next release. 652Known bugs will be fixed in the next release, except:
653
654 If you use a transaction in any request, and the request returns
655 with an operating system error or DB_LOCK_NOTGRANTED, the internal
656 TXN_DEADLOCK flag will be set on the transaction. See C<db_txn_finish>,
657 above.
327 658
328=head1 SEE ALSO 659=head1 SEE ALSO
329 660
330L<Coro::AIO>. 661L<Coro::AIO>.
331 662

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines