1 /*
2 * CDDL HEADER START
3 *
4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
7 *
8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 * or http://www.opensolaris.org/os/licensing.
10 * See the License for the specific language governing permissions
11 * and limitations under the License.
12 *
13 * When distributing Covered Code, include this CDDL HEADER in each
14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 * If applicable, add the following below this CDDL HEADER, with the
16 * fields enclosed by brackets "[]" replaced with your own identifying
17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 *
19 * CDDL HEADER END
20 */
21
22 /*
23 * Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved.
24 * Portions Copyright 2011 iXsystems, Inc
25 * Copyright (c) 2013 by Delphix. All rights reserved.
26 */
27
28 #include <sys/zfs_context.h>
29 #include <sys/types.h>
30 #include <sys/param.h>
31 #include <sys/systm.h>
32 #include <sys/sysmacros.h>
33 #include <sys/dmu.h>
34 #include <sys/dmu_impl.h>
35 #include <sys/dmu_objset.h>
36 #include <sys/dbuf.h>
37 #include <sys/dnode.h>
38 #include <sys/zap.h>
39 #include <sys/sa.h>
40 #include <sys/sunddi.h>
41 #include <sys/sa_impl.h>
42 #include <sys/dnode.h>
43 #include <sys/errno.h>
44 #include <sys/zfs_context.h>
45
46 /*
47 * ZFS System attributes:
48 *
49 * A generic mechanism to allow for arbitrary attributes
50 * to be stored in a dnode. The data will be stored in the bonus buffer of
51 * the dnode and if necessary a special "spill" block will be used to handle
52 * overflow situations. The spill block will be sized to fit the data
53 * from 512 - 128K. When a spill block is used the BP (blkptr_t) for the
54 * spill block is stored at the end of the current bonus buffer. Any
55 * attributes that would be in the way of the blkptr_t will be relocated
56 * into the spill block.
57 *
58 * Attribute registration:
59 *
60 * Stored persistently on a per dataset basis
61 * a mapping between attribute "string" names and their actual attribute
62 * numeric values, length, and byteswap function. The names are only used
63 * during registration. All attributes are known by their unique attribute
64 * id value. If an attribute can have a variable size then the value
65 * 0 will be used to indicate this.
66 *
67 * Attribute Layout:
68 *
69 * Attribute layouts are a way to compactly store multiple attributes, but
70 * without taking the overhead associated with managing each attribute
71 * individually. Since you will typically have the same set of attributes
72 * stored in the same order a single table will be used to represent that
73 * layout. The ZPL for example will usually have only about 10 different
74 * layouts (regular files, device files, symlinks,
75 * regular files + scanstamp, files/dir with extended attributes, and then
76 * you have the possibility of all of those minus ACL, because it would
77 * be kicked out into the spill block)
78 *
79 * Layouts are simply an array of the attributes and their
80 * ordering i.e. [0, 1, 4, 5, 2]
81 *
82 * Each distinct layout is given a unique layout number and that is whats
83 * stored in the header at the beginning of the SA data buffer.
84 *
85 * A layout only covers a single dbuf (bonus or spill). If a set of
86 * attributes is split up between the bonus buffer and a spill buffer then
87 * two different layouts will be used. This allows us to byteswap the
88 * spill without looking at the bonus buffer and keeps the on disk format of
89 * the bonus and spill buffer the same.
90 *
91 * Adding a single attribute will cause the entire set of attributes to
92 * be rewritten and could result in a new layout number being constructed
93 * as part of the rewrite if no such layout exists for the new set of
94 * attribues. The new attribute will be appended to the end of the already
95 * existing attributes.
96 *
97 * Both the attribute registration and attribute layout information are
98 * stored in normal ZAP attributes. Their should be a small number of
99 * known layouts and the set of attributes is assumed to typically be quite
100 * small.
101 *
102 * The registered attributes and layout "table" information is maintained
103 * in core and a special "sa_os_t" is attached to the objset_t.
104 *
105 * A special interface is provided to allow for quickly applying
106 * a large set of attributes at once. sa_replace_all_by_template() is
107 * used to set an array of attributes. This is used by the ZPL when
108 * creating a brand new file. The template that is passed into the function
109 * specifies the attribute, size for variable length attributes, location of
110 * data and special "data locator" function if the data isn't in a contiguous
111 * location.
112 *
113 * Byteswap implications:
114 * Since the SA attributes are not entirely self describing we can't do
115 * the normal byteswap processing. The special ZAP layout attribute and
116 * attribute registration attributes define the byteswap function and the
117 * size of the attributes, unless it is variable sized.
118 * The normal ZFS byteswapping infrastructure assumes you don't need
119 * to read any objects in order to do the necessary byteswapping. Whereas
120 * SA attributes can only be properly byteswapped if the dataset is opened
121 * and the layout/attribute ZAP attributes are available. Because of this
122 * the SA attributes will be byteswapped when they are first accessed by
123 * the SA code that will read the SA data.
124 */
125
126 typedef void (sa_iterfunc_t)(void *hdr, void *addr, sa_attr_type_t,
127 uint16_t length, int length_idx, boolean_t, void *userp);
128
129 static int sa_build_index(sa_handle_t *hdl, sa_buf_type_t buftype);
130 static void sa_idx_tab_hold(objset_t *os, sa_idx_tab_t *idx_tab);
131 static void *sa_find_idx_tab(objset_t *os, dmu_object_type_t bonustype,
132 void *data);
133 static void sa_idx_tab_rele(objset_t *os, void *arg);
134 static void sa_copy_data(sa_data_locator_t *func, void *start, void *target,
135 int buflen);
136 static int sa_modify_attrs(sa_handle_t *hdl, sa_attr_type_t newattr,
137 sa_data_op_t action, sa_data_locator_t *locator, void *datastart,
138 uint16_t buflen, dmu_tx_t *tx);
139
140 arc_byteswap_func_t *sa_bswap_table[] = {
141 byteswap_uint64_array,
142 byteswap_uint32_array,
143 byteswap_uint16_array,
144 byteswap_uint8_array,
145 zfs_acl_byteswap,
146 };
147
148 #define SA_COPY_DATA(f, s, t, l) \
149 { \
150 if (f == NULL) { \
151 if (l == 8) { \
152 *(uint64_t *)t = *(uint64_t *)s; \
153 } else if (l == 16) { \
154 *(uint64_t *)t = *(uint64_t *)s; \
155 *(uint64_t *)((uintptr_t)t + 8) = \
156 *(uint64_t *)((uintptr_t)s + 8); \
157 } else { \
158 bcopy(s, t, l); \
159 } \
160 } else \
161 sa_copy_data(f, s, t, l); \
162 }
163
164 /*
165 * This table is fixed and cannot be changed. Its purpose is to
166 * allow the SA code to work with both old/new ZPL file systems.
167 * It contains the list of legacy attributes. These attributes aren't
168 * stored in the "attribute" registry zap objects, since older ZPL file systems
169 * won't have the registry. Only objsets of type ZFS_TYPE_FILESYSTEM will
170 * use this static table.
171 */
172 sa_attr_reg_t sa_legacy_attrs[] = {
173 {"ZPL_ATIME", sizeof (uint64_t) * 2, SA_UINT64_ARRAY, 0},
174 {"ZPL_MTIME", sizeof (uint64_t) * 2, SA_UINT64_ARRAY, 1},
175 {"ZPL_CTIME", sizeof (uint64_t) * 2, SA_UINT64_ARRAY, 2},
176 {"ZPL_CRTIME", sizeof (uint64_t) * 2, SA_UINT64_ARRAY, 3},
177 {"ZPL_GEN", sizeof (uint64_t), SA_UINT64_ARRAY, 4},
178 {"ZPL_MODE", sizeof (uint64_t), SA_UINT64_ARRAY, 5},
179 {"ZPL_SIZE", sizeof (uint64_t), SA_UINT64_ARRAY, 6},
180 {"ZPL_PARENT", sizeof (uint64_t), SA_UINT64_ARRAY, 7},
181 {"ZPL_LINKS", sizeof (uint64_t), SA_UINT64_ARRAY, 8},
182 {"ZPL_XATTR", sizeof (uint64_t), SA_UINT64_ARRAY, 9},
183 {"ZPL_RDEV", sizeof (uint64_t), SA_UINT64_ARRAY, 10},
184 {"ZPL_FLAGS", sizeof (uint64_t), SA_UINT64_ARRAY, 11},
185 {"ZPL_UID", sizeof (uint64_t), SA_UINT64_ARRAY, 12},
186 {"ZPL_GID", sizeof (uint64_t), SA_UINT64_ARRAY, 13},
187 {"ZPL_PAD", sizeof (uint64_t) * 4, SA_UINT64_ARRAY, 14},
188 {"ZPL_ZNODE_ACL", 88, SA_UINT8_ARRAY, 15},
189 };
190
191 /*
192 * ZPL legacy layout
193 * This is only used for objects of type DMU_OT_ZNODE
194 */
195 sa_attr_type_t sa_legacy_zpl_layout[] = {
196 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
197 };
198
199 /*
200 * Special dummy layout used for buffers with no attributes.
201 */
202
203 sa_attr_type_t sa_dummy_zpl_layout[] = { 0 };
204
205 static int sa_legacy_attr_count = 16;
206 static kmem_cache_t *sa_cache = NULL;
207
208 /*ARGSUSED*/
209 static int
210 sa_cache_constructor(void *buf, void *unused, int kmflag)
211 {
212 sa_handle_t *hdl = buf;
213
214 hdl->sa_bonus_tab = NULL;
215 hdl->sa_spill_tab = NULL;
216 hdl->sa_os = NULL;
217 hdl->sa_userp = NULL;
218 hdl->sa_bonus = NULL;
219 hdl->sa_spill = NULL;
220 mutex_init(&hdl->sa_lock, NULL, MUTEX_DEFAULT, NULL);
221 return (0);
222 }
223
224 /*ARGSUSED*/
225 static void
226 sa_cache_destructor(void *buf, void *unused)
227 {
228 sa_handle_t *hdl = buf;
229 mutex_destroy(&hdl->sa_lock);
230 }
231
232 void
233 sa_cache_init(void)
234 {
235 sa_cache = kmem_cache_create("sa_cache",
236 sizeof (sa_handle_t), 0, sa_cache_constructor,
237 sa_cache_destructor, NULL, NULL, NULL, 0);
238 }
239
240 void
241 sa_cache_fini(void)
242 {
243 if (sa_cache)
244 kmem_cache_destroy(sa_cache);
245 }
246
247 static int
248 layout_num_compare(const void *arg1, const void *arg2)
249 {
250 const sa_lot_t *node1 = arg1;
251 const sa_lot_t *node2 = arg2;
252
253 if (node1->lot_num > node2->lot_num)
254 return (1);
255 else if (node1->lot_num < node2->lot_num)
256 return (-1);
257 return (0);
258 }
259
260 static int
261 layout_hash_compare(const void *arg1, const void *arg2)
262 {
263 const sa_lot_t *node1 = arg1;
264 const sa_lot_t *node2 = arg2;
265
266 if (node1->lot_hash > node2->lot_hash)
267 return (1);
268 if (node1->lot_hash < node2->lot_hash)
269 return (-1);
270 if (node1->lot_instance > node2->lot_instance)
271 return (1);
272 if (node1->lot_instance < node2->lot_instance)
273 return (-1);
274 return (0);
275 }
276
277 boolean_t
278 sa_layout_equal(sa_lot_t *tbf, sa_attr_type_t *attrs, int count)
279 {
280 int i;
281
282 if (count != tbf->lot_attr_count)
283 return (1);
284
285 for (i = 0; i != count; i++) {
286 if (attrs[i] != tbf->lot_attrs[i])
287 return (1);
288 }
289 return (0);
290 }
291
292 #define SA_ATTR_HASH(attr) (zfs_crc64_table[(-1ULL ^ attr) & 0xFF])
293
294 static uint64_t
295 sa_layout_info_hash(sa_attr_type_t *attrs, int attr_count)
296 {
297 int i;
298 uint64_t crc = -1ULL;
299
300 for (i = 0; i != attr_count; i++)
301 crc ^= SA_ATTR_HASH(attrs[i]);
302
303 return (crc);
304 }
305
306 static int
307 sa_get_spill(sa_handle_t *hdl)
308 {
309 int rc;
310 if (hdl->sa_spill == NULL) {
311 if ((rc = dmu_spill_hold_existing(hdl->sa_bonus, NULL,
312 &hdl->sa_spill)) == 0)
313 VERIFY(0 == sa_build_index(hdl, SA_SPILL));
314 } else {
315 rc = 0;
316 }
317
318 return (rc);
319 }
320
321 /*
322 * Main attribute lookup/update function
323 * returns 0 for success or non zero for failures
324 *
325 * Operates on bulk array, first failure will abort further processing
326 */
327 int
328 sa_attr_op(sa_handle_t *hdl, sa_bulk_attr_t *bulk, int count,
329 sa_data_op_t data_op, dmu_tx_t *tx)
330 {
331 sa_os_t *sa = hdl->sa_os->os_sa;
332 int i;
333 int error = 0;
334 sa_buf_type_t buftypes;
335
336 buftypes = 0;
337
338 ASSERT(count > 0);
339 for (i = 0; i != count; i++) {
340 ASSERT(bulk[i].sa_attr <= hdl->sa_os->os_sa->sa_num_attrs);
341
342 bulk[i].sa_addr = NULL;
343 /* First check the bonus buffer */
344
345 if (hdl->sa_bonus_tab && TOC_ATTR_PRESENT(
346 hdl->sa_bonus_tab->sa_idx_tab[bulk[i].sa_attr])) {
347 SA_ATTR_INFO(sa, hdl->sa_bonus_tab,
348 SA_GET_HDR(hdl, SA_BONUS),
349 bulk[i].sa_attr, bulk[i], SA_BONUS, hdl);
350 if (tx && !(buftypes & SA_BONUS)) {
351 dmu_buf_will_dirty(hdl->sa_bonus, tx);
352 buftypes |= SA_BONUS;
353 }
354 }
355 if (bulk[i].sa_addr == NULL &&
356 ((error = sa_get_spill(hdl)) == 0)) {
357 if (TOC_ATTR_PRESENT(
358 hdl->sa_spill_tab->sa_idx_tab[bulk[i].sa_attr])) {
359 SA_ATTR_INFO(sa, hdl->sa_spill_tab,
360 SA_GET_HDR(hdl, SA_SPILL),
361 bulk[i].sa_attr, bulk[i], SA_SPILL, hdl);
362 if (tx && !(buftypes & SA_SPILL) &&
363 bulk[i].sa_size == bulk[i].sa_length) {
364 dmu_buf_will_dirty(hdl->sa_spill, tx);
365 buftypes |= SA_SPILL;
366 }
367 }
368 }
369 if (error && error != ENOENT) {
370 return ((error == ECKSUM) ? EIO : error);
371 }
372
373 switch (data_op) {
374 case SA_LOOKUP:
375 if (bulk[i].sa_addr == NULL)
376 return (SET_ERROR(ENOENT));
377 if (bulk[i].sa_data) {
378 SA_COPY_DATA(bulk[i].sa_data_func,
379 bulk[i].sa_addr, bulk[i].sa_data,
380 bulk[i].sa_size);
381 }
382 continue;
383
384 case SA_UPDATE:
385 /* existing rewrite of attr */
386 if (bulk[i].sa_addr &&
387 bulk[i].sa_size == bulk[i].sa_length) {
388 SA_COPY_DATA(bulk[i].sa_data_func,
389 bulk[i].sa_data, bulk[i].sa_addr,
390 bulk[i].sa_length);
391 continue;
392 } else if (bulk[i].sa_addr) { /* attr size change */
393 error = sa_modify_attrs(hdl, bulk[i].sa_attr,
394 SA_REPLACE, bulk[i].sa_data_func,
395 bulk[i].sa_data, bulk[i].sa_length, tx);
396 } else { /* adding new attribute */
397 error = sa_modify_attrs(hdl, bulk[i].sa_attr,
398 SA_ADD, bulk[i].sa_data_func,
399 bulk[i].sa_data, bulk[i].sa_length, tx);
400 }
401 if (error)
402 return (error);
403 break;
404 }
405 }
406 return (error);
407 }
408
409 static sa_lot_t *
410 sa_add_layout_entry(objset_t *os, sa_attr_type_t *attrs, int attr_count,
411 uint64_t lot_num, uint64_t hash, boolean_t zapadd, dmu_tx_t *tx)
412 {
413 sa_os_t *sa = os->os_sa;
414 sa_lot_t *tb, *findtb;
415 int i;
416 avl_index_t loc;
417
418 ASSERT(MUTEX_HELD(&sa->sa_lock));
419 tb = kmem_zalloc(sizeof (sa_lot_t), KM_SLEEP);
420 tb->lot_attr_count = attr_count;
421 tb->lot_attrs = kmem_alloc(sizeof (sa_attr_type_t) * attr_count,
422 KM_SLEEP);
423 bcopy(attrs, tb->lot_attrs, sizeof (sa_attr_type_t) * attr_count);
424 tb->lot_num = lot_num;
425 tb->lot_hash = hash;
426 tb->lot_instance = 0;
427
428 if (zapadd) {
429 char attr_name[8];
430
431 if (sa->sa_layout_attr_obj == 0) {
432 sa->sa_layout_attr_obj = zap_create_link(os,
433 DMU_OT_SA_ATTR_LAYOUTS,
434 sa->sa_master_obj, SA_LAYOUTS, tx);
435 }
436
437 (void) snprintf(attr_name, sizeof (attr_name),
438 "%d", (int)lot_num);
439 VERIFY(0 == zap_update(os, os->os_sa->sa_layout_attr_obj,
440 attr_name, 2, attr_count, attrs, tx));
441 }
442
443 list_create(&tb->lot_idx_tab, sizeof (sa_idx_tab_t),
444 offsetof(sa_idx_tab_t, sa_next));
445
446 for (i = 0; i != attr_count; i++) {
447 if (sa->sa_attr_table[tb->lot_attrs[i]].sa_length == 0)
448 tb->lot_var_sizes++;
449 }
450
451 avl_add(&sa->sa_layout_num_tree, tb);
452
453 /* verify we don't have a hash collision */
454 if ((findtb = avl_find(&sa->sa_layout_hash_tree, tb, &loc)) != NULL) {
455 for (; findtb && findtb->lot_hash == hash;
456 findtb = AVL_NEXT(&sa->sa_layout_hash_tree, findtb)) {
457 if (findtb->lot_instance != tb->lot_instance)
458 break;
459 tb->lot_instance++;
460 }
461 }
462 avl_add(&sa->sa_layout_hash_tree, tb);
463 return (tb);
464 }
465
466 static void
467 sa_find_layout(objset_t *os, uint64_t hash, sa_attr_type_t *attrs,
468 int count, dmu_tx_t *tx, sa_lot_t **lot)
469 {
470 sa_lot_t *tb, tbsearch;
471 avl_index_t loc;
472 sa_os_t *sa = os->os_sa;
473 boolean_t found = B_FALSE;
474
475 mutex_enter(&sa->sa_lock);
476 tbsearch.lot_hash = hash;
477 tbsearch.lot_instance = 0;
478 tb = avl_find(&sa->sa_layout_hash_tree, &tbsearch, &loc);
479 if (tb) {
480 for (; tb && tb->lot_hash == hash;
481 tb = AVL_NEXT(&sa->sa_layout_hash_tree, tb)) {
482 if (sa_layout_equal(tb, attrs, count) == 0) {
483 found = B_TRUE;
484 break;
485 }
486 }
487 }
488 if (!found) {
489 tb = sa_add_layout_entry(os, attrs, count,
490 avl_numnodes(&sa->sa_layout_num_tree), hash, B_TRUE, tx);
491 }
492 mutex_exit(&sa->sa_lock);
493 *lot = tb;
494 }
495
496 static int
497 sa_resize_spill(sa_handle_t *hdl, uint32_t size, dmu_tx_t *tx)
498 {
499 int error;
500 uint32_t blocksize;
501
502 if (size == 0) {
503 blocksize = SPA_MINBLOCKSIZE;
504 } else if (size > SPA_MAXBLOCKSIZE) {
505 ASSERT(0);
506 return (SET_ERROR(EFBIG));
507 } else {
508 blocksize = P2ROUNDUP_TYPED(size, SPA_MINBLOCKSIZE, uint32_t);
509 }
510
511 error = dbuf_spill_set_blksz(hdl->sa_spill, blocksize, tx);
512 ASSERT(error == 0);
513 return (error);
514 }
515
516 static void
517 sa_copy_data(sa_data_locator_t *func, void *datastart, void *target, int buflen)
518 {
519 if (func == NULL) {
520 bcopy(datastart, target, buflen);
521 } else {
522 boolean_t start;
523 int bytes;
524 void *dataptr;
525 void *saptr = target;
526 uint32_t length;
527
528 start = B_TRUE;
529 bytes = 0;
530 while (bytes < buflen) {
531 func(&dataptr, &length, buflen, start, datastart);
532 bcopy(dataptr, saptr, length);
533 saptr = (void *)((caddr_t)saptr + length);
534 bytes += length;
535 start = B_FALSE;
536 }
537 }
538 }
539
540 /*
541 * Determine several different sizes
542 * first the sa header size
543 * the number of bytes to be stored
544 * if spill would occur the index in the attribute array is returned
545 *
546 * the boolean will_spill will be set when spilling is necessary. It
547 * is only set when the buftype is SA_BONUS
548 */
549 static int
550 sa_find_sizes(sa_os_t *sa, sa_bulk_attr_t *attr_desc, int attr_count,
551 dmu_buf_t *db, sa_buf_type_t buftype, int *index, int *total,
552 boolean_t *will_spill)
553 {
554 int var_size = 0;
555 int i;
556 int j = -1;
557 int full_space;
558 int hdrsize;
559 boolean_t done = B_FALSE;
560
561 if (buftype == SA_BONUS && sa->sa_force_spill) {
562 *total = 0;
563 *index = 0;
564 *will_spill = B_TRUE;
565 return (0);
566 }
567
568 *index = -1;
569 *total = 0;
570
571 if (buftype == SA_BONUS)
572 *will_spill = B_FALSE;
573
574 hdrsize = (SA_BONUSTYPE_FROM_DB(db) == DMU_OT_ZNODE) ? 0 :
575 sizeof (sa_hdr_phys_t);
576
577 full_space = (buftype == SA_BONUS) ? DN_MAX_BONUSLEN : db->db_size;
578 ASSERT(IS_P2ALIGNED(full_space, 8));
579
580 for (i = 0; i != attr_count; i++) {
581 boolean_t is_var_sz;
582
583 *total = P2ROUNDUP(*total, 8);
584 *total += attr_desc[i].sa_length;
585 if (done)
586 goto next;
587
588 is_var_sz = (SA_REGISTERED_LEN(sa, attr_desc[i].sa_attr) == 0);
589 if (is_var_sz) {
590 var_size++;
591 }
592
593 if (is_var_sz && var_size > 1) {
594 if (P2ROUNDUP(hdrsize + sizeof (uint16_t), 8) +
595 *total < full_space) {
596 /*
597 * Account for header space used by array of
598 * optional sizes of variable-length attributes.
599 * Record the index in case this increase needs
600 * to be reversed due to spill-over.
601 */
602 hdrsize += sizeof (uint16_t);
603 j = i;
604 } else {
605 done = B_TRUE;
606 *index = i;
607 if (buftype == SA_BONUS)
608 *will_spill = B_TRUE;
609 continue;
610 }
611 }
612
613 /*
614 * find index of where spill *could* occur.
615 * Then continue to count of remainder attribute
616 * space. The sum is used later for sizing bonus
617 * and spill buffer.
618 */
619 if (buftype == SA_BONUS && *index == -1 &&
620 *total + P2ROUNDUP(hdrsize, 8) >
621 (full_space - sizeof (blkptr_t))) {
622 *index = i;
623 done = B_TRUE;
624 }
625
626 next:
627 if (*total + P2ROUNDUP(hdrsize, 8) > full_space &&
628 buftype == SA_BONUS)
629 *will_spill = B_TRUE;
630 }
631
632 /*
633 * j holds the index of the last variable-sized attribute for
634 * which hdrsize was increased. Reverse the increase if that
635 * attribute will be relocated to the spill block.
636 */
637 if (*will_spill && j == *index)
638 hdrsize -= sizeof (uint16_t);
639
640 hdrsize = P2ROUNDUP(hdrsize, 8);
641 return (hdrsize);
642 }
643
644 #define BUF_SPACE_NEEDED(total, header) (total + header)
645
646 /*
647 * Find layout that corresponds to ordering of attributes
648 * If not found a new layout number is created and added to
649 * persistent layout tables.
650 */
651 static int
652 sa_build_layouts(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc, int attr_count,
653 dmu_tx_t *tx)
654 {
655 sa_os_t *sa = hdl->sa_os->os_sa;
656 uint64_t hash;
657 sa_buf_type_t buftype;
658 sa_hdr_phys_t *sahdr;
659 void *data_start;
660 int buf_space;
661 sa_attr_type_t *attrs, *attrs_start;
662 int i, lot_count;
663 int hdrsize;
664 int spillhdrsize = 0;
665 int used;
666 dmu_object_type_t bonustype;
667 sa_lot_t *lot;
668 int len_idx;
669 int spill_used;
670 boolean_t spilling;
671
672 dmu_buf_will_dirty(hdl->sa_bonus, tx);
673 bonustype = SA_BONUSTYPE_FROM_DB(hdl->sa_bonus);
674
675 /* first determine bonus header size and sum of all attributes */
676 hdrsize = sa_find_sizes(sa, attr_desc, attr_count, hdl->sa_bonus,
677 SA_BONUS, &i, &used, &spilling);
678
679 if (used > SPA_MAXBLOCKSIZE)
680 return (SET_ERROR(EFBIG));
681
682 VERIFY(0 == dmu_set_bonus(hdl->sa_bonus, spilling ?
683 MIN(DN_MAX_BONUSLEN - sizeof (blkptr_t), used + hdrsize) :
684 used + hdrsize, tx));
685
686 ASSERT((bonustype == DMU_OT_ZNODE && spilling == 0) ||
687 bonustype == DMU_OT_SA);
688
689 /* setup and size spill buffer when needed */
690 if (spilling) {
691 boolean_t dummy;
692
693 if (hdl->sa_spill == NULL) {
694 VERIFY(dmu_spill_hold_by_bonus(hdl->sa_bonus, NULL,
695 &hdl->sa_spill) == 0);
696 }
697 dmu_buf_will_dirty(hdl->sa_spill, tx);
698
699 spillhdrsize = sa_find_sizes(sa, &attr_desc[i],
700 attr_count - i, hdl->sa_spill, SA_SPILL, &i,
701 &spill_used, &dummy);
702
703 if (spill_used > SPA_MAXBLOCKSIZE)
704 return (SET_ERROR(EFBIG));
705
706 buf_space = hdl->sa_spill->db_size - spillhdrsize;
707 if (BUF_SPACE_NEEDED(spill_used, spillhdrsize) >
708 hdl->sa_spill->db_size)
709 VERIFY(0 == sa_resize_spill(hdl,
710 BUF_SPACE_NEEDED(spill_used, spillhdrsize), tx));
711 }
712
713 /* setup starting pointers to lay down data */
714 data_start = (void *)((uintptr_t)hdl->sa_bonus->db_data + hdrsize);
715 sahdr = (sa_hdr_phys_t *)hdl->sa_bonus->db_data;
716 buftype = SA_BONUS;
717
718 if (spilling)
719 buf_space = (sa->sa_force_spill) ?
720 0 : SA_BLKPTR_SPACE - hdrsize;
721 else
722 buf_space = hdl->sa_bonus->db_size - hdrsize;
723
724 attrs_start = attrs = kmem_alloc(sizeof (sa_attr_type_t) * attr_count,
725 KM_SLEEP);
726 lot_count = 0;
727
728 for (i = 0, len_idx = 0, hash = -1ULL; i != attr_count; i++) {
729 uint16_t length;
730
731 ASSERT(IS_P2ALIGNED(data_start, 8));
732 ASSERT(IS_P2ALIGNED(buf_space, 8));
733 attrs[i] = attr_desc[i].sa_attr;
734 length = SA_REGISTERED_LEN(sa, attrs[i]);
735 if (length == 0)
736 length = attr_desc[i].sa_length;
737
738 if (buf_space < length) { /* switch to spill buffer */
739 VERIFY(spilling);
740 VERIFY(bonustype == DMU_OT_SA);
741 if (buftype == SA_BONUS && !sa->sa_force_spill) {
742 sa_find_layout(hdl->sa_os, hash, attrs_start,
743 lot_count, tx, &lot);
744 SA_SET_HDR(sahdr, lot->lot_num, hdrsize);
745 }
746
747 buftype = SA_SPILL;
748 hash = -1ULL;
749 len_idx = 0;
750
751 sahdr = (sa_hdr_phys_t *)hdl->sa_spill->db_data;
752 sahdr->sa_magic = SA_MAGIC;
753 data_start = (void *)((uintptr_t)sahdr +
754 spillhdrsize);
755 attrs_start = &attrs[i];
756 buf_space = hdl->sa_spill->db_size - spillhdrsize;
757 lot_count = 0;
758 }
759 hash ^= SA_ATTR_HASH(attrs[i]);
760 attr_desc[i].sa_addr = data_start;
761 attr_desc[i].sa_size = length;
762 SA_COPY_DATA(attr_desc[i].sa_data_func, attr_desc[i].sa_data,
763 data_start, length);
764 if (sa->sa_attr_table[attrs[i]].sa_length == 0) {
765 sahdr->sa_lengths[len_idx++] = length;
766 }
767 data_start = (void *)P2ROUNDUP(((uintptr_t)data_start +
768 length), 8);
769 buf_space -= P2ROUNDUP(length, 8);
770 lot_count++;
771 }
772
773 sa_find_layout(hdl->sa_os, hash, attrs_start, lot_count, tx, &lot);
774
775 /*
776 * Verify that old znodes always have layout number 0.
777 * Must be DMU_OT_SA for arbitrary layouts
778 */
779 VERIFY((bonustype == DMU_OT_ZNODE && lot->lot_num == 0) ||
780 (bonustype == DMU_OT_SA && lot->lot_num > 1));
781
782 if (bonustype == DMU_OT_SA) {
783 SA_SET_HDR(sahdr, lot->lot_num,
784 buftype == SA_BONUS ? hdrsize : spillhdrsize);
785 }
786
787 kmem_free(attrs, sizeof (sa_attr_type_t) * attr_count);
788 if (hdl->sa_bonus_tab) {
789 sa_idx_tab_rele(hdl->sa_os, hdl->sa_bonus_tab);
790 hdl->sa_bonus_tab = NULL;
791 }
792 if (!sa->sa_force_spill)
793 VERIFY(0 == sa_build_index(hdl, SA_BONUS));
794 if (hdl->sa_spill) {
795 sa_idx_tab_rele(hdl->sa_os, hdl->sa_spill_tab);
796 if (!spilling) {
797 /*
798 * remove spill block that is no longer needed.
799 */
800 dmu_buf_rele(hdl->sa_spill, NULL);
801 hdl->sa_spill = NULL;
802 hdl->sa_spill_tab = NULL;
803 VERIFY(0 == dmu_rm_spill(hdl->sa_os,
804 sa_handle_object(hdl), tx));
805 } else {
806 VERIFY(0 == sa_build_index(hdl, SA_SPILL));
807 }
808 }
809
810 return (0);
811 }
812
813 static void
814 sa_free_attr_table(sa_os_t *sa)
815 {
816 int i;
817
818 if (sa->sa_attr_table == NULL)
819 return;
820
821 for (i = 0; i != sa->sa_num_attrs; i++) {
822 if (sa->sa_attr_table[i].sa_name)
823 kmem_free(sa->sa_attr_table[i].sa_name,
824 strlen(sa->sa_attr_table[i].sa_name) + 1);
825 }
826
827 kmem_free(sa->sa_attr_table,
828 sizeof (sa_attr_table_t) * sa->sa_num_attrs);
829
830 sa->sa_attr_table = NULL;
831 }
832
833 static int
834 sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
835 {
836 sa_os_t *sa = os->os_sa;
837 uint64_t sa_attr_count = 0;
838 uint64_t sa_reg_count = 0;
839 int error = 0;
840 uint64_t attr_value;
841 sa_attr_table_t *tb;
842 zap_cursor_t zc;
843 zap_attribute_t za;
844 int registered_count = 0;
845 int i;
846 dmu_objset_type_t ostype = dmu_objset_type(os);
847
848 sa->sa_user_table =
849 kmem_zalloc(count * sizeof (sa_attr_type_t), KM_SLEEP);
850 sa->sa_user_table_sz = count * sizeof (sa_attr_type_t);
851
852 if (sa->sa_reg_attr_obj != 0) {
853 error = zap_count(os, sa->sa_reg_attr_obj,
854 &sa_attr_count);
855
856 /*
857 * Make sure we retrieved a count and that it isn't zero
858 */
859 if (error || (error == 0 && sa_attr_count == 0)) {
860 if (error == 0)
861 error = SET_ERROR(EINVAL);
862 goto bail;
863 }
864 sa_reg_count = sa_attr_count;
865 }
866
867 if (ostype == DMU_OST_ZFS && sa_attr_count == 0)
868 sa_attr_count += sa_legacy_attr_count;
869
870 /* Allocate attribute numbers for attributes that aren't registered */
871 for (i = 0; i != count; i++) {
872 boolean_t found = B_FALSE;
873 int j;
874
875 if (ostype == DMU_OST_ZFS) {
876 for (j = 0; j != sa_legacy_attr_count; j++) {
877 if (strcmp(reg_attrs[i].sa_name,
878 sa_legacy_attrs[j].sa_name) == 0) {
879 sa->sa_user_table[i] =
880 sa_legacy_attrs[j].sa_attr;
881 found = B_TRUE;
882 }
883 }
884 }
885 if (found)
886 continue;
887
888 if (sa->sa_reg_attr_obj)
889 error = zap_lookup(os, sa->sa_reg_attr_obj,
890 reg_attrs[i].sa_name, 8, 1, &attr_value);
891 else
892 error = SET_ERROR(ENOENT);
893 switch (error) {
894 case ENOENT:
895 sa->sa_user_table[i] = (sa_attr_type_t)sa_attr_count;
896 sa_attr_count++;
897 break;
898 case 0:
899 sa->sa_user_table[i] = ATTR_NUM(attr_value);
900 break;
901 default:
902 goto bail;
903 }
904 }
905
906 sa->sa_num_attrs = sa_attr_count;
907 tb = sa->sa_attr_table =
908 kmem_zalloc(sizeof (sa_attr_table_t) * sa_attr_count, KM_SLEEP);
909
910 /*
911 * Attribute table is constructed from requested attribute list,
912 * previously foreign registered attributes, and also the legacy
913 * ZPL set of attributes.
914 */
915
916 if (sa->sa_reg_attr_obj) {
917 for (zap_cursor_init(&zc, os, sa->sa_reg_attr_obj);
918 (error = zap_cursor_retrieve(&zc, &za)) == 0;
919 zap_cursor_advance(&zc)) {
920 uint64_t value;
921 value = za.za_first_integer;
922
923 registered_count++;
924 tb[ATTR_NUM(value)].sa_attr = ATTR_NUM(value);
925 tb[ATTR_NUM(value)].sa_length = ATTR_LENGTH(value);
926 tb[ATTR_NUM(value)].sa_byteswap = ATTR_BSWAP(value);
927 tb[ATTR_NUM(value)].sa_registered = B_TRUE;
928
929 if (tb[ATTR_NUM(value)].sa_name) {
930 continue;
931 }
932 tb[ATTR_NUM(value)].sa_name =
933 kmem_zalloc(strlen(za.za_name) +1, KM_SLEEP);
934 (void) strlcpy(tb[ATTR_NUM(value)].sa_name, za.za_name,
935 strlen(za.za_name) +1);
936 }
937 zap_cursor_fini(&zc);
938 /*
939 * Make sure we processed the correct number of registered
940 * attributes
941 */
942 if (registered_count != sa_reg_count) {
943 ASSERT(error != 0);
944 goto bail;
945 }
946
947 }
948
949 if (ostype == DMU_OST_ZFS) {
950 for (i = 0; i != sa_legacy_attr_count; i++) {
951 if (tb[i].sa_name)
952 continue;
953 tb[i].sa_attr = sa_legacy_attrs[i].sa_attr;
954 tb[i].sa_length = sa_legacy_attrs[i].sa_length;
955 tb[i].sa_byteswap = sa_legacy_attrs[i].sa_byteswap;
956 tb[i].sa_registered = B_FALSE;
957 tb[i].sa_name =
958 kmem_zalloc(strlen(sa_legacy_attrs[i].sa_name) +1,
959 KM_SLEEP);
960 (void) strlcpy(tb[i].sa_name,
961 sa_legacy_attrs[i].sa_name,
962 strlen(sa_legacy_attrs[i].sa_name) + 1);
963 }
964 }
965
966 for (i = 0; i != count; i++) {
967 sa_attr_type_t attr_id;
968
969 attr_id = sa->sa_user_table[i];
970 if (tb[attr_id].sa_name)
971 continue;
972
973 tb[attr_id].sa_length = reg_attrs[i].sa_length;
974 tb[attr_id].sa_byteswap = reg_attrs[i].sa_byteswap;
975 tb[attr_id].sa_attr = attr_id;
976 tb[attr_id].sa_name =
977 kmem_zalloc(strlen(reg_attrs[i].sa_name) + 1, KM_SLEEP);
978 (void) strlcpy(tb[attr_id].sa_name, reg_attrs[i].sa_name,
979 strlen(reg_attrs[i].sa_name) + 1);
980 }
981
982 sa->sa_need_attr_registration =
983 (sa_attr_count != registered_count);
984
985 return (0);
986 bail:
987 kmem_free(sa->sa_user_table, count * sizeof (sa_attr_type_t));
988 sa->sa_user_table = NULL;
989 sa_free_attr_table(sa);
990 return ((error != 0) ? error : EINVAL);
991 }
992
993 int
994 sa_setup(objset_t *os, uint64_t sa_obj, sa_attr_reg_t *reg_attrs, int count,
995 sa_attr_type_t **user_table)
996 {
997 zap_cursor_t zc;
998 zap_attribute_t za;
999 sa_os_t *sa;
1000 dmu_objset_type_t ostype = dmu_objset_type(os);
1001 sa_attr_type_t *tb;
1002 int error;
1003
1004 mutex_enter(&os->os_user_ptr_lock);
1005 if (os->os_sa) {
1006 mutex_enter(&os->os_sa->sa_lock);
1007 mutex_exit(&os->os_user_ptr_lock);
1008 tb = os->os_sa->sa_user_table;
1009 mutex_exit(&os->os_sa->sa_lock);
1010 *user_table = tb;
1011 return (0);
1012 }
1013
1014 sa = kmem_zalloc(sizeof (sa_os_t), KM_SLEEP);
1015 mutex_init(&sa->sa_lock, NULL, MUTEX_DEFAULT, NULL);
1016 sa->sa_master_obj = sa_obj;
1017
1018 os->os_sa = sa;
1019 mutex_enter(&sa->sa_lock);
1020 mutex_exit(&os->os_user_ptr_lock);
1021 avl_create(&sa->sa_layout_num_tree, layout_num_compare,
1022 sizeof (sa_lot_t), offsetof(sa_lot_t, lot_num_node));
1023 avl_create(&sa->sa_layout_hash_tree, layout_hash_compare,
1024 sizeof (sa_lot_t), offsetof(sa_lot_t, lot_hash_node));
1025
1026 if (sa_obj) {
1027 error = zap_lookup(os, sa_obj, SA_LAYOUTS,
1028 8, 1, &sa->sa_layout_attr_obj);
1029 if (error != 0 && error != ENOENT)
1030 goto fail;
1031 error = zap_lookup(os, sa_obj, SA_REGISTRY,
1032 8, 1, &sa->sa_reg_attr_obj);
1033 if (error != 0 && error != ENOENT)
1034 goto fail;
1035 }
1036
1037 if ((error = sa_attr_table_setup(os, reg_attrs, count)) != 0)
1038 goto fail;
1039
1040 if (sa->sa_layout_attr_obj != 0) {
1041 uint64_t layout_count;
1042
1043 error = zap_count(os, sa->sa_layout_attr_obj,
1044 &layout_count);
1045
1046 /*
1047 * Layout number count should be > 0
1048 */
1049 if (error || (error == 0 && layout_count == 0)) {
1050 if (error == 0)
1051 error = SET_ERROR(EINVAL);
1052 goto fail;
1053 }
1054
1055 for (zap_cursor_init(&zc, os, sa->sa_layout_attr_obj);
1056 (error = zap_cursor_retrieve(&zc, &za)) == 0;
1057 zap_cursor_advance(&zc)) {
1058 sa_attr_type_t *lot_attrs;
1059 uint64_t lot_num;
1060
1061 lot_attrs = kmem_zalloc(sizeof (sa_attr_type_t) *
1062 za.za_num_integers, KM_SLEEP);
1063
1064 if ((error = (zap_lookup(os, sa->sa_layout_attr_obj,
1065 za.za_name, 2, za.za_num_integers,
1066 lot_attrs))) != 0) {
1067 kmem_free(lot_attrs, sizeof (sa_attr_type_t) *
1068 za.za_num_integers);
1069 break;
1070 }
1071 VERIFY(ddi_strtoull(za.za_name, NULL, 10,
1072 (unsigned long long *)&lot_num) == 0);
1073
1074 (void) sa_add_layout_entry(os, lot_attrs,
1075 za.za_num_integers, lot_num,
1076 sa_layout_info_hash(lot_attrs,
1077 za.za_num_integers), B_FALSE, NULL);
1078 kmem_free(lot_attrs, sizeof (sa_attr_type_t) *
1079 za.za_num_integers);
1080 }
1081 zap_cursor_fini(&zc);
1082
1083 /*
1084 * Make sure layout count matches number of entries added
1085 * to AVL tree
1086 */
1087 if (avl_numnodes(&sa->sa_layout_num_tree) != layout_count) {
1088 ASSERT(error != 0);
1089 goto fail;
1090 }
1091 }
1092
1093 /* Add special layout number for old ZNODES */
1094 if (ostype == DMU_OST_ZFS) {
1095 (void) sa_add_layout_entry(os, sa_legacy_zpl_layout,
1096 sa_legacy_attr_count, 0,
1097 sa_layout_info_hash(sa_legacy_zpl_layout,
1098 sa_legacy_attr_count), B_FALSE, NULL);
1099
1100 (void) sa_add_layout_entry(os, sa_dummy_zpl_layout, 0, 1,
1101 0, B_FALSE, NULL);
1102 }
1103 *user_table = os->os_sa->sa_user_table;
1104 mutex_exit(&sa->sa_lock);
1105 return (0);
1106 fail:
1107 os->os_sa = NULL;
1108 sa_free_attr_table(sa);
1109 if (sa->sa_user_table)
1110 kmem_free(sa->sa_user_table, sa->sa_user_table_sz);
1111 mutex_exit(&sa->sa_lock);
1112 kmem_free(sa, sizeof (sa_os_t));
1113 return ((error == ECKSUM) ? EIO : error);
1114 }
1115
1116 void
1117 sa_tear_down(objset_t *os)
1118 {
1119 sa_os_t *sa = os->os_sa;
1120 sa_lot_t *layout;
1121 void *cookie;
1122
1123 kmem_free(sa->sa_user_table, sa->sa_user_table_sz);
1124
1125 /* Free up attr table */
1126
1127 sa_free_attr_table(sa);
1128
1129 cookie = NULL;
1130 while (layout = avl_destroy_nodes(&sa->sa_layout_hash_tree, &cookie)) {
1131 sa_idx_tab_t *tab;
1132 while (tab = list_head(&layout->lot_idx_tab)) {
1133 ASSERT(refcount_count(&tab->sa_refcount));
1134 sa_idx_tab_rele(os, tab);
1135 }
1136 }
1137
1138 cookie = NULL;
1139 while (layout = avl_destroy_nodes(&sa->sa_layout_num_tree, &cookie)) {
1140 kmem_free(layout->lot_attrs,
1141 sizeof (sa_attr_type_t) * layout->lot_attr_count);
1142 kmem_free(layout, sizeof (sa_lot_t));
1143 }
1144
1145 avl_destroy(&sa->sa_layout_hash_tree);
1146 avl_destroy(&sa->sa_layout_num_tree);
1147
1148 kmem_free(sa, sizeof (sa_os_t));
1149 os->os_sa = NULL;
1150 }
1151
1152 void
1153 sa_build_idx_tab(void *hdr, void *attr_addr, sa_attr_type_t attr,
1154 uint16_t length, int length_idx, boolean_t var_length, void *userp)
1155 {
1156 sa_idx_tab_t *idx_tab = userp;
1157
1158 if (var_length) {
1159 ASSERT(idx_tab->sa_variable_lengths);
1160 idx_tab->sa_variable_lengths[length_idx] = length;
1161 }
1162 TOC_ATTR_ENCODE(idx_tab->sa_idx_tab[attr], length_idx,
1163 (uint32_t)((uintptr_t)attr_addr - (uintptr_t)hdr));
1164 }
1165
1166 static void
1167 sa_attr_iter(objset_t *os, sa_hdr_phys_t *hdr, dmu_object_type_t type,
1168 sa_iterfunc_t func, sa_lot_t *tab, void *userp)
1169 {
1170 void *data_start;
1171 sa_lot_t *tb = tab;
1172 sa_lot_t search;
1173 avl_index_t loc;
1174 sa_os_t *sa = os->os_sa;
1175 int i;
1176 uint16_t *length_start = NULL;
1177 uint8_t length_idx = 0;
1178
1179 if (tab == NULL) {
1180 search.lot_num = SA_LAYOUT_NUM(hdr, type);
1181 tb = avl_find(&sa->sa_layout_num_tree, &search, &loc);
1182 ASSERT(tb);
1183 }
1184
1185 if (IS_SA_BONUSTYPE(type)) {
1186 data_start = (void *)P2ROUNDUP(((uintptr_t)hdr +
1187 offsetof(sa_hdr_phys_t, sa_lengths) +
1188 (sizeof (uint16_t) * tb->lot_var_sizes)), 8);
1189 length_start = hdr->sa_lengths;
1190 } else {
1191 data_start = hdr;
1192 }
1193
1194 for (i = 0; i != tb->lot_attr_count; i++) {
1195 int attr_length, reg_length;
1196 uint8_t idx_len;
1197
1198 reg_length = sa->sa_attr_table[tb->lot_attrs[i]].sa_length;
1199 if (reg_length) {
1200 attr_length = reg_length;
1201 idx_len = 0;
1202 } else {
1203 attr_length = length_start[length_idx];
1204 idx_len = length_idx++;
1205 }
1206
1207 func(hdr, data_start, tb->lot_attrs[i], attr_length,
1208 idx_len, reg_length == 0 ? B_TRUE : B_FALSE, userp);
1209
1210 data_start = (void *)P2ROUNDUP(((uintptr_t)data_start +
1211 attr_length), 8);
1212 }
1213 }
1214
1215 /*ARGSUSED*/
1216 void
1217 sa_byteswap_cb(void *hdr, void *attr_addr, sa_attr_type_t attr,
1218 uint16_t length, int length_idx, boolean_t variable_length, void *userp)
1219 {
1220 sa_handle_t *hdl = userp;
1221 sa_os_t *sa = hdl->sa_os->os_sa;
1222
1223 sa_bswap_table[sa->sa_attr_table[attr].sa_byteswap](attr_addr, length);
1224 }
1225
1226 void
1227 sa_byteswap(sa_handle_t *hdl, sa_buf_type_t buftype)
1228 {
1229 sa_hdr_phys_t *sa_hdr_phys = SA_GET_HDR(hdl, buftype);
1230 dmu_buf_impl_t *db;
1231 sa_os_t *sa = hdl->sa_os->os_sa;
1232 int num_lengths = 1;
1233 int i;
1234
1235 ASSERT(MUTEX_HELD(&sa->sa_lock));
1236 if (sa_hdr_phys->sa_magic == SA_MAGIC)
1237 return;
1238
1239 db = SA_GET_DB(hdl, buftype);
1240
1241 if (buftype == SA_SPILL) {
1242 arc_release(db->db_buf, NULL);
1243 arc_buf_thaw(db->db_buf);
1244 }
1245
1246 sa_hdr_phys->sa_magic = BSWAP_32(sa_hdr_phys->sa_magic);
1247 sa_hdr_phys->sa_layout_info = BSWAP_16(sa_hdr_phys->sa_layout_info);
1248
1249 /*
1250 * Determine number of variable lenghts in header
1251 * The standard 8 byte header has one for free and a
1252 * 16 byte header would have 4 + 1;
1253 */
1254 if (SA_HDR_SIZE(sa_hdr_phys) > 8)
1255 num_lengths += (SA_HDR_SIZE(sa_hdr_phys) - 8) >> 1;
1256 for (i = 0; i != num_lengths; i++)
1257 sa_hdr_phys->sa_lengths[i] =
1258 BSWAP_16(sa_hdr_phys->sa_lengths[i]);
1259
1260 sa_attr_iter(hdl->sa_os, sa_hdr_phys, DMU_OT_SA,
1261 sa_byteswap_cb, NULL, hdl);
1262
1263 if (buftype == SA_SPILL)
1264 arc_buf_freeze(((dmu_buf_impl_t *)hdl->sa_spill)->db_buf);
1265 }
1266
1267 static int
1268 sa_build_index(sa_handle_t *hdl, sa_buf_type_t buftype)
1269 {
1270 sa_hdr_phys_t *sa_hdr_phys;
1271 dmu_buf_impl_t *db = SA_GET_DB(hdl, buftype);
1272 dmu_object_type_t bonustype = SA_BONUSTYPE_FROM_DB(db);
1273 sa_os_t *sa = hdl->sa_os->os_sa;
1274 sa_idx_tab_t *idx_tab;
1275
1276 sa_hdr_phys = SA_GET_HDR(hdl, buftype);
1277
1278 mutex_enter(&sa->sa_lock);
1279
1280 /* Do we need to byteswap? */
1281
1282 /* only check if not old znode */
1283 if (IS_SA_BONUSTYPE(bonustype) && sa_hdr_phys->sa_magic != SA_MAGIC &&
1284 sa_hdr_phys->sa_magic != 0) {
1285 VERIFY(BSWAP_32(sa_hdr_phys->sa_magic) == SA_MAGIC);
1286 sa_byteswap(hdl, buftype);
1287 }
1288
1289 idx_tab = sa_find_idx_tab(hdl->sa_os, bonustype, sa_hdr_phys);
1290
1291 if (buftype == SA_BONUS)
1292 hdl->sa_bonus_tab = idx_tab;
1293 else
1294 hdl->sa_spill_tab = idx_tab;
1295
1296 mutex_exit(&sa->sa_lock);
1297 return (0);
1298 }
1299
1300 /*ARGSUSED*/
1301 void
1302 sa_evict(dmu_buf_user_t *dbu)
1303 {
1304 panic("evicting sa dbuf\n");
1305 }
1306
1307 static void
1308 sa_idx_tab_rele(objset_t *os, void *arg)
1309 {
1310 sa_os_t *sa = os->os_sa;
1311 sa_idx_tab_t *idx_tab = arg;
1312
1313 if (idx_tab == NULL)
1314 return;
1315
1316 mutex_enter(&sa->sa_lock);
1317 if (refcount_remove(&idx_tab->sa_refcount, NULL) == 0) {
1318 list_remove(&idx_tab->sa_layout->lot_idx_tab, idx_tab);
1319 if (idx_tab->sa_variable_lengths)
1320 kmem_free(idx_tab->sa_variable_lengths,
1321 sizeof (uint16_t) *
1322 idx_tab->sa_layout->lot_var_sizes);
1323 refcount_destroy(&idx_tab->sa_refcount);
1324 kmem_free(idx_tab->sa_idx_tab,
1325 sizeof (uint32_t) * sa->sa_num_attrs);
1326 kmem_free(idx_tab, sizeof (sa_idx_tab_t));
1327 }
1328 mutex_exit(&sa->sa_lock);
1329 }
1330
1331 static void
1332 sa_idx_tab_hold(objset_t *os, sa_idx_tab_t *idx_tab)
1333 {
1334 sa_os_t *sa = os->os_sa;
1335
1336 ASSERT(MUTEX_HELD(&sa->sa_lock));
1337 (void) refcount_add(&idx_tab->sa_refcount, NULL);
1338 }
1339
1340 void
1341 sa_handle_destroy(sa_handle_t *hdl)
1342 {
1343 dmu_buf_t *db = hdl->sa_bonus;
1344
1345 mutex_enter(&hdl->sa_lock);
1346 (void) dmu_buf_remove_user(db, &hdl->db_evict);
1347
1348 if (hdl->sa_bonus_tab) {
1349 sa_idx_tab_rele(hdl->sa_os, hdl->sa_bonus_tab);
1350 hdl->sa_bonus_tab = NULL;
1351 }
1352 if (hdl->sa_spill_tab) {
1353 sa_idx_tab_rele(hdl->sa_os, hdl->sa_spill_tab);
1354 hdl->sa_spill_tab = NULL;
1355 }
1356
1357 dmu_buf_rele(hdl->sa_bonus, NULL);
1358
1359 if (hdl->sa_spill)
1360 dmu_buf_rele((dmu_buf_t *)hdl->sa_spill, NULL);
1361 mutex_exit(&hdl->sa_lock);
1362
1363 kmem_cache_free(sa_cache, hdl);
1364 }
1365
1366 int
1367 sa_handle_get_from_db(objset_t *os, dmu_buf_t *db, void *userp,
1368 sa_handle_type_t hdl_type, sa_handle_t **handlepp)
1369 {
1370 int error = 0;
1371 dmu_object_info_t doi;
1372 sa_handle_t *handle = NULL;
1373
1374 #ifdef ZFS_DEBUG
1375 dmu_object_info_from_db(db, &doi);
1376 ASSERT(doi.doi_bonus_type == DMU_OT_SA ||
1377 doi.doi_bonus_type == DMU_OT_ZNODE);
1378 #endif
1379 /* find handle, if it exists */
1380 /* if one doesn't exist then create a new one, and initialize it */
1381
1382 if (hdl_type == SA_HDL_SHARED)
1383 handle = (sa_handle_t *)dmu_buf_get_user(db);
1384
1385 if (handle == NULL) {
1386 sa_handle_t *winner = NULL;
1387
1388 bzero(&handle->db_evict, sizeof(dmu_buf_user_t));
1389 handle = kmem_cache_alloc(sa_cache, KM_SLEEP);
1390 handle->sa_userp = userp;
1391 handle->sa_bonus = db;
1392 handle->sa_os = os;
1393 handle->sa_spill = NULL;
1394
1395 error = sa_build_index(handle, SA_BONUS);
1396 if (hdl_type == SA_HDL_SHARED) {
1397 dmu_buf_init_user(&handle->db_evict, sa_evict);
1398 winner = (sa_handle_t *)
1399 dmu_buf_set_user_ie(db, &handle->db_evict);
1400 }
1401 if (winner != NULL) {
1402 kmem_cache_free(sa_cache, handle);
1403 handle = winner;
1404 }
1405 }
1406 *handlepp = handle;
1407
1408 return (error);
1409 }
1410
1411 int
1412 sa_handle_get(objset_t *objset, uint64_t objid, void *userp,
1413 sa_handle_type_t hdl_type, sa_handle_t **handlepp)
1414 {
1415 dmu_buf_t *db;
1416 int error;
1417
1418 if (error = dmu_bonus_hold(objset, objid, NULL, &db))
1419 return (error);
1420
1421 return (sa_handle_get_from_db(objset, db, userp, hdl_type,
1422 handlepp));
1423 }
1424
1425 int
1426 sa_buf_hold(objset_t *objset, uint64_t obj_num, void *tag, dmu_buf_t **db)
1427 {
1428 return (dmu_bonus_hold(objset, obj_num, tag, db));
1429 }
1430
1431 void
1432 sa_buf_rele(dmu_buf_t *db, void *tag)
1433 {
1434 dmu_buf_rele(db, tag);
1435 }
1436
1437 int
1438 sa_lookup_impl(sa_handle_t *hdl, sa_bulk_attr_t *bulk, int count)
1439 {
1440 ASSERT(hdl);
1441 ASSERT(MUTEX_HELD(&hdl->sa_lock));
1442 return (sa_attr_op(hdl, bulk, count, SA_LOOKUP, NULL));
1443 }
1444
1445 int
1446 sa_lookup(sa_handle_t *hdl, sa_attr_type_t attr, void *buf, uint32_t buflen)
1447 {
1448 int error;
1449 sa_bulk_attr_t bulk;
1450
1451 bulk.sa_attr = attr;
1452 bulk.sa_data = buf;
1453 bulk.sa_length = buflen;
1454 bulk.sa_data_func = NULL;
1455
1456 ASSERT(hdl);
1457 mutex_enter(&hdl->sa_lock);
1458 error = sa_lookup_impl(hdl, &bulk, 1);
1459 mutex_exit(&hdl->sa_lock);
1460 return (error);
1461 }
1462
1463 #ifdef _KERNEL
1464 int
1465 sa_lookup_uio(sa_handle_t *hdl, sa_attr_type_t attr, uio_t *uio)
1466 {
1467 int error;
1468 sa_bulk_attr_t bulk;
1469
1470 bulk.sa_data = NULL;
1471 bulk.sa_attr = attr;
1472 bulk.sa_data_func = NULL;
1473
1474 ASSERT(hdl);
1475
1476 mutex_enter(&hdl->sa_lock);
1477 if ((error = sa_attr_op(hdl, &bulk, 1, SA_LOOKUP, NULL)) == 0) {
1478 error = uiomove((void *)bulk.sa_addr, MIN(bulk.sa_size,
1479 uio->uio_resid), UIO_READ, uio);
1480 }
1481 mutex_exit(&hdl->sa_lock);
1482 return (error);
1483
1484 }
1485 #endif
1486
1487 void *
1488 sa_find_idx_tab(objset_t *os, dmu_object_type_t bonustype, void *data)
1489 {
1490 sa_idx_tab_t *idx_tab;
1491 sa_hdr_phys_t *hdr = (sa_hdr_phys_t *)data;
1492 sa_os_t *sa = os->os_sa;
1493 sa_lot_t *tb, search;
1494 avl_index_t loc;
1495
1496 /*
1497 * Deterimine layout number. If SA node and header == 0 then
1498 * force the index table to the dummy "1" empty layout.
1499 *
1500 * The layout number would only be zero for a newly created file
1501 * that has not added any attributes yet, or with crypto enabled which
1502 * doesn't write any attributes to the bonus buffer.
1503 */
1504
1505 search.lot_num = SA_LAYOUT_NUM(hdr, bonustype);
1506
1507 tb = avl_find(&sa->sa_layout_num_tree, &search, &loc);
1508
1509 /* Verify header size is consistent with layout information */
1510 ASSERT(tb);
1511 ASSERT(IS_SA_BONUSTYPE(bonustype) &&
1512 SA_HDR_SIZE_MATCH_LAYOUT(hdr, tb) || !IS_SA_BONUSTYPE(bonustype) ||
1513 (IS_SA_BONUSTYPE(bonustype) && hdr->sa_layout_info == 0));
1514
1515 /*
1516 * See if any of the already existing TOC entries can be reused?
1517 */
1518
1519 for (idx_tab = list_head(&tb->lot_idx_tab); idx_tab;
1520 idx_tab = list_next(&tb->lot_idx_tab, idx_tab)) {
1521 boolean_t valid_idx = B_TRUE;
1522 int i;
1523
1524 if (tb->lot_var_sizes != 0 &&
1525 idx_tab->sa_variable_lengths != NULL) {
1526 for (i = 0; i != tb->lot_var_sizes; i++) {
1527 if (hdr->sa_lengths[i] !=
1528 idx_tab->sa_variable_lengths[i]) {
1529 valid_idx = B_FALSE;
1530 break;
1531 }
1532 }
1533 }
1534 if (valid_idx) {
1535 sa_idx_tab_hold(os, idx_tab);
1536 return (idx_tab);
1537 }
1538 }
1539
1540 /* No such luck, create a new entry */
1541 idx_tab = kmem_zalloc(sizeof (sa_idx_tab_t), KM_SLEEP);
1542 idx_tab->sa_idx_tab =
1543 kmem_zalloc(sizeof (uint32_t) * sa->sa_num_attrs, KM_SLEEP);
1544 idx_tab->sa_layout = tb;
1545 refcount_create(&idx_tab->sa_refcount);
1546 if (tb->lot_var_sizes)
1547 idx_tab->sa_variable_lengths = kmem_alloc(sizeof (uint16_t) *
1548 tb->lot_var_sizes, KM_SLEEP);
1549
1550 sa_attr_iter(os, hdr, bonustype, sa_build_idx_tab,
1551 tb, idx_tab);
1552 sa_idx_tab_hold(os, idx_tab); /* one hold for consumer */
1553 sa_idx_tab_hold(os, idx_tab); /* one for layout */
1554 list_insert_tail(&tb->lot_idx_tab, idx_tab);
1555 return (idx_tab);
1556 }
1557
1558 void
1559 sa_default_locator(void **dataptr, uint32_t *len, uint32_t total_len,
1560 boolean_t start, void *userdata)
1561 {
1562 ASSERT(start);
1563
1564 *dataptr = userdata;
1565 *len = total_len;
1566 }
1567
1568 static void
1569 sa_attr_register_sync(sa_handle_t *hdl, dmu_tx_t *tx)
1570 {
1571 uint64_t attr_value = 0;
1572 sa_os_t *sa = hdl->sa_os->os_sa;
1573 sa_attr_table_t *tb = sa->sa_attr_table;
1574 int i;
1575
1576 mutex_enter(&sa->sa_lock);
1577
1578 if (!sa->sa_need_attr_registration || sa->sa_master_obj == NULL) {
1579 mutex_exit(&sa->sa_lock);
1580 return;
1581 }
1582
1583 if (sa->sa_reg_attr_obj == NULL) {
1584 sa->sa_reg_attr_obj = zap_create_link(hdl->sa_os,
1585 DMU_OT_SA_ATTR_REGISTRATION,
1586 sa->sa_master_obj, SA_REGISTRY, tx);
1587 }
1588 for (i = 0; i != sa->sa_num_attrs; i++) {
1589 if (sa->sa_attr_table[i].sa_registered)
1590 continue;
1591 ATTR_ENCODE(attr_value, tb[i].sa_attr, tb[i].sa_length,
1592 tb[i].sa_byteswap);
1593 VERIFY(0 == zap_update(hdl->sa_os, sa->sa_reg_attr_obj,
1594 tb[i].sa_name, 8, 1, &attr_value, tx));
1595 tb[i].sa_registered = B_TRUE;
1596 }
1597 sa->sa_need_attr_registration = B_FALSE;
1598 mutex_exit(&sa->sa_lock);
1599 }
1600
1601 /*
1602 * Replace all attributes with attributes specified in template.
1603 * If dnode had a spill buffer then those attributes will be
1604 * also be replaced, possibly with just an empty spill block
1605 *
1606 * This interface is intended to only be used for bulk adding of
1607 * attributes for a new file. It will also be used by the ZPL
1608 * when converting and old formatted znode to native SA support.
1609 */
1610 int
1611 sa_replace_all_by_template_locked(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc,
1612 int attr_count, dmu_tx_t *tx)
1613 {
1614 sa_os_t *sa = hdl->sa_os->os_sa;
1615
1616 if (sa->sa_need_attr_registration)
1617 sa_attr_register_sync(hdl, tx);
1618 return (sa_build_layouts(hdl, attr_desc, attr_count, tx));
1619 }
1620
1621 int
1622 sa_replace_all_by_template(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc,
1623 int attr_count, dmu_tx_t *tx)
1624 {
1625 int error;
1626
1627 mutex_enter(&hdl->sa_lock);
1628 error = sa_replace_all_by_template_locked(hdl, attr_desc,
1629 attr_count, tx);
1630 mutex_exit(&hdl->sa_lock);
1631 return (error);
1632 }
1633
1634 /*
1635 * add/remove/replace a single attribute and then rewrite the entire set
1636 * of attributes.
1637 */
1638 static int
1639 sa_modify_attrs(sa_handle_t *hdl, sa_attr_type_t newattr,
1640 sa_data_op_t action, sa_data_locator_t *locator, void *datastart,
1641 uint16_t buflen, dmu_tx_t *tx)
1642 {
1643 sa_os_t *sa = hdl->sa_os->os_sa;
1644 dmu_buf_impl_t *db = (dmu_buf_impl_t *)hdl->sa_bonus;
1645 dnode_t *dn;
1646 sa_bulk_attr_t *attr_desc;
1647 void *old_data[2];
1648 int bonus_attr_count = 0;
1649 int bonus_data_size = 0;
1650 int spill_data_size = 0;
1651 int spill_attr_count = 0;
1652 int error;
1653 uint16_t length;
1654 int i, j, k, length_idx;
1655 sa_hdr_phys_t *hdr;
1656 sa_idx_tab_t *idx_tab;
1657 int attr_count;
1658 int count;
1659
1660 ASSERT(MUTEX_HELD(&hdl->sa_lock));
1661
1662 /* First make of copy of the old data */
1663
1664 DB_DNODE_ENTER(db);
1665 dn = DB_DNODE(db);
1666 if (dn->dn_bonuslen != 0) {
1667 bonus_data_size = hdl->sa_bonus->db_size;
1668 old_data[0] = kmem_alloc(bonus_data_size, KM_SLEEP);
1669 bcopy(hdl->sa_bonus->db_data, old_data[0],
1670 hdl->sa_bonus->db_size);
1671 bonus_attr_count = hdl->sa_bonus_tab->sa_layout->lot_attr_count;
1672 } else {
1673 old_data[0] = NULL;
1674 }
1675 DB_DNODE_EXIT(db);
1676
1677 /* Bring spill buffer online if it isn't currently */
1678
1679 if ((error = sa_get_spill(hdl)) == 0) {
1680 spill_data_size = hdl->sa_spill->db_size;
1681 old_data[1] = kmem_alloc(spill_data_size, KM_SLEEP);
1682 bcopy(hdl->sa_spill->db_data, old_data[1],
1683 hdl->sa_spill->db_size);
1684 spill_attr_count =
1685 hdl->sa_spill_tab->sa_layout->lot_attr_count;
1686 } else if (error && error != ENOENT) {
1687 if (old_data[0])
1688 kmem_free(old_data[0], bonus_data_size);
1689 return (error);
1690 } else {
1691 old_data[1] = NULL;
1692 }
1693
1694 /* build descriptor of all attributes */
1695
1696 attr_count = bonus_attr_count + spill_attr_count;
1697 if (action == SA_ADD)
1698 attr_count++;
1699 else if (action == SA_REMOVE)
1700 attr_count--;
1701
1702 attr_desc = kmem_zalloc(sizeof (sa_bulk_attr_t) * attr_count, KM_SLEEP);
1703
1704 /*
1705 * loop through bonus and spill buffer if it exists, and
1706 * build up new attr_descriptor to reset the attributes
1707 */
1708 k = j = 0;
1709 count = bonus_attr_count;
1710 hdr = SA_GET_HDR(hdl, SA_BONUS);
1711 idx_tab = SA_IDX_TAB_GET(hdl, SA_BONUS);
1712 for (; k != 2; k++) {
1713 /* iterate over each attribute in layout */
1714 for (i = 0, length_idx = 0; i != count; i++) {
1715 sa_attr_type_t attr;
1716
1717 attr = idx_tab->sa_layout->lot_attrs[i];
1718 if (attr == newattr) {
1719 if (action == SA_REMOVE) {
1720 j++;
1721 continue;
1722 }
1723 ASSERT(SA_REGISTERED_LEN(sa, attr) == 0);
1724 ASSERT(action == SA_REPLACE);
1725 SA_ADD_BULK_ATTR(attr_desc, j, attr,
1726 locator, datastart, buflen);
1727 } else {
1728 length = SA_REGISTERED_LEN(sa, attr);
1729 if (length == 0) {
1730 length = hdr->sa_lengths[length_idx++];
1731 }
1732
1733 SA_ADD_BULK_ATTR(attr_desc, j, attr,
1734 NULL, (void *)
1735 (TOC_OFF(idx_tab->sa_idx_tab[attr]) +
1736 (uintptr_t)old_data[k]), length);
1737 }
1738 }
1739 if (k == 0 && hdl->sa_spill) {
1740 hdr = SA_GET_HDR(hdl, SA_SPILL);
1741 idx_tab = SA_IDX_TAB_GET(hdl, SA_SPILL);
1742 count = spill_attr_count;
1743 } else {
1744 break;
1745 }
1746 }
1747 if (action == SA_ADD) {
1748 length = SA_REGISTERED_LEN(sa, newattr);
1749 if (length == 0) {
1750 length = buflen;
1751 }
1752 SA_ADD_BULK_ATTR(attr_desc, j, newattr, locator,
1753 datastart, buflen);
1754 }
1755
1756 error = sa_build_layouts(hdl, attr_desc, attr_count, tx);
1757
1758 if (old_data[0])
1759 kmem_free(old_data[0], bonus_data_size);
1760 if (old_data[1])
1761 kmem_free(old_data[1], spill_data_size);
1762 kmem_free(attr_desc, sizeof (sa_bulk_attr_t) * attr_count);
1763
1764 return (error);
1765 }
1766
1767 static int
1768 sa_bulk_update_impl(sa_handle_t *hdl, sa_bulk_attr_t *bulk, int count,
1769 dmu_tx_t *tx)
1770 {
1771 int error;
1772 sa_os_t *sa = hdl->sa_os->os_sa;
1773 dmu_object_type_t bonustype;
1774
1775 bonustype = SA_BONUSTYPE_FROM_DB(SA_GET_DB(hdl, SA_BONUS));
1776
1777 ASSERT(hdl);
1778 ASSERT(MUTEX_HELD(&hdl->sa_lock));
1779
1780 /* sync out registration table if necessary */
1781 if (sa->sa_need_attr_registration)
1782 sa_attr_register_sync(hdl, tx);
1783
1784 error = sa_attr_op(hdl, bulk, count, SA_UPDATE, tx);
1785 if (error == 0 && !IS_SA_BONUSTYPE(bonustype) && sa->sa_update_cb)
1786 sa->sa_update_cb(hdl, tx);
1787
1788 return (error);
1789 }
1790
1791 /*
1792 * update or add new attribute
1793 */
1794 int
1795 sa_update(sa_handle_t *hdl, sa_attr_type_t type,
1796 void *buf, uint32_t buflen, dmu_tx_t *tx)
1797 {
1798 int error;
1799 sa_bulk_attr_t bulk;
1800
1801 bulk.sa_attr = type;
1802 bulk.sa_data_func = NULL;
1803 bulk.sa_length = buflen;
1804 bulk.sa_data = buf;
1805
1806 mutex_enter(&hdl->sa_lock);
1807 error = sa_bulk_update_impl(hdl, &bulk, 1, tx);
1808 mutex_exit(&hdl->sa_lock);
1809 return (error);
1810 }
1811
1812 int
1813 sa_update_from_cb(sa_handle_t *hdl, sa_attr_type_t attr,
1814 uint32_t buflen, sa_data_locator_t *locator, void *userdata, dmu_tx_t *tx)
1815 {
1816 int error;
1817 sa_bulk_attr_t bulk;
1818
1819 bulk.sa_attr = attr;
1820 bulk.sa_data = userdata;
1821 bulk.sa_data_func = locator;
1822 bulk.sa_length = buflen;
1823
1824 mutex_enter(&hdl->sa_lock);
1825 error = sa_bulk_update_impl(hdl, &bulk, 1, tx);
1826 mutex_exit(&hdl->sa_lock);
1827 return (error);
1828 }
1829
1830 /*
1831 * Return size of an attribute
1832 */
1833
1834 int
1835 sa_size(sa_handle_t *hdl, sa_attr_type_t attr, int *size)
1836 {
1837 sa_bulk_attr_t bulk;
1838 int error;
1839
1840 bulk.sa_data = NULL;
1841 bulk.sa_attr = attr;
1842 bulk.sa_data_func = NULL;
1843
1844 ASSERT(hdl);
1845 mutex_enter(&hdl->sa_lock);
1846 if ((error = sa_attr_op(hdl, &bulk, 1, SA_LOOKUP, NULL)) != 0) {
1847 mutex_exit(&hdl->sa_lock);
1848 return (error);
1849 }
1850 *size = bulk.sa_size;
1851
1852 mutex_exit(&hdl->sa_lock);
1853 return (0);
1854 }
1855
1856 int
1857 sa_bulk_lookup_locked(sa_handle_t *hdl, sa_bulk_attr_t *attrs, int count)
1858 {
1859 ASSERT(hdl);
1860 ASSERT(MUTEX_HELD(&hdl->sa_lock));
1861 return (sa_lookup_impl(hdl, attrs, count));
1862 }
1863
1864 int
1865 sa_bulk_lookup(sa_handle_t *hdl, sa_bulk_attr_t *attrs, int count)
1866 {
1867 int error;
1868
1869 ASSERT(hdl);
1870 mutex_enter(&hdl->sa_lock);
1871 error = sa_bulk_lookup_locked(hdl, attrs, count);
1872 mutex_exit(&hdl->sa_lock);
1873 return (error);
1874 }
1875
1876 int
1877 sa_bulk_update(sa_handle_t *hdl, sa_bulk_attr_t *attrs, int count, dmu_tx_t *tx)
1878 {
1879 int error;
1880
1881 ASSERT(hdl);
1882 mutex_enter(&hdl->sa_lock);
1883 error = sa_bulk_update_impl(hdl, attrs, count, tx);
1884 mutex_exit(&hdl->sa_lock);
1885 return (error);
1886 }
1887
1888 int
1889 sa_remove(sa_handle_t *hdl, sa_attr_type_t attr, dmu_tx_t *tx)
1890 {
1891 int error;
1892
1893 mutex_enter(&hdl->sa_lock);
1894 error = sa_modify_attrs(hdl, attr, SA_REMOVE, NULL,
1895 NULL, 0, tx);
1896 mutex_exit(&hdl->sa_lock);
1897 return (error);
1898 }
1899
1900 void
1901 sa_object_info(sa_handle_t *hdl, dmu_object_info_t *doi)
1902 {
1903 dmu_object_info_from_db((dmu_buf_t *)hdl->sa_bonus, doi);
1904 }
1905
1906 void
1907 sa_object_size(sa_handle_t *hdl, uint32_t *blksize, u_longlong_t *nblocks)
1908 {
1909 dmu_object_size_from_db((dmu_buf_t *)hdl->sa_bonus,
1910 blksize, nblocks);
1911 }
1912
1913 void
1914 sa_update_user(sa_handle_t *newhdl, sa_handle_t *oldhdl)
1915 {
1916 dmu_buf_user_t *new_user = &newhdl->db_evict;
1917 dmu_buf_user_t *old_user = &oldhdl->db_evict;
1918
1919 dmu_buf_init_user(new_user, sa_evict);
1920 VERIFY(dmu_buf_replace_user(newhdl->sa_bonus, old_user,
1921 new_user) == old_user);
1922 oldhdl->sa_bonus = NULL;
1923 }
1924
1925 void
1926 sa_set_userp(sa_handle_t *hdl, void *ptr)
1927 {
1928 hdl->sa_userp = ptr;
1929 }
1930
1931 dmu_buf_t *
1932 sa_get_db(sa_handle_t *hdl)
1933 {
1934 return ((dmu_buf_t *)hdl->sa_bonus);
1935 }
1936
1937 void *
1938 sa_get_userdata(sa_handle_t *hdl)
1939 {
1940 return (hdl->sa_userp);
1941 }
1942
1943 void
1944 sa_register_update_callback_locked(objset_t *os, sa_update_cb_t *func)
1945 {
1946 ASSERT(MUTEX_HELD(&os->os_sa->sa_lock));
1947 os->os_sa->sa_update_cb = func;
1948 }
1949
1950 void
1951 sa_register_update_callback(objset_t *os, sa_update_cb_t *func)
1952 {
1953
1954 mutex_enter(&os->os_sa->sa_lock);
1955 sa_register_update_callback_locked(os, func);
1956 mutex_exit(&os->os_sa->sa_lock);
1957 }
1958
1959 uint64_t
1960 sa_handle_object(sa_handle_t *hdl)
1961 {
1962 return (hdl->sa_bonus->db_object);
1963 }
1964
1965 boolean_t
1966 sa_enabled(objset_t *os)
1967 {
1968 return (os->os_sa == NULL);
1969 }
1970
1971 int
1972 sa_set_sa_object(objset_t *os, uint64_t sa_object)
1973 {
1974 sa_os_t *sa = os->os_sa;
1975
1976 if (sa->sa_master_obj)
1977 return (1);
1978
1979 sa->sa_master_obj = sa_object;
1980
1981 return (0);
1982 }
1983
1984 int
1985 sa_hdrsize(void *arg)
1986 {
1987 sa_hdr_phys_t *hdr = arg;
1988
1989 return (SA_HDR_SIZE(hdr));
1990 }
1991
1992 void
1993 sa_handle_lock(sa_handle_t *hdl)
1994 {
1995 ASSERT(hdl);
1996 mutex_enter(&hdl->sa_lock);
1997 }
1998
1999 void
2000 sa_handle_unlock(sa_handle_t *hdl)
2001 {
2002 ASSERT(hdl);
2003 mutex_exit(&hdl->sa_lock);
2004 }