blob: 4dcc0581999881063ece0e2f6d745b7d52da2a5b [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/*
2 * Copyright (C) International Business Machines Corp., 2000-2004
3 *
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
Dave Kleikamp63f83c92006-10-02 09:55:27 -05006 * the Free Software Foundation; either version 2 of the License, or
Linus Torvalds1da177e2005-04-16 15:20:36 -07007 * (at your option) any later version.
Dave Kleikamp63f83c92006-10-02 09:55:27 -05008 *
Linus Torvalds1da177e2005-04-16 15:20:36 -07009 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See
12 * the GNU General Public License for more details.
13 *
14 * You should have received a copy of the GNU General Public License
Dave Kleikamp63f83c92006-10-02 09:55:27 -050015 * along with this program; if not, write to the Free Software
Linus Torvalds1da177e2005-04-16 15:20:36 -070016 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17 */
18
19/*
20 * jfs_dtree.c: directory B+-tree manager
21 *
22 * B+-tree with variable length key directory:
23 *
24 * each directory page is structured as an array of 32-byte
25 * directory entry slots initialized as a freelist
26 * to avoid search/compaction of free space at insertion.
27 * when an entry is inserted, a number of slots are allocated
28 * from the freelist as required to store variable length data
29 * of the entry; when the entry is deleted, slots of the entry
30 * are returned to freelist.
31 *
32 * leaf entry stores full name as key and file serial number
33 * (aka inode number) as data.
34 * internal/router entry stores sufffix compressed name
35 * as key and simple extent descriptor as data.
36 *
37 * each directory page maintains a sorted entry index table
38 * which stores the start slot index of sorted entries
39 * to allow binary search on the table.
40 *
41 * directory starts as a root/leaf page in on-disk inode
42 * inline data area.
43 * when it becomes full, it starts a leaf of a external extent
44 * of length of 1 block. each time the first leaf becomes full,
45 * it is extended rather than split (its size is doubled),
46 * until its length becoms 4 KBytes, from then the extent is split
47 * with new 4 Kbyte extent when it becomes full
48 * to reduce external fragmentation of small directories.
49 *
50 * blah, blah, blah, for linear scan of directory in pieces by
51 * readdir().
52 *
53 *
54 * case-insensitive directory file system
55 *
56 * names are stored in case-sensitive way in leaf entry.
57 * but stored, searched and compared in case-insensitive (uppercase) order
58 * (i.e., both search key and entry key are folded for search/compare):
59 * (note that case-sensitive order is BROKEN in storage, e.g.,
60 * sensitive: Ad, aB, aC, aD -> insensitive: aB, aC, aD, Ad
61 *
62 * entries which folds to the same key makes up a equivalent class
63 * whose members are stored as contiguous cluster (may cross page boundary)
64 * but whose order is arbitrary and acts as duplicate, e.g.,
65 * abc, Abc, aBc, abC)
66 *
67 * once match is found at leaf, requires scan forward/backward
68 * either for, in case-insensitive search, duplicate
69 * or for, in case-sensitive search, for exact match
70 *
71 * router entry must be created/stored in case-insensitive way
72 * in internal entry:
73 * (right most key of left page and left most key of right page
74 * are folded, and its suffix compression is propagated as router
75 * key in parent)
76 * (e.g., if split occurs <abc> and <aBd>, <ABD> trather than <aB>
77 * should be made the router key for the split)
78 *
79 * case-insensitive search:
80 *
Dave Kleikamp63f83c92006-10-02 09:55:27 -050081 * fold search key;
Linus Torvalds1da177e2005-04-16 15:20:36 -070082 *
83 * case-insensitive search of B-tree:
84 * for internal entry, router key is already folded;
85 * for leaf entry, fold the entry key before comparison.
86 *
87 * if (leaf entry case-insensitive match found)
88 * if (next entry satisfies case-insensitive match)
89 * return EDUPLICATE;
90 * if (prev entry satisfies case-insensitive match)
91 * return EDUPLICATE;
92 * return match;
93 * else
94 * return no match;
95 *
Dave Kleikamp63f83c92006-10-02 09:55:27 -050096 * serialization:
Linus Torvalds1da177e2005-04-16 15:20:36 -070097 * target directory inode lock is being held on entry/exit
98 * of all main directory service routines.
99 *
100 * log based recovery:
101 */
102
103#include <linux/fs.h>
104#include <linux/quotaops.h>
105#include "jfs_incore.h"
106#include "jfs_superblock.h"
107#include "jfs_filsys.h"
108#include "jfs_metapage.h"
109#include "jfs_dmap.h"
110#include "jfs_unicode.h"
111#include "jfs_debug.h"
112
113/* dtree split parameter */
114struct dtsplit {
115 struct metapage *mp;
116 s16 index;
117 s16 nslot;
118 struct component_name *key;
119 ddata_t *data;
120 struct pxdlist *pxdlist;
121};
122
123#define DT_PAGE(IP, MP) BT_PAGE(IP, MP, dtpage_t, i_dtroot)
124
125/* get page buffer for specified block address */
126#define DT_GETPAGE(IP, BN, MP, SIZE, P, RC)\
127{\
128 BT_GETPAGE(IP, BN, MP, dtpage_t, SIZE, P, RC, i_dtroot)\
129 if (!(RC))\
130 {\
131 if (((P)->header.nextindex > (((BN)==0)?DTROOTMAXSLOT:(P)->header.maxslot)) ||\
132 ((BN) && ((P)->header.maxslot > DTPAGEMAXSLOT)))\
133 {\
134 BT_PUTPAGE(MP);\
135 jfs_error((IP)->i_sb, "DT_GETPAGE: dtree page corrupt");\
136 MP = NULL;\
137 RC = -EIO;\
138 }\
139 }\
140}
141
142/* for consistency */
143#define DT_PUTPAGE(MP) BT_PUTPAGE(MP)
144
145#define DT_GETSEARCH(IP, LEAF, BN, MP, P, INDEX) \
146 BT_GETSEARCH(IP, LEAF, BN, MP, dtpage_t, P, INDEX, i_dtroot)
147
148/*
149 * forward references
150 */
151static int dtSplitUp(tid_t tid, struct inode *ip,
152 struct dtsplit * split, struct btstack * btstack);
153
154static int dtSplitPage(tid_t tid, struct inode *ip, struct dtsplit * split,
155 struct metapage ** rmpp, dtpage_t ** rpp, pxd_t * rxdp);
156
157static int dtExtendPage(tid_t tid, struct inode *ip,
158 struct dtsplit * split, struct btstack * btstack);
159
160static int dtSplitRoot(tid_t tid, struct inode *ip,
161 struct dtsplit * split, struct metapage ** rmpp);
162
163static int dtDeleteUp(tid_t tid, struct inode *ip, struct metapage * fmp,
164 dtpage_t * fp, struct btstack * btstack);
165
166static int dtRelink(tid_t tid, struct inode *ip, dtpage_t * p);
167
168static int dtReadFirst(struct inode *ip, struct btstack * btstack);
169
170static int dtReadNext(struct inode *ip,
171 loff_t * offset, struct btstack * btstack);
172
173static int dtCompare(struct component_name * key, dtpage_t * p, int si);
174
175static int ciCompare(struct component_name * key, dtpage_t * p, int si,
176 int flag);
177
178static void dtGetKey(dtpage_t * p, int i, struct component_name * key,
179 int flag);
180
181static int ciGetLeafPrefixKey(dtpage_t * lp, int li, dtpage_t * rp,
182 int ri, struct component_name * key, int flag);
183
184static void dtInsertEntry(dtpage_t * p, int index, struct component_name * key,
185 ddata_t * data, struct dt_lock **);
186
187static void dtMoveEntry(dtpage_t * sp, int si, dtpage_t * dp,
188 struct dt_lock ** sdtlock, struct dt_lock ** ddtlock,
189 int do_index);
190
191static void dtDeleteEntry(dtpage_t * p, int fi, struct dt_lock ** dtlock);
192
193static void dtTruncateEntry(dtpage_t * p, int ti, struct dt_lock ** dtlock);
194
195static void dtLinelockFreelist(dtpage_t * p, int m, struct dt_lock ** dtlock);
196
197#define ciToUpper(c) UniStrupr((c)->name)
198
199/*
200 * read_index_page()
201 *
202 * Reads a page of a directory's index table.
203 * Having metadata mapped into the directory inode's address space
204 * presents a multitude of problems. We avoid this by mapping to
205 * the absolute address space outside of the *_metapage routines
206 */
207static struct metapage *read_index_page(struct inode *inode, s64 blkno)
208{
209 int rc;
210 s64 xaddr;
211 int xflag;
212 s32 xlen;
213
214 rc = xtLookup(inode, blkno, 1, &xflag, &xaddr, &xlen, 1);
Dave Kleikamp66284652005-05-02 12:25:13 -0600215 if (rc || (xaddr == 0))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700216 return NULL;
217
218 return read_metapage(inode, xaddr, PSIZE, 1);
219}
220
221/*
222 * get_index_page()
223 *
224 * Same as get_index_page(), but get's a new page without reading
225 */
226static struct metapage *get_index_page(struct inode *inode, s64 blkno)
227{
228 int rc;
229 s64 xaddr;
230 int xflag;
231 s32 xlen;
232
233 rc = xtLookup(inode, blkno, 1, &xflag, &xaddr, &xlen, 1);
Dave Kleikamp66284652005-05-02 12:25:13 -0600234 if (rc || (xaddr == 0))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700235 return NULL;
236
237 return get_metapage(inode, xaddr, PSIZE, 1);
238}
239
240/*
241 * find_index()
242 *
243 * Returns dtree page containing directory table entry for specified
244 * index and pointer to its entry.
245 *
246 * mp must be released by caller.
247 */
248static struct dir_table_slot *find_index(struct inode *ip, u32 index,
249 struct metapage ** mp, s64 *lblock)
250{
251 struct jfs_inode_info *jfs_ip = JFS_IP(ip);
252 s64 blkno;
253 s64 offset;
254 int page_offset;
255 struct dir_table_slot *slot;
256 static int maxWarnings = 10;
257
258 if (index < 2) {
259 if (maxWarnings) {
260 jfs_warn("find_entry called with index = %d", index);
261 maxWarnings--;
262 }
263 return NULL;
264 }
265
266 if (index >= jfs_ip->next_index) {
267 jfs_warn("find_entry called with index >= next_index");
268 return NULL;
269 }
270
271 if (jfs_dirtable_inline(ip)) {
272 /*
273 * Inline directory table
274 */
275 *mp = NULL;
276 slot = &jfs_ip->i_dirtable[index - 2];
277 } else {
278 offset = (index - 2) * sizeof(struct dir_table_slot);
279 page_offset = offset & (PSIZE - 1);
280 blkno = ((offset + 1) >> L2PSIZE) <<
281 JFS_SBI(ip->i_sb)->l2nbperpage;
282
283 if (*mp && (*lblock != blkno)) {
284 release_metapage(*mp);
285 *mp = NULL;
286 }
Joe Perches09aaa742007-11-13 22:16:08 -0600287 if (!(*mp)) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700288 *lblock = blkno;
289 *mp = read_index_page(ip, blkno);
290 }
Joe Perches09aaa742007-11-13 22:16:08 -0600291 if (!(*mp)) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700292 jfs_err("free_index: error reading directory table");
293 return NULL;
294 }
295
296 slot =
297 (struct dir_table_slot *) ((char *) (*mp)->data +
298 page_offset);
299 }
300 return slot;
301}
302
303static inline void lock_index(tid_t tid, struct inode *ip, struct metapage * mp,
304 u32 index)
305{
306 struct tlock *tlck;
307 struct linelock *llck;
308 struct lv *lv;
309
310 tlck = txLock(tid, ip, mp, tlckDATA);
311 llck = (struct linelock *) tlck->lock;
312
313 if (llck->index >= llck->maxcnt)
314 llck = txLinelock(llck);
315 lv = &llck->lv[llck->index];
316
317 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500318 * Linelock slot size is twice the size of directory table
319 * slot size. 512 entries per page.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700320 */
321 lv->offset = ((index - 2) & 511) >> 1;
322 lv->length = 1;
323 llck->index++;
324}
325
326/*
327 * add_index()
328 *
329 * Adds an entry to the directory index table. This is used to provide
330 * each directory entry with a persistent index in which to resume
331 * directory traversals
332 */
333static u32 add_index(tid_t tid, struct inode *ip, s64 bn, int slot)
334{
335 struct super_block *sb = ip->i_sb;
336 struct jfs_sb_info *sbi = JFS_SBI(sb);
337 struct jfs_inode_info *jfs_ip = JFS_IP(ip);
338 u64 blkno;
339 struct dir_table_slot *dirtab_slot;
340 u32 index;
341 struct linelock *llck;
342 struct lv *lv;
343 struct metapage *mp;
344 s64 offset;
345 uint page_offset;
346 struct tlock *tlck;
347 s64 xaddr;
348
349 ASSERT(DO_INDEX(ip));
350
351 if (jfs_ip->next_index < 2) {
352 jfs_warn("add_index: next_index = %d. Resetting!",
353 jfs_ip->next_index);
354 jfs_ip->next_index = 2;
355 }
356
357 index = jfs_ip->next_index++;
358
359 if (index <= MAX_INLINE_DIRTABLE_ENTRY) {
360 /*
361 * i_size reflects size of index table, or 8 bytes per entry.
362 */
363 ip->i_size = (loff_t) (index - 1) << 3;
364
365 /*
366 * dir table fits inline within inode
367 */
368 dirtab_slot = &jfs_ip->i_dirtable[index-2];
369 dirtab_slot->flag = DIR_INDEX_VALID;
370 dirtab_slot->slot = slot;
371 DTSaddress(dirtab_slot, bn);
372
373 set_cflag(COMMIT_Dirtable, ip);
374
375 return index;
376 }
377 if (index == (MAX_INLINE_DIRTABLE_ENTRY + 1)) {
378 struct dir_table_slot temp_table[12];
379
380 /*
381 * It's time to move the inline table to an external
382 * page and begin to build the xtree
383 */
Dave Kleikamp18190cc2005-07-26 09:29:13 -0500384 if (DQUOT_ALLOC_BLOCK(ip, sbi->nbperpage))
385 goto clean_up;
386 if (dbAlloc(ip, 0, sbi->nbperpage, &xaddr)) {
387 DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
388 goto clean_up;
389 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700390
391 /*
392 * Save the table, we're going to overwrite it with the
393 * xtree root
394 */
395 memcpy(temp_table, &jfs_ip->i_dirtable, sizeof(temp_table));
396
397 /*
398 * Initialize empty x-tree
399 */
400 xtInitRoot(tid, ip);
401
402 /*
Dave Kleikamp18190cc2005-07-26 09:29:13 -0500403 * Add the first block to the xtree
Linus Torvalds1da177e2005-04-16 15:20:36 -0700404 */
405 if (xtInsert(tid, ip, 0, 0, sbi->nbperpage, &xaddr, 0)) {
406 /* This really shouldn't fail */
407 jfs_warn("add_index: xtInsert failed!");
408 memcpy(&jfs_ip->i_dirtable, temp_table,
409 sizeof (temp_table));
Dave Kleikamp18190cc2005-07-26 09:29:13 -0500410 dbFree(ip, xaddr, sbi->nbperpage);
411 DQUOT_FREE_BLOCK(ip, sbi->nbperpage);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700412 goto clean_up;
413 }
414 ip->i_size = PSIZE;
415
Joe Perches09aaa742007-11-13 22:16:08 -0600416 mp = get_index_page(ip, 0);
417 if (!mp) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700418 jfs_err("add_index: get_metapage failed!");
419 xtTruncate(tid, ip, 0, COMMIT_PWMAP);
420 memcpy(&jfs_ip->i_dirtable, temp_table,
421 sizeof (temp_table));
422 goto clean_up;
423 }
424 tlck = txLock(tid, ip, mp, tlckDATA);
425 llck = (struct linelock *) & tlck->lock;
426 ASSERT(llck->index == 0);
427 lv = &llck->lv[0];
428
429 lv->offset = 0;
430 lv->length = 6; /* tlckDATA slot size is 16 bytes */
431 llck->index++;
432
433 memcpy(mp->data, temp_table, sizeof(temp_table));
434
435 mark_metapage_dirty(mp);
436 release_metapage(mp);
437
438 /*
439 * Logging is now directed by xtree tlocks
440 */
441 clear_cflag(COMMIT_Dirtable, ip);
442 }
443
444 offset = (index - 2) * sizeof(struct dir_table_slot);
445 page_offset = offset & (PSIZE - 1);
446 blkno = ((offset + 1) >> L2PSIZE) << sbi->l2nbperpage;
447 if (page_offset == 0) {
448 /*
449 * This will be the beginning of a new page
450 */
451 xaddr = 0;
452 if (xtInsert(tid, ip, 0, blkno, sbi->nbperpage, &xaddr, 0)) {
453 jfs_warn("add_index: xtInsert failed!");
454 goto clean_up;
455 }
456 ip->i_size += PSIZE;
457
458 if ((mp = get_index_page(ip, blkno)))
459 memset(mp->data, 0, PSIZE); /* Just looks better */
460 else
461 xtTruncate(tid, ip, offset, COMMIT_PWMAP);
462 } else
463 mp = read_index_page(ip, blkno);
464
Joe Perches09aaa742007-11-13 22:16:08 -0600465 if (!mp) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700466 jfs_err("add_index: get/read_metapage failed!");
467 goto clean_up;
468 }
469
470 lock_index(tid, ip, mp, index);
471
472 dirtab_slot =
473 (struct dir_table_slot *) ((char *) mp->data + page_offset);
474 dirtab_slot->flag = DIR_INDEX_VALID;
475 dirtab_slot->slot = slot;
476 DTSaddress(dirtab_slot, bn);
477
478 mark_metapage_dirty(mp);
479 release_metapage(mp);
480
481 return index;
482
483 clean_up:
484
485 jfs_ip->next_index--;
486
487 return 0;
488}
489
490/*
491 * free_index()
492 *
493 * Marks an entry to the directory index table as free.
494 */
495static void free_index(tid_t tid, struct inode *ip, u32 index, u32 next)
496{
497 struct dir_table_slot *dirtab_slot;
498 s64 lblock;
499 struct metapage *mp = NULL;
500
501 dirtab_slot = find_index(ip, index, &mp, &lblock);
502
Joe Perches09aaa742007-11-13 22:16:08 -0600503 if (!dirtab_slot)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700504 return;
505
506 dirtab_slot->flag = DIR_INDEX_FREE;
507 dirtab_slot->slot = dirtab_slot->addr1 = 0;
508 dirtab_slot->addr2 = cpu_to_le32(next);
509
510 if (mp) {
511 lock_index(tid, ip, mp, index);
512 mark_metapage_dirty(mp);
513 release_metapage(mp);
514 } else
515 set_cflag(COMMIT_Dirtable, ip);
516}
517
518/*
519 * modify_index()
520 *
521 * Changes an entry in the directory index table
522 */
523static void modify_index(tid_t tid, struct inode *ip, u32 index, s64 bn,
Al Viro5ba25332007-10-14 19:35:50 +0100524 int slot, struct metapage ** mp, s64 *lblock)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700525{
526 struct dir_table_slot *dirtab_slot;
527
528 dirtab_slot = find_index(ip, index, mp, lblock);
529
Joe Perches09aaa742007-11-13 22:16:08 -0600530 if (!dirtab_slot)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700531 return;
532
533 DTSaddress(dirtab_slot, bn);
534 dirtab_slot->slot = slot;
535
536 if (*mp) {
537 lock_index(tid, ip, *mp, index);
538 mark_metapage_dirty(*mp);
539 } else
540 set_cflag(COMMIT_Dirtable, ip);
541}
542
543/*
544 * read_index()
545 *
546 * reads a directory table slot
547 */
548static int read_index(struct inode *ip, u32 index,
549 struct dir_table_slot * dirtab_slot)
550{
551 s64 lblock;
552 struct metapage *mp = NULL;
553 struct dir_table_slot *slot;
554
555 slot = find_index(ip, index, &mp, &lblock);
Joe Perches09aaa742007-11-13 22:16:08 -0600556 if (!slot) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700557 return -EIO;
558 }
559
560 memcpy(dirtab_slot, slot, sizeof(struct dir_table_slot));
561
562 if (mp)
563 release_metapage(mp);
564
565 return 0;
566}
567
568/*
569 * dtSearch()
570 *
571 * function:
572 * Search for the entry with specified key
573 *
574 * parameter:
575 *
576 * return: 0 - search result on stack, leaf page pinned;
577 * errno - I/O error
578 */
579int dtSearch(struct inode *ip, struct component_name * key, ino_t * data,
580 struct btstack * btstack, int flag)
581{
582 int rc = 0;
583 int cmp = 1; /* init for empty page */
584 s64 bn;
585 struct metapage *mp;
586 dtpage_t *p;
587 s8 *stbl;
588 int base, index, lim;
589 struct btframe *btsp;
590 pxd_t *pxd;
591 int psize = 288; /* initial in-line directory */
592 ino_t inumber;
593 struct component_name ciKey;
594 struct super_block *sb = ip->i_sb;
595
Jack Stone1eb3a712007-07-31 09:36:53 -0500596 ciKey.name = kmalloc((JFS_NAME_MAX + 1) * sizeof(wchar_t), GFP_NOFS);
Joe Perches09aaa742007-11-13 22:16:08 -0600597 if (!ciKey.name) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700598 rc = -ENOMEM;
599 goto dtSearch_Exit2;
600 }
601
602
603 /* uppercase search key for c-i directory */
604 UniStrcpy(ciKey.name, key->name);
605 ciKey.namlen = key->namlen;
606
607 /* only uppercase if case-insensitive support is on */
608 if ((JFS_SBI(sb)->mntflag & JFS_OS2) == JFS_OS2) {
609 ciToUpper(&ciKey);
610 }
611 BT_CLR(btstack); /* reset stack */
612
613 /* init level count for max pages to split */
614 btstack->nsplit = 1;
615
616 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500617 * search down tree from root:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700618 *
619 * between two consecutive entries of <Ki, Pi> and <Kj, Pj> of
620 * internal page, child page Pi contains entry with k, Ki <= K < Kj.
621 *
622 * if entry with search key K is not found
623 * internal page search find the entry with largest key Ki
624 * less than K which point to the child page to search;
625 * leaf page search find the entry with smallest key Kj
626 * greater than K so that the returned index is the position of
627 * the entry to be shifted right for insertion of new entry.
628 * for empty tree, search key is greater than any key of the tree.
629 *
630 * by convention, root bn = 0.
631 */
632 for (bn = 0;;) {
633 /* get/pin the page to search */
634 DT_GETPAGE(ip, bn, mp, psize, p, rc);
635 if (rc)
636 goto dtSearch_Exit1;
637
638 /* get sorted entry table of the page */
639 stbl = DT_GETSTBL(p);
640
641 /*
642 * binary search with search key K on the current page.
643 */
644 for (base = 0, lim = p->header.nextindex; lim; lim >>= 1) {
645 index = base + (lim >> 1);
646
647 if (p->header.flag & BT_LEAF) {
648 /* uppercase leaf name to compare */
649 cmp =
650 ciCompare(&ciKey, p, stbl[index],
651 JFS_SBI(sb)->mntflag);
652 } else {
653 /* router key is in uppercase */
654
655 cmp = dtCompare(&ciKey, p, stbl[index]);
656
657
658 }
659 if (cmp == 0) {
660 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500661 * search hit
Linus Torvalds1da177e2005-04-16 15:20:36 -0700662 */
663 /* search hit - leaf page:
664 * return the entry found
665 */
666 if (p->header.flag & BT_LEAF) {
667 inumber = le32_to_cpu(
668 ((struct ldtentry *) & p->slot[stbl[index]])->inumber);
669
670 /*
671 * search for JFS_LOOKUP
672 */
673 if (flag == JFS_LOOKUP) {
674 *data = inumber;
675 rc = 0;
676 goto out;
677 }
678
679 /*
680 * search for JFS_CREATE
681 */
682 if (flag == JFS_CREATE) {
683 *data = inumber;
684 rc = -EEXIST;
685 goto out;
686 }
687
688 /*
689 * search for JFS_REMOVE or JFS_RENAME
690 */
691 if ((flag == JFS_REMOVE ||
692 flag == JFS_RENAME) &&
693 *data != inumber) {
694 rc = -ESTALE;
695 goto out;
696 }
697
698 /*
699 * JFS_REMOVE|JFS_FINDDIR|JFS_RENAME
700 */
701 /* save search result */
702 *data = inumber;
703 btsp = btstack->top;
704 btsp->bn = bn;
705 btsp->index = index;
706 btsp->mp = mp;
707
708 rc = 0;
709 goto dtSearch_Exit1;
710 }
711
712 /* search hit - internal page:
713 * descend/search its child page
714 */
715 goto getChild;
716 }
717
718 if (cmp > 0) {
719 base = index + 1;
720 --lim;
721 }
722 }
723
724 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500725 * search miss
Linus Torvalds1da177e2005-04-16 15:20:36 -0700726 *
727 * base is the smallest index with key (Kj) greater than
728 * search key (K) and may be zero or (maxindex + 1) index.
729 */
730 /*
731 * search miss - leaf page
732 *
733 * return location of entry (base) where new entry with
734 * search key K is to be inserted.
735 */
736 if (p->header.flag & BT_LEAF) {
737 /*
738 * search for JFS_LOOKUP, JFS_REMOVE, or JFS_RENAME
739 */
740 if (flag == JFS_LOOKUP || flag == JFS_REMOVE ||
741 flag == JFS_RENAME) {
742 rc = -ENOENT;
743 goto out;
744 }
745
746 /*
747 * search for JFS_CREATE|JFS_FINDDIR:
748 *
749 * save search result
750 */
751 *data = 0;
752 btsp = btstack->top;
753 btsp->bn = bn;
754 btsp->index = base;
755 btsp->mp = mp;
756
757 rc = 0;
758 goto dtSearch_Exit1;
759 }
760
761 /*
762 * search miss - internal page
763 *
764 * if base is non-zero, decrement base by one to get the parent
765 * entry of the child page to search.
766 */
767 index = base ? base - 1 : base;
768
769 /*
770 * go down to child page
771 */
772 getChild:
773 /* update max. number of pages to split */
774 if (BT_STACK_FULL(btstack)) {
Robert P. J. Day3a4fa0a2007-10-19 23:10:43 +0200775 /* Something's corrupted, mark filesystem dirty so
Linus Torvalds1da177e2005-04-16 15:20:36 -0700776 * chkdsk will fix it.
777 */
778 jfs_error(sb, "stack overrun in dtSearch!");
779 BT_STACK_DUMP(btstack);
780 rc = -EIO;
781 goto out;
782 }
783 btstack->nsplit++;
784
785 /* push (bn, index) of the parent page/entry */
786 BT_PUSH(btstack, bn, index);
787
788 /* get the child page block number */
789 pxd = (pxd_t *) & p->slot[stbl[index]];
790 bn = addressPXD(pxd);
791 psize = lengthPXD(pxd) << JFS_SBI(ip->i_sb)->l2bsize;
792
793 /* unpin the parent page */
794 DT_PUTPAGE(mp);
795 }
796
797 out:
798 DT_PUTPAGE(mp);
799
800 dtSearch_Exit1:
801
802 kfree(ciKey.name);
803
804 dtSearch_Exit2:
805
806 return rc;
807}
808
809
810/*
811 * dtInsert()
812 *
813 * function: insert an entry to directory tree
814 *
815 * parameter:
816 *
817 * return: 0 - success;
818 * errno - failure;
819 */
820int dtInsert(tid_t tid, struct inode *ip,
821 struct component_name * name, ino_t * fsn, struct btstack * btstack)
822{
823 int rc = 0;
824 struct metapage *mp; /* meta-page buffer */
825 dtpage_t *p; /* base B+-tree index page */
826 s64 bn;
827 int index;
828 struct dtsplit split; /* split information */
829 ddata_t data;
830 struct dt_lock *dtlck;
831 int n;
832 struct tlock *tlck;
833 struct lv *lv;
834
835 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500836 * retrieve search result
Linus Torvalds1da177e2005-04-16 15:20:36 -0700837 *
838 * dtSearch() returns (leaf page pinned, index at which to insert).
839 * n.b. dtSearch() may return index of (maxindex + 1) of
840 * the full page.
841 */
842 DT_GETSEARCH(ip, btstack->top, bn, mp, p, index);
843
844 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500845 * insert entry for new key
Linus Torvalds1da177e2005-04-16 15:20:36 -0700846 */
847 if (DO_INDEX(ip)) {
848 if (JFS_IP(ip)->next_index == DIREND) {
849 DT_PUTPAGE(mp);
850 return -EMLINK;
851 }
852 n = NDTLEAF(name->namlen);
853 data.leaf.tid = tid;
854 data.leaf.ip = ip;
855 } else {
856 n = NDTLEAF_LEGACY(name->namlen);
857 data.leaf.ip = NULL; /* signifies legacy directory format */
858 }
859 data.leaf.ino = *fsn;
860
861 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500862 * leaf page does not have enough room for new entry:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700863 *
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500864 * extend/split the leaf page;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700865 *
866 * dtSplitUp() will insert the entry and unpin the leaf page.
867 */
868 if (n > p->header.freecnt) {
869 split.mp = mp;
870 split.index = index;
871 split.nslot = n;
872 split.key = name;
873 split.data = &data;
874 rc = dtSplitUp(tid, ip, &split, btstack);
875 return rc;
876 }
877
878 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500879 * leaf page does have enough room for new entry:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700880 *
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500881 * insert the new data entry into the leaf page;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700882 */
883 BT_MARK_DIRTY(mp, ip);
884 /*
885 * acquire a transaction lock on the leaf page
886 */
887 tlck = txLock(tid, ip, mp, tlckDTREE | tlckENTRY);
888 dtlck = (struct dt_lock *) & tlck->lock;
889 ASSERT(dtlck->index == 0);
890 lv = & dtlck->lv[0];
891
892 /* linelock header */
893 lv->offset = 0;
894 lv->length = 1;
895 dtlck->index++;
896
897 dtInsertEntry(p, index, name, &data, &dtlck);
898
899 /* linelock stbl of non-root leaf page */
900 if (!(p->header.flag & BT_ROOT)) {
901 if (dtlck->index >= dtlck->maxcnt)
902 dtlck = (struct dt_lock *) txLinelock(dtlck);
903 lv = & dtlck->lv[dtlck->index];
904 n = index >> L2DTSLOTSIZE;
905 lv->offset = p->header.stblindex + n;
906 lv->length =
907 ((p->header.nextindex - 1) >> L2DTSLOTSIZE) - n + 1;
908 dtlck->index++;
909 }
910
911 /* unpin the leaf page */
912 DT_PUTPAGE(mp);
913
914 return 0;
915}
916
917
918/*
919 * dtSplitUp()
920 *
921 * function: propagate insertion bottom up;
922 *
923 * parameter:
924 *
925 * return: 0 - success;
926 * errno - failure;
Dave Kleikamp63f83c92006-10-02 09:55:27 -0500927 * leaf page unpinned;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700928 */
929static int dtSplitUp(tid_t tid,
930 struct inode *ip, struct dtsplit * split, struct btstack * btstack)
931{
932 struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
933 int rc = 0;
934 struct metapage *smp;
935 dtpage_t *sp; /* split page */
936 struct metapage *rmp;
937 dtpage_t *rp; /* new right page split from sp */
938 pxd_t rpxd; /* new right page extent descriptor */
939 struct metapage *lmp;
940 dtpage_t *lp; /* left child page */
941 int skip; /* index of entry of insertion */
942 struct btframe *parent; /* parent page entry on traverse stack */
943 s64 xaddr, nxaddr;
944 int xlen, xsize;
945 struct pxdlist pxdlist;
946 pxd_t *pxd;
947 struct component_name key = { 0, NULL };
948 ddata_t *data = split->data;
949 int n;
950 struct dt_lock *dtlck;
951 struct tlock *tlck;
952 struct lv *lv;
953 int quota_allocation = 0;
954
955 /* get split page */
956 smp = split->mp;
957 sp = DT_PAGE(ip, smp);
958
Jack Stone1eb3a712007-07-31 09:36:53 -0500959 key.name = kmalloc((JFS_NAME_MAX + 2) * sizeof(wchar_t), GFP_NOFS);
Joe Perches09aaa742007-11-13 22:16:08 -0600960 if (!key.name) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700961 DT_PUTPAGE(smp);
962 rc = -ENOMEM;
963 goto dtSplitUp_Exit;
964 }
965
966 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500967 * split leaf page
Linus Torvalds1da177e2005-04-16 15:20:36 -0700968 *
969 * The split routines insert the new entry, and
970 * acquire txLock as appropriate.
971 */
972 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -0500973 * split root leaf page:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700974 */
975 if (sp->header.flag & BT_ROOT) {
976 /*
977 * allocate a single extent child page
978 */
979 xlen = 1;
980 n = sbi->bsize >> L2DTSLOTSIZE;
981 n -= (n + 31) >> L2DTSLOTSIZE; /* stbl size */
982 n -= DTROOTMAXSLOT - sp->header.freecnt; /* header + entries */
983 if (n <= split->nslot)
984 xlen++;
985 if ((rc = dbAlloc(ip, 0, (s64) xlen, &xaddr))) {
986 DT_PUTPAGE(smp);
987 goto freeKeyName;
988 }
989
990 pxdlist.maxnpxd = 1;
991 pxdlist.npxd = 0;
992 pxd = &pxdlist.pxd[0];
993 PXDaddress(pxd, xaddr);
994 PXDlength(pxd, xlen);
995 split->pxdlist = &pxdlist;
996 rc = dtSplitRoot(tid, ip, split, &rmp);
997
998 if (rc)
999 dbFree(ip, xaddr, xlen);
1000 else
1001 DT_PUTPAGE(rmp);
1002
1003 DT_PUTPAGE(smp);
1004
Dave Kleikampdd8a3062005-11-10 07:50:03 -06001005 if (!DO_INDEX(ip))
1006 ip->i_size = xlen << sbi->l2bsize;
1007
Linus Torvalds1da177e2005-04-16 15:20:36 -07001008 goto freeKeyName;
1009 }
1010
1011 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001012 * extend first leaf page
Linus Torvalds1da177e2005-04-16 15:20:36 -07001013 *
1014 * extend the 1st extent if less than buffer page size
1015 * (dtExtendPage() reurns leaf page unpinned)
1016 */
1017 pxd = &sp->header.self;
1018 xlen = lengthPXD(pxd);
1019 xsize = xlen << sbi->l2bsize;
1020 if (xsize < PSIZE) {
1021 xaddr = addressPXD(pxd);
1022 n = xsize >> L2DTSLOTSIZE;
1023 n -= (n + 31) >> L2DTSLOTSIZE; /* stbl size */
1024 if ((n + sp->header.freecnt) <= split->nslot)
1025 n = xlen + (xlen << 1);
1026 else
1027 n = xlen;
1028
1029 /* Allocate blocks to quota. */
1030 if (DQUOT_ALLOC_BLOCK(ip, n)) {
1031 rc = -EDQUOT;
1032 goto extendOut;
1033 }
1034 quota_allocation += n;
1035
1036 if ((rc = dbReAlloc(sbi->ipbmap, xaddr, (s64) xlen,
1037 (s64) n, &nxaddr)))
1038 goto extendOut;
1039
1040 pxdlist.maxnpxd = 1;
1041 pxdlist.npxd = 0;
1042 pxd = &pxdlist.pxd[0];
1043 PXDaddress(pxd, nxaddr)
1044 PXDlength(pxd, xlen + n);
1045 split->pxdlist = &pxdlist;
1046 if ((rc = dtExtendPage(tid, ip, split, btstack))) {
1047 nxaddr = addressPXD(pxd);
1048 if (xaddr != nxaddr) {
1049 /* free relocated extent */
1050 xlen = lengthPXD(pxd);
1051 dbFree(ip, nxaddr, (s64) xlen);
1052 } else {
1053 /* free extended delta */
1054 xlen = lengthPXD(pxd) - n;
1055 xaddr = addressPXD(pxd) + xlen;
1056 dbFree(ip, xaddr, (s64) n);
1057 }
Dave Kleikampdd8a3062005-11-10 07:50:03 -06001058 } else if (!DO_INDEX(ip))
1059 ip->i_size = lengthPXD(pxd) << sbi->l2bsize;
1060
Linus Torvalds1da177e2005-04-16 15:20:36 -07001061
1062 extendOut:
1063 DT_PUTPAGE(smp);
1064 goto freeKeyName;
1065 }
1066
1067 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001068 * split leaf page <sp> into <sp> and a new right page <rp>.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001069 *
1070 * return <rp> pinned and its extent descriptor <rpxd>
1071 */
1072 /*
1073 * allocate new directory page extent and
1074 * new index page(s) to cover page split(s)
1075 *
1076 * allocation hint: ?
1077 */
1078 n = btstack->nsplit;
1079 pxdlist.maxnpxd = pxdlist.npxd = 0;
1080 xlen = sbi->nbperpage;
1081 for (pxd = pxdlist.pxd; n > 0; n--, pxd++) {
1082 if ((rc = dbAlloc(ip, 0, (s64) xlen, &xaddr)) == 0) {
1083 PXDaddress(pxd, xaddr);
1084 PXDlength(pxd, xlen);
1085 pxdlist.maxnpxd++;
1086 continue;
1087 }
1088
1089 DT_PUTPAGE(smp);
1090
1091 /* undo allocation */
1092 goto splitOut;
1093 }
1094
1095 split->pxdlist = &pxdlist;
1096 if ((rc = dtSplitPage(tid, ip, split, &rmp, &rp, &rpxd))) {
1097 DT_PUTPAGE(smp);
1098
1099 /* undo allocation */
1100 goto splitOut;
1101 }
1102
Dave Kleikampdd8a3062005-11-10 07:50:03 -06001103 if (!DO_INDEX(ip))
1104 ip->i_size += PSIZE;
1105
Linus Torvalds1da177e2005-04-16 15:20:36 -07001106 /*
1107 * propagate up the router entry for the leaf page just split
1108 *
1109 * insert a router entry for the new page into the parent page,
1110 * propagate the insert/split up the tree by walking back the stack
1111 * of (bn of parent page, index of child page entry in parent page)
1112 * that were traversed during the search for the page that split.
1113 *
1114 * the propagation of insert/split up the tree stops if the root
1115 * splits or the page inserted into doesn't have to split to hold
1116 * the new entry.
1117 *
1118 * the parent entry for the split page remains the same, and
1119 * a new entry is inserted at its right with the first key and
1120 * block number of the new right page.
1121 *
1122 * There are a maximum of 4 pages pinned at any time:
1123 * two children, left parent and right parent (when the parent splits).
1124 * keep the child pages pinned while working on the parent.
1125 * make sure that all pins are released at exit.
1126 */
1127 while ((parent = BT_POP(btstack)) != NULL) {
1128 /* parent page specified by stack frame <parent> */
1129
1130 /* keep current child pages (<lp>, <rp>) pinned */
1131 lmp = smp;
1132 lp = sp;
1133
1134 /*
1135 * insert router entry in parent for new right child page <rp>
1136 */
1137 /* get the parent page <sp> */
1138 DT_GETPAGE(ip, parent->bn, smp, PSIZE, sp, rc);
1139 if (rc) {
1140 DT_PUTPAGE(lmp);
1141 DT_PUTPAGE(rmp);
1142 goto splitOut;
1143 }
1144
1145 /*
1146 * The new key entry goes ONE AFTER the index of parent entry,
1147 * because the split was to the right.
1148 */
1149 skip = parent->index + 1;
1150
1151 /*
1152 * compute the key for the router entry
1153 *
1154 * key suffix compression:
1155 * for internal pages that have leaf pages as children,
1156 * retain only what's needed to distinguish between
1157 * the new entry and the entry on the page to its left.
1158 * If the keys compare equal, retain the entire key.
1159 *
1160 * note that compression is performed only at computing
1161 * router key at the lowest internal level.
1162 * further compression of the key between pairs of higher
1163 * level internal pages loses too much information and
1164 * the search may fail.
1165 * (e.g., two adjacent leaf pages of {a, ..., x} {xx, ...,}
1166 * results in two adjacent parent entries (a)(xx).
1167 * if split occurs between these two entries, and
1168 * if compression is applied, the router key of parent entry
1169 * of right page (x) will divert search for x into right
1170 * subtree and miss x in the left subtree.)
1171 *
1172 * the entire key must be retained for the next-to-leftmost
1173 * internal key at any level of the tree, or search may fail
1174 * (e.g., ?)
1175 */
1176 switch (rp->header.flag & BT_TYPE) {
1177 case BT_LEAF:
1178 /*
1179 * compute the length of prefix for suffix compression
1180 * between last entry of left page and first entry
1181 * of right page
1182 */
1183 if ((sp->header.flag & BT_ROOT && skip > 1) ||
1184 sp->header.prev != 0 || skip > 1) {
1185 /* compute uppercase router prefix key */
1186 rc = ciGetLeafPrefixKey(lp,
1187 lp->header.nextindex-1,
1188 rp, 0, &key,
1189 sbi->mntflag);
1190 if (rc) {
1191 DT_PUTPAGE(lmp);
1192 DT_PUTPAGE(rmp);
1193 DT_PUTPAGE(smp);
1194 goto splitOut;
1195 }
1196 } else {
1197 /* next to leftmost entry of
1198 lowest internal level */
1199
1200 /* compute uppercase router key */
1201 dtGetKey(rp, 0, &key, sbi->mntflag);
1202 key.name[key.namlen] = 0;
1203
1204 if ((sbi->mntflag & JFS_OS2) == JFS_OS2)
1205 ciToUpper(&key);
1206 }
1207
1208 n = NDTINTERNAL(key.namlen);
1209 break;
1210
1211 case BT_INTERNAL:
1212 dtGetKey(rp, 0, &key, sbi->mntflag);
1213 n = NDTINTERNAL(key.namlen);
1214 break;
1215
1216 default:
1217 jfs_err("dtSplitUp(): UFO!");
1218 break;
1219 }
1220
1221 /* unpin left child page */
1222 DT_PUTPAGE(lmp);
1223
1224 /*
1225 * compute the data for the router entry
1226 */
1227 data->xd = rpxd; /* child page xd */
1228
1229 /*
1230 * parent page is full - split the parent page
1231 */
1232 if (n > sp->header.freecnt) {
1233 /* init for parent page split */
1234 split->mp = smp;
1235 split->index = skip; /* index at insert */
1236 split->nslot = n;
1237 split->key = &key;
1238 /* split->data = data; */
1239
1240 /* unpin right child page */
1241 DT_PUTPAGE(rmp);
1242
1243 /* The split routines insert the new entry,
1244 * acquire txLock as appropriate.
1245 * return <rp> pinned and its block number <rbn>.
1246 */
1247 rc = (sp->header.flag & BT_ROOT) ?
1248 dtSplitRoot(tid, ip, split, &rmp) :
1249 dtSplitPage(tid, ip, split, &rmp, &rp, &rpxd);
1250 if (rc) {
1251 DT_PUTPAGE(smp);
1252 goto splitOut;
1253 }
1254
1255 /* smp and rmp are pinned */
1256 }
1257 /*
1258 * parent page is not full - insert router entry in parent page
1259 */
1260 else {
1261 BT_MARK_DIRTY(smp, ip);
1262 /*
1263 * acquire a transaction lock on the parent page
1264 */
1265 tlck = txLock(tid, ip, smp, tlckDTREE | tlckENTRY);
1266 dtlck = (struct dt_lock *) & tlck->lock;
1267 ASSERT(dtlck->index == 0);
1268 lv = & dtlck->lv[0];
1269
1270 /* linelock header */
1271 lv->offset = 0;
1272 lv->length = 1;
1273 dtlck->index++;
1274
1275 /* linelock stbl of non-root parent page */
1276 if (!(sp->header.flag & BT_ROOT)) {
1277 lv++;
1278 n = skip >> L2DTSLOTSIZE;
1279 lv->offset = sp->header.stblindex + n;
1280 lv->length =
1281 ((sp->header.nextindex -
1282 1) >> L2DTSLOTSIZE) - n + 1;
1283 dtlck->index++;
1284 }
1285
1286 dtInsertEntry(sp, skip, &key, data, &dtlck);
1287
1288 /* exit propagate up */
1289 break;
1290 }
1291 }
1292
1293 /* unpin current split and its right page */
1294 DT_PUTPAGE(smp);
1295 DT_PUTPAGE(rmp);
1296
1297 /*
1298 * free remaining extents allocated for split
1299 */
1300 splitOut:
1301 n = pxdlist.npxd;
1302 pxd = &pxdlist.pxd[n];
1303 for (; n < pxdlist.maxnpxd; n++, pxd++)
1304 dbFree(ip, addressPXD(pxd), (s64) lengthPXD(pxd));
1305
1306 freeKeyName:
1307 kfree(key.name);
1308
1309 /* Rollback quota allocation */
1310 if (rc && quota_allocation)
1311 DQUOT_FREE_BLOCK(ip, quota_allocation);
1312
1313 dtSplitUp_Exit:
1314
1315 return rc;
1316}
1317
1318
1319/*
1320 * dtSplitPage()
1321 *
1322 * function: Split a non-root page of a btree.
1323 *
1324 * parameter:
1325 *
1326 * return: 0 - success;
1327 * errno - failure;
1328 * return split and new page pinned;
1329 */
1330static int dtSplitPage(tid_t tid, struct inode *ip, struct dtsplit * split,
1331 struct metapage ** rmpp, dtpage_t ** rpp, pxd_t * rpxdp)
1332{
1333 int rc = 0;
1334 struct metapage *smp;
1335 dtpage_t *sp;
1336 struct metapage *rmp;
1337 dtpage_t *rp; /* new right page allocated */
1338 s64 rbn; /* new right page block number */
1339 struct metapage *mp;
1340 dtpage_t *p;
1341 s64 nextbn;
1342 struct pxdlist *pxdlist;
1343 pxd_t *pxd;
1344 int skip, nextindex, half, left, nxt, off, si;
1345 struct ldtentry *ldtentry;
1346 struct idtentry *idtentry;
1347 u8 *stbl;
1348 struct dtslot *f;
1349 int fsi, stblsize;
1350 int n;
1351 struct dt_lock *sdtlck, *rdtlck;
1352 struct tlock *tlck;
1353 struct dt_lock *dtlck;
1354 struct lv *slv, *rlv, *lv;
1355
1356 /* get split page */
1357 smp = split->mp;
1358 sp = DT_PAGE(ip, smp);
1359
1360 /*
1361 * allocate the new right page for the split
1362 */
1363 pxdlist = split->pxdlist;
1364 pxd = &pxdlist->pxd[pxdlist->npxd];
1365 pxdlist->npxd++;
1366 rbn = addressPXD(pxd);
1367 rmp = get_metapage(ip, rbn, PSIZE, 1);
1368 if (rmp == NULL)
1369 return -EIO;
1370
1371 /* Allocate blocks to quota. */
1372 if (DQUOT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
1373 release_metapage(rmp);
1374 return -EDQUOT;
1375 }
1376
1377 jfs_info("dtSplitPage: ip:0x%p smp:0x%p rmp:0x%p", ip, smp, rmp);
1378
1379 BT_MARK_DIRTY(rmp, ip);
1380 /*
1381 * acquire a transaction lock on the new right page
1382 */
1383 tlck = txLock(tid, ip, rmp, tlckDTREE | tlckNEW);
1384 rdtlck = (struct dt_lock *) & tlck->lock;
1385
1386 rp = (dtpage_t *) rmp->data;
1387 *rpp = rp;
1388 rp->header.self = *pxd;
1389
1390 BT_MARK_DIRTY(smp, ip);
1391 /*
1392 * acquire a transaction lock on the split page
1393 *
1394 * action:
1395 */
1396 tlck = txLock(tid, ip, smp, tlckDTREE | tlckENTRY);
1397 sdtlck = (struct dt_lock *) & tlck->lock;
1398
1399 /* linelock header of split page */
1400 ASSERT(sdtlck->index == 0);
1401 slv = & sdtlck->lv[0];
1402 slv->offset = 0;
1403 slv->length = 1;
1404 sdtlck->index++;
1405
1406 /*
1407 * initialize/update sibling pointers between sp and rp
1408 */
1409 nextbn = le64_to_cpu(sp->header.next);
1410 rp->header.next = cpu_to_le64(nextbn);
1411 rp->header.prev = cpu_to_le64(addressPXD(&sp->header.self));
1412 sp->header.next = cpu_to_le64(rbn);
1413
1414 /*
1415 * initialize new right page
1416 */
1417 rp->header.flag = sp->header.flag;
1418
1419 /* compute sorted entry table at start of extent data area */
1420 rp->header.nextindex = 0;
1421 rp->header.stblindex = 1;
1422
1423 n = PSIZE >> L2DTSLOTSIZE;
1424 rp->header.maxslot = n;
1425 stblsize = (n + 31) >> L2DTSLOTSIZE; /* in unit of slot */
1426
1427 /* init freelist */
1428 fsi = rp->header.stblindex + stblsize;
1429 rp->header.freelist = fsi;
1430 rp->header.freecnt = rp->header.maxslot - fsi;
1431
1432 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001433 * sequential append at tail: append without split
Linus Torvalds1da177e2005-04-16 15:20:36 -07001434 *
1435 * If splitting the last page on a level because of appending
1436 * a entry to it (skip is maxentry), it's likely that the access is
1437 * sequential. Adding an empty page on the side of the level is less
1438 * work and can push the fill factor much higher than normal.
1439 * If we're wrong it's no big deal, we'll just do the split the right
1440 * way next time.
1441 * (It may look like it's equally easy to do a similar hack for
1442 * reverse sorted data, that is, split the tree left,
1443 * but it's not. Be my guest.)
1444 */
1445 if (nextbn == 0 && split->index == sp->header.nextindex) {
1446 /* linelock header + stbl (first slot) of new page */
1447 rlv = & rdtlck->lv[rdtlck->index];
1448 rlv->offset = 0;
1449 rlv->length = 2;
1450 rdtlck->index++;
1451
1452 /*
1453 * initialize freelist of new right page
1454 */
1455 f = &rp->slot[fsi];
1456 for (fsi++; fsi < rp->header.maxslot; f++, fsi++)
1457 f->next = fsi;
1458 f->next = -1;
1459
1460 /* insert entry at the first entry of the new right page */
1461 dtInsertEntry(rp, 0, split->key, split->data, &rdtlck);
1462
1463 goto out;
1464 }
1465
1466 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001467 * non-sequential insert (at possibly middle page)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001468 */
1469
1470 /*
1471 * update prev pointer of previous right sibling page;
1472 */
1473 if (nextbn != 0) {
1474 DT_GETPAGE(ip, nextbn, mp, PSIZE, p, rc);
1475 if (rc) {
1476 discard_metapage(rmp);
1477 return rc;
1478 }
1479
1480 BT_MARK_DIRTY(mp, ip);
1481 /*
1482 * acquire a transaction lock on the next page
1483 */
1484 tlck = txLock(tid, ip, mp, tlckDTREE | tlckRELINK);
1485 jfs_info("dtSplitPage: tlck = 0x%p, ip = 0x%p, mp=0x%p",
1486 tlck, ip, mp);
1487 dtlck = (struct dt_lock *) & tlck->lock;
1488
1489 /* linelock header of previous right sibling page */
1490 lv = & dtlck->lv[dtlck->index];
1491 lv->offset = 0;
1492 lv->length = 1;
1493 dtlck->index++;
1494
1495 p->header.prev = cpu_to_le64(rbn);
1496
1497 DT_PUTPAGE(mp);
1498 }
1499
1500 /*
1501 * split the data between the split and right pages.
1502 */
1503 skip = split->index;
1504 half = (PSIZE >> L2DTSLOTSIZE) >> 1; /* swag */
1505 left = 0;
1506
1507 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001508 * compute fill factor for split pages
Linus Torvalds1da177e2005-04-16 15:20:36 -07001509 *
1510 * <nxt> traces the next entry to move to rp
1511 * <off> traces the next entry to stay in sp
1512 */
1513 stbl = (u8 *) & sp->slot[sp->header.stblindex];
1514 nextindex = sp->header.nextindex;
1515 for (nxt = off = 0; nxt < nextindex; ++off) {
1516 if (off == skip)
1517 /* check for fill factor with new entry size */
1518 n = split->nslot;
1519 else {
1520 si = stbl[nxt];
1521 switch (sp->header.flag & BT_TYPE) {
1522 case BT_LEAF:
1523 ldtentry = (struct ldtentry *) & sp->slot[si];
1524 if (DO_INDEX(ip))
1525 n = NDTLEAF(ldtentry->namlen);
1526 else
1527 n = NDTLEAF_LEGACY(ldtentry->
1528 namlen);
1529 break;
1530
1531 case BT_INTERNAL:
1532 idtentry = (struct idtentry *) & sp->slot[si];
1533 n = NDTINTERNAL(idtentry->namlen);
1534 break;
1535
1536 default:
1537 break;
1538 }
1539
1540 ++nxt; /* advance to next entry to move in sp */
1541 }
1542
1543 left += n;
1544 if (left >= half)
1545 break;
1546 }
1547
1548 /* <nxt> poins to the 1st entry to move */
1549
1550 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001551 * move entries to right page
Linus Torvalds1da177e2005-04-16 15:20:36 -07001552 *
1553 * dtMoveEntry() initializes rp and reserves entry for insertion
1554 *
1555 * split page moved out entries are linelocked;
1556 * new/right page moved in entries are linelocked;
1557 */
1558 /* linelock header + stbl of new right page */
1559 rlv = & rdtlck->lv[rdtlck->index];
1560 rlv->offset = 0;
1561 rlv->length = 5;
1562 rdtlck->index++;
1563
1564 dtMoveEntry(sp, nxt, rp, &sdtlck, &rdtlck, DO_INDEX(ip));
1565
1566 sp->header.nextindex = nxt;
1567
1568 /*
1569 * finalize freelist of new right page
1570 */
1571 fsi = rp->header.freelist;
1572 f = &rp->slot[fsi];
1573 for (fsi++; fsi < rp->header.maxslot; f++, fsi++)
1574 f->next = fsi;
1575 f->next = -1;
1576
1577 /*
1578 * Update directory index table for entries now in right page
1579 */
1580 if ((rp->header.flag & BT_LEAF) && DO_INDEX(ip)) {
1581 s64 lblock;
1582
1583 mp = NULL;
1584 stbl = DT_GETSTBL(rp);
1585 for (n = 0; n < rp->header.nextindex; n++) {
1586 ldtentry = (struct ldtentry *) & rp->slot[stbl[n]];
1587 modify_index(tid, ip, le32_to_cpu(ldtentry->index),
1588 rbn, n, &mp, &lblock);
1589 }
1590 if (mp)
1591 release_metapage(mp);
1592 }
1593
1594 /*
1595 * the skipped index was on the left page,
1596 */
1597 if (skip <= off) {
1598 /* insert the new entry in the split page */
1599 dtInsertEntry(sp, skip, split->key, split->data, &sdtlck);
1600
1601 /* linelock stbl of split page */
1602 if (sdtlck->index >= sdtlck->maxcnt)
1603 sdtlck = (struct dt_lock *) txLinelock(sdtlck);
1604 slv = & sdtlck->lv[sdtlck->index];
1605 n = skip >> L2DTSLOTSIZE;
1606 slv->offset = sp->header.stblindex + n;
1607 slv->length =
1608 ((sp->header.nextindex - 1) >> L2DTSLOTSIZE) - n + 1;
1609 sdtlck->index++;
1610 }
1611 /*
1612 * the skipped index was on the right page,
1613 */
1614 else {
1615 /* adjust the skip index to reflect the new position */
1616 skip -= nxt;
1617
1618 /* insert the new entry in the right page */
1619 dtInsertEntry(rp, skip, split->key, split->data, &rdtlck);
1620 }
1621
1622 out:
1623 *rmpp = rmp;
1624 *rpxdp = *pxd;
1625
1626 return rc;
1627}
1628
1629
1630/*
1631 * dtExtendPage()
1632 *
1633 * function: extend 1st/only directory leaf page
1634 *
1635 * parameter:
1636 *
1637 * return: 0 - success;
1638 * errno - failure;
1639 * return extended page pinned;
1640 */
1641static int dtExtendPage(tid_t tid,
1642 struct inode *ip, struct dtsplit * split, struct btstack * btstack)
1643{
1644 struct super_block *sb = ip->i_sb;
1645 int rc;
1646 struct metapage *smp, *pmp, *mp;
1647 dtpage_t *sp, *pp;
1648 struct pxdlist *pxdlist;
1649 pxd_t *pxd, *tpxd;
1650 int xlen, xsize;
1651 int newstblindex, newstblsize;
1652 int oldstblindex, oldstblsize;
1653 int fsi, last;
1654 struct dtslot *f;
1655 struct btframe *parent;
1656 int n;
1657 struct dt_lock *dtlck;
1658 s64 xaddr, txaddr;
1659 struct tlock *tlck;
1660 struct pxd_lock *pxdlock;
1661 struct lv *lv;
1662 uint type;
1663 struct ldtentry *ldtentry;
1664 u8 *stbl;
1665
1666 /* get page to extend */
1667 smp = split->mp;
1668 sp = DT_PAGE(ip, smp);
1669
1670 /* get parent/root page */
1671 parent = BT_POP(btstack);
1672 DT_GETPAGE(ip, parent->bn, pmp, PSIZE, pp, rc);
1673 if (rc)
1674 return (rc);
1675
1676 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001677 * extend the extent
Linus Torvalds1da177e2005-04-16 15:20:36 -07001678 */
1679 pxdlist = split->pxdlist;
1680 pxd = &pxdlist->pxd[pxdlist->npxd];
1681 pxdlist->npxd++;
1682
1683 xaddr = addressPXD(pxd);
1684 tpxd = &sp->header.self;
1685 txaddr = addressPXD(tpxd);
1686 /* in-place extension */
1687 if (xaddr == txaddr) {
1688 type = tlckEXTEND;
1689 }
1690 /* relocation */
1691 else {
1692 type = tlckNEW;
1693
1694 /* save moved extent descriptor for later free */
1695 tlck = txMaplock(tid, ip, tlckDTREE | tlckRELOCATE);
1696 pxdlock = (struct pxd_lock *) & tlck->lock;
1697 pxdlock->flag = mlckFREEPXD;
1698 pxdlock->pxd = sp->header.self;
1699 pxdlock->index = 1;
1700
1701 /*
1702 * Update directory index table to reflect new page address
1703 */
1704 if (DO_INDEX(ip)) {
1705 s64 lblock;
1706
1707 mp = NULL;
1708 stbl = DT_GETSTBL(sp);
1709 for (n = 0; n < sp->header.nextindex; n++) {
1710 ldtentry =
1711 (struct ldtentry *) & sp->slot[stbl[n]];
1712 modify_index(tid, ip,
1713 le32_to_cpu(ldtentry->index),
1714 xaddr, n, &mp, &lblock);
1715 }
1716 if (mp)
1717 release_metapage(mp);
1718 }
1719 }
1720
1721 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001722 * extend the page
Linus Torvalds1da177e2005-04-16 15:20:36 -07001723 */
1724 sp->header.self = *pxd;
1725
1726 jfs_info("dtExtendPage: ip:0x%p smp:0x%p sp:0x%p", ip, smp, sp);
1727
1728 BT_MARK_DIRTY(smp, ip);
1729 /*
1730 * acquire a transaction lock on the extended/leaf page
1731 */
1732 tlck = txLock(tid, ip, smp, tlckDTREE | type);
1733 dtlck = (struct dt_lock *) & tlck->lock;
1734 lv = & dtlck->lv[0];
1735
1736 /* update buffer extent descriptor of extended page */
1737 xlen = lengthPXD(pxd);
1738 xsize = xlen << JFS_SBI(sb)->l2bsize;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001739
1740 /*
1741 * copy old stbl to new stbl at start of extended area
1742 */
1743 oldstblindex = sp->header.stblindex;
1744 oldstblsize = (sp->header.maxslot + 31) >> L2DTSLOTSIZE;
1745 newstblindex = sp->header.maxslot;
1746 n = xsize >> L2DTSLOTSIZE;
1747 newstblsize = (n + 31) >> L2DTSLOTSIZE;
1748 memcpy(&sp->slot[newstblindex], &sp->slot[oldstblindex],
1749 sp->header.nextindex);
1750
1751 /*
1752 * in-line extension: linelock old area of extended page
1753 */
1754 if (type == tlckEXTEND) {
1755 /* linelock header */
1756 lv->offset = 0;
1757 lv->length = 1;
1758 dtlck->index++;
1759 lv++;
1760
1761 /* linelock new stbl of extended page */
1762 lv->offset = newstblindex;
1763 lv->length = newstblsize;
1764 }
1765 /*
1766 * relocation: linelock whole relocated area
1767 */
1768 else {
1769 lv->offset = 0;
1770 lv->length = sp->header.maxslot + newstblsize;
1771 }
1772
1773 dtlck->index++;
1774
1775 sp->header.maxslot = n;
1776 sp->header.stblindex = newstblindex;
1777 /* sp->header.nextindex remains the same */
1778
1779 /*
1780 * add old stbl region at head of freelist
1781 */
1782 fsi = oldstblindex;
1783 f = &sp->slot[fsi];
1784 last = sp->header.freelist;
1785 for (n = 0; n < oldstblsize; n++, fsi++, f++) {
1786 f->next = last;
1787 last = fsi;
1788 }
1789 sp->header.freelist = last;
1790 sp->header.freecnt += oldstblsize;
1791
1792 /*
1793 * append free region of newly extended area at tail of freelist
1794 */
1795 /* init free region of newly extended area */
1796 fsi = n = newstblindex + newstblsize;
1797 f = &sp->slot[fsi];
1798 for (fsi++; fsi < sp->header.maxslot; f++, fsi++)
1799 f->next = fsi;
1800 f->next = -1;
1801
1802 /* append new free region at tail of old freelist */
1803 fsi = sp->header.freelist;
1804 if (fsi == -1)
1805 sp->header.freelist = n;
1806 else {
1807 do {
1808 f = &sp->slot[fsi];
1809 fsi = f->next;
1810 } while (fsi != -1);
1811
1812 f->next = n;
1813 }
1814
1815 sp->header.freecnt += sp->header.maxslot - n;
1816
1817 /*
1818 * insert the new entry
1819 */
1820 dtInsertEntry(sp, split->index, split->key, split->data, &dtlck);
1821
1822 BT_MARK_DIRTY(pmp, ip);
1823 /*
1824 * linelock any freeslots residing in old extent
1825 */
1826 if (type == tlckEXTEND) {
1827 n = sp->header.maxslot >> 2;
1828 if (sp->header.freelist < n)
1829 dtLinelockFreelist(sp, n, &dtlck);
1830 }
1831
1832 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001833 * update parent entry on the parent/root page
Linus Torvalds1da177e2005-04-16 15:20:36 -07001834 */
1835 /*
1836 * acquire a transaction lock on the parent/root page
1837 */
1838 tlck = txLock(tid, ip, pmp, tlckDTREE | tlckENTRY);
1839 dtlck = (struct dt_lock *) & tlck->lock;
1840 lv = & dtlck->lv[dtlck->index];
1841
1842 /* linelock parent entry - 1st slot */
1843 lv->offset = 1;
1844 lv->length = 1;
1845 dtlck->index++;
1846
1847 /* update the parent pxd for page extension */
1848 tpxd = (pxd_t *) & pp->slot[1];
1849 *tpxd = *pxd;
1850
1851 DT_PUTPAGE(pmp);
1852 return 0;
1853}
1854
1855
1856/*
1857 * dtSplitRoot()
1858 *
1859 * function:
1860 * split the full root page into
1861 * original/root/split page and new right page
1862 * i.e., root remains fixed in tree anchor (inode) and
1863 * the root is copied to a single new right child page
1864 * since root page << non-root page, and
1865 * the split root page contains a single entry for the
1866 * new right child page.
1867 *
1868 * parameter:
1869 *
1870 * return: 0 - success;
1871 * errno - failure;
1872 * return new page pinned;
1873 */
1874static int dtSplitRoot(tid_t tid,
1875 struct inode *ip, struct dtsplit * split, struct metapage ** rmpp)
1876{
1877 struct super_block *sb = ip->i_sb;
1878 struct metapage *smp;
1879 dtroot_t *sp;
1880 struct metapage *rmp;
1881 dtpage_t *rp;
1882 s64 rbn;
1883 int xlen;
1884 int xsize;
1885 struct dtslot *f;
1886 s8 *stbl;
1887 int fsi, stblsize, n;
1888 struct idtentry *s;
1889 pxd_t *ppxd;
1890 struct pxdlist *pxdlist;
1891 pxd_t *pxd;
1892 struct dt_lock *dtlck;
1893 struct tlock *tlck;
1894 struct lv *lv;
1895
1896 /* get split root page */
1897 smp = split->mp;
1898 sp = &JFS_IP(ip)->i_dtroot;
1899
1900 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001901 * allocate/initialize a single (right) child page
Linus Torvalds1da177e2005-04-16 15:20:36 -07001902 *
1903 * N.B. at first split, a one (or two) block to fit new entry
1904 * is allocated; at subsequent split, a full page is allocated;
1905 */
1906 pxdlist = split->pxdlist;
1907 pxd = &pxdlist->pxd[pxdlist->npxd];
1908 pxdlist->npxd++;
1909 rbn = addressPXD(pxd);
1910 xlen = lengthPXD(pxd);
1911 xsize = xlen << JFS_SBI(sb)->l2bsize;
1912 rmp = get_metapage(ip, rbn, xsize, 1);
1913 if (!rmp)
1914 return -EIO;
1915
1916 rp = rmp->data;
1917
1918 /* Allocate blocks to quota. */
1919 if (DQUOT_ALLOC_BLOCK(ip, lengthPXD(pxd))) {
1920 release_metapage(rmp);
1921 return -EDQUOT;
1922 }
1923
1924 BT_MARK_DIRTY(rmp, ip);
1925 /*
1926 * acquire a transaction lock on the new right page
1927 */
1928 tlck = txLock(tid, ip, rmp, tlckDTREE | tlckNEW);
1929 dtlck = (struct dt_lock *) & tlck->lock;
1930
1931 rp->header.flag =
1932 (sp->header.flag & BT_LEAF) ? BT_LEAF : BT_INTERNAL;
1933 rp->header.self = *pxd;
1934
1935 /* initialize sibling pointers */
1936 rp->header.next = 0;
1937 rp->header.prev = 0;
1938
1939 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05001940 * move in-line root page into new right page extent
Linus Torvalds1da177e2005-04-16 15:20:36 -07001941 */
1942 /* linelock header + copied entries + new stbl (1st slot) in new page */
1943 ASSERT(dtlck->index == 0);
1944 lv = & dtlck->lv[0];
1945 lv->offset = 0;
1946 lv->length = 10; /* 1 + 8 + 1 */
1947 dtlck->index++;
1948
1949 n = xsize >> L2DTSLOTSIZE;
1950 rp->header.maxslot = n;
1951 stblsize = (n + 31) >> L2DTSLOTSIZE;
1952
1953 /* copy old stbl to new stbl at start of extended area */
1954 rp->header.stblindex = DTROOTMAXSLOT;
1955 stbl = (s8 *) & rp->slot[DTROOTMAXSLOT];
1956 memcpy(stbl, sp->header.stbl, sp->header.nextindex);
1957 rp->header.nextindex = sp->header.nextindex;
1958
1959 /* copy old data area to start of new data area */
1960 memcpy(&rp->slot[1], &sp->slot[1], IDATASIZE);
1961
1962 /*
1963 * append free region of newly extended area at tail of freelist
1964 */
1965 /* init free region of newly extended area */
1966 fsi = n = DTROOTMAXSLOT + stblsize;
1967 f = &rp->slot[fsi];
1968 for (fsi++; fsi < rp->header.maxslot; f++, fsi++)
1969 f->next = fsi;
1970 f->next = -1;
1971
1972 /* append new free region at tail of old freelist */
1973 fsi = sp->header.freelist;
1974 if (fsi == -1)
1975 rp->header.freelist = n;
1976 else {
1977 rp->header.freelist = fsi;
1978
1979 do {
1980 f = &rp->slot[fsi];
1981 fsi = f->next;
1982 } while (fsi != -1);
1983
1984 f->next = n;
1985 }
1986
1987 rp->header.freecnt = sp->header.freecnt + rp->header.maxslot - n;
1988
1989 /*
1990 * Update directory index table for entries now in right page
1991 */
1992 if ((rp->header.flag & BT_LEAF) && DO_INDEX(ip)) {
1993 s64 lblock;
1994 struct metapage *mp = NULL;
1995 struct ldtentry *ldtentry;
1996
1997 stbl = DT_GETSTBL(rp);
1998 for (n = 0; n < rp->header.nextindex; n++) {
1999 ldtentry = (struct ldtentry *) & rp->slot[stbl[n]];
2000 modify_index(tid, ip, le32_to_cpu(ldtentry->index),
2001 rbn, n, &mp, &lblock);
2002 }
2003 if (mp)
2004 release_metapage(mp);
2005 }
2006 /*
2007 * insert the new entry into the new right/child page
2008 * (skip index in the new right page will not change)
2009 */
2010 dtInsertEntry(rp, split->index, split->key, split->data, &dtlck);
2011
2012 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002013 * reset parent/root page
Linus Torvalds1da177e2005-04-16 15:20:36 -07002014 *
2015 * set the 1st entry offset to 0, which force the left-most key
2016 * at any level of the tree to be less than any search key.
2017 *
2018 * The btree comparison code guarantees that the left-most key on any
2019 * level of the tree is never used, so it doesn't need to be filled in.
2020 */
2021 BT_MARK_DIRTY(smp, ip);
2022 /*
2023 * acquire a transaction lock on the root page (in-memory inode)
2024 */
2025 tlck = txLock(tid, ip, smp, tlckDTREE | tlckNEW | tlckBTROOT);
2026 dtlck = (struct dt_lock *) & tlck->lock;
2027
2028 /* linelock root */
2029 ASSERT(dtlck->index == 0);
2030 lv = & dtlck->lv[0];
2031 lv->offset = 0;
2032 lv->length = DTROOTMAXSLOT;
2033 dtlck->index++;
2034
2035 /* update page header of root */
2036 if (sp->header.flag & BT_LEAF) {
2037 sp->header.flag &= ~BT_LEAF;
2038 sp->header.flag |= BT_INTERNAL;
2039 }
2040
2041 /* init the first entry */
2042 s = (struct idtentry *) & sp->slot[DTENTRYSTART];
2043 ppxd = (pxd_t *) s;
2044 *ppxd = *pxd;
2045 s->next = -1;
2046 s->namlen = 0;
2047
2048 stbl = sp->header.stbl;
2049 stbl[0] = DTENTRYSTART;
2050 sp->header.nextindex = 1;
2051
2052 /* init freelist */
2053 fsi = DTENTRYSTART + 1;
2054 f = &sp->slot[fsi];
2055
2056 /* init free region of remaining area */
2057 for (fsi++; fsi < DTROOTMAXSLOT; f++, fsi++)
2058 f->next = fsi;
2059 f->next = -1;
2060
2061 sp->header.freelist = DTENTRYSTART + 1;
2062 sp->header.freecnt = DTROOTMAXSLOT - (DTENTRYSTART + 1);
2063
2064 *rmpp = rmp;
2065
2066 return 0;
2067}
2068
2069
2070/*
2071 * dtDelete()
2072 *
2073 * function: delete the entry(s) referenced by a key.
2074 *
2075 * parameter:
2076 *
2077 * return:
2078 */
2079int dtDelete(tid_t tid,
2080 struct inode *ip, struct component_name * key, ino_t * ino, int flag)
2081{
2082 int rc = 0;
2083 s64 bn;
2084 struct metapage *mp, *imp;
2085 dtpage_t *p;
2086 int index;
2087 struct btstack btstack;
2088 struct dt_lock *dtlck;
2089 struct tlock *tlck;
2090 struct lv *lv;
2091 int i;
2092 struct ldtentry *ldtentry;
2093 u8 *stbl;
2094 u32 table_index, next_index;
2095 struct metapage *nmp;
2096 dtpage_t *np;
2097
2098 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002099 * search for the entry to delete:
Linus Torvalds1da177e2005-04-16 15:20:36 -07002100 *
2101 * dtSearch() returns (leaf page pinned, index at which to delete).
2102 */
2103 if ((rc = dtSearch(ip, key, ino, &btstack, flag)))
2104 return rc;
2105
2106 /* retrieve search result */
2107 DT_GETSEARCH(ip, btstack.top, bn, mp, p, index);
2108
2109 /*
2110 * We need to find put the index of the next entry into the
2111 * directory index table in order to resume a readdir from this
2112 * entry.
2113 */
2114 if (DO_INDEX(ip)) {
2115 stbl = DT_GETSTBL(p);
2116 ldtentry = (struct ldtentry *) & p->slot[stbl[index]];
2117 table_index = le32_to_cpu(ldtentry->index);
2118 if (index == (p->header.nextindex - 1)) {
2119 /*
2120 * Last entry in this leaf page
2121 */
2122 if ((p->header.flag & BT_ROOT)
2123 || (p->header.next == 0))
2124 next_index = -1;
2125 else {
2126 /* Read next leaf page */
2127 DT_GETPAGE(ip, le64_to_cpu(p->header.next),
2128 nmp, PSIZE, np, rc);
2129 if (rc)
2130 next_index = -1;
2131 else {
2132 stbl = DT_GETSTBL(np);
2133 ldtentry =
2134 (struct ldtentry *) & np->
2135 slot[stbl[0]];
2136 next_index =
2137 le32_to_cpu(ldtentry->index);
2138 DT_PUTPAGE(nmp);
2139 }
2140 }
2141 } else {
2142 ldtentry =
2143 (struct ldtentry *) & p->slot[stbl[index + 1]];
2144 next_index = le32_to_cpu(ldtentry->index);
2145 }
2146 free_index(tid, ip, table_index, next_index);
2147 }
2148 /*
2149 * the leaf page becomes empty, delete the page
2150 */
2151 if (p->header.nextindex == 1) {
2152 /* delete empty page */
2153 rc = dtDeleteUp(tid, ip, mp, p, &btstack);
2154 }
2155 /*
2156 * the leaf page has other entries remaining:
2157 *
2158 * delete the entry from the leaf page.
2159 */
2160 else {
2161 BT_MARK_DIRTY(mp, ip);
2162 /*
2163 * acquire a transaction lock on the leaf page
2164 */
2165 tlck = txLock(tid, ip, mp, tlckDTREE | tlckENTRY);
2166 dtlck = (struct dt_lock *) & tlck->lock;
2167
2168 /*
2169 * Do not assume that dtlck->index will be zero. During a
2170 * rename within a directory, this transaction may have
2171 * modified this page already when adding the new entry.
2172 */
2173
2174 /* linelock header */
2175 if (dtlck->index >= dtlck->maxcnt)
2176 dtlck = (struct dt_lock *) txLinelock(dtlck);
2177 lv = & dtlck->lv[dtlck->index];
2178 lv->offset = 0;
2179 lv->length = 1;
2180 dtlck->index++;
2181
2182 /* linelock stbl of non-root leaf page */
2183 if (!(p->header.flag & BT_ROOT)) {
2184 if (dtlck->index >= dtlck->maxcnt)
2185 dtlck = (struct dt_lock *) txLinelock(dtlck);
2186 lv = & dtlck->lv[dtlck->index];
2187 i = index >> L2DTSLOTSIZE;
2188 lv->offset = p->header.stblindex + i;
2189 lv->length =
2190 ((p->header.nextindex - 1) >> L2DTSLOTSIZE) -
2191 i + 1;
2192 dtlck->index++;
2193 }
2194
2195 /* free the leaf entry */
2196 dtDeleteEntry(p, index, &dtlck);
2197
2198 /*
2199 * Update directory index table for entries moved in stbl
2200 */
2201 if (DO_INDEX(ip) && index < p->header.nextindex) {
2202 s64 lblock;
2203
2204 imp = NULL;
2205 stbl = DT_GETSTBL(p);
2206 for (i = index; i < p->header.nextindex; i++) {
2207 ldtentry =
2208 (struct ldtentry *) & p->slot[stbl[i]];
2209 modify_index(tid, ip,
2210 le32_to_cpu(ldtentry->index),
2211 bn, i, &imp, &lblock);
2212 }
2213 if (imp)
2214 release_metapage(imp);
2215 }
2216
2217 DT_PUTPAGE(mp);
2218 }
2219
2220 return rc;
2221}
2222
2223
2224/*
2225 * dtDeleteUp()
2226 *
2227 * function:
2228 * free empty pages as propagating deletion up the tree
2229 *
2230 * parameter:
2231 *
2232 * return:
2233 */
2234static int dtDeleteUp(tid_t tid, struct inode *ip,
2235 struct metapage * fmp, dtpage_t * fp, struct btstack * btstack)
2236{
2237 int rc = 0;
2238 struct metapage *mp;
2239 dtpage_t *p;
2240 int index, nextindex;
2241 int xlen;
2242 struct btframe *parent;
2243 struct dt_lock *dtlck;
2244 struct tlock *tlck;
2245 struct lv *lv;
2246 struct pxd_lock *pxdlock;
2247 int i;
2248
2249 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002250 * keep the root leaf page which has become empty
Linus Torvalds1da177e2005-04-16 15:20:36 -07002251 */
2252 if (BT_IS_ROOT(fmp)) {
2253 /*
2254 * reset the root
2255 *
2256 * dtInitRoot() acquires txlock on the root
2257 */
2258 dtInitRoot(tid, ip, PARENT(ip));
2259
2260 DT_PUTPAGE(fmp);
2261
2262 return 0;
2263 }
2264
2265 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002266 * free the non-root leaf page
Linus Torvalds1da177e2005-04-16 15:20:36 -07002267 */
2268 /*
2269 * acquire a transaction lock on the page
2270 *
2271 * write FREEXTENT|NOREDOPAGE log record
2272 * N.B. linelock is overlaid as freed extent descriptor, and
2273 * the buffer page is freed;
2274 */
2275 tlck = txMaplock(tid, ip, tlckDTREE | tlckFREE);
2276 pxdlock = (struct pxd_lock *) & tlck->lock;
2277 pxdlock->flag = mlckFREEPXD;
2278 pxdlock->pxd = fp->header.self;
2279 pxdlock->index = 1;
2280
2281 /* update sibling pointers */
2282 if ((rc = dtRelink(tid, ip, fp))) {
2283 BT_PUTPAGE(fmp);
2284 return rc;
2285 }
2286
2287 xlen = lengthPXD(&fp->header.self);
2288
2289 /* Free quota allocation. */
2290 DQUOT_FREE_BLOCK(ip, xlen);
2291
2292 /* free/invalidate its buffer page */
2293 discard_metapage(fmp);
2294
2295 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002296 * propagate page deletion up the directory tree
Linus Torvalds1da177e2005-04-16 15:20:36 -07002297 *
2298 * If the delete from the parent page makes it empty,
2299 * continue all the way up the tree.
2300 * stop if the root page is reached (which is never deleted) or
2301 * if the entry deletion does not empty the page.
2302 */
2303 while ((parent = BT_POP(btstack)) != NULL) {
2304 /* pin the parent page <sp> */
2305 DT_GETPAGE(ip, parent->bn, mp, PSIZE, p, rc);
2306 if (rc)
2307 return rc;
2308
2309 /*
2310 * free the extent of the child page deleted
2311 */
2312 index = parent->index;
2313
2314 /*
2315 * delete the entry for the child page from parent
2316 */
2317 nextindex = p->header.nextindex;
2318
2319 /*
2320 * the parent has the single entry being deleted:
2321 *
2322 * free the parent page which has become empty.
2323 */
2324 if (nextindex == 1) {
2325 /*
2326 * keep the root internal page which has become empty
2327 */
2328 if (p->header.flag & BT_ROOT) {
2329 /*
2330 * reset the root
2331 *
2332 * dtInitRoot() acquires txlock on the root
2333 */
2334 dtInitRoot(tid, ip, PARENT(ip));
2335
2336 DT_PUTPAGE(mp);
2337
2338 return 0;
2339 }
2340 /*
2341 * free the parent page
2342 */
2343 else {
2344 /*
2345 * acquire a transaction lock on the page
2346 *
2347 * write FREEXTENT|NOREDOPAGE log record
2348 */
2349 tlck =
2350 txMaplock(tid, ip,
2351 tlckDTREE | tlckFREE);
2352 pxdlock = (struct pxd_lock *) & tlck->lock;
2353 pxdlock->flag = mlckFREEPXD;
2354 pxdlock->pxd = p->header.self;
2355 pxdlock->index = 1;
2356
2357 /* update sibling pointers */
2358 if ((rc = dtRelink(tid, ip, p))) {
2359 DT_PUTPAGE(mp);
2360 return rc;
2361 }
2362
2363 xlen = lengthPXD(&p->header.self);
2364
2365 /* Free quota allocation */
2366 DQUOT_FREE_BLOCK(ip, xlen);
2367
2368 /* free/invalidate its buffer page */
2369 discard_metapage(mp);
2370
2371 /* propagate up */
2372 continue;
2373 }
2374 }
2375
2376 /*
2377 * the parent has other entries remaining:
2378 *
2379 * delete the router entry from the parent page.
2380 */
2381 BT_MARK_DIRTY(mp, ip);
2382 /*
2383 * acquire a transaction lock on the page
2384 *
2385 * action: router entry deletion
2386 */
2387 tlck = txLock(tid, ip, mp, tlckDTREE | tlckENTRY);
2388 dtlck = (struct dt_lock *) & tlck->lock;
2389
2390 /* linelock header */
2391 if (dtlck->index >= dtlck->maxcnt)
2392 dtlck = (struct dt_lock *) txLinelock(dtlck);
2393 lv = & dtlck->lv[dtlck->index];
2394 lv->offset = 0;
2395 lv->length = 1;
2396 dtlck->index++;
2397
2398 /* linelock stbl of non-root leaf page */
2399 if (!(p->header.flag & BT_ROOT)) {
2400 if (dtlck->index < dtlck->maxcnt)
2401 lv++;
2402 else {
2403 dtlck = (struct dt_lock *) txLinelock(dtlck);
2404 lv = & dtlck->lv[0];
2405 }
2406 i = index >> L2DTSLOTSIZE;
2407 lv->offset = p->header.stblindex + i;
2408 lv->length =
2409 ((p->header.nextindex - 1) >> L2DTSLOTSIZE) -
2410 i + 1;
2411 dtlck->index++;
2412 }
2413
2414 /* free the router entry */
2415 dtDeleteEntry(p, index, &dtlck);
2416
2417 /* reset key of new leftmost entry of level (for consistency) */
2418 if (index == 0 &&
2419 ((p->header.flag & BT_ROOT) || p->header.prev == 0))
2420 dtTruncateEntry(p, 0, &dtlck);
2421
2422 /* unpin the parent page */
2423 DT_PUTPAGE(mp);
2424
2425 /* exit propagation up */
2426 break;
2427 }
2428
Dave Kleikampdd8a3062005-11-10 07:50:03 -06002429 if (!DO_INDEX(ip))
2430 ip->i_size -= PSIZE;
2431
Linus Torvalds1da177e2005-04-16 15:20:36 -07002432 return 0;
2433}
2434
2435#ifdef _NOTYET
2436/*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002437 * NAME: dtRelocate()
Linus Torvalds1da177e2005-04-16 15:20:36 -07002438 *
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002439 * FUNCTION: relocate dtpage (internal or leaf) of directory;
2440 * This function is mainly used by defragfs utility.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002441 */
2442int dtRelocate(tid_t tid, struct inode *ip, s64 lmxaddr, pxd_t * opxd,
2443 s64 nxaddr)
2444{
2445 int rc = 0;
2446 struct metapage *mp, *pmp, *lmp, *rmp;
2447 dtpage_t *p, *pp, *rp = 0, *lp= 0;
2448 s64 bn;
2449 int index;
2450 struct btstack btstack;
2451 pxd_t *pxd;
2452 s64 oxaddr, nextbn, prevbn;
2453 int xlen, xsize;
2454 struct tlock *tlck;
2455 struct dt_lock *dtlck;
2456 struct pxd_lock *pxdlock;
2457 s8 *stbl;
2458 struct lv *lv;
2459
2460 oxaddr = addressPXD(opxd);
2461 xlen = lengthPXD(opxd);
2462
2463 jfs_info("dtRelocate: lmxaddr:%Ld xaddr:%Ld:%Ld xlen:%d",
2464 (long long)lmxaddr, (long long)oxaddr, (long long)nxaddr,
2465 xlen);
2466
2467 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002468 * 1. get the internal parent dtpage covering
2469 * router entry for the tartget page to be relocated;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002470 */
2471 rc = dtSearchNode(ip, lmxaddr, opxd, &btstack);
2472 if (rc)
2473 return rc;
2474
2475 /* retrieve search result */
2476 DT_GETSEARCH(ip, btstack.top, bn, pmp, pp, index);
2477 jfs_info("dtRelocate: parent router entry validated.");
2478
2479 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002480 * 2. relocate the target dtpage
Linus Torvalds1da177e2005-04-16 15:20:36 -07002481 */
2482 /* read in the target page from src extent */
2483 DT_GETPAGE(ip, oxaddr, mp, PSIZE, p, rc);
2484 if (rc) {
2485 /* release the pinned parent page */
2486 DT_PUTPAGE(pmp);
2487 return rc;
2488 }
2489
2490 /*
2491 * read in sibling pages if any to update sibling pointers;
2492 */
2493 rmp = NULL;
2494 if (p->header.next) {
2495 nextbn = le64_to_cpu(p->header.next);
2496 DT_GETPAGE(ip, nextbn, rmp, PSIZE, rp, rc);
2497 if (rc) {
2498 DT_PUTPAGE(mp);
2499 DT_PUTPAGE(pmp);
2500 return (rc);
2501 }
2502 }
2503
2504 lmp = NULL;
2505 if (p->header.prev) {
2506 prevbn = le64_to_cpu(p->header.prev);
2507 DT_GETPAGE(ip, prevbn, lmp, PSIZE, lp, rc);
2508 if (rc) {
2509 DT_PUTPAGE(mp);
2510 DT_PUTPAGE(pmp);
2511 if (rmp)
2512 DT_PUTPAGE(rmp);
2513 return (rc);
2514 }
2515 }
2516
2517 /* at this point, all xtpages to be updated are in memory */
2518
2519 /*
2520 * update sibling pointers of sibling dtpages if any;
2521 */
2522 if (lmp) {
2523 tlck = txLock(tid, ip, lmp, tlckDTREE | tlckRELINK);
2524 dtlck = (struct dt_lock *) & tlck->lock;
2525 /* linelock header */
2526 ASSERT(dtlck->index == 0);
2527 lv = & dtlck->lv[0];
2528 lv->offset = 0;
2529 lv->length = 1;
2530 dtlck->index++;
2531
2532 lp->header.next = cpu_to_le64(nxaddr);
2533 DT_PUTPAGE(lmp);
2534 }
2535
2536 if (rmp) {
2537 tlck = txLock(tid, ip, rmp, tlckDTREE | tlckRELINK);
2538 dtlck = (struct dt_lock *) & tlck->lock;
2539 /* linelock header */
2540 ASSERT(dtlck->index == 0);
2541 lv = & dtlck->lv[0];
2542 lv->offset = 0;
2543 lv->length = 1;
2544 dtlck->index++;
2545
2546 rp->header.prev = cpu_to_le64(nxaddr);
2547 DT_PUTPAGE(rmp);
2548 }
2549
2550 /*
2551 * update the target dtpage to be relocated
2552 *
2553 * write LOG_REDOPAGE of LOG_NEW type for dst page
2554 * for the whole target page (logredo() will apply
2555 * after image and update bmap for allocation of the
2556 * dst extent), and update bmap for allocation of
2557 * the dst extent;
2558 */
2559 tlck = txLock(tid, ip, mp, tlckDTREE | tlckNEW);
2560 dtlck = (struct dt_lock *) & tlck->lock;
2561 /* linelock header */
2562 ASSERT(dtlck->index == 0);
2563 lv = & dtlck->lv[0];
2564
2565 /* update the self address in the dtpage header */
2566 pxd = &p->header.self;
2567 PXDaddress(pxd, nxaddr);
2568
2569 /* the dst page is the same as the src page, i.e.,
2570 * linelock for afterimage of the whole page;
2571 */
2572 lv->offset = 0;
2573 lv->length = p->header.maxslot;
2574 dtlck->index++;
2575
2576 /* update the buffer extent descriptor of the dtpage */
2577 xsize = xlen << JFS_SBI(ip->i_sb)->l2bsize;
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002578
Linus Torvalds1da177e2005-04-16 15:20:36 -07002579 /* unpin the relocated page */
2580 DT_PUTPAGE(mp);
2581 jfs_info("dtRelocate: target dtpage relocated.");
2582
2583 /* the moved extent is dtpage, then a LOG_NOREDOPAGE log rec
2584 * needs to be written (in logredo(), the LOG_NOREDOPAGE log rec
2585 * will also force a bmap update ).
2586 */
2587
2588 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002589 * 3. acquire maplock for the source extent to be freed;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002590 */
2591 /* for dtpage relocation, write a LOG_NOREDOPAGE record
2592 * for the source dtpage (logredo() will init NoRedoPage
2593 * filter and will also update bmap for free of the source
2594 * dtpage), and upadte bmap for free of the source dtpage;
2595 */
2596 tlck = txMaplock(tid, ip, tlckDTREE | tlckFREE);
2597 pxdlock = (struct pxd_lock *) & tlck->lock;
2598 pxdlock->flag = mlckFREEPXD;
2599 PXDaddress(&pxdlock->pxd, oxaddr);
2600 PXDlength(&pxdlock->pxd, xlen);
2601 pxdlock->index = 1;
2602
2603 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002604 * 4. update the parent router entry for relocation;
Linus Torvalds1da177e2005-04-16 15:20:36 -07002605 *
2606 * acquire tlck for the parent entry covering the target dtpage;
2607 * write LOG_REDOPAGE to apply after image only;
2608 */
2609 jfs_info("dtRelocate: update parent router entry.");
2610 tlck = txLock(tid, ip, pmp, tlckDTREE | tlckENTRY);
2611 dtlck = (struct dt_lock *) & tlck->lock;
2612 lv = & dtlck->lv[dtlck->index];
2613
2614 /* update the PXD with the new address */
2615 stbl = DT_GETSTBL(pp);
2616 pxd = (pxd_t *) & pp->slot[stbl[index]];
2617 PXDaddress(pxd, nxaddr);
2618 lv->offset = stbl[index];
2619 lv->length = 1;
2620 dtlck->index++;
2621
2622 /* unpin the parent dtpage */
2623 DT_PUTPAGE(pmp);
2624
2625 return rc;
2626}
2627
2628/*
2629 * NAME: dtSearchNode()
2630 *
2631 * FUNCTION: Search for an dtpage containing a specified address
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002632 * This function is mainly used by defragfs utility.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002633 *
2634 * NOTE: Search result on stack, the found page is pinned at exit.
2635 * The result page must be an internal dtpage.
2636 * lmxaddr give the address of the left most page of the
2637 * dtree level, in which the required dtpage resides.
2638 */
2639static int dtSearchNode(struct inode *ip, s64 lmxaddr, pxd_t * kpxd,
2640 struct btstack * btstack)
2641{
2642 int rc = 0;
2643 s64 bn;
2644 struct metapage *mp;
2645 dtpage_t *p;
2646 int psize = 288; /* initial in-line directory */
2647 s8 *stbl;
2648 int i;
2649 pxd_t *pxd;
2650 struct btframe *btsp;
2651
2652 BT_CLR(btstack); /* reset stack */
2653
2654 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002655 * descend tree to the level with specified leftmost page
Linus Torvalds1da177e2005-04-16 15:20:36 -07002656 *
2657 * by convention, root bn = 0.
2658 */
2659 for (bn = 0;;) {
2660 /* get/pin the page to search */
2661 DT_GETPAGE(ip, bn, mp, psize, p, rc);
2662 if (rc)
2663 return rc;
2664
2665 /* does the xaddr of leftmost page of the levevl
2666 * matches levevl search key ?
2667 */
2668 if (p->header.flag & BT_ROOT) {
2669 if (lmxaddr == 0)
2670 break;
2671 } else if (addressPXD(&p->header.self) == lmxaddr)
2672 break;
2673
2674 /*
2675 * descend down to leftmost child page
2676 */
2677 if (p->header.flag & BT_LEAF) {
2678 DT_PUTPAGE(mp);
2679 return -ESTALE;
2680 }
2681
2682 /* get the leftmost entry */
2683 stbl = DT_GETSTBL(p);
2684 pxd = (pxd_t *) & p->slot[stbl[0]];
2685
2686 /* get the child page block address */
2687 bn = addressPXD(pxd);
2688 psize = lengthPXD(pxd) << JFS_SBI(ip->i_sb)->l2bsize;
2689 /* unpin the parent page */
2690 DT_PUTPAGE(mp);
2691 }
2692
2693 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05002694 * search each page at the current levevl
Linus Torvalds1da177e2005-04-16 15:20:36 -07002695 */
2696 loop:
2697 stbl = DT_GETSTBL(p);
2698 for (i = 0; i < p->header.nextindex; i++) {
2699 pxd = (pxd_t *) & p->slot[stbl[i]];
2700
2701 /* found the specified router entry */
2702 if (addressPXD(pxd) == addressPXD(kpxd) &&
2703 lengthPXD(pxd) == lengthPXD(kpxd)) {
2704 btsp = btstack->top;
2705 btsp->bn = bn;
2706 btsp->index = i;
2707 btsp->mp = mp;
2708
2709 return 0;
2710 }
2711 }
2712
2713 /* get the right sibling page if any */
2714 if (p->header.next)
2715 bn = le64_to_cpu(p->header.next);
2716 else {
2717 DT_PUTPAGE(mp);
2718 return -ESTALE;
2719 }
2720
2721 /* unpin current page */
2722 DT_PUTPAGE(mp);
2723
2724 /* get the right sibling page */
2725 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
2726 if (rc)
2727 return rc;
2728
2729 goto loop;
2730}
2731#endif /* _NOTYET */
2732
2733/*
2734 * dtRelink()
2735 *
2736 * function:
2737 * link around a freed page.
2738 *
2739 * parameter:
2740 * fp: page to be freed
2741 *
2742 * return:
2743 */
2744static int dtRelink(tid_t tid, struct inode *ip, dtpage_t * p)
2745{
2746 int rc;
2747 struct metapage *mp;
2748 s64 nextbn, prevbn;
2749 struct tlock *tlck;
2750 struct dt_lock *dtlck;
2751 struct lv *lv;
2752
2753 nextbn = le64_to_cpu(p->header.next);
2754 prevbn = le64_to_cpu(p->header.prev);
2755
2756 /* update prev pointer of the next page */
2757 if (nextbn != 0) {
2758 DT_GETPAGE(ip, nextbn, mp, PSIZE, p, rc);
2759 if (rc)
2760 return rc;
2761
2762 BT_MARK_DIRTY(mp, ip);
2763 /*
2764 * acquire a transaction lock on the next page
2765 *
2766 * action: update prev pointer;
2767 */
2768 tlck = txLock(tid, ip, mp, tlckDTREE | tlckRELINK);
2769 jfs_info("dtRelink nextbn: tlck = 0x%p, ip = 0x%p, mp=0x%p",
2770 tlck, ip, mp);
2771 dtlck = (struct dt_lock *) & tlck->lock;
2772
2773 /* linelock header */
2774 if (dtlck->index >= dtlck->maxcnt)
2775 dtlck = (struct dt_lock *) txLinelock(dtlck);
2776 lv = & dtlck->lv[dtlck->index];
2777 lv->offset = 0;
2778 lv->length = 1;
2779 dtlck->index++;
2780
2781 p->header.prev = cpu_to_le64(prevbn);
2782 DT_PUTPAGE(mp);
2783 }
2784
2785 /* update next pointer of the previous page */
2786 if (prevbn != 0) {
2787 DT_GETPAGE(ip, prevbn, mp, PSIZE, p, rc);
2788 if (rc)
2789 return rc;
2790
2791 BT_MARK_DIRTY(mp, ip);
2792 /*
2793 * acquire a transaction lock on the prev page
2794 *
2795 * action: update next pointer;
2796 */
2797 tlck = txLock(tid, ip, mp, tlckDTREE | tlckRELINK);
2798 jfs_info("dtRelink prevbn: tlck = 0x%p, ip = 0x%p, mp=0x%p",
2799 tlck, ip, mp);
2800 dtlck = (struct dt_lock *) & tlck->lock;
2801
2802 /* linelock header */
2803 if (dtlck->index >= dtlck->maxcnt)
2804 dtlck = (struct dt_lock *) txLinelock(dtlck);
2805 lv = & dtlck->lv[dtlck->index];
2806 lv->offset = 0;
2807 lv->length = 1;
2808 dtlck->index++;
2809
2810 p->header.next = cpu_to_le64(nextbn);
2811 DT_PUTPAGE(mp);
2812 }
2813
2814 return 0;
2815}
2816
2817
2818/*
2819 * dtInitRoot()
2820 *
2821 * initialize directory root (inline in inode)
2822 */
2823void dtInitRoot(tid_t tid, struct inode *ip, u32 idotdot)
2824{
2825 struct jfs_inode_info *jfs_ip = JFS_IP(ip);
2826 dtroot_t *p;
2827 int fsi;
2828 struct dtslot *f;
2829 struct tlock *tlck;
2830 struct dt_lock *dtlck;
2831 struct lv *lv;
2832 u16 xflag_save;
2833
2834 /*
2835 * If this was previously an non-empty directory, we need to remove
2836 * the old directory table.
2837 */
2838 if (DO_INDEX(ip)) {
2839 if (!jfs_dirtable_inline(ip)) {
2840 struct tblock *tblk = tid_to_tblock(tid);
2841 /*
2842 * We're playing games with the tid's xflag. If
2843 * we're removing a regular file, the file's xtree
2844 * is committed with COMMIT_PMAP, but we always
2845 * commit the directories xtree with COMMIT_PWMAP.
2846 */
2847 xflag_save = tblk->xflag;
2848 tblk->xflag = 0;
2849 /*
2850 * xtTruncate isn't guaranteed to fully truncate
2851 * the xtree. The caller needs to check i_size
2852 * after committing the transaction to see if
2853 * additional truncation is needed. The
2854 * COMMIT_Stale flag tells caller that we
2855 * initiated the truncation.
2856 */
2857 xtTruncate(tid, ip, 0, COMMIT_PWMAP);
2858 set_cflag(COMMIT_Stale, ip);
2859
2860 tblk->xflag = xflag_save;
2861 } else
2862 ip->i_size = 1;
2863
2864 jfs_ip->next_index = 2;
2865 } else
2866 ip->i_size = IDATASIZE;
2867
2868 /*
2869 * acquire a transaction lock on the root
2870 *
2871 * action: directory initialization;
2872 */
2873 tlck = txLock(tid, ip, (struct metapage *) & jfs_ip->bxflag,
2874 tlckDTREE | tlckENTRY | tlckBTROOT);
2875 dtlck = (struct dt_lock *) & tlck->lock;
2876
2877 /* linelock root */
2878 ASSERT(dtlck->index == 0);
2879 lv = & dtlck->lv[0];
2880 lv->offset = 0;
2881 lv->length = DTROOTMAXSLOT;
2882 dtlck->index++;
2883
2884 p = &jfs_ip->i_dtroot;
2885
2886 p->header.flag = DXD_INDEX | BT_ROOT | BT_LEAF;
2887
2888 p->header.nextindex = 0;
2889
2890 /* init freelist */
2891 fsi = 1;
2892 f = &p->slot[fsi];
2893
2894 /* init data area of root */
2895 for (fsi++; fsi < DTROOTMAXSLOT; f++, fsi++)
2896 f->next = fsi;
2897 f->next = -1;
2898
2899 p->header.freelist = 1;
2900 p->header.freecnt = 8;
2901
2902 /* init '..' entry */
2903 p->header.idotdot = cpu_to_le32(idotdot);
2904
2905 return;
2906}
2907
2908/*
2909 * add_missing_indices()
2910 *
2911 * function: Fix dtree page in which one or more entries has an invalid index.
2912 * fsck.jfs should really fix this, but it currently does not.
2913 * Called from jfs_readdir when bad index is detected.
2914 */
2915static void add_missing_indices(struct inode *inode, s64 bn)
2916{
2917 struct ldtentry *d;
2918 struct dt_lock *dtlck;
2919 int i;
2920 uint index;
2921 struct lv *lv;
2922 struct metapage *mp;
2923 dtpage_t *p;
2924 int rc;
2925 s8 *stbl;
2926 tid_t tid;
2927 struct tlock *tlck;
2928
2929 tid = txBegin(inode->i_sb, 0);
2930
2931 DT_GETPAGE(inode, bn, mp, PSIZE, p, rc);
2932
2933 if (rc) {
2934 printk(KERN_ERR "DT_GETPAGE failed!\n");
2935 goto end;
2936 }
2937 BT_MARK_DIRTY(mp, inode);
2938
2939 ASSERT(p->header.flag & BT_LEAF);
2940
2941 tlck = txLock(tid, inode, mp, tlckDTREE | tlckENTRY);
Dave Kleikampc2731502005-06-02 12:18:20 -05002942 if (BT_IS_ROOT(mp))
2943 tlck->type |= tlckBTROOT;
2944
Linus Torvalds1da177e2005-04-16 15:20:36 -07002945 dtlck = (struct dt_lock *) &tlck->lock;
2946
2947 stbl = DT_GETSTBL(p);
2948 for (i = 0; i < p->header.nextindex; i++) {
2949 d = (struct ldtentry *) &p->slot[stbl[i]];
2950 index = le32_to_cpu(d->index);
2951 if ((index < 2) || (index >= JFS_IP(inode)->next_index)) {
2952 d->index = cpu_to_le32(add_index(tid, inode, bn, i));
2953 if (dtlck->index >= dtlck->maxcnt)
2954 dtlck = (struct dt_lock *) txLinelock(dtlck);
2955 lv = &dtlck->lv[dtlck->index];
2956 lv->offset = stbl[i];
2957 lv->length = 1;
2958 dtlck->index++;
2959 }
2960 }
2961
2962 DT_PUTPAGE(mp);
2963 (void) txCommit(tid, 1, &inode, 0);
2964end:
2965 txEnd(tid);
2966}
2967
2968/*
2969 * Buffer to hold directory entry info while traversing a dtree page
2970 * before being fed to the filldir function
2971 */
2972struct jfs_dirent {
2973 loff_t position;
2974 int ino;
2975 u16 name_len;
2976 char name[0];
2977};
2978
2979/*
2980 * function to determine next variable-sized jfs_dirent in buffer
2981 */
2982static inline struct jfs_dirent *next_jfs_dirent(struct jfs_dirent *dirent)
2983{
2984 return (struct jfs_dirent *)
2985 ((char *)dirent +
2986 ((sizeof (struct jfs_dirent) + dirent->name_len + 1 +
2987 sizeof (loff_t) - 1) &
2988 ~(sizeof (loff_t) - 1)));
2989}
2990
2991/*
2992 * jfs_readdir()
2993 *
2994 * function: read directory entries sequentially
2995 * from the specified entry offset
2996 *
2997 * parameter:
2998 *
2999 * return: offset = (pn, index) of start entry
3000 * of next jfs_readdir()/dtRead()
3001 */
3002int jfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
3003{
Josef Sipekff273772006-12-08 02:37:16 -08003004 struct inode *ip = filp->f_path.dentry->d_inode;
Linus Torvalds1da177e2005-04-16 15:20:36 -07003005 struct nls_table *codepage = JFS_SBI(ip->i_sb)->nls_tab;
3006 int rc = 0;
3007 loff_t dtpos; /* legacy OS/2 style position */
3008 struct dtoffset {
3009 s16 pn;
3010 s16 index;
3011 s32 unused;
3012 } *dtoffset = (struct dtoffset *) &dtpos;
3013 s64 bn;
3014 struct metapage *mp;
3015 dtpage_t *p;
3016 int index;
3017 s8 *stbl;
3018 struct btstack btstack;
3019 int i, next;
3020 struct ldtentry *d;
3021 struct dtslot *t;
3022 int d_namleft, len, outlen;
3023 unsigned long dirent_buf;
3024 char *name_ptr;
3025 u32 dir_index;
3026 int do_index = 0;
3027 uint loop_count = 0;
3028 struct jfs_dirent *jfs_dirent;
3029 int jfs_dirents;
3030 int overflow, fix_page, page_fixed = 0;
3031 static int unique_pos = 2; /* If we can't fix broken index */
3032
3033 if (filp->f_pos == DIREND)
3034 return 0;
3035
3036 if (DO_INDEX(ip)) {
3037 /*
3038 * persistent index is stored in directory entries.
Dave Kleikampf720e3b2007-06-06 15:28:35 -05003039 * Special cases: 0 = .
3040 * 1 = ..
3041 * -1 = End of directory
Linus Torvalds1da177e2005-04-16 15:20:36 -07003042 */
3043 do_index = 1;
3044
3045 dir_index = (u32) filp->f_pos;
3046
3047 if (dir_index > 1) {
3048 struct dir_table_slot dirtab_slot;
3049
3050 if (dtEmpty(ip) ||
3051 (dir_index >= JFS_IP(ip)->next_index)) {
3052 /* Stale position. Directory has shrunk */
3053 filp->f_pos = DIREND;
3054 return 0;
3055 }
3056 repeat:
3057 rc = read_index(ip, dir_index, &dirtab_slot);
3058 if (rc) {
3059 filp->f_pos = DIREND;
3060 return rc;
3061 }
3062 if (dirtab_slot.flag == DIR_INDEX_FREE) {
3063 if (loop_count++ > JFS_IP(ip)->next_index) {
3064 jfs_err("jfs_readdir detected "
3065 "infinite loop!");
3066 filp->f_pos = DIREND;
3067 return 0;
3068 }
3069 dir_index = le32_to_cpu(dirtab_slot.addr2);
3070 if (dir_index == -1) {
3071 filp->f_pos = DIREND;
3072 return 0;
3073 }
3074 goto repeat;
3075 }
3076 bn = addressDTS(&dirtab_slot);
3077 index = dirtab_slot.slot;
3078 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3079 if (rc) {
3080 filp->f_pos = DIREND;
3081 return 0;
3082 }
3083 if (p->header.flag & BT_INTERNAL) {
3084 jfs_err("jfs_readdir: bad index table");
3085 DT_PUTPAGE(mp);
3086 filp->f_pos = -1;
3087 return 0;
3088 }
3089 } else {
3090 if (dir_index == 0) {
3091 /*
3092 * self "."
3093 */
3094 filp->f_pos = 0;
3095 if (filldir(dirent, ".", 1, 0, ip->i_ino,
3096 DT_DIR))
3097 return 0;
3098 }
3099 /*
3100 * parent ".."
3101 */
3102 filp->f_pos = 1;
3103 if (filldir(dirent, "..", 2, 1, PARENT(ip), DT_DIR))
3104 return 0;
3105
3106 /*
3107 * Find first entry of left-most leaf
3108 */
3109 if (dtEmpty(ip)) {
3110 filp->f_pos = DIREND;
3111 return 0;
3112 }
3113
3114 if ((rc = dtReadFirst(ip, &btstack)))
3115 return rc;
3116
3117 DT_GETSEARCH(ip, btstack.top, bn, mp, p, index);
3118 }
3119 } else {
3120 /*
3121 * Legacy filesystem - OS/2 & Linux JFS < 0.3.6
3122 *
Dave Kleikampf720e3b2007-06-06 15:28:35 -05003123 * pn = index = 0: First entry "."
3124 * pn = 0; index = 1: Second entry ".."
3125 * pn > 0: Real entries, pn=1 -> leftmost page
3126 * pn = index = -1: No more entries
Linus Torvalds1da177e2005-04-16 15:20:36 -07003127 */
3128 dtpos = filp->f_pos;
3129 if (dtpos == 0) {
3130 /* build "." entry */
3131
3132 if (filldir(dirent, ".", 1, filp->f_pos, ip->i_ino,
3133 DT_DIR))
3134 return 0;
3135 dtoffset->index = 1;
3136 filp->f_pos = dtpos;
3137 }
3138
3139 if (dtoffset->pn == 0) {
3140 if (dtoffset->index == 1) {
3141 /* build ".." entry */
3142
3143 if (filldir(dirent, "..", 2, filp->f_pos,
3144 PARENT(ip), DT_DIR))
3145 return 0;
3146 } else {
3147 jfs_err("jfs_readdir called with "
3148 "invalid offset!");
3149 }
3150 dtoffset->pn = 1;
3151 dtoffset->index = 0;
3152 filp->f_pos = dtpos;
3153 }
3154
3155 if (dtEmpty(ip)) {
3156 filp->f_pos = DIREND;
3157 return 0;
3158 }
3159
3160 if ((rc = dtReadNext(ip, &filp->f_pos, &btstack))) {
3161 jfs_err("jfs_readdir: unexpected rc = %d "
3162 "from dtReadNext", rc);
3163 filp->f_pos = DIREND;
3164 return 0;
3165 }
3166 /* get start leaf page and index */
3167 DT_GETSEARCH(ip, btstack.top, bn, mp, p, index);
3168
3169 /* offset beyond directory eof ? */
3170 if (bn < 0) {
3171 filp->f_pos = DIREND;
3172 return 0;
3173 }
3174 }
3175
3176 dirent_buf = __get_free_page(GFP_KERNEL);
3177 if (dirent_buf == 0) {
3178 DT_PUTPAGE(mp);
3179 jfs_warn("jfs_readdir: __get_free_page failed!");
3180 filp->f_pos = DIREND;
3181 return -ENOMEM;
3182 }
3183
3184 while (1) {
3185 jfs_dirent = (struct jfs_dirent *) dirent_buf;
3186 jfs_dirents = 0;
3187 overflow = fix_page = 0;
3188
3189 stbl = DT_GETSTBL(p);
3190
3191 for (i = index; i < p->header.nextindex; i++) {
3192 d = (struct ldtentry *) & p->slot[stbl[i]];
3193
3194 if (((long) jfs_dirent + d->namlen + 1) >
Dave Kleikampdc5798d2005-05-02 12:24:57 -06003195 (dirent_buf + PAGE_SIZE)) {
Linus Torvalds1da177e2005-04-16 15:20:36 -07003196 /* DBCS codepages could overrun dirent_buf */
3197 index = i;
3198 overflow = 1;
3199 break;
3200 }
3201
3202 d_namleft = d->namlen;
3203 name_ptr = jfs_dirent->name;
3204 jfs_dirent->ino = le32_to_cpu(d->inumber);
3205
3206 if (do_index) {
3207 len = min(d_namleft, DTLHDRDATALEN);
3208 jfs_dirent->position = le32_to_cpu(d->index);
3209 /*
3210 * d->index should always be valid, but it
3211 * isn't. fsck.jfs doesn't create the
3212 * directory index for the lost+found
3213 * directory. Rather than let it go,
3214 * we can try to fix it.
3215 */
3216 if ((jfs_dirent->position < 2) ||
3217 (jfs_dirent->position >=
3218 JFS_IP(ip)->next_index)) {
3219 if (!page_fixed && !isReadOnly(ip)) {
3220 fix_page = 1;
3221 /*
3222 * setting overflow and setting
3223 * index to i will cause the
3224 * same page to be processed
3225 * again starting here
3226 */
3227 overflow = 1;
3228 index = i;
3229 break;
3230 }
3231 jfs_dirent->position = unique_pos++;
3232 }
3233 } else {
3234 jfs_dirent->position = dtpos;
3235 len = min(d_namleft, DTLHDRDATALEN_LEGACY);
3236 }
3237
3238 /* copy the name of head/only segment */
3239 outlen = jfs_strfromUCS_le(name_ptr, d->name, len,
3240 codepage);
3241 jfs_dirent->name_len = outlen;
3242
3243 /* copy name in the additional segment(s) */
3244 next = d->next;
3245 while (next >= 0) {
3246 t = (struct dtslot *) & p->slot[next];
3247 name_ptr += outlen;
3248 d_namleft -= len;
3249 /* Sanity Check */
3250 if (d_namleft == 0) {
3251 jfs_error(ip->i_sb,
3252 "JFS:Dtree error: ino = "
3253 "%ld, bn=%Ld, index = %d",
3254 (long)ip->i_ino,
3255 (long long)bn,
3256 i);
3257 goto skip_one;
3258 }
3259 len = min(d_namleft, DTSLOTDATALEN);
3260 outlen = jfs_strfromUCS_le(name_ptr, t->name,
3261 len, codepage);
3262 jfs_dirent->name_len += outlen;
3263
3264 next = t->next;
3265 }
3266
3267 jfs_dirents++;
3268 jfs_dirent = next_jfs_dirent(jfs_dirent);
3269skip_one:
3270 if (!do_index)
3271 dtoffset->index++;
3272 }
3273
3274 if (!overflow) {
3275 /* Point to next leaf page */
3276 if (p->header.flag & BT_ROOT)
3277 bn = 0;
3278 else {
3279 bn = le64_to_cpu(p->header.next);
3280 index = 0;
3281 /* update offset (pn:index) for new page */
3282 if (!do_index) {
3283 dtoffset->pn++;
3284 dtoffset->index = 0;
3285 }
3286 }
3287 page_fixed = 0;
3288 }
3289
3290 /* unpin previous leaf page */
3291 DT_PUTPAGE(mp);
3292
3293 jfs_dirent = (struct jfs_dirent *) dirent_buf;
3294 while (jfs_dirents--) {
3295 filp->f_pos = jfs_dirent->position;
3296 if (filldir(dirent, jfs_dirent->name,
3297 jfs_dirent->name_len, filp->f_pos,
3298 jfs_dirent->ino, DT_UNKNOWN))
3299 goto out;
3300 jfs_dirent = next_jfs_dirent(jfs_dirent);
3301 }
3302
3303 if (fix_page) {
3304 add_missing_indices(ip, bn);
3305 page_fixed = 1;
3306 }
3307
3308 if (!overflow && (bn == 0)) {
3309 filp->f_pos = DIREND;
3310 break;
3311 }
3312
3313 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3314 if (rc) {
3315 free_page(dirent_buf);
3316 return rc;
3317 }
3318 }
3319
3320 out:
3321 free_page(dirent_buf);
3322
3323 return rc;
3324}
3325
3326
3327/*
3328 * dtReadFirst()
3329 *
3330 * function: get the leftmost page of the directory
3331 */
3332static int dtReadFirst(struct inode *ip, struct btstack * btstack)
3333{
3334 int rc = 0;
3335 s64 bn;
3336 int psize = 288; /* initial in-line directory */
3337 struct metapage *mp;
3338 dtpage_t *p;
3339 s8 *stbl;
3340 struct btframe *btsp;
3341 pxd_t *xd;
3342
3343 BT_CLR(btstack); /* reset stack */
3344
3345 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05003346 * descend leftmost path of the tree
Linus Torvalds1da177e2005-04-16 15:20:36 -07003347 *
3348 * by convention, root bn = 0.
3349 */
3350 for (bn = 0;;) {
3351 DT_GETPAGE(ip, bn, mp, psize, p, rc);
3352 if (rc)
3353 return rc;
3354
3355 /*
3356 * leftmost leaf page
3357 */
3358 if (p->header.flag & BT_LEAF) {
3359 /* return leftmost entry */
3360 btsp = btstack->top;
3361 btsp->bn = bn;
3362 btsp->index = 0;
3363 btsp->mp = mp;
3364
3365 return 0;
3366 }
3367
3368 /*
3369 * descend down to leftmost child page
3370 */
3371 if (BT_STACK_FULL(btstack)) {
3372 DT_PUTPAGE(mp);
3373 jfs_error(ip->i_sb, "dtReadFirst: btstack overrun");
3374 BT_STACK_DUMP(btstack);
3375 return -EIO;
3376 }
3377 /* push (bn, index) of the parent page/entry */
3378 BT_PUSH(btstack, bn, 0);
3379
3380 /* get the leftmost entry */
3381 stbl = DT_GETSTBL(p);
3382 xd = (pxd_t *) & p->slot[stbl[0]];
3383
3384 /* get the child page block address */
3385 bn = addressPXD(xd);
3386 psize = lengthPXD(xd) << JFS_SBI(ip->i_sb)->l2bsize;
3387
3388 /* unpin the parent page */
3389 DT_PUTPAGE(mp);
3390 }
3391}
3392
3393
3394/*
3395 * dtReadNext()
3396 *
3397 * function: get the page of the specified offset (pn:index)
3398 *
3399 * return: if (offset > eof), bn = -1;
3400 *
3401 * note: if index > nextindex of the target leaf page,
3402 * start with 1st entry of next leaf page;
3403 */
3404static int dtReadNext(struct inode *ip, loff_t * offset,
3405 struct btstack * btstack)
3406{
3407 int rc = 0;
3408 struct dtoffset {
3409 s16 pn;
3410 s16 index;
3411 s32 unused;
3412 } *dtoffset = (struct dtoffset *) offset;
3413 s64 bn;
3414 struct metapage *mp;
3415 dtpage_t *p;
3416 int index;
3417 int pn;
3418 s8 *stbl;
3419 struct btframe *btsp, *parent;
3420 pxd_t *xd;
3421
3422 /*
3423 * get leftmost leaf page pinned
3424 */
3425 if ((rc = dtReadFirst(ip, btstack)))
3426 return rc;
3427
3428 /* get leaf page */
3429 DT_GETSEARCH(ip, btstack->top, bn, mp, p, index);
3430
3431 /* get the start offset (pn:index) */
3432 pn = dtoffset->pn - 1; /* Now pn = 0 represents leftmost leaf */
3433 index = dtoffset->index;
3434
3435 /* start at leftmost page ? */
3436 if (pn == 0) {
3437 /* offset beyond eof ? */
3438 if (index < p->header.nextindex)
3439 goto out;
3440
3441 if (p->header.flag & BT_ROOT) {
3442 bn = -1;
3443 goto out;
3444 }
3445
3446 /* start with 1st entry of next leaf page */
3447 dtoffset->pn++;
3448 dtoffset->index = index = 0;
3449 goto a;
3450 }
3451
3452 /* start at non-leftmost page: scan parent pages for large pn */
3453 if (p->header.flag & BT_ROOT) {
3454 bn = -1;
3455 goto out;
3456 }
3457
3458 /* start after next leaf page ? */
3459 if (pn > 1)
3460 goto b;
3461
3462 /* get leaf page pn = 1 */
3463 a:
3464 bn = le64_to_cpu(p->header.next);
3465
3466 /* unpin leaf page */
3467 DT_PUTPAGE(mp);
3468
3469 /* offset beyond eof ? */
3470 if (bn == 0) {
3471 bn = -1;
3472 goto out;
3473 }
3474
3475 goto c;
3476
3477 /*
3478 * scan last internal page level to get target leaf page
3479 */
3480 b:
3481 /* unpin leftmost leaf page */
3482 DT_PUTPAGE(mp);
3483
3484 /* get left most parent page */
3485 btsp = btstack->top;
3486 parent = btsp - 1;
3487 bn = parent->bn;
3488 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3489 if (rc)
3490 return rc;
3491
3492 /* scan parent pages at last internal page level */
3493 while (pn >= p->header.nextindex) {
3494 pn -= p->header.nextindex;
3495
3496 /* get next parent page address */
3497 bn = le64_to_cpu(p->header.next);
3498
3499 /* unpin current parent page */
3500 DT_PUTPAGE(mp);
3501
3502 /* offset beyond eof ? */
3503 if (bn == 0) {
3504 bn = -1;
3505 goto out;
3506 }
3507
3508 /* get next parent page */
3509 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3510 if (rc)
3511 return rc;
3512
3513 /* update parent page stack frame */
3514 parent->bn = bn;
3515 }
3516
3517 /* get leaf page address */
3518 stbl = DT_GETSTBL(p);
3519 xd = (pxd_t *) & p->slot[stbl[pn]];
3520 bn = addressPXD(xd);
3521
3522 /* unpin parent page */
3523 DT_PUTPAGE(mp);
3524
3525 /*
3526 * get target leaf page
3527 */
3528 c:
3529 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3530 if (rc)
3531 return rc;
3532
3533 /*
3534 * leaf page has been completed:
3535 * start with 1st entry of next leaf page
3536 */
3537 if (index >= p->header.nextindex) {
3538 bn = le64_to_cpu(p->header.next);
3539
3540 /* unpin leaf page */
3541 DT_PUTPAGE(mp);
3542
3543 /* offset beyond eof ? */
3544 if (bn == 0) {
3545 bn = -1;
3546 goto out;
3547 }
3548
3549 /* get next leaf page */
3550 DT_GETPAGE(ip, bn, mp, PSIZE, p, rc);
3551 if (rc)
3552 return rc;
3553
3554 /* start with 1st entry of next leaf page */
3555 dtoffset->pn++;
3556 dtoffset->index = 0;
3557 }
3558
3559 out:
3560 /* return target leaf page pinned */
3561 btsp = btstack->top;
3562 btsp->bn = bn;
3563 btsp->index = dtoffset->index;
3564 btsp->mp = mp;
3565
3566 return 0;
3567}
3568
3569
3570/*
3571 * dtCompare()
3572 *
3573 * function: compare search key with an internal entry
3574 *
3575 * return:
3576 * < 0 if k is < record
3577 * = 0 if k is = record
3578 * > 0 if k is > record
3579 */
3580static int dtCompare(struct component_name * key, /* search key */
3581 dtpage_t * p, /* directory page */
3582 int si)
3583{ /* entry slot index */
3584 wchar_t *kname;
3585 __le16 *name;
3586 int klen, namlen, len, rc;
3587 struct idtentry *ih;
3588 struct dtslot *t;
3589
3590 /*
3591 * force the left-most key on internal pages, at any level of
3592 * the tree, to be less than any search key.
3593 * this obviates having to update the leftmost key on an internal
3594 * page when the user inserts a new key in the tree smaller than
3595 * anything that has been stored.
3596 *
3597 * (? if/when dtSearch() narrows down to 1st entry (index = 0),
3598 * at any internal page at any level of the tree,
3599 * it descends to child of the entry anyway -
3600 * ? make the entry as min size dummy entry)
3601 *
3602 * if (e->index == 0 && h->prevpg == P_INVALID && !(h->flags & BT_LEAF))
3603 * return (1);
3604 */
3605
3606 kname = key->name;
3607 klen = key->namlen;
3608
3609 ih = (struct idtentry *) & p->slot[si];
3610 si = ih->next;
3611 name = ih->name;
3612 namlen = ih->namlen;
3613 len = min(namlen, DTIHDRDATALEN);
3614
3615 /* compare with head/only segment */
3616 len = min(klen, len);
3617 if ((rc = UniStrncmp_le(kname, name, len)))
3618 return rc;
3619
3620 klen -= len;
3621 namlen -= len;
3622
3623 /* compare with additional segment(s) */
3624 kname += len;
3625 while (klen > 0 && namlen > 0) {
3626 /* compare with next name segment */
3627 t = (struct dtslot *) & p->slot[si];
3628 len = min(namlen, DTSLOTDATALEN);
3629 len = min(klen, len);
3630 name = t->name;
3631 if ((rc = UniStrncmp_le(kname, name, len)))
3632 return rc;
3633
3634 klen -= len;
3635 namlen -= len;
3636 kname += len;
3637 si = t->next;
3638 }
3639
3640 return (klen - namlen);
3641}
3642
3643
3644
3645
3646/*
3647 * ciCompare()
3648 *
3649 * function: compare search key with an (leaf/internal) entry
3650 *
3651 * return:
3652 * < 0 if k is < record
3653 * = 0 if k is = record
3654 * > 0 if k is > record
3655 */
3656static int ciCompare(struct component_name * key, /* search key */
3657 dtpage_t * p, /* directory page */
3658 int si, /* entry slot index */
3659 int flag)
3660{
3661 wchar_t *kname, x;
3662 __le16 *name;
3663 int klen, namlen, len, rc;
3664 struct ldtentry *lh;
3665 struct idtentry *ih;
3666 struct dtslot *t;
3667 int i;
3668
3669 /*
3670 * force the left-most key on internal pages, at any level of
3671 * the tree, to be less than any search key.
3672 * this obviates having to update the leftmost key on an internal
3673 * page when the user inserts a new key in the tree smaller than
3674 * anything that has been stored.
3675 *
3676 * (? if/when dtSearch() narrows down to 1st entry (index = 0),
3677 * at any internal page at any level of the tree,
3678 * it descends to child of the entry anyway -
3679 * ? make the entry as min size dummy entry)
3680 *
3681 * if (e->index == 0 && h->prevpg == P_INVALID && !(h->flags & BT_LEAF))
3682 * return (1);
3683 */
3684
3685 kname = key->name;
3686 klen = key->namlen;
3687
3688 /*
3689 * leaf page entry
3690 */
3691 if (p->header.flag & BT_LEAF) {
3692 lh = (struct ldtentry *) & p->slot[si];
3693 si = lh->next;
3694 name = lh->name;
3695 namlen = lh->namlen;
3696 if (flag & JFS_DIR_INDEX)
3697 len = min(namlen, DTLHDRDATALEN);
3698 else
3699 len = min(namlen, DTLHDRDATALEN_LEGACY);
3700 }
3701 /*
3702 * internal page entry
3703 */
3704 else {
3705 ih = (struct idtentry *) & p->slot[si];
3706 si = ih->next;
3707 name = ih->name;
3708 namlen = ih->namlen;
3709 len = min(namlen, DTIHDRDATALEN);
3710 }
3711
3712 /* compare with head/only segment */
3713 len = min(klen, len);
3714 for (i = 0; i < len; i++, kname++, name++) {
3715 /* only uppercase if case-insensitive support is on */
3716 if ((flag & JFS_OS2) == JFS_OS2)
3717 x = UniToupper(le16_to_cpu(*name));
3718 else
3719 x = le16_to_cpu(*name);
3720 if ((rc = *kname - x))
3721 return rc;
3722 }
3723
3724 klen -= len;
3725 namlen -= len;
3726
3727 /* compare with additional segment(s) */
3728 while (klen > 0 && namlen > 0) {
3729 /* compare with next name segment */
3730 t = (struct dtslot *) & p->slot[si];
3731 len = min(namlen, DTSLOTDATALEN);
3732 len = min(klen, len);
3733 name = t->name;
3734 for (i = 0; i < len; i++, kname++, name++) {
3735 /* only uppercase if case-insensitive support is on */
3736 if ((flag & JFS_OS2) == JFS_OS2)
3737 x = UniToupper(le16_to_cpu(*name));
3738 else
3739 x = le16_to_cpu(*name);
3740
3741 if ((rc = *kname - x))
3742 return rc;
3743 }
3744
3745 klen -= len;
3746 namlen -= len;
3747 si = t->next;
3748 }
3749
3750 return (klen - namlen);
3751}
3752
3753
3754/*
3755 * ciGetLeafPrefixKey()
3756 *
3757 * function: compute prefix of suffix compression
3758 * from two adjacent leaf entries
3759 * across page boundary
3760 *
3761 * return: non-zero on error
Dave Kleikamp63f83c92006-10-02 09:55:27 -05003762 *
Linus Torvalds1da177e2005-04-16 15:20:36 -07003763 */
3764static int ciGetLeafPrefixKey(dtpage_t * lp, int li, dtpage_t * rp,
3765 int ri, struct component_name * key, int flag)
3766{
3767 int klen, namlen;
3768 wchar_t *pl, *pr, *kname;
3769 struct component_name lkey;
3770 struct component_name rkey;
3771
Robert P. J. Day5cbded52006-12-13 00:35:56 -08003772 lkey.name = kmalloc((JFS_NAME_MAX + 1) * sizeof(wchar_t),
Linus Torvalds1da177e2005-04-16 15:20:36 -07003773 GFP_KERNEL);
3774 if (lkey.name == NULL)
Akinobu Mita087387f2006-09-14 09:22:38 -05003775 return -ENOMEM;
Linus Torvalds1da177e2005-04-16 15:20:36 -07003776
Robert P. J. Day5cbded52006-12-13 00:35:56 -08003777 rkey.name = kmalloc((JFS_NAME_MAX + 1) * sizeof(wchar_t),
Linus Torvalds1da177e2005-04-16 15:20:36 -07003778 GFP_KERNEL);
3779 if (rkey.name == NULL) {
3780 kfree(lkey.name);
Akinobu Mita087387f2006-09-14 09:22:38 -05003781 return -ENOMEM;
Linus Torvalds1da177e2005-04-16 15:20:36 -07003782 }
3783
3784 /* get left and right key */
3785 dtGetKey(lp, li, &lkey, flag);
3786 lkey.name[lkey.namlen] = 0;
3787
3788 if ((flag & JFS_OS2) == JFS_OS2)
3789 ciToUpper(&lkey);
3790
3791 dtGetKey(rp, ri, &rkey, flag);
3792 rkey.name[rkey.namlen] = 0;
3793
3794
3795 if ((flag & JFS_OS2) == JFS_OS2)
3796 ciToUpper(&rkey);
3797
3798 /* compute prefix */
3799 klen = 0;
3800 kname = key->name;
3801 namlen = min(lkey.namlen, rkey.namlen);
3802 for (pl = lkey.name, pr = rkey.name;
3803 namlen; pl++, pr++, namlen--, klen++, kname++) {
3804 *kname = *pr;
3805 if (*pl != *pr) {
3806 key->namlen = klen + 1;
3807 goto free_names;
3808 }
3809 }
3810
3811 /* l->namlen <= r->namlen since l <= r */
3812 if (lkey.namlen < rkey.namlen) {
3813 *kname = *pr;
3814 key->namlen = klen + 1;
3815 } else /* l->namelen == r->namelen */
3816 key->namlen = klen;
3817
3818free_names:
3819 kfree(lkey.name);
3820 kfree(rkey.name);
3821 return 0;
3822}
3823
3824
3825
3826/*
3827 * dtGetKey()
3828 *
3829 * function: get key of the entry
3830 */
3831static void dtGetKey(dtpage_t * p, int i, /* entry index */
3832 struct component_name * key, int flag)
3833{
3834 int si;
3835 s8 *stbl;
3836 struct ldtentry *lh;
3837 struct idtentry *ih;
3838 struct dtslot *t;
3839 int namlen, len;
3840 wchar_t *kname;
3841 __le16 *name;
3842
3843 /* get entry */
3844 stbl = DT_GETSTBL(p);
3845 si = stbl[i];
3846 if (p->header.flag & BT_LEAF) {
3847 lh = (struct ldtentry *) & p->slot[si];
3848 si = lh->next;
3849 namlen = lh->namlen;
3850 name = lh->name;
3851 if (flag & JFS_DIR_INDEX)
3852 len = min(namlen, DTLHDRDATALEN);
3853 else
3854 len = min(namlen, DTLHDRDATALEN_LEGACY);
3855 } else {
3856 ih = (struct idtentry *) & p->slot[si];
3857 si = ih->next;
3858 namlen = ih->namlen;
3859 name = ih->name;
3860 len = min(namlen, DTIHDRDATALEN);
3861 }
3862
3863 key->namlen = namlen;
3864 kname = key->name;
3865
3866 /*
3867 * move head/only segment
3868 */
3869 UniStrncpy_from_le(kname, name, len);
3870
3871 /*
3872 * move additional segment(s)
3873 */
3874 while (si >= 0) {
3875 /* get next segment */
3876 t = &p->slot[si];
3877 kname += len;
3878 namlen -= len;
3879 len = min(namlen, DTSLOTDATALEN);
3880 UniStrncpy_from_le(kname, t->name, len);
3881
3882 si = t->next;
3883 }
3884}
3885
3886
3887/*
3888 * dtInsertEntry()
3889 *
3890 * function: allocate free slot(s) and
3891 * write a leaf/internal entry
3892 *
3893 * return: entry slot index
3894 */
3895static void dtInsertEntry(dtpage_t * p, int index, struct component_name * key,
3896 ddata_t * data, struct dt_lock ** dtlock)
3897{
3898 struct dtslot *h, *t;
3899 struct ldtentry *lh = NULL;
3900 struct idtentry *ih = NULL;
3901 int hsi, fsi, klen, len, nextindex;
3902 wchar_t *kname;
3903 __le16 *name;
3904 s8 *stbl;
3905 pxd_t *xd;
3906 struct dt_lock *dtlck = *dtlock;
3907 struct lv *lv;
3908 int xsi, n;
3909 s64 bn = 0;
3910 struct metapage *mp = NULL;
3911
3912 klen = key->namlen;
3913 kname = key->name;
3914
3915 /* allocate a free slot */
3916 hsi = fsi = p->header.freelist;
3917 h = &p->slot[fsi];
3918 p->header.freelist = h->next;
3919 --p->header.freecnt;
3920
3921 /* open new linelock */
3922 if (dtlck->index >= dtlck->maxcnt)
3923 dtlck = (struct dt_lock *) txLinelock(dtlck);
3924
3925 lv = & dtlck->lv[dtlck->index];
3926 lv->offset = hsi;
3927
3928 /* write head/only segment */
3929 if (p->header.flag & BT_LEAF) {
3930 lh = (struct ldtentry *) h;
3931 lh->next = h->next;
3932 lh->inumber = cpu_to_le32(data->leaf.ino);
3933 lh->namlen = klen;
3934 name = lh->name;
3935 if (data->leaf.ip) {
3936 len = min(klen, DTLHDRDATALEN);
3937 if (!(p->header.flag & BT_ROOT))
3938 bn = addressPXD(&p->header.self);
3939 lh->index = cpu_to_le32(add_index(data->leaf.tid,
3940 data->leaf.ip,
3941 bn, index));
3942 } else
3943 len = min(klen, DTLHDRDATALEN_LEGACY);
3944 } else {
3945 ih = (struct idtentry *) h;
3946 ih->next = h->next;
3947 xd = (pxd_t *) ih;
3948 *xd = data->xd;
3949 ih->namlen = klen;
3950 name = ih->name;
3951 len = min(klen, DTIHDRDATALEN);
3952 }
3953
3954 UniStrncpy_to_le(name, kname, len);
3955
3956 n = 1;
3957 xsi = hsi;
3958
3959 /* write additional segment(s) */
3960 t = h;
3961 klen -= len;
3962 while (klen) {
3963 /* get free slot */
3964 fsi = p->header.freelist;
3965 t = &p->slot[fsi];
3966 p->header.freelist = t->next;
3967 --p->header.freecnt;
3968
3969 /* is next slot contiguous ? */
3970 if (fsi != xsi + 1) {
3971 /* close current linelock */
3972 lv->length = n;
3973 dtlck->index++;
3974
3975 /* open new linelock */
3976 if (dtlck->index < dtlck->maxcnt)
3977 lv++;
3978 else {
3979 dtlck = (struct dt_lock *) txLinelock(dtlck);
3980 lv = & dtlck->lv[0];
3981 }
3982
3983 lv->offset = fsi;
3984 n = 0;
3985 }
3986
3987 kname += len;
3988 len = min(klen, DTSLOTDATALEN);
3989 UniStrncpy_to_le(t->name, kname, len);
3990
3991 n++;
3992 xsi = fsi;
3993 klen -= len;
3994 }
3995
3996 /* close current linelock */
3997 lv->length = n;
3998 dtlck->index++;
3999
4000 *dtlock = dtlck;
4001
4002 /* terminate last/only segment */
4003 if (h == t) {
4004 /* single segment entry */
4005 if (p->header.flag & BT_LEAF)
4006 lh->next = -1;
4007 else
4008 ih->next = -1;
4009 } else
4010 /* multi-segment entry */
4011 t->next = -1;
4012
4013 /* if insert into middle, shift right succeeding entries in stbl */
4014 stbl = DT_GETSTBL(p);
4015 nextindex = p->header.nextindex;
4016 if (index < nextindex) {
4017 memmove(stbl + index + 1, stbl + index, nextindex - index);
4018
4019 if ((p->header.flag & BT_LEAF) && data->leaf.ip) {
4020 s64 lblock;
4021
4022 /*
4023 * Need to update slot number for entries that moved
4024 * in the stbl
4025 */
4026 mp = NULL;
4027 for (n = index + 1; n <= nextindex; n++) {
4028 lh = (struct ldtentry *) & (p->slot[stbl[n]]);
4029 modify_index(data->leaf.tid, data->leaf.ip,
4030 le32_to_cpu(lh->index), bn, n,
4031 &mp, &lblock);
4032 }
4033 if (mp)
4034 release_metapage(mp);
4035 }
4036 }
4037
4038 stbl[index] = hsi;
4039
4040 /* advance next available entry index of stbl */
4041 ++p->header.nextindex;
4042}
4043
4044
4045/*
4046 * dtMoveEntry()
4047 *
4048 * function: move entries from split/left page to new/right page
4049 *
4050 * nextindex of dst page and freelist/freecnt of both pages
4051 * are updated.
4052 */
4053static void dtMoveEntry(dtpage_t * sp, int si, dtpage_t * dp,
4054 struct dt_lock ** sdtlock, struct dt_lock ** ddtlock,
4055 int do_index)
4056{
4057 int ssi, next; /* src slot index */
4058 int di; /* dst entry index */
4059 int dsi; /* dst slot index */
4060 s8 *sstbl, *dstbl; /* sorted entry table */
4061 int snamlen, len;
4062 struct ldtentry *slh, *dlh = NULL;
4063 struct idtentry *sih, *dih = NULL;
4064 struct dtslot *h, *s, *d;
4065 struct dt_lock *sdtlck = *sdtlock, *ddtlck = *ddtlock;
4066 struct lv *slv, *dlv;
4067 int xssi, ns, nd;
4068 int sfsi;
4069
4070 sstbl = (s8 *) & sp->slot[sp->header.stblindex];
4071 dstbl = (s8 *) & dp->slot[dp->header.stblindex];
4072
4073 dsi = dp->header.freelist; /* first (whole page) free slot */
4074 sfsi = sp->header.freelist;
4075
4076 /* linelock destination entry slot */
4077 dlv = & ddtlck->lv[ddtlck->index];
4078 dlv->offset = dsi;
4079
4080 /* linelock source entry slot */
4081 slv = & sdtlck->lv[sdtlck->index];
4082 slv->offset = sstbl[si];
4083 xssi = slv->offset - 1;
4084
4085 /*
4086 * move entries
4087 */
4088 ns = nd = 0;
4089 for (di = 0; si < sp->header.nextindex; si++, di++) {
4090 ssi = sstbl[si];
4091 dstbl[di] = dsi;
4092
4093 /* is next slot contiguous ? */
4094 if (ssi != xssi + 1) {
4095 /* close current linelock */
4096 slv->length = ns;
4097 sdtlck->index++;
4098
4099 /* open new linelock */
4100 if (sdtlck->index < sdtlck->maxcnt)
4101 slv++;
4102 else {
4103 sdtlck = (struct dt_lock *) txLinelock(sdtlck);
4104 slv = & sdtlck->lv[0];
4105 }
4106
4107 slv->offset = ssi;
4108 ns = 0;
4109 }
4110
4111 /*
4112 * move head/only segment of an entry
4113 */
4114 /* get dst slot */
4115 h = d = &dp->slot[dsi];
4116
4117 /* get src slot and move */
4118 s = &sp->slot[ssi];
4119 if (sp->header.flag & BT_LEAF) {
4120 /* get source entry */
4121 slh = (struct ldtentry *) s;
4122 dlh = (struct ldtentry *) h;
4123 snamlen = slh->namlen;
4124
4125 if (do_index) {
4126 len = min(snamlen, DTLHDRDATALEN);
4127 dlh->index = slh->index; /* little-endian */
4128 } else
4129 len = min(snamlen, DTLHDRDATALEN_LEGACY);
4130
4131 memcpy(dlh, slh, 6 + len * 2);
4132
4133 next = slh->next;
4134
4135 /* update dst head/only segment next field */
4136 dsi++;
4137 dlh->next = dsi;
4138 } else {
4139 sih = (struct idtentry *) s;
4140 snamlen = sih->namlen;
4141
4142 len = min(snamlen, DTIHDRDATALEN);
4143 dih = (struct idtentry *) h;
4144 memcpy(dih, sih, 10 + len * 2);
4145 next = sih->next;
4146
4147 dsi++;
4148 dih->next = dsi;
4149 }
4150
4151 /* free src head/only segment */
4152 s->next = sfsi;
4153 s->cnt = 1;
4154 sfsi = ssi;
4155
4156 ns++;
4157 nd++;
4158 xssi = ssi;
4159
4160 /*
4161 * move additional segment(s) of the entry
4162 */
4163 snamlen -= len;
4164 while ((ssi = next) >= 0) {
4165 /* is next slot contiguous ? */
4166 if (ssi != xssi + 1) {
4167 /* close current linelock */
4168 slv->length = ns;
4169 sdtlck->index++;
4170
4171 /* open new linelock */
4172 if (sdtlck->index < sdtlck->maxcnt)
4173 slv++;
4174 else {
4175 sdtlck =
4176 (struct dt_lock *)
4177 txLinelock(sdtlck);
4178 slv = & sdtlck->lv[0];
4179 }
4180
4181 slv->offset = ssi;
4182 ns = 0;
4183 }
4184
4185 /* get next source segment */
4186 s = &sp->slot[ssi];
4187
4188 /* get next destination free slot */
4189 d++;
4190
4191 len = min(snamlen, DTSLOTDATALEN);
4192 UniStrncpy_le(d->name, s->name, len);
4193
4194 ns++;
4195 nd++;
4196 xssi = ssi;
4197
4198 dsi++;
4199 d->next = dsi;
4200
4201 /* free source segment */
4202 next = s->next;
4203 s->next = sfsi;
4204 s->cnt = 1;
4205 sfsi = ssi;
4206
4207 snamlen -= len;
4208 } /* end while */
4209
4210 /* terminate dst last/only segment */
4211 if (h == d) {
4212 /* single segment entry */
4213 if (dp->header.flag & BT_LEAF)
4214 dlh->next = -1;
4215 else
4216 dih->next = -1;
4217 } else
4218 /* multi-segment entry */
4219 d->next = -1;
4220 } /* end for */
4221
4222 /* close current linelock */
4223 slv->length = ns;
4224 sdtlck->index++;
4225 *sdtlock = sdtlck;
4226
4227 dlv->length = nd;
4228 ddtlck->index++;
4229 *ddtlock = ddtlck;
4230
4231 /* update source header */
4232 sp->header.freelist = sfsi;
4233 sp->header.freecnt += nd;
4234
4235 /* update destination header */
4236 dp->header.nextindex = di;
4237
4238 dp->header.freelist = dsi;
4239 dp->header.freecnt -= nd;
4240}
4241
4242
4243/*
4244 * dtDeleteEntry()
4245 *
4246 * function: free a (leaf/internal) entry
4247 *
4248 * log freelist header, stbl, and each segment slot of entry
4249 * (even though last/only segment next field is modified,
4250 * physical image logging requires all segment slots of
4251 * the entry logged to avoid applying previous updates
4252 * to the same slots)
4253 */
4254static void dtDeleteEntry(dtpage_t * p, int fi, struct dt_lock ** dtlock)
4255{
4256 int fsi; /* free entry slot index */
4257 s8 *stbl;
4258 struct dtslot *t;
4259 int si, freecnt;
4260 struct dt_lock *dtlck = *dtlock;
4261 struct lv *lv;
4262 int xsi, n;
4263
4264 /* get free entry slot index */
4265 stbl = DT_GETSTBL(p);
4266 fsi = stbl[fi];
4267
4268 /* open new linelock */
4269 if (dtlck->index >= dtlck->maxcnt)
4270 dtlck = (struct dt_lock *) txLinelock(dtlck);
4271 lv = & dtlck->lv[dtlck->index];
4272
4273 lv->offset = fsi;
4274
4275 /* get the head/only segment */
4276 t = &p->slot[fsi];
4277 if (p->header.flag & BT_LEAF)
4278 si = ((struct ldtentry *) t)->next;
4279 else
4280 si = ((struct idtentry *) t)->next;
4281 t->next = si;
4282 t->cnt = 1;
4283
4284 n = freecnt = 1;
4285 xsi = fsi;
4286
4287 /* find the last/only segment */
4288 while (si >= 0) {
4289 /* is next slot contiguous ? */
4290 if (si != xsi + 1) {
4291 /* close current linelock */
4292 lv->length = n;
4293 dtlck->index++;
4294
4295 /* open new linelock */
4296 if (dtlck->index < dtlck->maxcnt)
4297 lv++;
4298 else {
4299 dtlck = (struct dt_lock *) txLinelock(dtlck);
4300 lv = & dtlck->lv[0];
4301 }
4302
4303 lv->offset = si;
4304 n = 0;
4305 }
4306
4307 n++;
4308 xsi = si;
4309 freecnt++;
4310
4311 t = &p->slot[si];
4312 t->cnt = 1;
4313 si = t->next;
4314 }
4315
4316 /* close current linelock */
4317 lv->length = n;
4318 dtlck->index++;
4319
4320 *dtlock = dtlck;
4321
4322 /* update freelist */
4323 t->next = p->header.freelist;
4324 p->header.freelist = fsi;
4325 p->header.freecnt += freecnt;
4326
4327 /* if delete from middle,
4328 * shift left the succedding entries in the stbl
4329 */
4330 si = p->header.nextindex;
4331 if (fi < si - 1)
4332 memmove(&stbl[fi], &stbl[fi + 1], si - fi - 1);
4333
4334 p->header.nextindex--;
4335}
4336
4337
4338/*
4339 * dtTruncateEntry()
4340 *
4341 * function: truncate a (leaf/internal) entry
4342 *
4343 * log freelist header, stbl, and each segment slot of entry
4344 * (even though last/only segment next field is modified,
4345 * physical image logging requires all segment slots of
4346 * the entry logged to avoid applying previous updates
4347 * to the same slots)
4348 */
4349static void dtTruncateEntry(dtpage_t * p, int ti, struct dt_lock ** dtlock)
4350{
4351 int tsi; /* truncate entry slot index */
4352 s8 *stbl;
4353 struct dtslot *t;
4354 int si, freecnt;
4355 struct dt_lock *dtlck = *dtlock;
4356 struct lv *lv;
4357 int fsi, xsi, n;
4358
4359 /* get free entry slot index */
4360 stbl = DT_GETSTBL(p);
4361 tsi = stbl[ti];
4362
4363 /* open new linelock */
4364 if (dtlck->index >= dtlck->maxcnt)
4365 dtlck = (struct dt_lock *) txLinelock(dtlck);
4366 lv = & dtlck->lv[dtlck->index];
4367
4368 lv->offset = tsi;
4369
4370 /* get the head/only segment */
4371 t = &p->slot[tsi];
4372 ASSERT(p->header.flag & BT_INTERNAL);
4373 ((struct idtentry *) t)->namlen = 0;
4374 si = ((struct idtentry *) t)->next;
4375 ((struct idtentry *) t)->next = -1;
4376
4377 n = 1;
4378 freecnt = 0;
4379 fsi = si;
4380 xsi = tsi;
4381
4382 /* find the last/only segment */
4383 while (si >= 0) {
4384 /* is next slot contiguous ? */
4385 if (si != xsi + 1) {
4386 /* close current linelock */
4387 lv->length = n;
4388 dtlck->index++;
4389
4390 /* open new linelock */
4391 if (dtlck->index < dtlck->maxcnt)
4392 lv++;
4393 else {
4394 dtlck = (struct dt_lock *) txLinelock(dtlck);
4395 lv = & dtlck->lv[0];
4396 }
4397
4398 lv->offset = si;
4399 n = 0;
4400 }
4401
4402 n++;
4403 xsi = si;
4404 freecnt++;
4405
4406 t = &p->slot[si];
4407 t->cnt = 1;
4408 si = t->next;
4409 }
4410
4411 /* close current linelock */
4412 lv->length = n;
4413 dtlck->index++;
4414
4415 *dtlock = dtlck;
4416
4417 /* update freelist */
4418 if (freecnt == 0)
4419 return;
4420 t->next = p->header.freelist;
4421 p->header.freelist = fsi;
4422 p->header.freecnt += freecnt;
4423}
4424
4425
4426/*
4427 * dtLinelockFreelist()
4428 */
4429static void dtLinelockFreelist(dtpage_t * p, /* directory page */
4430 int m, /* max slot index */
4431 struct dt_lock ** dtlock)
4432{
4433 int fsi; /* free entry slot index */
4434 struct dtslot *t;
4435 int si;
4436 struct dt_lock *dtlck = *dtlock;
4437 struct lv *lv;
4438 int xsi, n;
4439
4440 /* get free entry slot index */
4441 fsi = p->header.freelist;
4442
4443 /* open new linelock */
4444 if (dtlck->index >= dtlck->maxcnt)
4445 dtlck = (struct dt_lock *) txLinelock(dtlck);
4446 lv = & dtlck->lv[dtlck->index];
4447
4448 lv->offset = fsi;
4449
4450 n = 1;
4451 xsi = fsi;
4452
4453 t = &p->slot[fsi];
4454 si = t->next;
4455
4456 /* find the last/only segment */
4457 while (si < m && si >= 0) {
4458 /* is next slot contiguous ? */
4459 if (si != xsi + 1) {
4460 /* close current linelock */
4461 lv->length = n;
4462 dtlck->index++;
4463
4464 /* open new linelock */
4465 if (dtlck->index < dtlck->maxcnt)
4466 lv++;
4467 else {
4468 dtlck = (struct dt_lock *) txLinelock(dtlck);
4469 lv = & dtlck->lv[0];
4470 }
4471
4472 lv->offset = si;
4473 n = 0;
4474 }
4475
4476 n++;
4477 xsi = si;
4478
4479 t = &p->slot[si];
4480 si = t->next;
4481 }
4482
4483 /* close current linelock */
4484 lv->length = n;
4485 dtlck->index++;
4486
4487 *dtlock = dtlck;
4488}
4489
4490
4491/*
4492 * NAME: dtModify
4493 *
4494 * FUNCTION: Modify the inode number part of a directory entry
4495 *
4496 * PARAMETERS:
4497 * tid - Transaction id
4498 * ip - Inode of parent directory
4499 * key - Name of entry to be modified
4500 * orig_ino - Original inode number expected in entry
4501 * new_ino - New inode number to put into entry
4502 * flag - JFS_RENAME
4503 *
4504 * RETURNS:
4505 * -ESTALE - If entry found does not match orig_ino passed in
4506 * -ENOENT - If no entry can be found to match key
4507 * 0 - If successfully modified entry
4508 */
4509int dtModify(tid_t tid, struct inode *ip,
4510 struct component_name * key, ino_t * orig_ino, ino_t new_ino, int flag)
4511{
4512 int rc;
4513 s64 bn;
4514 struct metapage *mp;
4515 dtpage_t *p;
4516 int index;
4517 struct btstack btstack;
4518 struct tlock *tlck;
4519 struct dt_lock *dtlck;
4520 struct lv *lv;
4521 s8 *stbl;
4522 int entry_si; /* entry slot index */
4523 struct ldtentry *entry;
4524
4525 /*
Dave Kleikampf720e3b2007-06-06 15:28:35 -05004526 * search for the entry to modify:
Linus Torvalds1da177e2005-04-16 15:20:36 -07004527 *
4528 * dtSearch() returns (leaf page pinned, index at which to modify).
4529 */
4530 if ((rc = dtSearch(ip, key, orig_ino, &btstack, flag)))
4531 return rc;
4532
4533 /* retrieve search result */
4534 DT_GETSEARCH(ip, btstack.top, bn, mp, p, index);
4535
4536 BT_MARK_DIRTY(mp, ip);
4537 /*
4538 * acquire a transaction lock on the leaf page of named entry
4539 */
4540 tlck = txLock(tid, ip, mp, tlckDTREE | tlckENTRY);
4541 dtlck = (struct dt_lock *) & tlck->lock;
4542
4543 /* get slot index of the entry */
4544 stbl = DT_GETSTBL(p);
4545 entry_si = stbl[index];
4546
4547 /* linelock entry */
4548 ASSERT(dtlck->index == 0);
4549 lv = & dtlck->lv[0];
4550 lv->offset = entry_si;
4551 lv->length = 1;
4552 dtlck->index++;
4553
4554 /* get the head/only segment */
4555 entry = (struct ldtentry *) & p->slot[entry_si];
4556
4557 /* substitute the inode number of the entry */
4558 entry->inumber = cpu_to_le32(new_ino);
4559
4560 /* unpin the leaf page */
4561 DT_PUTPAGE(mp);
4562
4563 return 0;
4564}