xref: /openbsd/lib/libc/db/btree/bt_split.c (revision 53b37aa9)
1 /*	$OpenBSD: bt_split.c,v 1.13 2005/08/05 13:03:00 espie Exp $	*/
2 
3 /*-
4  * Copyright (c) 1990, 1993, 1994
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Mike Olson.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 #include <sys/types.h>
36 
37 #include <limits.h>
38 #include <stdio.h>
39 #include <stdlib.h>
40 #include <string.h>
41 
42 #include <db.h>
43 #include "btree.h"
44 
45 static int	 bt_broot(BTREE *, PAGE *, PAGE *, PAGE *);
46 static PAGE	*bt_page(BTREE *, PAGE *, PAGE **, PAGE **, indx_t *, size_t);
47 static int	 bt_preserve(BTREE *, pgno_t);
48 static PAGE	*bt_psplit(BTREE *, PAGE *, PAGE *, PAGE *, indx_t *, size_t);
49 static PAGE	*bt_root(BTREE *, PAGE *, PAGE **, PAGE **, indx_t *, size_t);
50 static int	 bt_rroot(BTREE *, PAGE *, PAGE *, PAGE *);
51 static recno_t	 rec_total(PAGE *);
52 
53 #ifdef STATISTICS
54 u_long	bt_rootsplit, bt_split, bt_sortsplit, bt_pfxsaved;
55 #endif
56 
57 /*
58  * __BT_SPLIT -- Split the tree.
59  *
60  * Parameters:
61  *	t:	tree
62  *	sp:	page to split
63  *	key:	key to insert
64  *	data:	data to insert
65  *	flags:	BIGKEY/BIGDATA flags
66  *	ilen:	insert length
67  *	skip:	index to leave open
68  *
69  * Returns:
70  *	RET_ERROR, RET_SUCCESS
71  */
72 int
__bt_split(BTREE * t,PAGE * sp,const DBT * key,const DBT * data,int flags,size_t ilen,u_int32_t argskip)73 __bt_split(BTREE *t, PAGE *sp, const DBT *key, const DBT *data, int flags,
74     size_t ilen, u_int32_t argskip)
75 {
76 	BINTERNAL *bi;
77 	BLEAF *bl, *tbl;
78 	DBT a, b;
79 	EPGNO *parent;
80 	PAGE *h, *l, *r, *lchild, *rchild;
81 	indx_t nxtindex;
82 	u_int16_t skip;
83 	u_int32_t n, nbytes, nksize;
84 	int parentsplit;
85 	char *dest;
86 
87 	/*
88 	 * Split the page into two pages, l and r.  The split routines return
89 	 * a pointer to the page into which the key should be inserted and with
90 	 * skip set to the offset which should be used.  Additionally, l and r
91 	 * are pinned.
92 	 */
93 	skip = argskip;
94 	h = sp->pgno == P_ROOT ?
95 	    bt_root(t, sp, &l, &r, &skip, ilen) :
96 	    bt_page(t, sp, &l, &r, &skip, ilen);
97 	if (h == NULL)
98 		return (RET_ERROR);
99 
100 	/*
101 	 * Insert the new key/data pair into the leaf page.  (Key inserts
102 	 * always cause a leaf page to split first.)
103 	 */
104 	h->linp[skip] = h->upper -= ilen;
105 	dest = (char *)h + h->upper;
106 	if (F_ISSET(t, R_RECNO))
107 		WR_RLEAF(dest, data, flags)
108 	else
109 		WR_BLEAF(dest, key, data, flags)
110 
111 	/* If the root page was split, make it look right. */
112 	if (sp->pgno == P_ROOT &&
113 	    (F_ISSET(t, R_RECNO) ?
114 	    bt_rroot(t, sp, l, r) : bt_broot(t, sp, l, r)) == RET_ERROR)
115 		goto err2;
116 
117 	/*
118 	 * Now we walk the parent page stack -- a LIFO stack of the pages that
119 	 * were traversed when we searched for the page that split.  Each stack
120 	 * entry is a page number and a page index offset.  The offset is for
121 	 * the page traversed on the search.  We've just split a page, so we
122 	 * have to insert a new key into the parent page.
123 	 *
124 	 * If the insert into the parent page causes it to split, may have to
125 	 * continue splitting all the way up the tree.  We stop if the root
126 	 * splits or the page inserted into didn't have to split to hold the
127 	 * new key.  Some algorithms replace the key for the old page as well
128 	 * as the new page.  We don't, as there's no reason to believe that the
129 	 * first key on the old page is any better than the key we have, and,
130 	 * in the case of a key being placed at index 0 causing the split, the
131 	 * key is unavailable.
132 	 *
133 	 * There are a maximum of 5 pages pinned at any time.  We keep the left
134 	 * and right pages pinned while working on the parent.   The 5 are the
135 	 * two children, left parent and right parent (when the parent splits)
136 	 * and the root page or the overflow key page when calling bt_preserve.
137 	 * This code must make sure that all pins are released other than the
138 	 * root page or overflow page which is unlocked elsewhere.
139 	 */
140 	while ((parent = BT_POP(t)) != NULL) {
141 		lchild = l;
142 		rchild = r;
143 
144 		/* Get the parent page. */
145 		if ((h = mpool_get(t->bt_mp, parent->pgno, 0)) == NULL)
146 			goto err2;
147 
148 	 	/*
149 		 * The new key goes ONE AFTER the index, because the split
150 		 * was to the right.
151 		 */
152 		skip = parent->index + 1;
153 
154 		/*
155 		 * Calculate the space needed on the parent page.
156 		 *
157 		 * Prefix trees: space hack when inserting into BINTERNAL
158 		 * pages.  Retain only what's needed to distinguish between
159 		 * the new entry and the LAST entry on the page to its left.
160 		 * If the keys compare equal, retain the entire key.  Note,
161 		 * we don't touch overflow keys, and the entire key must be
162 		 * retained for the next-to-left most key on the leftmost
163 		 * page of each level, or the search will fail.  Applicable
164 		 * ONLY to internal pages that have leaf pages as children.
165 		 * Further reduction of the key between pairs of internal
166 		 * pages loses too much information.
167 		 */
168 		switch (rchild->flags & P_TYPE) {
169 		case P_BINTERNAL:
170 			bi = GETBINTERNAL(rchild, 0);
171 			nbytes = NBINTERNAL(bi->ksize);
172 			break;
173 		case P_BLEAF:
174 			bl = GETBLEAF(rchild, 0);
175 			nbytes = NBINTERNAL(bl->ksize);
176 			if (t->bt_pfx && !(bl->flags & P_BIGKEY) &&
177 			    (h->prevpg != P_INVALID || skip > 1)) {
178 				tbl = GETBLEAF(lchild, NEXTINDEX(lchild) - 1);
179 				a.size = tbl->ksize;
180 				a.data = tbl->bytes;
181 				b.size = bl->ksize;
182 				b.data = bl->bytes;
183 				nksize = t->bt_pfx(&a, &b);
184 				n = NBINTERNAL(nksize);
185 				if (n < nbytes) {
186 #ifdef STATISTICS
187 					bt_pfxsaved += nbytes - n;
188 #endif
189 					nbytes = n;
190 				} else
191 					nksize = 0;
192 			} else
193 				nksize = 0;
194 			break;
195 		case P_RINTERNAL:
196 		case P_RLEAF:
197 			nbytes = NRINTERNAL;
198 			break;
199 		default:
200 			abort();
201 		}
202 
203 		/* Split the parent page if necessary or shift the indices. */
204 		if (h->upper - h->lower < nbytes + sizeof(indx_t)) {
205 			sp = h;
206 			h = h->pgno == P_ROOT ?
207 			    bt_root(t, h, &l, &r, &skip, nbytes) :
208 			    bt_page(t, h, &l, &r, &skip, nbytes);
209 			if (h == NULL)
210 				goto err1;
211 			parentsplit = 1;
212 		} else {
213 			if (skip < (nxtindex = NEXTINDEX(h)))
214 				memmove(h->linp + skip + 1, h->linp + skip,
215 				    (nxtindex - skip) * sizeof(indx_t));
216 			h->lower += sizeof(indx_t);
217 			parentsplit = 0;
218 		}
219 
220 		/* Insert the key into the parent page. */
221 		switch (rchild->flags & P_TYPE) {
222 		case P_BINTERNAL:
223 			h->linp[skip] = h->upper -= nbytes;
224 			dest = (char *)h + h->linp[skip];
225 			memmove(dest, bi, nbytes);
226 			((BINTERNAL *)dest)->pgno = rchild->pgno;
227 			break;
228 		case P_BLEAF:
229 			h->linp[skip] = h->upper -= nbytes;
230 			dest = (char *)h + h->linp[skip];
231 			WR_BINTERNAL(dest, nksize ? nksize : bl->ksize,
232 			    rchild->pgno, bl->flags & P_BIGKEY);
233 			memmove(dest, bl->bytes, nksize ? nksize : bl->ksize);
234 			if (bl->flags & P_BIGKEY &&
235 			    bt_preserve(t, *(pgno_t *)bl->bytes) == RET_ERROR)
236 				goto err1;
237 			break;
238 		case P_RINTERNAL:
239 			/*
240 			 * Update the left page count.  If split
241 			 * added at index 0, fix the correct page.
242 			 */
243 			if (skip > 0)
244 				dest = (char *)h + h->linp[skip - 1];
245 			else
246 				dest = (char *)l + l->linp[NEXTINDEX(l) - 1];
247 			((RINTERNAL *)dest)->nrecs = rec_total(lchild);
248 			((RINTERNAL *)dest)->pgno = lchild->pgno;
249 
250 			/* Update the right page count. */
251 			h->linp[skip] = h->upper -= nbytes;
252 			dest = (char *)h + h->linp[skip];
253 			((RINTERNAL *)dest)->nrecs = rec_total(rchild);
254 			((RINTERNAL *)dest)->pgno = rchild->pgno;
255 			break;
256 		case P_RLEAF:
257 			/*
258 			 * Update the left page count.  If split
259 			 * added at index 0, fix the correct page.
260 			 */
261 			if (skip > 0)
262 				dest = (char *)h + h->linp[skip - 1];
263 			else
264 				dest = (char *)l + l->linp[NEXTINDEX(l) - 1];
265 			((RINTERNAL *)dest)->nrecs = NEXTINDEX(lchild);
266 			((RINTERNAL *)dest)->pgno = lchild->pgno;
267 
268 			/* Update the right page count. */
269 			h->linp[skip] = h->upper -= nbytes;
270 			dest = (char *)h + h->linp[skip];
271 			((RINTERNAL *)dest)->nrecs = NEXTINDEX(rchild);
272 			((RINTERNAL *)dest)->pgno = rchild->pgno;
273 			break;
274 		default:
275 			abort();
276 		}
277 
278 		/* Unpin the held pages. */
279 		if (!parentsplit) {
280 			mpool_put(t->bt_mp, h, MPOOL_DIRTY);
281 			break;
282 		}
283 
284 		/* If the root page was split, make it look right. */
285 		if (sp->pgno == P_ROOT &&
286 		    (F_ISSET(t, R_RECNO) ?
287 		    bt_rroot(t, sp, l, r) : bt_broot(t, sp, l, r)) == RET_ERROR)
288 			goto err1;
289 
290 		mpool_put(t->bt_mp, lchild, MPOOL_DIRTY);
291 		mpool_put(t->bt_mp, rchild, MPOOL_DIRTY);
292 	}
293 
294 	/* Unpin the held pages. */
295 	mpool_put(t->bt_mp, l, MPOOL_DIRTY);
296 	mpool_put(t->bt_mp, r, MPOOL_DIRTY);
297 
298 	/* Clear any pages left on the stack. */
299 	return (RET_SUCCESS);
300 
301 	/*
302 	 * If something fails in the above loop we were already walking back
303 	 * up the tree and the tree is now inconsistent.  Nothing much we can
304 	 * do about it but release any memory we're holding.
305 	 */
306 err1:	mpool_put(t->bt_mp, lchild, MPOOL_DIRTY);
307 	mpool_put(t->bt_mp, rchild, MPOOL_DIRTY);
308 
309 err2:	mpool_put(t->bt_mp, l, 0);
310 	mpool_put(t->bt_mp, r, 0);
311 	__dbpanic(t->bt_dbp);
312 	return (RET_ERROR);
313 }
314 
315 /*
316  * BT_PAGE -- Split a non-root page of a btree.
317  *
318  * Parameters:
319  *	t:	tree
320  *	h:	root page
321  *	lp:	pointer to left page pointer
322  *	rp:	pointer to right page pointer
323  *	skip:	pointer to index to leave open
324  *	ilen:	insert length
325  *
326  * Returns:
327  *	Pointer to page in which to insert or NULL on error.
328  */
329 static PAGE *
bt_page(BTREE * t,PAGE * h,PAGE ** lp,PAGE ** rp,indx_t * skip,size_t ilen)330 bt_page(BTREE *t, PAGE *h, PAGE **lp, PAGE **rp, indx_t *skip, size_t ilen)
331 {
332 	PAGE *l, *r, *tp;
333 	pgno_t npg;
334 
335 #ifdef STATISTICS
336 	++bt_split;
337 #endif
338 	/* Put the new right page for the split into place. */
339 	if ((r = __bt_new(t, &npg)) == NULL)
340 		return (NULL);
341 	r->pgno = npg;
342 	r->lower = BTDATAOFF;
343 	r->upper = t->bt_psize;
344 	r->nextpg = h->nextpg;
345 	r->prevpg = h->pgno;
346 	r->flags = h->flags & P_TYPE;
347 
348 	/*
349 	 * If we're splitting the last page on a level because we're appending
350 	 * a key to it (skip is NEXTINDEX()), it's likely that the data is
351 	 * sorted.  Adding an empty page on the side of the level is less work
352 	 * and can push the fill factor much higher than normal.  If we're
353 	 * wrong it's no big deal, we'll just do the split the right way next
354 	 * time.  It may look like it's equally easy to do a similar hack for
355 	 * reverse sorted data, that is, split the tree left, but it's not.
356 	 * Don't even try.
357 	 */
358 	if (h->nextpg == P_INVALID && *skip == NEXTINDEX(h)) {
359 #ifdef STATISTICS
360 		++bt_sortsplit;
361 #endif
362 		h->nextpg = r->pgno;
363 		r->lower = BTDATAOFF + sizeof(indx_t);
364 		*skip = 0;
365 		*lp = h;
366 		*rp = r;
367 		return (r);
368 	}
369 
370 	/* Put the new left page for the split into place. */
371 	if ((l = (PAGE *)malloc(t->bt_psize)) == NULL) {
372 		mpool_put(t->bt_mp, r, 0);
373 		return (NULL);
374 	}
375 	memset(l, 0xff, t->bt_psize);
376 	l->pgno = h->pgno;
377 	l->nextpg = r->pgno;
378 	l->prevpg = h->prevpg;
379 	l->lower = BTDATAOFF;
380 	l->upper = t->bt_psize;
381 	l->flags = h->flags & P_TYPE;
382 
383 	/* Fix up the previous pointer of the page after the split page. */
384 	if (h->nextpg != P_INVALID) {
385 		if ((tp = mpool_get(t->bt_mp, h->nextpg, 0)) == NULL) {
386 			free(l);
387 			/* XXX mpool_free(t->bt_mp, r->pgno); */
388 			return (NULL);
389 		}
390 		tp->prevpg = r->pgno;
391 		mpool_put(t->bt_mp, tp, MPOOL_DIRTY);
392 	}
393 
394 	/*
395 	 * Split right.  The key/data pairs aren't sorted in the btree page so
396 	 * it's simpler to copy the data from the split page onto two new pages
397 	 * instead of copying half the data to the right page and compacting
398 	 * the left page in place.  Since the left page can't change, we have
399 	 * to swap the original and the allocated left page after the split.
400 	 */
401 	tp = bt_psplit(t, h, l, r, skip, ilen);
402 
403 	/* Move the new left page onto the old left page. */
404 	memmove(h, l, t->bt_psize);
405 	if (tp == l)
406 		tp = h;
407 	free(l);
408 
409 	*lp = h;
410 	*rp = r;
411 	return (tp);
412 }
413 
414 /*
415  * BT_ROOT -- Split the root page of a btree.
416  *
417  * Parameters:
418  *	t:	tree
419  *	h:	root page
420  *	lp:	pointer to left page pointer
421  *	rp:	pointer to right page pointer
422  *	skip:	pointer to index to leave open
423  *	ilen:	insert length
424  *
425  * Returns:
426  *	Pointer to page in which to insert or NULL on error.
427  */
428 static PAGE *
bt_root(BTREE * t,PAGE * h,PAGE ** lp,PAGE ** rp,indx_t * skip,size_t ilen)429 bt_root(BTREE *t, PAGE *h, PAGE **lp, PAGE **rp, indx_t *skip, size_t ilen)
430 {
431 	PAGE *l, *r, *tp;
432 	pgno_t lnpg, rnpg;
433 
434 #ifdef STATISTICS
435 	++bt_split;
436 	++bt_rootsplit;
437 #endif
438 	/* Put the new left and right pages for the split into place. */
439 	if ((l = __bt_new(t, &lnpg)) == NULL ||
440 	    (r = __bt_new(t, &rnpg)) == NULL)
441 		return (NULL);
442 	l->pgno = lnpg;
443 	r->pgno = rnpg;
444 	l->nextpg = r->pgno;
445 	r->prevpg = l->pgno;
446 	l->prevpg = r->nextpg = P_INVALID;
447 	l->lower = r->lower = BTDATAOFF;
448 	l->upper = r->upper = t->bt_psize;
449 	l->flags = r->flags = h->flags & P_TYPE;
450 
451 	/* Split the root page. */
452 	tp = bt_psplit(t, h, l, r, skip, ilen);
453 
454 	*lp = l;
455 	*rp = r;
456 	return (tp);
457 }
458 
459 /*
460  * BT_RROOT -- Fix up the recno root page after it has been split.
461  *
462  * Parameters:
463  *	t:	tree
464  *	h:	root page
465  *	l:	left page
466  *	r:	right page
467  *
468  * Returns:
469  *	RET_ERROR, RET_SUCCESS
470  */
471 static int
bt_rroot(BTREE * t,PAGE * h,PAGE * l,PAGE * r)472 bt_rroot(BTREE *t, PAGE *h, PAGE *l, PAGE *r)
473 {
474 	char *dest;
475 
476 	/* Insert the left and right keys, set the header information. */
477 	h->linp[0] = h->upper = t->bt_psize - NRINTERNAL;
478 	dest = (char *)h + h->upper;
479 	WR_RINTERNAL(dest,
480 	    l->flags & P_RLEAF ? NEXTINDEX(l) : rec_total(l), l->pgno);
481 
482 	h->linp[1] = h->upper -= NRINTERNAL;
483 	dest = (char *)h + h->upper;
484 	WR_RINTERNAL(dest,
485 	    r->flags & P_RLEAF ? NEXTINDEX(r) : rec_total(r), r->pgno);
486 
487 	h->lower = BTDATAOFF + 2 * sizeof(indx_t);
488 
489 	/* Unpin the root page, set to recno internal page. */
490 	h->flags &= ~P_TYPE;
491 	h->flags |= P_RINTERNAL;
492 	mpool_put(t->bt_mp, h, MPOOL_DIRTY);
493 
494 	return (RET_SUCCESS);
495 }
496 
497 /*
498  * BT_BROOT -- Fix up the btree root page after it has been split.
499  *
500  * Parameters:
501  *	t:	tree
502  *	h:	root page
503  *	l:	left page
504  *	r:	right page
505  *
506  * Returns:
507  *	RET_ERROR, RET_SUCCESS
508  */
509 static int
bt_broot(BTREE * t,PAGE * h,PAGE * l,PAGE * r)510 bt_broot(BTREE *t, PAGE *h, PAGE *l, PAGE *r)
511 {
512 	BINTERNAL *bi;
513 	BLEAF *bl;
514 	u_int32_t nbytes;
515 	char *dest;
516 
517 	/*
518 	 * If the root page was a leaf page, change it into an internal page.
519 	 * We copy the key we split on (but not the key's data, in the case of
520 	 * a leaf page) to the new root page.
521 	 *
522 	 * The btree comparison code guarantees that the left-most key on any
523 	 * level of the tree is never used, so it doesn't need to be filled in.
524 	 */
525 	nbytes = NBINTERNAL(0);
526 	h->linp[0] = h->upper = t->bt_psize - nbytes;
527 	dest = (char *)h + h->upper;
528 	WR_BINTERNAL(dest, 0, l->pgno, 0);
529 
530 	switch (h->flags & P_TYPE) {
531 	case P_BLEAF:
532 		bl = GETBLEAF(r, 0);
533 		nbytes = NBINTERNAL(bl->ksize);
534 		h->linp[1] = h->upper -= nbytes;
535 		dest = (char *)h + h->upper;
536 		WR_BINTERNAL(dest, bl->ksize, r->pgno, 0);
537 		memmove(dest, bl->bytes, bl->ksize);
538 
539 		/*
540 		 * If the key is on an overflow page, mark the overflow chain
541 		 * so it isn't deleted when the leaf copy of the key is deleted.
542 		 */
543 		if (bl->flags & P_BIGKEY &&
544 		    bt_preserve(t, *(pgno_t *)bl->bytes) == RET_ERROR)
545 			return (RET_ERROR);
546 		break;
547 	case P_BINTERNAL:
548 		bi = GETBINTERNAL(r, 0);
549 		nbytes = NBINTERNAL(bi->ksize);
550 		h->linp[1] = h->upper -= nbytes;
551 		dest = (char *)h + h->upper;
552 		memmove(dest, bi, nbytes);
553 		((BINTERNAL *)dest)->pgno = r->pgno;
554 		break;
555 	default:
556 		abort();
557 	}
558 
559 	/* There are two keys on the page. */
560 	h->lower = BTDATAOFF + 2 * sizeof(indx_t);
561 
562 	/* Unpin the root page, set to btree internal page. */
563 	h->flags &= ~P_TYPE;
564 	h->flags |= P_BINTERNAL;
565 	mpool_put(t->bt_mp, h, MPOOL_DIRTY);
566 
567 	return (RET_SUCCESS);
568 }
569 
570 /*
571  * BT_PSPLIT -- Do the real work of splitting the page.
572  *
573  * Parameters:
574  *	t:	tree
575  *	h:	page to be split
576  *	l:	page to put lower half of data
577  *	r:	page to put upper half of data
578  *	pskip:	pointer to index to leave open
579  *	ilen:	insert length
580  *
581  * Returns:
582  *	Pointer to page in which to insert.
583  */
584 static PAGE *
bt_psplit(BTREE * t,PAGE * h,PAGE * l,PAGE * r,indx_t * pskip,size_t ilen)585 bt_psplit(BTREE *t, PAGE *h, PAGE *l, PAGE *r, indx_t *pskip, size_t ilen)
586 {
587 	BINTERNAL *bi;
588 	BLEAF *bl;
589 	CURSOR *c;
590 	RLEAF *rl;
591 	PAGE *rval;
592 	void *src;
593 	indx_t full, half, nxt, off, skip, top, used;
594 	u_int32_t nbytes;
595 	int bigkeycnt, isbigkey;
596 
597 	/*
598 	 * Split the data to the left and right pages.  Leave the skip index
599 	 * open.  Additionally, make some effort not to split on an overflow
600 	 * key.  This makes internal page processing faster and can save
601 	 * space as overflow keys used by internal pages are never deleted.
602 	 */
603 	bigkeycnt = 0;
604 	skip = *pskip;
605 	full = t->bt_psize - BTDATAOFF;
606 	half = full / 2;
607 	used = 0;
608 	for (nxt = off = 0, top = NEXTINDEX(h); nxt < top; ++off) {
609 		if (skip == off) {
610 			nbytes = ilen;
611 			isbigkey = 0;		/* XXX: not really known. */
612 		} else
613 			switch (h->flags & P_TYPE) {
614 			case P_BINTERNAL:
615 				src = bi = GETBINTERNAL(h, nxt);
616 				nbytes = NBINTERNAL(bi->ksize);
617 				isbigkey = bi->flags & P_BIGKEY;
618 				break;
619 			case P_BLEAF:
620 				src = bl = GETBLEAF(h, nxt);
621 				nbytes = NBLEAF(bl);
622 				isbigkey = bl->flags & P_BIGKEY;
623 				break;
624 			case P_RINTERNAL:
625 				src = GETRINTERNAL(h, nxt);
626 				nbytes = NRINTERNAL;
627 				isbigkey = 0;
628 				break;
629 			case P_RLEAF:
630 				src = rl = GETRLEAF(h, nxt);
631 				nbytes = NRLEAF(rl);
632 				isbigkey = 0;
633 				break;
634 			default:
635 				abort();
636 			}
637 
638 		/*
639 		 * If the key/data pairs are substantial fractions of the max
640 		 * possible size for the page, it's possible to get situations
641 		 * where we decide to try and copy too much onto the left page.
642 		 * Make sure that doesn't happen.
643 		 */
644 		if ((skip <= off && used + nbytes + sizeof(indx_t) >= full) ||
645 		    nxt == top - 1) {
646 			--off;
647 			break;
648 		}
649 
650 		/* Copy the key/data pair, if not the skipped index. */
651 		if (skip != off) {
652 			++nxt;
653 
654 			l->linp[off] = l->upper -= nbytes;
655 			memmove((char *)l + l->upper, src, nbytes);
656 		}
657 
658 		used += nbytes + sizeof(indx_t);
659 		if (used >= half) {
660 			if (!isbigkey || bigkeycnt == 3)
661 				break;
662 			else
663 				++bigkeycnt;
664 		}
665 	}
666 
667 	/*
668 	 * Off is the last offset that's valid for the left page.
669 	 * Nxt is the first offset to be placed on the right page.
670 	 */
671 	l->lower += (off + 1) * sizeof(indx_t);
672 
673 	/*
674 	 * If splitting the page that the cursor was on, the cursor has to be
675 	 * adjusted to point to the same record as before the split.  If the
676 	 * cursor is at or past the skipped slot, the cursor is incremented by
677 	 * one.  If the cursor is on the right page, it is decremented by the
678 	 * number of records split to the left page.
679 	 */
680 	c = &t->bt_cursor;
681 	if (F_ISSET(c, CURS_INIT) && c->pg.pgno == h->pgno) {
682 		if (c->pg.index >= skip)
683 			++c->pg.index;
684 		if (c->pg.index < nxt)			/* Left page. */
685 			c->pg.pgno = l->pgno;
686 		else {					/* Right page. */
687 			c->pg.pgno = r->pgno;
688 			c->pg.index -= nxt;
689 		}
690 	}
691 
692 	/*
693 	 * If the skipped index was on the left page, just return that page.
694 	 * Otherwise, adjust the skip index to reflect the new position on
695 	 * the right page.
696 	 */
697 	if (skip <= off) {
698 		skip = MAX_PAGE_OFFSET;
699 		rval = l;
700 	} else {
701 		rval = r;
702 		*pskip -= nxt;
703 	}
704 
705 	for (off = 0; nxt < top; ++off) {
706 		if (skip == nxt) {
707 			++off;
708 			skip = MAX_PAGE_OFFSET;
709 		}
710 		switch (h->flags & P_TYPE) {
711 		case P_BINTERNAL:
712 			src = bi = GETBINTERNAL(h, nxt);
713 			nbytes = NBINTERNAL(bi->ksize);
714 			break;
715 		case P_BLEAF:
716 			src = bl = GETBLEAF(h, nxt);
717 			nbytes = NBLEAF(bl);
718 			break;
719 		case P_RINTERNAL:
720 			src = GETRINTERNAL(h, nxt);
721 			nbytes = NRINTERNAL;
722 			break;
723 		case P_RLEAF:
724 			src = rl = GETRLEAF(h, nxt);
725 			nbytes = NRLEAF(rl);
726 			break;
727 		default:
728 			abort();
729 		}
730 		++nxt;
731 		r->linp[off] = r->upper -= nbytes;
732 		memmove((char *)r + r->upper, src, nbytes);
733 	}
734 	r->lower += off * sizeof(indx_t);
735 
736 	/* If the key is being appended to the page, adjust the index. */
737 	if (skip == top)
738 		r->lower += sizeof(indx_t);
739 
740 	return (rval);
741 }
742 
743 /*
744  * BT_PRESERVE -- Mark a chain of pages as used by an internal node.
745  *
746  * Chains of indirect blocks pointed to by leaf nodes get reclaimed when the
747  * record that references them gets deleted.  Chains pointed to by internal
748  * pages never get deleted.  This routine marks a chain as pointed to by an
749  * internal page.
750  *
751  * Parameters:
752  *	t:	tree
753  *	pg:	page number of first page in the chain.
754  *
755  * Returns:
756  *	RET_SUCCESS, RET_ERROR.
757  */
758 static int
bt_preserve(BTREE * t,pgno_t pg)759 bt_preserve(BTREE *t, pgno_t pg)
760 {
761 	PAGE *h;
762 
763 	if ((h = mpool_get(t->bt_mp, pg, 0)) == NULL)
764 		return (RET_ERROR);
765 	h->flags |= P_PRESERVE;
766 	mpool_put(t->bt_mp, h, MPOOL_DIRTY);
767 	return (RET_SUCCESS);
768 }
769 
770 /*
771  * REC_TOTAL -- Return the number of recno entries below a page.
772  *
773  * Parameters:
774  *	h:	page
775  *
776  * Returns:
777  *	The number of recno entries below a page.
778  *
779  * XXX
780  * These values could be set by the bt_psplit routine.  The problem is that the
781  * entry has to be popped off of the stack etc. or the values have to be passed
782  * all the way back to bt_split/bt_rroot and it's not very clean.
783  */
784 static recno_t
rec_total(PAGE * h)785 rec_total(PAGE *h)
786 {
787 	recno_t recs;
788 	indx_t nxt, top;
789 
790 	for (recs = 0, nxt = 0, top = NEXTINDEX(h); nxt < top; ++nxt)
791 		recs += GETRINTERNAL(h, nxt)->nrecs;
792 	return (recs);
793 }
794