xref: /freebsd/lib/libc/db/btree/bt_open.c (revision 4cf49a43559ed9fdad601bdcccd2c55963008675)
1 /*-
2  * Copyright (c) 1990, 1993, 1994
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * This code is derived from software contributed to Berkeley by
6  * Mike Olson.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. All advertising materials mentioning features or use of this software
17  *    must display the following acknowledgement:
18  *	This product includes software developed by the University of
19  *	California, Berkeley and its contributors.
20  * 4. Neither the name of the University nor the names of its contributors
21  *    may be used to endorse or promote products derived from this software
22  *    without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34  * SUCH DAMAGE.
35  */
36 
37 #if defined(LIBC_SCCS) && !defined(lint)
38 static char sccsid[] = "@(#)bt_open.c	8.10 (Berkeley) 8/17/94";
39 #endif /* LIBC_SCCS and not lint */
40 
41 /*
42  * Implementation of btree access method for 4.4BSD.
43  *
44  * The design here was originally based on that of the btree access method
45  * used in the Postgres database system at UC Berkeley.  This implementation
46  * is wholly independent of the Postgres code.
47  */
48 
49 #include <sys/param.h>
50 #include <sys/stat.h>
51 
52 #include <errno.h>
53 #include <fcntl.h>
54 #include <limits.h>
55 #include <signal.h>
56 #include <stdio.h>
57 #include <stdlib.h>
58 #include <string.h>
59 #include <unistd.h>
60 
61 #include <db.h>
62 #include "btree.h"
63 
64 #ifdef DEBUG
65 #undef	MINPSIZE
66 #define	MINPSIZE	128
67 #endif
68 
69 static int byteorder __P((void));
70 static int nroot __P((BTREE *));
71 static int tmp __P((void));
72 
73 /*
74  * __BT_OPEN -- Open a btree.
75  *
76  * Creates and fills a DB struct, and calls the routine that actually
77  * opens the btree.
78  *
79  * Parameters:
80  *	fname:	filename (NULL for in-memory trees)
81  *	flags:	open flag bits
82  *	mode:	open permission bits
83  *	b:	BTREEINFO pointer
84  *
85  * Returns:
86  *	NULL on failure, pointer to DB on success.
87  *
88  */
89 DB *
90 __bt_open(fname, flags, mode, openinfo, dflags)
91 	const char *fname;
92 	int flags, mode, dflags;
93 	const BTREEINFO *openinfo;
94 {
95 	struct stat sb;
96 	BTMETA m;
97 	BTREE *t;
98 	BTREEINFO b;
99 	DB *dbp;
100 	pgno_t ncache;
101 	ssize_t nr;
102 	int machine_lorder;
103 
104 	t = NULL;
105 
106 	/*
107 	 * Intention is to make sure all of the user's selections are okay
108 	 * here and then use them without checking.  Can't be complete, since
109 	 * we don't know the right page size, lorder or flags until the backing
110 	 * file is opened.  Also, the file's page size can cause the cachesize
111 	 * to change.
112 	 */
113 	machine_lorder = byteorder();
114 	if (openinfo) {
115 		b = *openinfo;
116 
117 		/* Flags: R_DUP. */
118 		if (b.flags & ~(R_DUP))
119 			goto einval;
120 
121 		/*
122 		 * Page size must be indx_t aligned and >= MINPSIZE.  Default
123 		 * page size is set farther on, based on the underlying file
124 		 * transfer size.
125 		 */
126 		if (b.psize &&
127 		    (b.psize < MINPSIZE || b.psize > MAX_PAGE_OFFSET + 1 ||
128 		    b.psize & (sizeof(indx_t) - 1) ))
129 			goto einval;
130 
131 		/* Minimum number of keys per page; absolute minimum is 2. */
132 		if (b.minkeypage) {
133 			if (b.minkeypage < 2)
134 				goto einval;
135 		} else
136 			b.minkeypage = DEFMINKEYPAGE;
137 
138 		/* If no comparison, use default comparison and prefix. */
139 		if (b.compare == NULL) {
140 			b.compare = __bt_defcmp;
141 			if (b.prefix == NULL)
142 				b.prefix = __bt_defpfx;
143 		}
144 
145 		if (b.lorder == 0)
146 			b.lorder = machine_lorder;
147 	} else {
148 		b.compare = __bt_defcmp;
149 		b.cachesize = 0;
150 		b.flags = 0;
151 		b.lorder = machine_lorder;
152 		b.minkeypage = DEFMINKEYPAGE;
153 		b.prefix = __bt_defpfx;
154 		b.psize = 0;
155 	}
156 
157 	/* Check for the ubiquitous PDP-11. */
158 	if (b.lorder != BIG_ENDIAN && b.lorder != LITTLE_ENDIAN)
159 		goto einval;
160 
161 	/* Allocate and initialize DB and BTREE structures. */
162 	if ((t = (BTREE *)malloc(sizeof(BTREE))) == NULL)
163 		goto err;
164 	memset(t, 0, sizeof(BTREE));
165 	t->bt_fd = -1;			/* Don't close unopened fd on error. */
166 	t->bt_lorder = b.lorder;
167 	t->bt_order = NOT;
168 	t->bt_cmp = b.compare;
169 	t->bt_pfx = b.prefix;
170 	t->bt_rfd = -1;
171 
172 	if ((t->bt_dbp = dbp = (DB *)malloc(sizeof(DB))) == NULL)
173 		goto err;
174 	memset(t->bt_dbp, 0, sizeof(DB));
175 	if (t->bt_lorder != machine_lorder)
176 		F_SET(t, B_NEEDSWAP);
177 
178 	dbp->type = DB_BTREE;
179 	dbp->internal = t;
180 	dbp->close = __bt_close;
181 	dbp->del = __bt_delete;
182 	dbp->fd = __bt_fd;
183 	dbp->get = __bt_get;
184 	dbp->put = __bt_put;
185 	dbp->seq = __bt_seq;
186 	dbp->sync = __bt_sync;
187 
188 	/*
189 	 * If no file name was supplied, this is an in-memory btree and we
190 	 * open a backing temporary file.  Otherwise, it's a disk-based tree.
191 	 */
192 	if (fname) {
193 		switch (flags & O_ACCMODE) {
194 		case O_RDONLY:
195 			F_SET(t, B_RDONLY);
196 			break;
197 		case O_RDWR:
198 			break;
199 		case O_WRONLY:
200 		default:
201 			goto einval;
202 		}
203 
204 		if ((t->bt_fd = open(fname, flags, mode)) < 0)
205 			goto err;
206 
207 	} else {
208 		if ((flags & O_ACCMODE) != O_RDWR)
209 			goto einval;
210 		if ((t->bt_fd = tmp()) == -1)
211 			goto err;
212 		F_SET(t, B_INMEM);
213 	}
214 
215 	if (fcntl(t->bt_fd, F_SETFD, 1) == -1)
216 		goto err;
217 
218 	if (fstat(t->bt_fd, &sb))
219 		goto err;
220 	if (sb.st_size) {
221 		if ((nr = read(t->bt_fd, &m, sizeof(BTMETA))) < 0)
222 			goto err;
223 		if (nr != sizeof(BTMETA))
224 			goto eftype;
225 
226 		/*
227 		 * Read in the meta-data.  This can change the notion of what
228 		 * the lorder, page size and flags are, and, when the page size
229 		 * changes, the cachesize value can change too.  If the user
230 		 * specified the wrong byte order for an existing database, we
231 		 * don't bother to return an error, we just clear the NEEDSWAP
232 		 * bit.
233 		 */
234 		if (m.magic == BTREEMAGIC)
235 			F_CLR(t, B_NEEDSWAP);
236 		else {
237 			F_SET(t, B_NEEDSWAP);
238 			M_32_SWAP(m.magic);
239 			M_32_SWAP(m.version);
240 			M_32_SWAP(m.psize);
241 			M_32_SWAP(m.free);
242 			M_32_SWAP(m.nrecs);
243 			M_32_SWAP(m.flags);
244 		}
245 		if (m.magic != BTREEMAGIC || m.version != BTREEVERSION)
246 			goto eftype;
247 		if (m.psize < MINPSIZE || m.psize > MAX_PAGE_OFFSET + 1 ||
248 		    m.psize & (sizeof(indx_t) - 1) )
249 			goto eftype;
250 		if (m.flags & ~SAVEMETA)
251 			goto eftype;
252 		b.psize = m.psize;
253 		F_SET(t, m.flags);
254 		t->bt_free = m.free;
255 		t->bt_nrecs = m.nrecs;
256 	} else {
257 		/*
258 		 * Set the page size to the best value for I/O to this file.
259 		 * Don't overflow the page offset type.
260 		 */
261 		if (b.psize == 0) {
262 			b.psize = sb.st_blksize;
263 			if (b.psize < MINPSIZE)
264 				b.psize = MINPSIZE;
265 			if (b.psize > MAX_PAGE_OFFSET + 1)
266 				b.psize = MAX_PAGE_OFFSET + 1;
267 		}
268 
269 		/* Set flag if duplicates permitted. */
270 		if (!(b.flags & R_DUP))
271 			F_SET(t, B_NODUPS);
272 
273 		t->bt_free = P_INVALID;
274 		t->bt_nrecs = 0;
275 		F_SET(t, B_METADIRTY);
276 	}
277 
278 	t->bt_psize = b.psize;
279 
280 	/* Set the cache size; must be a multiple of the page size. */
281 	if (b.cachesize && b.cachesize & (b.psize - 1) )
282 		b.cachesize += (~b.cachesize & (b.psize - 1) ) + 1;
283 	if (b.cachesize < b.psize * MINCACHE)
284 		b.cachesize = b.psize * MINCACHE;
285 
286 	/* Calculate number of pages to cache. */
287 	ncache = (b.cachesize + t->bt_psize - 1) / t->bt_psize;
288 
289 	/*
290 	 * The btree data structure requires that at least two keys can fit on
291 	 * a page, but other than that there's no fixed requirement.  The user
292 	 * specified a minimum number per page, and we translated that into the
293 	 * number of bytes a key/data pair can use before being placed on an
294 	 * overflow page.  This calculation includes the page header, the size
295 	 * of the index referencing the leaf item and the size of the leaf item
296 	 * structure.  Also, don't let the user specify a minkeypage such that
297 	 * a key/data pair won't fit even if both key and data are on overflow
298 	 * pages.
299 	 */
300 	t->bt_ovflsize = (t->bt_psize - BTDATAOFF) / b.minkeypage -
301 	    (sizeof(indx_t) + NBLEAFDBT(0, 0));
302 	if (t->bt_ovflsize < NBLEAFDBT(NOVFLSIZE, NOVFLSIZE) + sizeof(indx_t))
303 		t->bt_ovflsize =
304 		    NBLEAFDBT(NOVFLSIZE, NOVFLSIZE) + sizeof(indx_t);
305 
306 	/* Initialize the buffer pool. */
307 	if ((t->bt_mp =
308 	    mpool_open(NULL, t->bt_fd, t->bt_psize, ncache)) == NULL)
309 		goto err;
310 	if (!F_ISSET(t, B_INMEM))
311 		mpool_filter(t->bt_mp, __bt_pgin, __bt_pgout, t);
312 
313 	/* Create a root page if new tree. */
314 	if (nroot(t) == RET_ERROR)
315 		goto err;
316 
317 	/* Global flags. */
318 	if (dflags & DB_LOCK)
319 		F_SET(t, B_DB_LOCK);
320 	if (dflags & DB_SHMEM)
321 		F_SET(t, B_DB_SHMEM);
322 	if (dflags & DB_TXN)
323 		F_SET(t, B_DB_TXN);
324 
325 	return (dbp);
326 
327 einval:	errno = EINVAL;
328 	goto err;
329 
330 eftype:	errno = EFTYPE;
331 	goto err;
332 
333 err:	if (t) {
334 		if (t->bt_dbp)
335 			free(t->bt_dbp);
336 		if (t->bt_fd != -1)
337 			(void)close(t->bt_fd);
338 		free(t);
339 	}
340 	return (NULL);
341 }
342 
343 /*
344  * NROOT -- Create the root of a new tree.
345  *
346  * Parameters:
347  *	t:	tree
348  *
349  * Returns:
350  *	RET_ERROR, RET_SUCCESS
351  */
352 static int
353 nroot(t)
354 	BTREE *t;
355 {
356 	PAGE *meta, *root;
357 	pgno_t npg;
358 
359 	if ((meta = mpool_get(t->bt_mp, 0, 0)) != NULL) {
360 		mpool_put(t->bt_mp, meta, 0);
361 		return (RET_SUCCESS);
362 	}
363 	if (errno != EINVAL)		/* It's OK to not exist. */
364 		return (RET_ERROR);
365 	errno = 0;
366 
367 	if ((meta = mpool_new(t->bt_mp, &npg)) == NULL)
368 		return (RET_ERROR);
369 
370 	if ((root = mpool_new(t->bt_mp, &npg)) == NULL)
371 		return (RET_ERROR);
372 
373 	if (npg != P_ROOT)
374 		return (RET_ERROR);
375 	root->pgno = npg;
376 	root->prevpg = root->nextpg = P_INVALID;
377 	root->lower = BTDATAOFF;
378 	root->upper = t->bt_psize;
379 	root->flags = P_BLEAF;
380 	memset(meta, 0, t->bt_psize);
381 	mpool_put(t->bt_mp, meta, MPOOL_DIRTY);
382 	mpool_put(t->bt_mp, root, MPOOL_DIRTY);
383 	return (RET_SUCCESS);
384 }
385 
386 static int
387 tmp()
388 {
389 	sigset_t set, oset;
390 	int fd;
391 	char *envtmp = NULL;
392 	char path[MAXPATHLEN];
393 
394 	if (issetugid() == 0)
395 		envtmp = getenv("TMPDIR");
396 	(void)snprintf(path,
397 	    sizeof(path), "%s/bt.XXXXXX", envtmp ? envtmp : "/tmp");
398 
399 	(void)sigfillset(&set);
400 	(void)sigprocmask(SIG_BLOCK, &set, &oset);
401 	if ((fd = mkstemp(path)) != -1)
402 		(void)unlink(path);
403 	(void)sigprocmask(SIG_SETMASK, &oset, NULL);
404 	return(fd);
405 }
406 
407 static int
408 byteorder()
409 {
410 	u_int32_t x;
411 	u_char *p;
412 
413 	x = 0x01020304;
414 	p = (u_char *)&x;
415 	switch (*p) {
416 	case 1:
417 		return (BIG_ENDIAN);
418 	case 4:
419 		return (LITTLE_ENDIAN);
420 	default:
421 		return (0);
422 	}
423 }
424 
425 int
426 __bt_fd(dbp)
427         const DB *dbp;
428 {
429 	BTREE *t;
430 
431 	t = dbp->internal;
432 
433 	/* Toss any page pinned across calls. */
434 	if (t->bt_pinned != NULL) {
435 		mpool_put(t->bt_mp, t->bt_pinned, 0);
436 		t->bt_pinned = NULL;
437 	}
438 
439 	/* In-memory database can't have a file descriptor. */
440 	if (F_ISSET(t, B_INMEM)) {
441 		errno = ENOENT;
442 		return (-1);
443 	}
444 	return (t->bt_fd);
445 }
446