2 * Copyright (c) 2007 The DragonFly Project. All rights reserved.
4 * This code is derived from software contributed to The DragonFly Project
5 * by Matthew Dillon <dillon@backplane.com>
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in
15 * the documentation and/or other materials provided with the
17 * 3. Neither the name of The DragonFly Project nor the names of its
18 * contributors may be used to endorse or promote products derived
19 * from this software without specific, prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
24 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
25 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
26 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
27 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
28 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
29 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
30 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
31 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * $DragonFly: src/sys/vfs/hammer/hammer_object.c,v 1.20 2008/01/17 05:06:09 dillon Exp $
39 static int hammer_mem_add(hammer_transaction_t trans,
40 hammer_record_t record);
41 static int hammer_mem_lookup(hammer_cursor_t cursor, hammer_inode_t ip);
42 static int hammer_mem_first(hammer_cursor_t cursor, hammer_inode_t ip);
45 * Red-black tree support.
48 hammer_rec_rb_compare(hammer_record_t rec1, hammer_record_t rec2)
50 if (rec1->rec.base.base.rec_type < rec2->rec.base.base.rec_type)
52 if (rec1->rec.base.base.rec_type > rec2->rec.base.base.rec_type)
55 if (rec1->rec.base.base.key < rec2->rec.base.base.key)
57 if (rec1->rec.base.base.key > rec2->rec.base.base.key)
60 if (rec1->rec.base.base.delete_tid == 0) {
61 if (rec2->rec.base.base.delete_tid == 0)
65 if (rec2->rec.base.base.delete_tid == 0)
68 if (rec1->rec.base.base.delete_tid < rec2->rec.base.base.delete_tid)
70 if (rec1->rec.base.base.delete_tid > rec2->rec.base.base.delete_tid)
76 hammer_rec_compare(hammer_base_elm_t info, hammer_record_t rec)
78 if (info->rec_type < rec->rec.base.base.rec_type)
80 if (info->rec_type > rec->rec.base.base.rec_type)
83 if (info->key < rec->rec.base.base.key)
85 if (info->key > rec->rec.base.base.key)
88 if (info->delete_tid == 0) {
89 if (rec->rec.base.base.delete_tid == 0)
93 if (rec->rec.base.base.delete_tid == 0)
95 if (info->delete_tid < rec->rec.base.base.delete_tid)
97 if (info->delete_tid > rec->rec.base.base.delete_tid)
103 * RB_SCAN comparison code for hammer_mem_first(). The argument order
104 * is reversed so the comparison result has to be negated. key_beg and
105 * key_end are both range-inclusive.
107 * The creation timestamp can cause hammer_rec_compare() to return -1 or +1.
108 * These do not stop the scan.
110 * Localized deletions are not cached in-memory.
114 hammer_rec_scan_cmp(hammer_record_t rec, void *data)
116 hammer_cursor_t cursor = data;
119 r = hammer_rec_compare(&cursor->key_beg, rec);
122 r = hammer_rec_compare(&cursor->key_end, rec);
128 RB_GENERATE(hammer_rec_rb_tree, hammer_record, rb_node, hammer_rec_rb_compare);
129 RB_GENERATE_XLOOKUP(hammer_rec_rb_tree, INFO, hammer_record, rb_node,
130 hammer_rec_compare, hammer_base_elm_t);
133 * Allocate a record for the caller to finish filling in. The record is
134 * returned referenced.
137 hammer_alloc_mem_record(hammer_inode_t ip)
139 hammer_record_t record;
141 ++hammer_count_records;
142 record = kmalloc(sizeof(*record), M_HAMMER, M_WAITOK|M_ZERO);
144 record->rec.base.base.btype = HAMMER_BTREE_TYPE_RECORD;
145 hammer_ref(&record->lock);
150 * Release a memory record. Records marked for deletion are immediately
151 * removed from the RB-Tree but otherwise left intact until the last ref
155 hammer_rel_mem_record(struct hammer_record *record)
157 hammer_unref(&record->lock);
158 if (record->flags & HAMMER_RECF_DELETED) {
159 if (record->flags & HAMMER_RECF_ONRBTREE) {
160 RB_REMOVE(hammer_rec_rb_tree, &record->ip->rec_tree,
162 record->flags &= ~HAMMER_RECF_ONRBTREE;
164 if (record->lock.refs == 0) {
165 if (record->flags & HAMMER_RECF_ALLOCDATA) {
166 --hammer_count_record_datas;
167 kfree(record->data, M_HAMMER);
168 record->flags &= ~HAMMER_RECF_ALLOCDATA;
171 --hammer_count_records;
172 kfree(record, M_HAMMER);
178 * Lookup an in-memory record given the key specified in the cursor. Works
179 * just like hammer_btree_lookup() but operates on an inode's in-memory
182 * The lookup must fail if the record is marked for deferred deletion.
186 hammer_mem_lookup(hammer_cursor_t cursor, hammer_inode_t ip)
191 hammer_rel_mem_record(cursor->iprec);
192 cursor->iprec = NULL;
195 hammer_rec_rb_tree_scan_info_done(&cursor->scan,
196 &cursor->ip->rec_tree);
199 hammer_rec_rb_tree_scan_info_link(&cursor->scan, &ip->rec_tree);
200 cursor->scan.node = NULL;
201 cursor->iprec = hammer_rec_rb_tree_RB_LOOKUP_INFO(
202 &ip->rec_tree, &cursor->key_beg);
203 if (cursor->iprec == NULL) {
206 hammer_ref(&cursor->iprec->lock);
213 * hammer_mem_first() - locate the first in-memory record matching the
216 * The RB_SCAN function we use is designed as a callback. We terminate it
217 * (return -1) as soon as we get a match.
221 hammer_rec_scan_callback(hammer_record_t rec, void *data)
223 hammer_cursor_t cursor = data;
226 * Skip if not visible due to our as-of TID
228 if (cursor->flags & HAMMER_CURSOR_ASOF) {
229 if (cursor->asof < rec->rec.base.base.create_tid)
231 if (rec->rec.base.base.delete_tid &&
232 cursor->asof >= rec->rec.base.base.delete_tid) {
238 * Return the first matching record and stop the scan
240 if (cursor->iprec == NULL) {
242 hammer_ref(&rec->lock);
250 hammer_mem_first(hammer_cursor_t cursor, hammer_inode_t ip)
253 hammer_rel_mem_record(cursor->iprec);
254 cursor->iprec = NULL;
257 hammer_rec_rb_tree_scan_info_done(&cursor->scan,
258 &cursor->ip->rec_tree);
261 hammer_rec_rb_tree_scan_info_link(&cursor->scan, &ip->rec_tree);
263 cursor->scan.node = NULL;
264 hammer_rec_rb_tree_RB_SCAN(&ip->rec_tree, hammer_rec_scan_cmp,
265 hammer_rec_scan_callback, cursor);
268 * Adjust scan.node and keep it linked into the RB-tree so we can
269 * hold the cursor through third party modifications of the RB-tree.
272 cursor->scan.node = hammer_rec_rb_tree_RB_NEXT(cursor->iprec);
279 hammer_mem_done(hammer_cursor_t cursor)
282 hammer_rec_rb_tree_scan_info_done(&cursor->scan,
283 &cursor->ip->rec_tree);
287 hammer_rel_mem_record(cursor->iprec);
288 cursor->iprec = NULL;
292 /************************************************************************
293 * HAMMER IN-MEMORY RECORD FUNCTIONS *
294 ************************************************************************
296 * These functions manipulate in-memory records. Such records typically
297 * exist prior to being committed to disk or indexed via the on-disk B-Tree.
301 * Add a directory entry (dip,ncp) which references inode (ip).
303 * Note that the low 32 bits of the namekey are set temporarily to create
304 * a unique in-memory record, and may be modified a second time when the
305 * record is synchronized to disk. In particular, the low 32 bits cannot be
306 * all 0's when synching to disk, which is not handled here.
309 hammer_ip_add_directory(struct hammer_transaction *trans,
310 struct hammer_inode *dip, struct namecache *ncp,
311 struct hammer_inode *ip)
313 hammer_record_t record;
317 record = hammer_alloc_mem_record(dip);
319 bytes = ncp->nc_nlen; /* NOTE: terminating \0 is NOT included */
320 if (++trans->hmp->namekey_iterator == 0)
321 ++trans->hmp->namekey_iterator;
323 record->rec.entry.base.base.obj_id = dip->obj_id;
324 record->rec.entry.base.base.key =
325 hammer_directory_namekey(ncp->nc_name, bytes);
326 record->rec.entry.base.base.key += trans->hmp->namekey_iterator;
327 record->rec.entry.base.base.create_tid = trans->tid;
328 record->rec.entry.base.base.rec_type = HAMMER_RECTYPE_DIRENTRY;
329 record->rec.entry.base.base.obj_type = ip->ino_rec.base.base.obj_type;
330 record->rec.entry.obj_id = ip->obj_id;
331 if (bytes <= sizeof(record->rec.entry.den_name)) {
332 record->data = (void *)record->rec.entry.den_name;
333 record->flags |= HAMMER_RECF_EMBEDDED_DATA;
335 ++hammer_count_record_datas;
336 record->data = kmalloc(bytes, M_HAMMER, M_WAITOK);
337 record->flags |= HAMMER_RECF_ALLOCDATA;
339 bcopy(ncp->nc_name, record->data, bytes);
340 record->rec.entry.base.data_len = bytes;
341 ++ip->ino_rec.ino_nlinks;
342 hammer_modify_inode(trans, ip, HAMMER_INODE_RDIRTY);
343 error = hammer_mem_add(trans, record);
348 * Delete the directory entry and update the inode link count. The
349 * cursor must be seeked to the directory entry record being deleted.
351 * NOTE: HAMMER_CURSOR_DELETE may not have been set. XXX remove flag.
354 hammer_ip_del_directory(struct hammer_transaction *trans,
355 hammer_cursor_t cursor, struct hammer_inode *dip,
356 struct hammer_inode *ip)
360 error = hammer_ip_delete_record(cursor, trans->tid);
363 * One less link. The file may still be open in the OS even after
364 * all links have gone away so we only try to sync if the OS has
365 * no references and nlinks falls to 0.
368 --ip->ino_rec.ino_nlinks;
369 hammer_modify_inode(trans, ip, HAMMER_INODE_RDIRTY);
370 if (ip->ino_rec.ino_nlinks == 0 &&
371 (ip->vp == NULL || (ip->vp->v_flag & VINACTIVE))) {
372 hammer_sync_inode(ip, MNT_NOWAIT, 1);
380 * Add a record to an inode.
382 * The caller must allocate the record with hammer_alloc_mem_record(ip) and
383 * initialize the following additional fields:
385 * record->rec.entry.base.base.key
386 * record->rec.entry.base.base.rec_type
387 * record->rec.entry.base.base.data_len
388 * record->data (a copy will be kmalloc'd if not embedded)
391 hammer_ip_add_record(struct hammer_transaction *trans, hammer_record_t record)
393 hammer_inode_t ip = record->ip;
398 record->rec.base.base.obj_id = ip->obj_id;
399 record->rec.base.base.create_tid = trans->tid;
400 record->rec.base.base.obj_type = ip->ino_rec.base.base.obj_type;
401 bytes = record->rec.base.data_len;
404 if ((char *)record->data < (char *)&record->rec ||
405 (char *)record->data >= (char *)(&record->rec + 1)) {
406 ++hammer_count_record_datas;
407 data = kmalloc(bytes, M_HAMMER, M_WAITOK);
408 record->flags |= HAMMER_RECF_ALLOCDATA;
409 bcopy(record->data, data, bytes);
412 record->flags |= HAMMER_RECF_EMBEDDED_DATA;
415 hammer_modify_inode(trans, ip, HAMMER_INODE_RDIRTY);
416 error = hammer_mem_add(trans, record);
421 * Sync data from a buffer cache buffer (typically) to the filesystem. This
422 * is called via the strategy called from a cached data source. This code
423 * is responsible for actually writing a data record out to the disk.
426 hammer_ip_sync_data(hammer_transaction_t trans, hammer_inode_t ip,
427 int64_t offset, void *data, int bytes,
428 struct hammer_cursor **spike)
430 struct hammer_cursor cursor;
431 hammer_record_ondisk_t rec;
432 union hammer_btree_elm elm;
436 error = hammer_init_cursor_hmp(&cursor, &ip->cache[0], ip->hmp);
439 cursor.key_beg.obj_id = ip->obj_id;
440 cursor.key_beg.key = offset + bytes;
441 cursor.key_beg.create_tid = 0;
442 cursor.key_beg.delete_tid = 0;
443 cursor.key_beg.rec_type = HAMMER_RECTYPE_DATA;
444 cursor.asof = trans->tid;
445 cursor.flags |= HAMMER_CURSOR_INSERT | HAMMER_CURSOR_ASOF;
448 * Issue a lookup to position the cursor and locate the cluster
450 error = hammer_btree_lookup(&cursor);
452 kprintf("hammer_ip_sync_data: duplicate data at (%lld,%d)\n",
454 hammer_print_btree_elm(&cursor.node->ondisk->elms[cursor.index],
455 HAMMER_BTREE_TYPE_LEAF, cursor.index);
462 * Allocate record and data space now that we know which cluster
463 * the B-Tree node ended up in.
465 bdata = hammer_alloc_data(cursor.node->cluster, bytes, &error,
466 &cursor.data_buffer);
469 rec = hammer_alloc_record(cursor.node->cluster, &error,
470 &cursor.record_buffer);
475 * Fill everything in and insert our B-Tree node.
477 hammer_modify_buffer(cursor.record_buffer);
478 rec->base.base.btype = HAMMER_BTREE_TYPE_RECORD;
479 rec->base.base.obj_id = ip->obj_id;
480 rec->base.base.key = offset + bytes;
481 rec->base.base.create_tid = trans->tid;
482 rec->base.base.delete_tid = 0;
483 rec->base.base.rec_type = HAMMER_RECTYPE_DATA;
484 rec->base.data_crc = crc32(data, bytes);
485 rec->base.rec_id = 0; /* XXX */
486 rec->base.data_offset = hammer_bclu_offset(cursor.data_buffer, bdata);
487 rec->base.data_len = bytes;
489 hammer_modify_buffer(cursor.data_buffer);
490 bcopy(data, bdata, bytes);
492 elm.leaf.base = rec->base.base;
493 elm.leaf.rec_offset = hammer_bclu_offset(cursor.record_buffer, rec);
494 elm.leaf.data_offset = rec->base.data_offset;
495 elm.leaf.data_len = bytes;
496 elm.leaf.data_crc = rec->base.data_crc;
499 * Data records can wind up on-disk before the inode itself is
500 * on-disk. One must assume data records may be on-disk if either
501 * HAMMER_INODE_DONDISK or HAMMER_INODE_ONDISK is set
503 ip->flags |= HAMMER_INODE_DONDISK;
505 error = hammer_btree_insert(&cursor, &elm);
507 hammer_update_syncid(cursor.record_buffer->cluster, trans->tid);
511 hammer_free_record_ptr(cursor.record_buffer, rec);
513 hammer_free_data_ptr(cursor.data_buffer, bdata, bytes);
516 * If ENOSPC in cluster fill in the spike structure and return
520 hammer_load_spike(&cursor, spike);
521 hammer_done_cursor(&cursor);
526 * Sync an in-memory record to the disk. this is typically called via fsync
527 * from a cached record source. This code is responsible for actually
528 * writing a record out to the disk.
531 hammer_ip_sync_record(hammer_record_t record, struct hammer_cursor **spike)
533 struct hammer_cursor cursor;
534 hammer_record_ondisk_t rec;
536 union hammer_btree_elm elm;
540 error = hammer_init_cursor_hmp(&cursor, &record->ip->cache[0],
544 cursor.key_beg = record->rec.base.base;
545 cursor.flags |= HAMMER_CURSOR_INSERT;
548 * Issue a lookup to position the cursor and locate the cluster. The
549 * target key should not exist. If we are creating a directory entry
550 * we may have to iterate the low 32 bits of the key to find an unused
553 * If we run out of space trying to adjust the B-Tree for the
554 * insert, re-lookup without the insert flag so the cursor
555 * is properly positioned for the spike.
558 error = hammer_btree_lookup(&cursor);
560 if (record->rec.base.base.rec_type == HAMMER_RECTYPE_DIRENTRY) {
561 hmp = cursor.node->cluster->volume->hmp;
562 if (++hmp->namekey_iterator == 0)
563 ++hmp->namekey_iterator;
564 record->rec.base.base.key &= ~(0xFFFFFFFFLL);
565 record->rec.base.base.key |= hmp->namekey_iterator;
566 cursor.key_beg.key = record->rec.base.base.key;
569 kprintf("hammer_ip_sync_record: duplicate rec at (%016llx)\n",
570 record->rec.base.base.key);
571 Debugger("duplicate record1");
578 * Mark the record as undergoing synchronization. Our cursor is
579 * holding a locked B-Tree node for the insertion which interlocks
580 * anyone trying to access this record.
582 * XXX There is still a race present related to iterations. An
583 * iteration may process the record, a sync may occur, and then
584 * later process the B-Tree element for the same record.
586 * We do not try to synchronize a deleted record.
588 if (record->flags & (HAMMER_RECF_DELETED | HAMMER_RECF_SYNCING)) {
592 record->flags |= HAMMER_RECF_SYNCING;
595 * Allocate record and data space now that we know which cluster
596 * the B-Tree node ended up in.
598 if (record->data == NULL ||
599 (record->flags & HAMMER_RECF_EMBEDDED_DATA)) {
600 bdata = record->data;
602 bdata = hammer_alloc_data(cursor.node->cluster,
603 record->rec.base.data_len, &error,
604 &cursor.data_buffer);
608 rec = hammer_alloc_record(cursor.node->cluster, &error,
609 &cursor.record_buffer);
614 * Fill everything in and insert our B-Tree node.
616 * XXX assign rec_id here
618 hammer_modify_buffer(cursor.record_buffer);
621 rec->base.data_crc = crc32(record->data,
622 record->rec.base.data_len);
623 if (record->flags & HAMMER_RECF_EMBEDDED_DATA) {
625 * Data embedded in record
627 rec->base.data_offset = ((char *)bdata -
628 (char *)&record->rec);
629 KKASSERT(rec->base.data_offset >= 0 &&
630 rec->base.data_offset + rec->base.data_len <=
632 rec->base.data_offset += hammer_bclu_offset(cursor.record_buffer, rec);
635 * Data separate from record
637 rec->base.data_offset = hammer_bclu_offset(cursor.data_buffer,bdata);
638 hammer_modify_buffer(cursor.data_buffer);
639 bcopy(record->data, bdata, rec->base.data_len);
642 rec->base.rec_id = 0; /* XXX */
644 elm.leaf.base = record->rec.base.base;
645 elm.leaf.rec_offset = hammer_bclu_offset(cursor.record_buffer, rec);
646 elm.leaf.data_offset = rec->base.data_offset;
647 elm.leaf.data_len = rec->base.data_len;
648 elm.leaf.data_crc = rec->base.data_crc;
650 error = hammer_btree_insert(&cursor, &elm);
653 * Clean up on success, or fall through on error.
656 record->flags |= HAMMER_RECF_DELETED;
657 record->flags &= ~HAMMER_RECF_SYNCING;
658 hammer_update_syncid(cursor.record_buffer->cluster,
659 record->rec.base.base.create_tid);
663 hammer_free_record_ptr(cursor.record_buffer, rec);
665 if (record->data && (record->flags & HAMMER_RECF_EMBEDDED_DATA) == 0) {
666 hammer_free_data_ptr(cursor.data_buffer, bdata,
667 record->rec.base.data_len);
670 record->flags &= ~HAMMER_RECF_SYNCING;
673 * If ENOSPC in cluster fill in the spike structure and return
677 hammer_load_spike(&cursor, spike);
678 hammer_done_cursor(&cursor);
683 * Write out a record using the specified cursor. The caller does not have
684 * to seek the cursor. The flags are used to determine whether the data
685 * (if any) is embedded in the record or not.
687 * The target cursor will be modified by this call. Note in particular
688 * that HAMMER_CURSOR_INSERT is set.
691 hammer_write_record(hammer_cursor_t cursor, hammer_record_ondisk_t orec,
692 void *data, int cursor_flags)
694 union hammer_btree_elm elm;
695 hammer_record_ondisk_t nrec;
699 cursor->key_beg = orec->base.base;
700 cursor->flags |= HAMMER_CURSOR_INSERT;
703 * Issue a lookup to position the cursor and locate the cluster. The
704 * target key should not exist.
706 * If we run out of space trying to adjust the B-Tree for the
707 * insert, re-lookup without the insert flag so the cursor
708 * is properly positioned for the spike.
710 error = hammer_btree_lookup(cursor);
712 kprintf("hammer_ip_sync_record: duplicate rec at (%016llx)\n",
713 orec->base.base.key);
714 Debugger("duplicate record2");
721 * Allocate record and data space now that we know which cluster
722 * the B-Tree node ended up in.
725 (cursor_flags & HAMMER_RECF_EMBEDDED_DATA)) {
728 bdata = hammer_alloc_data(cursor->node->cluster,
729 orec->base.data_len, &error,
730 &cursor->data_buffer);
734 nrec = hammer_alloc_record(cursor->node->cluster, &error,
735 &cursor->record_buffer);
740 * Fill everything in and insert our B-Tree node.
742 * XXX assign rec_id here
744 hammer_modify_buffer(cursor->record_buffer);
746 nrec->base.data_offset = 0;
748 nrec->base.data_crc = crc32(bdata, nrec->base.data_len);
749 if (cursor_flags & HAMMER_RECF_EMBEDDED_DATA) {
751 * Data embedded in record
753 nrec->base.data_offset = ((char *)bdata - (char *)orec);
754 KKASSERT(nrec->base.data_offset >= 0 &&
755 nrec->base.data_offset + nrec->base.data_len <
757 nrec->base.data_offset += hammer_bclu_offset(cursor->record_buffer, nrec);
760 * Data separate from record
762 nrec->base.data_offset = hammer_bclu_offset(cursor->data_buffer, bdata);
763 hammer_modify_buffer(cursor->data_buffer);
764 bcopy(data, bdata, nrec->base.data_len);
767 nrec->base.rec_id = 0; /* XXX */
769 elm.leaf.base = nrec->base.base;
770 elm.leaf.rec_offset = hammer_bclu_offset(cursor->record_buffer, nrec);
771 elm.leaf.data_offset = nrec->base.data_offset;
772 elm.leaf.data_len = nrec->base.data_len;
773 elm.leaf.data_crc = nrec->base.data_crc;
775 error = hammer_btree_insert(cursor, &elm);
777 hammer_update_syncid(cursor->record_buffer->cluster,
778 nrec->base.base.create_tid);
782 hammer_free_record_ptr(cursor->record_buffer, nrec);
784 if (data && (cursor_flags & HAMMER_RECF_EMBEDDED_DATA) == 0) {
785 hammer_free_data_ptr(cursor->data_buffer, bdata,
786 orec->base.data_len);
789 /* leave cursor intact */
794 * Add the record to the inode's rec_tree. The low 32 bits of a directory
795 * entry's key is used to deal with hash collisions in the upper 32 bits.
796 * A unique 64 bit key is generated in-memory and may be regenerated a
797 * second time when the directory record is flushed to the on-disk B-Tree.
799 * A referenced record is passed to this function. This function
800 * eats the reference. If an error occurs the record will be deleted.
804 hammer_mem_add(struct hammer_transaction *trans, hammer_record_t record)
806 while (RB_INSERT(hammer_rec_rb_tree, &record->ip->rec_tree, record)) {
807 if (record->rec.base.base.rec_type != HAMMER_RECTYPE_DIRENTRY){
808 record->flags |= HAMMER_RECF_DELETED;
809 hammer_rel_mem_record(record);
812 if (++trans->hmp->namekey_iterator == 0)
813 ++trans->hmp->namekey_iterator;
814 record->rec.base.base.key &= ~(0xFFFFFFFFLL);
815 record->rec.base.base.key |= trans->hmp->namekey_iterator;
817 record->flags |= HAMMER_RECF_ONRBTREE;
818 hammer_modify_inode(trans, record->ip, HAMMER_INODE_XDIRTY);
819 hammer_rel_mem_record(record);
823 /************************************************************************
824 * HAMMER INODE MERGED-RECORD FUNCTIONS *
825 ************************************************************************
827 * These functions augment the B-Tree scanning functions in hammer_btree.c
828 * by merging in-memory records with on-disk records.
832 * Locate a particular record either in-memory or on-disk.
834 * NOTE: This is basically a standalone routine, hammer_ip_next() may
835 * NOT be called to iterate results.
838 hammer_ip_lookup(hammer_cursor_t cursor, struct hammer_inode *ip)
843 * If the element is in-memory return it without searching the
846 error = hammer_mem_lookup(cursor, ip);
848 cursor->record = &cursor->iprec->rec;
855 * If the inode has on-disk components search the on-disk B-Tree.
857 if ((ip->flags & (HAMMER_INODE_ONDISK|HAMMER_INODE_DONDISK)) == 0)
859 error = hammer_btree_lookup(cursor);
861 error = hammer_btree_extract(cursor, HAMMER_CURSOR_GET_RECORD);
866 * Locate the first record within the cursor's key_beg/key_end range,
867 * restricted to a particular inode. 0 is returned on success, ENOENT
868 * if no records matched the requested range, or some other error.
870 * When 0 is returned hammer_ip_next() may be used to iterate additional
871 * records within the requested range.
874 hammer_ip_first(hammer_cursor_t cursor, struct hammer_inode *ip)
879 * Clean up fields and setup for merged scan
881 cursor->flags &= ~HAMMER_CURSOR_DELBTREE;
882 cursor->flags |= HAMMER_CURSOR_ATEDISK | HAMMER_CURSOR_ATEMEM;
883 cursor->flags |= HAMMER_CURSOR_DISKEOF | HAMMER_CURSOR_MEMEOF;
885 hammer_rel_mem_record(cursor->iprec);
886 cursor->iprec = NULL;
890 * Search the on-disk B-Tree. hammer_btree_lookup() only does an
891 * exact lookup so if we get ENOENT we have to call the iterate
892 * function to validate the first record after the begin key.
894 * The ATEDISK flag is used by hammer_btree_iterate to determine
895 * whether it must index forwards or not. It is also used here
896 * to select the next record from in-memory or on-disk.
898 if (ip->flags & (HAMMER_INODE_ONDISK|HAMMER_INODE_DONDISK)) {
899 error = hammer_btree_lookup(cursor);
900 if (error == ENOENT) {
901 cursor->flags &= ~HAMMER_CURSOR_ATEDISK;
902 error = hammer_btree_iterate(cursor);
904 if (error && error != ENOENT)
907 cursor->flags &= ~HAMMER_CURSOR_DISKEOF;
908 cursor->flags &= ~HAMMER_CURSOR_ATEDISK;
910 cursor->flags |= HAMMER_CURSOR_ATEDISK;
915 * Search the in-memory record list (Red-Black tree). Unlike the
916 * B-Tree search, mem_first checks for records in the range.
918 error = hammer_mem_first(cursor, ip);
919 if (error && error != ENOENT)
922 cursor->flags &= ~HAMMER_CURSOR_MEMEOF;
923 cursor->flags &= ~HAMMER_CURSOR_ATEMEM;
927 * This will return the first matching record.
929 return(hammer_ip_next(cursor));
933 * Retrieve the next record in a merged iteration within the bounds of the
934 * cursor. This call may be made multiple times after the cursor has been
935 * initially searched with hammer_ip_first().
937 * 0 is returned on success, ENOENT if no further records match the
938 * requested range, or some other error code is returned.
941 hammer_ip_next(hammer_cursor_t cursor)
943 hammer_btree_elm_t elm;
949 * Load the current on-disk and in-memory record. If we ate any
950 * records we have to get the next one.
952 * If we deleted the last on-disk record we had scanned ATEDISK will
953 * be clear and DELBTREE will be set, forcing a call to iterate. The
954 * fact that ATEDISK is clear causes iterate to re-test the 'current'
955 * element. If ATEDISK is set, iterate will skip the 'current'
958 * Get the next on-disk record
960 if (cursor->flags & (HAMMER_CURSOR_ATEDISK|HAMMER_CURSOR_DELBTREE)) {
961 if ((cursor->flags & HAMMER_CURSOR_DISKEOF) == 0) {
962 error = hammer_btree_iterate(cursor);
963 cursor->flags &= ~HAMMER_CURSOR_DELBTREE;
965 cursor->flags &= ~HAMMER_CURSOR_ATEDISK;
967 cursor->flags |= HAMMER_CURSOR_DISKEOF |
968 HAMMER_CURSOR_ATEDISK;
973 * Get the next in-memory record. The record can be ripped out
974 * of the RB tree so we maintain a scan_info structure to track
977 * hammer_rec_scan_cmp: Is the record still in our general range,
978 * (non-inclusive of snapshot exclusions)?
979 * hammer_rec_scan_callback: Is the record in our snapshot?
981 if (cursor->flags & HAMMER_CURSOR_ATEMEM) {
982 if ((cursor->flags & HAMMER_CURSOR_MEMEOF) == 0) {
984 hammer_rel_mem_record(cursor->iprec);
985 cursor->iprec = NULL;
987 rec = cursor->scan.node; /* next node */
989 if (hammer_rec_scan_cmp(rec, cursor) != 0)
991 if (hammer_rec_scan_callback(rec, cursor) != 0)
993 rec = hammer_rec_rb_tree_RB_NEXT(rec);
996 KKASSERT(cursor->iprec == rec);
997 cursor->flags &= ~HAMMER_CURSOR_ATEMEM;
999 hammer_rec_rb_tree_RB_NEXT(rec);
1001 cursor->flags |= HAMMER_CURSOR_MEMEOF;
1007 * Extract either the disk or memory record depending on their
1008 * relative position.
1011 switch(cursor->flags & (HAMMER_CURSOR_ATEDISK | HAMMER_CURSOR_ATEMEM)) {
1014 * Both entries valid
1016 elm = &cursor->node->ondisk->elms[cursor->index];
1017 r = hammer_btree_cmp(&elm->base, &cursor->iprec->rec.base.base);
1019 error = hammer_btree_extract(cursor,
1020 HAMMER_CURSOR_GET_RECORD);
1021 cursor->flags |= HAMMER_CURSOR_ATEDISK;
1024 /* fall through to the memory entry */
1025 case HAMMER_CURSOR_ATEDISK:
1027 * Only the memory entry is valid
1029 cursor->record = &cursor->iprec->rec;
1030 cursor->flags |= HAMMER_CURSOR_ATEMEM;
1032 case HAMMER_CURSOR_ATEMEM:
1034 * Only the disk entry is valid
1036 error = hammer_btree_extract(cursor, HAMMER_CURSOR_GET_RECORD);
1037 cursor->flags |= HAMMER_CURSOR_ATEDISK;
1041 * Neither entry is valid
1043 * XXX error not set properly
1045 cursor->record = NULL;
1053 * Resolve the cursor->data pointer for the current cursor position in
1054 * a merged iteration.
1057 hammer_ip_resolve_data(hammer_cursor_t cursor)
1061 if (cursor->iprec && cursor->record == &cursor->iprec->rec) {
1062 cursor->data = cursor->iprec->data;
1065 error = hammer_btree_extract(cursor, HAMMER_CURSOR_GET_DATA);
1071 * Delete all records within the specified range for inode ip.
1073 * NOTE: An unaligned range will cause new records to be added to cover
1074 * the edge cases. (XXX not implemented yet).
1076 * NOTE: ran_end is inclusive (e.g. 0,1023 instead of 0,1024).
1078 * NOTE: Record keys for regular file data have to be special-cased since
1079 * they indicate the end of the range (key = base + bytes).
1081 * NOTE: The spike structure must be filled in if we return ENOSPC.
1084 hammer_ip_delete_range(hammer_transaction_t trans, hammer_inode_t ip,
1085 int64_t ran_beg, int64_t ran_end,
1086 struct hammer_cursor **spike)
1088 struct hammer_cursor cursor;
1089 hammer_record_ondisk_t rec;
1090 hammer_base_elm_t base;
1094 hammer_init_cursor_hmp(&cursor, &ip->cache[0], ip->hmp);
1096 cursor.key_beg.obj_id = ip->obj_id;
1097 cursor.key_beg.create_tid = 0;
1098 cursor.key_beg.delete_tid = 0;
1099 cursor.key_beg.obj_type = 0;
1100 cursor.asof = ip->obj_asof;
1101 cursor.flags |= HAMMER_CURSOR_ASOF;
1103 cursor.key_end = cursor.key_beg;
1104 if (ip->ino_rec.base.base.obj_type == HAMMER_OBJTYPE_DBFILE) {
1105 cursor.key_beg.key = ran_beg;
1106 cursor.key_beg.rec_type = HAMMER_RECTYPE_DB;
1107 cursor.key_end.rec_type = HAMMER_RECTYPE_DB;
1108 cursor.key_end.key = ran_end;
1111 * The key in the B-Tree is (base+bytes), so the first possible
1112 * matching key is ran_beg + 1.
1116 cursor.key_beg.key = ran_beg + 1;
1117 cursor.key_beg.rec_type = HAMMER_RECTYPE_DATA;
1118 cursor.key_end.rec_type = HAMMER_RECTYPE_DATA;
1120 tmp64 = ran_end + MAXPHYS + 1; /* work around GCC-4 bug */
1121 if (tmp64 < ran_end)
1122 cursor.key_end.key = 0x7FFFFFFFFFFFFFFFLL;
1124 cursor.key_end.key = ran_end + MAXPHYS + 1;
1126 cursor.flags |= HAMMER_CURSOR_END_INCLUSIVE;
1128 error = hammer_ip_first(&cursor, ip);
1131 * Iterate through matching records and mark them as deleted.
1133 while (error == 0) {
1134 rec = cursor.record;
1135 base = &rec->base.base;
1137 KKASSERT(base->delete_tid == 0);
1140 * There may be overlap cases for regular file data. Also
1141 * remember the key for a regular file record is the offset
1142 * of the last byte of the record (base + len - 1), NOT the
1146 kprintf("delete_range rec_type %02x\n", base->rec_type);
1148 if (base->rec_type == HAMMER_RECTYPE_DATA) {
1150 kprintf("delete_range loop key %016llx\n",
1151 base->key - rec->base.data_len);
1153 off = base->key - rec->base.data_len;
1155 * Check the left edge case. We currently do not
1156 * split existing records.
1158 if (off < ran_beg) {
1159 panic("hammer left edge case %016llx %d\n",
1160 base->key, rec->base.data_len);
1164 * Check the right edge case. Note that the
1165 * record can be completely out of bounds, which
1166 * terminates the search.
1168 * base->key is exclusive of the right edge while
1169 * ran_end is inclusive of the right edge. The
1170 * (key - data_len) left boundary is inclusive.
1172 * XXX theory-check this test at some point, are
1173 * we missing a + 1 somewhere? Note that ran_end
1176 if (base->key - 1 > ran_end) {
1177 if (base->key - rec->base.data_len > ran_end)
1179 panic("hammer right edge case\n");
1184 * Mark the record and B-Tree entry as deleted. This will
1185 * also physically delete the B-Tree entry, record, and
1186 * data if the retention policy dictates. The function
1187 * will set HAMMER_CURSOR_DELBTREE which hammer_ip_next()
1188 * uses to perform a fixup.
1190 error = hammer_ip_delete_record(&cursor, trans->tid);
1193 error = hammer_ip_next(&cursor);
1195 hammer_done_cursor(&cursor);
1196 if (error == ENOENT)
1202 * Delete all records associated with an inode except the inode record
1206 hammer_ip_delete_range_all(hammer_transaction_t trans, hammer_inode_t ip)
1208 struct hammer_cursor cursor;
1209 hammer_record_ondisk_t rec;
1210 hammer_base_elm_t base;
1213 hammer_init_cursor_hmp(&cursor, &ip->cache[0], ip->hmp);
1215 cursor.key_beg.obj_id = ip->obj_id;
1216 cursor.key_beg.create_tid = 0;
1217 cursor.key_beg.delete_tid = 0;
1218 cursor.key_beg.obj_type = 0;
1219 cursor.key_beg.rec_type = HAMMER_RECTYPE_INODE + 1;
1220 cursor.key_beg.key = HAMMER_MIN_KEY;
1222 cursor.key_end = cursor.key_beg;
1223 cursor.key_end.rec_type = 0xFFFF;
1224 cursor.key_end.key = HAMMER_MAX_KEY;
1226 cursor.asof = ip->obj_asof;
1227 cursor.flags |= HAMMER_CURSOR_END_INCLUSIVE | HAMMER_CURSOR_ASOF;
1229 error = hammer_ip_first(&cursor, ip);
1232 * Iterate through matching records and mark them as deleted.
1234 while (error == 0) {
1235 rec = cursor.record;
1236 base = &rec->base.base;
1238 KKASSERT(base->delete_tid == 0);
1241 * Mark the record and B-Tree entry as deleted. This will
1242 * also physically delete the B-Tree entry, record, and
1243 * data if the retention policy dictates. The function
1244 * will set HAMMER_CURSOR_DELBTREE which hammer_ip_next()
1245 * uses to perform a fixup.
1247 error = hammer_ip_delete_record(&cursor, trans->tid);
1250 error = hammer_ip_next(&cursor);
1252 hammer_done_cursor(&cursor);
1253 if (error == ENOENT)
1259 * Delete the record at the current cursor
1262 hammer_ip_delete_record(hammer_cursor_t cursor, hammer_tid_t tid)
1264 hammer_btree_elm_t elm;
1269 * In-memory (unsynchronized) records can simply be freed.
1271 if (cursor->record == &cursor->iprec->rec) {
1272 cursor->iprec->flags |= HAMMER_RECF_DELETED;
1277 * On-disk records are marked as deleted by updating their delete_tid.
1279 error = hammer_btree_extract(cursor, HAMMER_CURSOR_GET_RECORD);
1281 hmp = cursor->node->cluster->volume->hmp;
1284 hammer_modify_buffer(cursor->record_buffer);
1285 cursor->record->base.base.delete_tid = tid;
1287 hammer_modify_node(cursor->node);
1288 elm = &cursor->node->ondisk->elms[cursor->index];
1289 elm->leaf.base.delete_tid = tid;
1290 hammer_update_syncid(cursor->record_buffer->cluster, tid);
1294 * If we were mounted with the nohistory option, we physically
1295 * delete the record.
1297 if (error == 0 && (hmp->hflags & HMNT_NOHISTORY)) {
1299 int32_t data_offset;
1301 hammer_cluster_t cluster;
1303 rec_offset = elm->leaf.rec_offset;
1304 data_offset = elm->leaf.data_offset;
1305 data_len = elm->leaf.data_len;
1307 kprintf("hammer_ip_delete_record: %08x %08x/%d\n",
1308 rec_offset, data_offset, data_len);
1310 cluster = cursor->node->cluster;
1311 hammer_ref_cluster(cluster);
1313 error = hammer_btree_delete(cursor);
1316 * This forces a fixup for the iteration because
1317 * the cursor is now either sitting at the 'next'
1318 * element or sitting at the end of a leaf.
1320 if ((cursor->flags & HAMMER_CURSOR_DISKEOF) == 0) {
1321 cursor->flags |= HAMMER_CURSOR_DELBTREE;
1322 cursor->flags &= ~HAMMER_CURSOR_ATEDISK;
1324 hammer_free_record(cluster, rec_offset);
1325 if (data_offset && (data_offset - rec_offset < 0 ||
1326 data_offset - rec_offset >= HAMMER_RECORD_SIZE)) {
1327 hammer_free_data(cluster, data_offset,data_len);
1330 hammer_rel_cluster(cluster, 0);
1332 panic("hammer_ip_delete_record: unable to physically delete the record!\n");
1340 * Determine whether a directory is empty or not. Returns 0 if the directory
1341 * is empty, ENOTEMPTY if it isn't, plus other possible errors.
1344 hammer_ip_check_directory_empty(hammer_transaction_t trans, hammer_inode_t ip)
1346 struct hammer_cursor cursor;
1349 hammer_init_cursor_hmp(&cursor, &ip->cache[0], ip->hmp);
1351 cursor.key_beg.obj_id = ip->obj_id;
1352 cursor.key_beg.create_tid = 0;
1353 cursor.key_beg.delete_tid = 0;
1354 cursor.key_beg.obj_type = 0;
1355 cursor.key_beg.rec_type = HAMMER_RECTYPE_INODE + 1;
1356 cursor.key_beg.key = HAMMER_MIN_KEY;
1358 cursor.key_end = cursor.key_beg;
1359 cursor.key_end.rec_type = 0xFFFF;
1360 cursor.key_end.key = HAMMER_MAX_KEY;
1362 cursor.asof = ip->obj_asof;
1363 cursor.flags |= HAMMER_CURSOR_END_INCLUSIVE | HAMMER_CURSOR_ASOF;
1365 error = hammer_ip_first(&cursor, ip);
1366 if (error == ENOENT)
1368 else if (error == 0)
1370 hammer_done_cursor(&cursor);