1 /*-
2 * See the file LICENSE for redistribution information.
3 *
4 * Copyright (c) 1996, 1997, 1998
5 * Sleepycat Software. All rights reserved.
6 */
7 /*
8 * Copyright (c) 1995, 1996
9 * The President and Fellows of Harvard University. All rights reserved.
10 *
11 * This code is derived from software contributed to Berkeley by
12 * Margo Seltzer.
13 *
14 * Redistribution and use in source and binary forms, with or without
15 * modification, are permitted provided that the following conditions
16 * are met:
17 * 1. Redistributions of source code must retain the above copyright
18 * notice, this list of conditions and the following disclaimer.
19 * 2. Redistributions in binary form must reproduce the above copyright
20 * notice, this list of conditions and the following disclaimer in the
21 * documentation and/or other materials provided with the distribution.
22 * 3. All advertising materials mentioning features or use of this software
23 * must display the following acknowledgement:
24 * This product includes software developed by the University of
25 * California, Berkeley and its contributors.
26 * 4. Neither the name of the University nor the names of its contributors
27 * may be used to endorse or promote products derived from this software
28 * without specific prior written permission.
29 *
30 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
31 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
32 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
33 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
34 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
35 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
36 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
37 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
38 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
39 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
40 * SUCH DAMAGE.
41 */
42
43 #include "config.h"
44
45 #ifndef lint
46 static const char sccsid[] = "@(#)db_dispatch.c 10.20 (Sleepycat) 10/10/98";
47 #endif /* not lint */
48
49 #ifndef NO_SYSTEM_INCLUDES
50 #include <sys/types.h>
51
52 #include <errno.h>
53 #include <shqueue.h>
54 #include <stddef.h>
55 #include <stdlib.h>
56 #include <string.h>
57 #endif
58
59 #include "db_int.h"
60 #include "db_page.h"
61 #include "db_dispatch.h"
62 #include "db_am.h"
63 #include "common_ext.h"
64 #include "log_auto.h"
65 #include "txn.h"
66 #include "txn_auto.h"
67
68 /*
69 * Data structures to manage the DB dispatch table. The dispatch table
70 * is a dynamically allocated array of pointers to dispatch functions.
71 * The dispatch_size is the number of entries possible in the current
72 * dispatch table and the dispatch_valid is the number of valid entries
73 * in the dispatch table.
74 */
75 static int (**dispatch_table) __P((DB_LOG *, DBT *, DB_LSN *, int, void *));
76 static u_int32_t dispatch_size = 0;
77
78 /*
79 * __db_dispatch --
80 *
81 * This is the transaction dispatch function used by the db access methods.
82 * It is designed to handle the record format used by all the access
83 * methods (the one automatically generated by the db_{h,log,read}.sh
84 * scripts in the tools directory). An application using a different
85 * recovery paradigm will supply a different dispatch function to txn_open.
86 *
87 * PUBLIC: int __db_dispatch __P((DB_LOG *, DBT *, DB_LSN *, int, void *));
88 */
89 int
__db_dispatch(logp,db,lsnp,redo,info)90 __db_dispatch(logp, db, lsnp, redo, info)
91 DB_LOG *logp; /* The log file. */
92 DBT *db; /* The log record upon which to dispatch. */
93 DB_LSN *lsnp; /* The lsn of the record being dispatched. */
94 int redo; /* Redo this op (or undo it). */
95 void *info;
96 {
97 u_int32_t rectype, txnid;
98
99 memcpy(&rectype, db->data, sizeof(rectype));
100 memcpy(&txnid, (u_int8_t *)db->data + sizeof(rectype), sizeof(txnid));
101
102 switch (redo) {
103 case TXN_REDO:
104 case TXN_UNDO:
105 return ((dispatch_table[rectype])(logp, db, lsnp, redo, info));
106 case TXN_OPENFILES:
107 if (rectype < DB_txn_BEGIN )
108 return ((dispatch_table[rectype])(logp,
109 db, lsnp, redo, info));
110 break;
111 case TXN_BACKWARD_ROLL:
112 /*
113 * Running full recovery in the backward pass. If we've
114 * seen this txnid before and added to it our commit list,
115 * then we do nothing during this pass. If we've never
116 * seen it, then we call the appropriate recovery routine
117 * in "abort mode".
118 */
119 if (rectype == DB_log_register || rectype == DB_txn_ckp ||
120 (__db_txnlist_find(info, txnid) == DB_NOTFOUND &&
121 txnid != 0))
122 return ((dispatch_table[rectype])(logp,
123 db, lsnp, TXN_UNDO, info));
124 break;
125 case TXN_FORWARD_ROLL:
126 /*
127 * In the forward pass, if we haven't seen the transaction,
128 * do nothing, else recovery it.
129 */
130 if (rectype == DB_log_register || rectype == DB_txn_ckp ||
131 __db_txnlist_find(info, txnid) != DB_NOTFOUND)
132 return ((dispatch_table[rectype])(logp,
133 db, lsnp, TXN_REDO, info));
134 break;
135 default:
136 abort();
137 }
138 return (0);
139 }
140
141 /*
142 * __db_add_recovery --
143 *
144 * PUBLIC: int __db_add_recovery __P((DB_ENV *,
145 * PUBLIC: int (*)(DB_LOG *, DBT *, DB_LSN *, int, void *), u_int32_t));
146 */
147 int
__db_add_recovery(dbenv,func,ndx)148 __db_add_recovery(dbenv, func, ndx)
149 DB_ENV *dbenv;
150 int (*func) __P((DB_LOG *, DBT *, DB_LSN *, int, void *));
151 u_int32_t ndx;
152 {
153 u_int32_t i;
154 int ret;
155
156 COMPQUIET(dbenv, NULL); /* !!!: not currently used. */
157
158 /* Check if we have to grow the table. */
159 if (ndx >= dispatch_size) {
160 if ((ret = __os_realloc(&dispatch_table,
161 (DB_user_BEGIN + dispatch_size) *
162 sizeof(dispatch_table[0]))) != 0)
163 return (ret);
164 for (i = dispatch_size,
165 dispatch_size += DB_user_BEGIN; i < dispatch_size; ++i)
166 dispatch_table[i] = NULL;
167 }
168
169 dispatch_table[ndx] = func;
170 return (0);
171 }
172
173 /*
174 * __db_txnlist_init --
175 * Initialize transaction linked list.
176 *
177 * PUBLIC: int __db_txnlist_init __P((void *));
178 */
179 int
__db_txnlist_init(retp)180 __db_txnlist_init(retp)
181 void *retp;
182 {
183 DB_TXNHEAD *headp;
184 int ret;
185
186 if ((ret = __os_malloc(sizeof(DB_TXNHEAD), NULL, &headp)) != 0)
187 return (ret);
188
189 LIST_INIT(&headp->head);
190 headp->maxid = 0;
191 headp->generation = 1;
192
193 *(void **)retp = headp;
194 return (0);
195 }
196
197 /*
198 * __db_txnlist_add --
199 * Add an element to our transaction linked list.
200 *
201 * PUBLIC: int __db_txnlist_add __P((void *, u_int32_t));
202 */
203 int
__db_txnlist_add(listp,txnid)204 __db_txnlist_add(listp, txnid)
205 void *listp;
206 u_int32_t txnid;
207 {
208 DB_TXNHEAD *hp;
209 DB_TXNLIST *elp;
210 int ret;
211
212 if ((ret = __os_malloc(sizeof(DB_TXNLIST), NULL, &elp)) != 0)
213 return (ret);
214
215 elp->txnid = txnid;
216 hp = (DB_TXNHEAD *)listp;
217 LIST_INSERT_HEAD(&hp->head, elp, links);
218 if (txnid > hp->maxid)
219 hp->maxid = txnid;
220 elp->generation = hp->generation;
221
222 return (0);
223 }
224
225 /*
226 * __db_txnlist_find --
227 * Checks to see if a txnid with the current generation is in the
228 * txnid list.
229 *
230 * PUBLIC: int __db_txnlist_find __P((void *, u_int32_t));
231 */
232 int
__db_txnlist_find(listp,txnid)233 __db_txnlist_find(listp, txnid)
234 void *listp;
235 u_int32_t txnid;
236 {
237 DB_TXNHEAD *hp;
238 DB_TXNLIST *p;
239
240 if ((hp = (DB_TXNHEAD *)listp) == NULL)
241 return (DB_NOTFOUND);
242
243 for (p = hp->head.lh_first; p != NULL; p = p->links.le_next)
244 if (p->txnid == txnid && hp->generation == p->generation)
245 return (0);
246
247 return (DB_NOTFOUND);
248 }
249
250 /*
251 * __db_txnlist_end --
252 * Discard transaction linked list.
253 *
254 * PUBLIC: void __db_txnlist_end __P((void *));
255 */
256 void
__db_txnlist_end(listp)257 __db_txnlist_end(listp)
258 void *listp;
259 {
260 DB_TXNHEAD *hp;
261 DB_TXNLIST *p;
262
263 hp = (DB_TXNHEAD *)listp;
264 while ((p = LIST_FIRST(&hp->head)) != LIST_END(&hp->head)) {
265 LIST_REMOVE(p, links);
266 __os_free(p, 0);
267 }
268 __os_free(listp, sizeof(DB_TXNHEAD));
269 }
270
271 /*
272 * __db_txnlist_gen --
273 * Change the current generation number.
274 *
275 * PUBLIC: void __db_txnlist_gen __P((void *, int));
276 */
277 void
__db_txnlist_gen(listp,incr)278 __db_txnlist_gen(listp, incr)
279 void *listp;
280 int incr;
281 {
282 DB_TXNHEAD *hp;
283
284 /*
285 * During recovery generation numbers keep track of how many "restart"
286 * checkpoints we've seen. Restart checkpoints occur whenever we take
287 * a checkpoint and there are no outstanding transactions. When that
288 * happens, we can reset transaction IDs back to 1. It always happens
289 * at recovery and it prevents us from exhausting the transaction IDs
290 * name space.
291 */
292 hp = (DB_TXNHEAD *)listp;
293 hp->generation += incr;
294 }
295
296 #ifdef DEBUG
297 /*
298 * __db_txnlist_print --
299 * Print out the transaction list.
300 *
301 * PUBLIC: void __db_txnlist_print __P((void *));
302 */
303 void
__db_txnlist_print(listp)304 __db_txnlist_print(listp)
305 void *listp;
306 {
307 DB_TXNHEAD *hp;
308 DB_TXNLIST *p;
309
310 hp = (DB_TXNHEAD *)listp;
311 printf("Maxid: %lu Generation: %lu\n", (u_long)hp->maxid,
312 (u_long)hp->generation);
313 for (p = hp->head.lh_first; p != NULL; p = p->links.le_next)
314 printf("TXNID: %lu(%lu)\n", (u_long)p->txnid,
315 (u_long)p->generation);
316 }
317 #endif
318