libdb-sys 0.1.1

statically linked ffi bindings for Berkeley DB
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
/*-
 * See the file LICENSE for redistribution information.
 *
 * Copyright (c) 2005, 2012 Oracle and/or its affiliates.  All rights reserved.
 *
 * $Id$
 */

#include "db_config.h"

#include "db_int.h"
#ifndef HAVE_SIMPLE_THREAD_TYPE
#include "dbinc/db_page.h"
#include "dbinc/hash.h"			/* Needed for call to __ham_func5. */
#endif
#include "dbinc/lock.h"
#include "dbinc/log.h"
#include "dbinc/mp.h"
#include "dbinc/txn.h"

static int __env_in_api __P((ENV *));
static void __env_clear_state __P((ENV *));

/*
 * __env_failchk_pp --
 *	ENV->failchk pre/post processing.
 *
 * PUBLIC: int __env_failchk_pp __P((DB_ENV *, u_int32_t));
 */
int
__env_failchk_pp(dbenv, flags)
	DB_ENV *dbenv;
	u_int32_t flags;
{
	DB_THREAD_INFO *ip;
	ENV *env;
	int ret;

	env = dbenv->env;

	ENV_ILLEGAL_BEFORE_OPEN(env, "DB_ENV->failchk");

	/*
	 * ENV->failchk requires self and is-alive functions.  We
	 * have a default self function, but no is-alive function.
	 */
	if (!ALIVE_ON(env)) {
		__db_errx(env, DB_STR("1503",
	"DB_ENV->failchk requires DB_ENV->is_alive be configured"));
		return (EINVAL);
	}

	if (flags != 0)
		return (__db_ferr(env, "DB_ENV->failchk", 0));

	ENV_ENTER(env, ip);
	FAILCHK_THREAD(env, ip);	/* mark as failchk thread */
	ret = __env_failchk_int(dbenv);
	ENV_LEAVE(env, ip);
	return (ret);
}
/*
 * __env_failchk_int --
 *	Process the subsystem failchk routines
 *
 * PUBLIC: int __env_failchk_int __P((DB_ENV *));
 */
int
__env_failchk_int(dbenv)
	DB_ENV *dbenv;
{
	ENV *env;
	int ret;

	env = dbenv->env;
	F_SET(dbenv, DB_ENV_FAILCHK);

	/*
	 * We check for dead threads in the API first as this would be likely
	 * to hang other things we try later, like locks and transactions.
	 */
	if ((ret = __env_in_api(env)) != 0)
		goto err;

	if (LOCKING_ON(env) && (ret = __lock_failchk(env)) != 0)
		goto err;

	if (TXN_ON(env) &&
	    ((ret = __txn_failchk(env)) != 0 ||
	    (ret = __dbreg_failchk(env)) != 0))
		goto err;

	if ((ret = __memp_failchk(env)) != 0)
		goto err;

#ifdef HAVE_REPLICATION_THREADS
	if (REP_ON(env) && (ret = __repmgr_failchk(env)) != 0)
		goto err;
#endif

	/* Mark any dead blocked threads as dead. */
	__env_clear_state(env);

#ifdef HAVE_MUTEX_SUPPORT
	ret = __mut_failchk(env);
#endif

err:	F_CLR(dbenv, DB_ENV_FAILCHK);
	return (ret);
}

/*
 * __env_thread_size --
 *	Initial amount of memory for thread info blocks.
 * PUBLIC: size_t __env_thread_size __P((ENV *, size_t));
 */
size_t
__env_thread_size(env, other_alloc)
	ENV *env;
	size_t other_alloc;
{
	DB_ENV *dbenv;
	size_t size;
	u_int32_t max;

	dbenv = env->dbenv;
	size = 0;

	max = dbenv->thr_max;
	if (dbenv->thr_init != 0) {
		size =
		    dbenv->thr_init * __env_alloc_size(sizeof(DB_THREAD_INFO));
		if (max < dbenv->thr_init)
			max = dbenv->thr_init;
	} else if (max == 0 && ALIVE_ON(env)) {
		if ((max = dbenv->tx_init) == 0) {
			/*
			 * They want thread tracking, but don't say how much.
			 * Arbitrarily assume 1/10 of the remaining memory
			 * or at least 100.  We just use this to size
			 * the hash table.
			 */
			if (dbenv->memory_max != 0)
				max = (u_int32_t)
				    (((dbenv->memory_max - other_alloc) / 10) /
					sizeof(DB_THREAD_INFO));
			if (max < 100)
				max = 100;
		}
	}
	/*
	 * Set the number of buckets to be 1/8th the number of
	 * thread control blocks.  This is rather arbitrary.
	 */
	dbenv->thr_max = max;
	if (max != 0)
		size += __env_alloc_size(sizeof(DB_HASHTAB) *
		    __db_tablesize(max / 8));
	return (size);
}

/*
 * __env_thread_max --
 *	Return the amount of extra memory to hold thread information.
 * PUBLIC: size_t __env_thread_max __P((ENV *));
 */
size_t
__env_thread_max(env)
	ENV *env;
{
	DB_ENV *dbenv;
	size_t size;

	dbenv = env->dbenv;

	/*
	 * Allocate space for thread info blocks.  Max is only advisory,
	 * so we allocate 25% more.
	 */
	if (dbenv->thr_max > dbenv->thr_init) {
		size = dbenv->thr_max - dbenv->thr_init;
		size += size / 4;
	} else {
		dbenv->thr_max = dbenv->thr_init;
		size = dbenv->thr_init / 4;
	}

	size = size * __env_alloc_size(sizeof(DB_THREAD_INFO));
	return (size);
}

/*
 * __env_thread_init --
 *	Initialize the thread control block table.
 *
 * PUBLIC: int __env_thread_init __P((ENV *, int));
 */
int
__env_thread_init(env, during_creation)
	ENV *env;
	int during_creation;
{
	DB_ENV *dbenv;
	DB_HASHTAB *htab;
	REGENV *renv;
	REGINFO *infop;
	THREAD_INFO *thread;
	int ret;

	dbenv = env->dbenv;
	infop = env->reginfo;
	renv = infop->primary;

	if (renv->thread_off == INVALID_ROFF) {
		if (dbenv->thr_max == 0) {
			env->thr_hashtab = NULL;
			if (ALIVE_ON(env)) {
				__db_errx(env, DB_STR("1504",
		"is_alive method specified but no thread region allocated"));
				return (EINVAL);
			}
			return (0);
		}

		if (!during_creation) {
			__db_errx(env, DB_STR("1505",
"thread table must be allocated when the database environment is created"));
			return (EINVAL);
		}

		if ((ret =
		    __env_alloc(infop, sizeof(THREAD_INFO), &thread)) != 0) {
			__db_err(env, ret, DB_STR("1506",
			    "unable to allocate a thread status block"));
			return (ret);
		}
		memset(thread, 0, sizeof(*thread));
		renv->thread_off = R_OFFSET(infop, thread);
		thread->thr_nbucket = __db_tablesize(dbenv->thr_max / 8);
		if ((ret = __env_alloc(infop,
		     thread->thr_nbucket * sizeof(DB_HASHTAB), &htab)) != 0)
			return (ret);
		thread->thr_hashoff = R_OFFSET(infop, htab);
		__db_hashinit(htab, thread->thr_nbucket);
		thread->thr_max = dbenv->thr_max;
		thread->thr_init = dbenv->thr_init;
	} else {
		thread = R_ADDR(infop, renv->thread_off);
		htab = R_ADDR(infop, thread->thr_hashoff);
	}

	env->thr_hashtab = htab;
	env->thr_nbucket = thread->thr_nbucket;
	dbenv->thr_max = thread->thr_max;
	dbenv->thr_init = thread->thr_init;
	return (0);
}

/*
 * __env_thread_destroy --
 *	Destroy the thread control block table.
 *
 * PUBLIC: void __env_thread_destroy __P((ENV *));
 */
void
__env_thread_destroy(env)
	ENV *env;
{
	DB_HASHTAB *htab;
	DB_THREAD_INFO *ip, *np;
	REGENV *renv;
	REGINFO *infop;
	THREAD_INFO *thread;
	u_int32_t i;

	infop = env->reginfo;
	renv = infop->primary;
	if (renv->thread_off == INVALID_ROFF)
		return;

	thread = R_ADDR(infop, renv->thread_off);
	if ((htab = env->thr_hashtab) != NULL) {
		for (i = 0; i < env->thr_nbucket; i++) {
			ip = SH_TAILQ_FIRST(&htab[i], __db_thread_info);
			for (; ip != NULL; ip = np) {
				np = SH_TAILQ_NEXT(ip,
				    dbth_links, __db_thread_info);
				__env_alloc_free(infop, ip);
			}
		}
		__env_alloc_free(infop, htab);
	}

	__env_alloc_free(infop, thread);
	return;
}

/*
 * __env_in_api --
 *	Look for threads which died in the api and complain.
 *	If no threads died but there are blocked threads unpin
 *	any buffers they may have locked.
 */
static int
__env_in_api(env)
	ENV *env;
{
	DB_ENV *dbenv;
	DB_HASHTAB *htab;
	DB_THREAD_INFO *ip;
	REGENV *renv;
	REGINFO *infop;
	THREAD_INFO *thread;
	u_int32_t i;
	int unpin, ret;

	if ((htab = env->thr_hashtab) == NULL)
		return (EINVAL);

	dbenv = env->dbenv;
	infop = env->reginfo;
	renv = infop->primary;
	thread = R_ADDR(infop, renv->thread_off);
	unpin = 0;

	for (i = 0; i < env->thr_nbucket; i++)
		SH_TAILQ_FOREACH(ip, &htab[i], dbth_links, __db_thread_info) {
			if (ip->dbth_state == THREAD_SLOT_NOT_IN_USE ||
			    (ip->dbth_state == THREAD_OUT &&
			    thread->thr_count <  thread->thr_max))
				continue;
			if (dbenv->is_alive(
			    dbenv, ip->dbth_pid, ip->dbth_tid, 0))
				continue;
			if (ip->dbth_state == THREAD_BLOCKED) {
				ip->dbth_state = THREAD_BLOCKED_DEAD;
				unpin = 1;
				continue;
			}
			if (ip->dbth_state == THREAD_OUT) {
				ip->dbth_state = THREAD_SLOT_NOT_IN_USE;
				continue;
			}
			return (__db_failed(env, DB_STR("1507",
			    "Thread died in Berkeley DB library"),
			    ip->dbth_pid, ip->dbth_tid));
		}

	if (unpin == 0)
		return (0);

	for (i = 0; i < env->thr_nbucket; i++)
		SH_TAILQ_FOREACH(ip, &htab[i], dbth_links, __db_thread_info)
			if (ip->dbth_state == THREAD_BLOCKED_DEAD &&
			    (ret = __memp_unpin_buffers(env, ip)) != 0)
				return (ret);

	return (0);
}

/*
 * __env_clear_state --
 *	Look for threads which died while blockedi and clear them..
 */
static void
__env_clear_state(env)
	ENV *env;
{
	DB_HASHTAB *htab;
	DB_THREAD_INFO *ip;
	u_int32_t i;

	htab = env->thr_hashtab;
	for (i = 0; i < env->thr_nbucket; i++)
		SH_TAILQ_FOREACH(ip, &htab[i], dbth_links, __db_thread_info)
			if (ip->dbth_state == THREAD_BLOCKED_DEAD)
				ip->dbth_state = THREAD_SLOT_NOT_IN_USE;
}

struct __db_threadid {
	pid_t pid;
	db_threadid_t tid;
};

/*
 * PUBLIC: int __env_set_state __P((ENV *, DB_THREAD_INFO **, DB_THREAD_STATE));
 */
int
__env_set_state(env, ipp, state)
	ENV *env;
	DB_THREAD_INFO **ipp;
	DB_THREAD_STATE state;
{
	struct __db_threadid id;
	DB_ENV *dbenv;
	DB_HASHTAB *htab;
	DB_THREAD_INFO *ip;
	REGENV *renv;
	REGINFO *infop;
	THREAD_INFO *thread;
	u_int32_t indx;
	int ret;

	dbenv = env->dbenv;
	htab = env->thr_hashtab;

	if (F_ISSET(dbenv, DB_ENV_NOLOCKING)) {
		*ipp = NULL;
		return (0);
	}
	dbenv->thread_id(dbenv, &id.pid, &id.tid);

	/*
	 * Hashing of thread ids.  This is simple but could be replaced with
	 * something more expensive if needed.
	 */
#ifdef HAVE_SIMPLE_THREAD_TYPE
	/*
	 * A thread ID may be a pointer, so explicitly cast to a pointer of
	 * the appropriate size before doing the bitwise XOR.
	 */
	indx = (u_int32_t)((uintptr_t)id.pid ^ (uintptr_t)id.tid);
#else
	indx = __ham_func5(NULL, &id.tid, sizeof(id.tid));
#endif
	indx %= env->thr_nbucket;
	SH_TAILQ_FOREACH(ip, &htab[indx], dbth_links, __db_thread_info) {
#ifdef HAVE_SIMPLE_THREAD_TYPE
		if (id.pid == ip->dbth_pid && id.tid == ip->dbth_tid)
			break;
#else
		if (memcmp(&id.pid, &ip->dbth_pid, sizeof(id.pid)) != 0)
			continue;
#ifdef HAVE_MUTEX_PTHREADS
		if (pthread_equal(id.tid, ip->dbth_tid) == 0)
#else
		if (memcmp(&id.tid, &ip->dbth_tid, sizeof(id.tid)) != 0)
#endif
			continue;
		break;
#endif
	}

	/*
	 * If ipp is not null,  return the thread control block if found.
	 * Check to ensure the thread of control has been registered.
	 */
	if (state == THREAD_VERIFY) {
		DB_ASSERT(env, ip != NULL && ip->dbth_state != THREAD_OUT);
		if (ipp != NULL) {
			if (ip == NULL) /* The control block wasn't found */
				return (EINVAL);
			*ipp = ip;
		}
		return (0);
	}

	*ipp = NULL;
	ret = 0;
	if (ip == NULL) {
		infop = env->reginfo;
		renv = infop->primary;
		thread = R_ADDR(infop, renv->thread_off);
		MUTEX_LOCK(env, renv->mtx_regenv);

		/*
		 * If we are passed the specified max, try to reclaim one from
		 * our queue.  If failcheck has marked the slot not in use, we
		 * can take it, otherwise we must call is_alive before freeing
		 * it.
		 */
		if (thread->thr_count >= thread->thr_max) {
			SH_TAILQ_FOREACH(
			    ip, &htab[indx], dbth_links, __db_thread_info)
				if (ip->dbth_state == THREAD_SLOT_NOT_IN_USE ||
				    (ip->dbth_state == THREAD_OUT &&
				    ALIVE_ON(env) && !dbenv->is_alive(
				    dbenv, ip->dbth_pid, ip->dbth_tid, 0)))
					break;

			if (ip != NULL) {
				DB_ASSERT(env, ip->dbth_pincount == 0);
				goto init;
			}
		}

		thread->thr_count++;
		if ((ret = __env_alloc(infop,
		     sizeof(DB_THREAD_INFO), &ip)) == 0) {
			memset(ip, 0, sizeof(*ip));
			/*
			 * This assumes we can link atomically since we do
			 * no locking here.  We never use the backpointer
			 * so we only need to be able to write an offset
			 * atomically.
			 */
			SH_TAILQ_INSERT_HEAD(
			    &htab[indx], ip, dbth_links, __db_thread_info);
			ip->dbth_pincount = 0;
			ip->dbth_pinmax = PINMAX;
			ip->dbth_pinlist = R_OFFSET(infop, ip->dbth_pinarray);

init:			ip->dbth_pid = id.pid;
			ip->dbth_tid = id.tid;
			ip->dbth_state = state;
			SH_TAILQ_INIT(&ip->dbth_xatxn);
		}
		MUTEX_UNLOCK(env, renv->mtx_regenv);
	} else
		ip->dbth_state = state;
	*ipp = ip;

	DB_ASSERT(env, ret == 0);
	if (ret != 0)
		__db_errx(env, DB_STR("1508",
		    "Unable to allocate thread control block"));
	return (ret);
}

/*
 * __env_thread_id_string --
 *	Convert a thread id to a string.
 *
 * PUBLIC: char *__env_thread_id_string
 * PUBLIC:     __P((DB_ENV *, pid_t, db_threadid_t, char *));
 */
char *
__env_thread_id_string(dbenv, pid, tid, buf)
	DB_ENV *dbenv;
	pid_t pid;
	db_threadid_t tid;
	char *buf;
{
#ifdef HAVE_SIMPLE_THREAD_TYPE
#ifdef UINT64_FMT
	char fmt[20];

	snprintf(fmt, sizeof(fmt), "%s/%s", UINT64_FMT, UINT64_FMT);
	snprintf(buf,
	    DB_THREADID_STRLEN, fmt, (u_int64_t)pid, (u_int64_t)(uintptr_t)tid);
#else
	snprintf(buf, DB_THREADID_STRLEN, "%lu/%lu", (u_long)pid, (u_long)tid);
#endif
#else
#ifdef UINT64_FMT
	char fmt[20];

	snprintf(fmt, sizeof(fmt), "%s/TID", UINT64_FMT);
	snprintf(buf, DB_THREADID_STRLEN, fmt, (u_int64_t)pid);
#else
	snprintf(buf, DB_THREADID_STRLEN, "%lu/TID", (u_long)pid);
#endif
#endif
	COMPQUIET(dbenv, NULL);
	COMPQUIET(*(u_int8_t *)&tid, 0);

	return (buf);
}