*
* CDDL HEADER END
*/
+
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2013 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2017 by Delphix. All rights reserved.
+ * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
+ * Copyright 2017 Nexenta Systems, Inc.
*/
#include <sys/zio.h>
#include <sys/zap_leaf.h>
#include <sys/avl.h>
#include <sys/arc.h>
+#include <sys/dmu_objset.h>
#ifdef _KERNEL
#include <sys/sunddi.h>
#endif
-static int mzap_upgrade(zap_t **zapp, dmu_tx_t *tx, zap_flags_t flags);
+extern inline mzap_phys_t *zap_m_phys(zap_t *zap);
+
+static int mzap_upgrade(zap_t **zapp,
+ void *tag, dmu_tx_t *tx, zap_flags_t flags);
uint64_t
zap_getflags(zap_t *zap)
{
if (zap->zap_ismicro)
return (0);
- return (zap->zap_u.zap_fat.zap_phys->zap_flags);
+ return (zap_f_phys(zap)->zap_flags);
}
int
ASSERT(zfs_crc64_table[128] == ZFS_CRC64_POLY);
if (zap_getflags(zap) & ZAP_FLAG_UINT64_KEY) {
- int i;
const uint64_t *wp = zn->zn_key_norm;
ASSERT(zn->zn_key_intlen == 8);
- for (i = 0; i < zn->zn_key_norm_numints; wp++, i++) {
- int j;
+ for (int i = 0; i < zn->zn_key_norm_numints;
+ wp++, i++) {
uint64_t word = *wp;
- for (j = 0; j < zn->zn_key_intlen; j++) {
+ for (int j = 0; j < zn->zn_key_intlen; j++) {
h = (h >> 8) ^
zfs_crc64_table[(h ^ word) & 0xFF];
word >>= NBBY;
}
}
} else {
- int i, len;
const uint8_t *cp = zn->zn_key_norm;
/*
* zn_key_*_numints includes the terminating
* null for non-binary keys.)
*/
- len = zn->zn_key_norm_numints - 1;
+ int len = zn->zn_key_norm_numints - 1;
ASSERT(zn->zn_key_intlen == 1);
- for (i = 0; i < len; cp++, i++) {
+ for (int i = 0; i < len; cp++, i++) {
h = (h >> 8) ^
zfs_crc64_table[(h ^ *cp) & 0xFF];
}
* Don't use all 64 bits, since we need some in the cookie for
* the collision differentiator. We MUST use the high bits,
* since those are the ones that we first pay attention to when
- * chosing the bucket.
+ * choosing the bucket.
*/
h &= ~((1ULL << (64 - zap_hashbits(zap))) - 1);
}
static int
-zap_normalize(zap_t *zap, const char *name, char *namenorm)
+zap_normalize(zap_t *zap, const char *name, char *namenorm, int normflags)
{
- size_t inlen, outlen;
- int err;
-
ASSERT(!(zap_getflags(zap) & ZAP_FLAG_UINT64_KEY));
- inlen = strlen(name) + 1;
- outlen = ZAP_MAXNAMELEN;
+ size_t inlen = strlen(name) + 1;
+ size_t outlen = ZAP_MAXNAMELEN;
- err = 0;
+ int err = 0;
(void) u8_textprep_str((char *)name, &inlen, namenorm, &outlen,
- zap->zap_normflags | U8_TEXTPREP_IGNORE_NULL |
- U8_TEXTPREP_IGNORE_INVALID, U8_UNICODE_LATEST, &err);
+ normflags | U8_TEXTPREP_IGNORE_NULL | U8_TEXTPREP_IGNORE_INVALID,
+ U8_UNICODE_LATEST, &err);
return (err);
}
{
ASSERT(!(zap_getflags(zn->zn_zap) & ZAP_FLAG_UINT64_KEY));
- if (zn->zn_matchtype == MT_FIRST) {
+ if (zn->zn_matchtype & MT_NORMALIZE) {
char norm[ZAP_MAXNAMELEN];
- if (zap_normalize(zn->zn_zap, matchname, norm) != 0)
+ if (zap_normalize(zn->zn_zap, matchname, norm,
+ zn->zn_normflags) != 0)
return (B_FALSE);
return (strcmp(zn->zn_key_norm, norm) == 0);
} else {
- /* MT_BEST or MT_EXACT */
return (strcmp(zn->zn_key_orig, matchname) == 0);
}
}
zap_name_t *
zap_name_alloc(zap_t *zap, const char *key, matchtype_t mt)
{
- zap_name_t *zn = kmem_alloc(sizeof (zap_name_t), KM_PUSHPAGE);
+ zap_name_t *zn = kmem_alloc(sizeof (zap_name_t), KM_SLEEP);
zn->zn_zap = zap;
zn->zn_key_intlen = sizeof (*key);
zn->zn_key_orig = key;
zn->zn_key_orig_numints = strlen(zn->zn_key_orig) + 1;
zn->zn_matchtype = mt;
+ zn->zn_normflags = zap->zap_normflags;
+
+ /*
+ * If we're dealing with a case sensitive lookup on a mixed or
+ * insensitive fs, remove U8_TEXTPREP_TOUPPER or the lookup
+ * will fold case to all caps overriding the lookup request.
+ */
+ if (mt & MT_MATCH_CASE)
+ zn->zn_normflags &= ~U8_TEXTPREP_TOUPPER;
+
if (zap->zap_normflags) {
- if (zap_normalize(zap, key, zn->zn_normbuf) != 0) {
+ /*
+ * We *must* use zap_normflags because this normalization is
+ * what the hash is computed from.
+ */
+ if (zap_normalize(zap, key, zn->zn_normbuf,
+ zap->zap_normflags) != 0) {
zap_name_free(zn);
return (NULL);
}
zn->zn_key_norm = zn->zn_normbuf;
zn->zn_key_norm_numints = strlen(zn->zn_key_norm) + 1;
} else {
- if (mt != MT_EXACT) {
+ if (mt != 0) {
zap_name_free(zn);
return (NULL);
}
}
zn->zn_hash = zap_hash(zn);
+
+ if (zap->zap_normflags != zn->zn_normflags) {
+ /*
+ * We *must* use zn_normflags because this normalization is
+ * what the matching is based on. (Not the hash!)
+ */
+ if (zap_normalize(zap, key, zn->zn_normbuf,
+ zn->zn_normflags) != 0) {
+ zap_name_free(zn);
+ return (NULL);
+ }
+ zn->zn_key_norm_numints = strlen(zn->zn_key_norm) + 1;
+ }
+
return (zn);
}
zap_name_t *
zap_name_alloc_uint64(zap_t *zap, const uint64_t *key, int numints)
{
- zap_name_t *zn = kmem_alloc(sizeof (zap_name_t), KM_PUSHPAGE);
+ zap_name_t *zn = kmem_alloc(sizeof (zap_name_t), KM_SLEEP);
ASSERT(zap->zap_normflags == 0);
zn->zn_zap = zap;
zn->zn_key_intlen = sizeof (*key);
zn->zn_key_orig = zn->zn_key_norm = key;
zn->zn_key_orig_numints = zn->zn_key_norm_numints = numints;
- zn->zn_matchtype = MT_EXACT;
+ zn->zn_matchtype = 0;
zn->zn_hash = zap_hash(zn);
return (zn);
static void
mzap_byteswap(mzap_phys_t *buf, size_t size)
{
- int i, max;
buf->mz_block_type = BSWAP_64(buf->mz_block_type);
buf->mz_salt = BSWAP_64(buf->mz_salt);
buf->mz_normflags = BSWAP_64(buf->mz_normflags);
- max = (size / MZAP_ENT_LEN) - 1;
- for (i = 0; i < max; i++) {
+ int max = (size / MZAP_ENT_LEN) - 1;
+ for (int i = 0; i < max; i++) {
buf->mz_chunk[i].mze_value =
BSWAP_64(buf->mz_chunk[i].mze_value);
buf->mz_chunk[i].mze_cd =
void
zap_byteswap(void *buf, size_t size)
{
- uint64_t block_type;
-
- block_type = *(uint64_t *)buf;
+ uint64_t block_type = *(uint64_t *)buf;
if (block_type == ZBT_MICRO || block_type == BSWAP_64(ZBT_MICRO)) {
/* ASSERT(magic == ZAP_LEAF_MAGIC); */
const mzap_ent_t *mze1 = arg1;
const mzap_ent_t *mze2 = arg2;
- if (mze1->mze_hash > mze2->mze_hash)
- return (+1);
- if (mze1->mze_hash < mze2->mze_hash)
- return (-1);
- if (mze1->mze_cd > mze2->mze_cd)
- return (+1);
- if (mze1->mze_cd < mze2->mze_cd)
- return (-1);
- return (0);
+ int cmp = AVL_CMP(mze1->mze_hash, mze2->mze_hash);
+ if (likely(cmp))
+ return (cmp);
+
+ return (AVL_CMP(mze1->mze_cd, mze2->mze_cd));
}
static void
mze_insert(zap_t *zap, int chunkid, uint64_t hash)
{
- mzap_ent_t *mze;
-
ASSERT(zap->zap_ismicro);
ASSERT(RW_WRITE_HELD(&zap->zap_rwlock));
- mze = kmem_alloc(sizeof (mzap_ent_t), KM_PUSHPAGE);
+ mzap_ent_t *mze = kmem_alloc(sizeof (mzap_ent_t), KM_SLEEP);
mze->mze_chunkid = chunkid;
mze->mze_hash = hash;
mze->mze_cd = MZE_PHYS(zap, mze)->mze_cd;
mze_tofind.mze_hash = zn->zn_hash;
mze_tofind.mze_cd = 0;
-again:
mze = avl_find(avl, &mze_tofind, &idx);
if (mze == NULL)
mze = avl_nearest(avl, idx, AVL_AFTER);
if (zap_match(zn, MZE_PHYS(zn->zn_zap, mze)->mze_name))
return (mze);
}
- if (zn->zn_matchtype == MT_BEST) {
- zn->zn_matchtype = MT_FIRST;
- goto again;
- }
+
return (NULL);
}
mze_find_unused_cd(zap_t *zap, uint64_t hash)
{
mzap_ent_t mze_tofind;
- mzap_ent_t *mze;
avl_index_t idx;
avl_tree_t *avl = &zap->zap_m.zap_avl;
- uint32_t cd;
ASSERT(zap->zap_ismicro);
ASSERT(RW_LOCK_HELD(&zap->zap_rwlock));
mze_tofind.mze_hash = hash;
mze_tofind.mze_cd = 0;
- cd = 0;
- for (mze = avl_find(avl, &mze_tofind, &idx);
+ uint32_t cd = 0;
+ for (mzap_ent_t *mze = avl_find(avl, &mze_tofind, &idx);
mze && mze->mze_hash == hash; mze = AVL_NEXT(avl, mze)) {
if (mze->mze_cd != cd)
break;
return (cd);
}
+/*
+ * Each mzap entry requires at max : 4 chunks
+ * 3 chunks for names + 1 chunk for value.
+ */
+#define MZAP_ENT_CHUNKS (1 + ZAP_LEAF_ARRAY_NCHUNKS(MZAP_NAME_LEN) + \
+ ZAP_LEAF_ARRAY_NCHUNKS(sizeof (uint64_t)))
+
+/*
+ * Check if the current entry keeps the colliding entries under the fatzap leaf
+ * size.
+ */
+static boolean_t
+mze_canfit_fzap_leaf(zap_name_t *zn, uint64_t hash)
+{
+ zap_t *zap = zn->zn_zap;
+ mzap_ent_t mze_tofind;
+ mzap_ent_t *mze;
+ avl_index_t idx;
+ avl_tree_t *avl = &zap->zap_m.zap_avl;
+ uint32_t mzap_ents = 0;
+
+ mze_tofind.mze_hash = hash;
+ mze_tofind.mze_cd = 0;
+
+ for (mze = avl_find(avl, &mze_tofind, &idx);
+ mze && mze->mze_hash == hash; mze = AVL_NEXT(avl, mze)) {
+ mzap_ents++;
+ }
+
+ /* Include the new entry being added */
+ mzap_ents++;
+
+ return (ZAP_LEAF_NUMCHUNKS_DEF > (mzap_ents * MZAP_ENT_CHUNKS));
+}
+
static void
mze_remove(zap_t *zap, mzap_ent_t *mze)
{
mzap_open(objset_t *os, uint64_t obj, dmu_buf_t *db)
{
zap_t *winner;
- zap_t *zap;
- int i;
+ uint64_t *zap_hdr = (uint64_t *)db->db_data;
+ uint64_t zap_block_type = zap_hdr[0];
+ uint64_t zap_magic = zap_hdr[1];
ASSERT3U(MZAP_ENT_LEN, ==, sizeof (mzap_ent_phys_t));
- zap = kmem_zalloc(sizeof (zap_t), KM_PUSHPAGE);
+ zap_t *zap = kmem_zalloc(sizeof (zap_t), KM_SLEEP);
rw_init(&zap->zap_rwlock, NULL, RW_DEFAULT, NULL);
rw_enter(&zap->zap_rwlock, RW_WRITER);
zap->zap_objset = os;
zap->zap_object = obj;
zap->zap_dbuf = db;
- if (*(uint64_t *)db->db_data != ZBT_MICRO) {
- mutex_init(&zap->zap_f.zap_num_entries_mtx, 0, 0, 0);
- zap->zap_f.zap_block_shift = highbit(db->db_size) - 1;
+ if (zap_block_type != ZBT_MICRO) {
+ mutex_init(&zap->zap_f.zap_num_entries_mtx, 0, MUTEX_DEFAULT,
+ 0);
+ zap->zap_f.zap_block_shift = highbit64(db->db_size) - 1;
+ if (zap_block_type != ZBT_HEADER || zap_magic != ZAP_MAGIC) {
+ winner = NULL; /* No actual winner here... */
+ goto handle_winner;
+ }
} else {
zap->zap_ismicro = TRUE;
}
* it, because zap_lockdir() checks zap_ismicro without the lock
* held.
*/
- winner = dmu_buf_set_user(db, zap, &zap->zap_m.zap_phys, zap_evict);
-
- if (winner != NULL) {
- rw_exit(&zap->zap_rwlock);
- rw_destroy(&zap->zap_rwlock);
- if (!zap->zap_ismicro)
- mutex_destroy(&zap->zap_f.zap_num_entries_mtx);
- kmem_free(zap, sizeof (zap_t));
- return (winner);
- }
+ dmu_buf_init_user(&zap->zap_dbu, zap_evict_sync, NULL, &zap->zap_dbuf);
+ winner = dmu_buf_set_user(db, &zap->zap_dbu);
+
+ if (winner != NULL)
+ goto handle_winner;
if (zap->zap_ismicro) {
- zap->zap_salt = zap->zap_m.zap_phys->mz_salt;
- zap->zap_normflags = zap->zap_m.zap_phys->mz_normflags;
+ zap->zap_salt = zap_m_phys(zap)->mz_salt;
+ zap->zap_normflags = zap_m_phys(zap)->mz_normflags;
zap->zap_m.zap_num_chunks = db->db_size / MZAP_ENT_LEN - 1;
avl_create(&zap->zap_m.zap_avl, mze_compare,
sizeof (mzap_ent_t), offsetof(mzap_ent_t, mze_node));
- for (i = 0; i < zap->zap_m.zap_num_chunks; i++) {
+ for (int i = 0; i < zap->zap_m.zap_num_chunks; i++) {
mzap_ent_phys_t *mze =
- &zap->zap_m.zap_phys->mz_chunk[i];
+ &zap_m_phys(zap)->mz_chunk[i];
if (mze->mze_name[0]) {
zap_name_t *zn;
zap->zap_m.zap_num_entries++;
- zn = zap_name_alloc(zap, mze->mze_name,
- MT_EXACT);
+ zn = zap_name_alloc(zap, mze->mze_name, 0);
mze_insert(zap, i, zn->zn_hash);
zap_name_free(zn);
}
}
} else {
- zap->zap_salt = zap->zap_f.zap_phys->zap_salt;
- zap->zap_normflags = zap->zap_f.zap_phys->zap_normflags;
+ zap->zap_salt = zap_f_phys(zap)->zap_salt;
+ zap->zap_normflags = zap_f_phys(zap)->zap_normflags;
ASSERT3U(sizeof (struct zap_leaf_header), ==,
2*ZAP_LEAF_CHUNKSIZE);
* other members.
*/
ASSERT3P(&ZAP_EMBEDDED_PTRTBL_ENT(zap, 0), >,
- &zap->zap_f.zap_phys->zap_salt);
+ &zap_f_phys(zap)->zap_salt);
/*
* The embedded pointer table should end at the end of
*/
ASSERT3U((uintptr_t)&ZAP_EMBEDDED_PTRTBL_ENT(zap,
1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap)) -
- (uintptr_t)zap->zap_f.zap_phys, ==,
+ (uintptr_t)zap_f_phys(zap), ==,
zap->zap_dbuf->db_size);
}
rw_exit(&zap->zap_rwlock);
return (zap);
+
+handle_winner:
+ rw_exit(&zap->zap_rwlock);
+ rw_destroy(&zap->zap_rwlock);
+ if (!zap->zap_ismicro)
+ mutex_destroy(&zap->zap_f.zap_num_entries_mtx);
+ kmem_free(zap, sizeof (zap_t));
+ return (winner);
}
-int
-zap_lockdir(objset_t *os, uint64_t obj, dmu_tx_t *tx,
+/*
+ * This routine "consumes" the caller's hold on the dbuf, which must
+ * have the specified tag.
+ */
+static int
+zap_lockdir_impl(dmu_buf_t *db, void *tag, dmu_tx_t *tx,
krw_t lti, boolean_t fatreader, boolean_t adding, zap_t **zapp)
{
- zap_t *zap;
- dmu_buf_t *db;
- krw_t lt;
- int err;
+ ASSERT0(db->db_offset);
+ objset_t *os = dmu_buf_get_objset(db);
+ uint64_t obj = db->db_object;
+ dmu_object_info_t doi;
*zapp = NULL;
- err = dmu_buf_hold(os, obj, 0, NULL, &db, DMU_READ_NO_PREFETCH);
- if (err)
- return (err);
-
-#ifdef ZFS_DEBUG
- {
- dmu_object_info_t doi;
- dmu_object_info_from_db(db, &doi);
- ASSERT3U(DMU_OT_BYTESWAP(doi.doi_type), ==, DMU_BSWAP_ZAP);
- }
-#endif
+ dmu_object_info_from_db(db, &doi);
+ if (DMU_OT_BYTESWAP(doi.doi_type) != DMU_BSWAP_ZAP)
+ return (SET_ERROR(EINVAL));
- zap = dmu_buf_get_user(db);
- if (zap == NULL)
+ zap_t *zap = dmu_buf_get_user(db);
+ if (zap == NULL) {
zap = mzap_open(os, obj, db);
+ if (zap == NULL) {
+ /*
+ * mzap_open() didn't like what it saw on-disk.
+ * Check for corruption!
+ */
+ return (SET_ERROR(EIO));
+ }
+ }
/*
* We're checking zap_ismicro without the lock held, in order to
* can only be different if it was upgraded from micro to fat,
* and micro wanted WRITER but fat only needs READER.
*/
- lt = (!zap->zap_ismicro && fatreader) ? RW_READER : lti;
+ krw_t lt = (!zap->zap_ismicro && fatreader) ? RW_READER : lti;
rw_enter(&zap->zap_rwlock, lt);
if (lt != ((!zap->zap_ismicro && fatreader) ? RW_READER : lti)) {
/* it was upgraded, now we only need reader */
ASSERT(lt == RW_WRITER);
ASSERT(RW_READER ==
- (!zap->zap_ismicro && fatreader) ? RW_READER : lti);
+ ((!zap->zap_ismicro && fatreader) ? RW_READER : lti));
rw_downgrade(&zap->zap_rwlock);
lt = RW_READER;
}
dprintf("upgrading obj %llu: num_entries=%u\n",
obj, zap->zap_m.zap_num_entries);
*zapp = zap;
- return (mzap_upgrade(zapp, tx, 0));
+ int err = mzap_upgrade(zapp, tag, tx, 0);
+ if (err != 0)
+ rw_exit(&zap->zap_rwlock);
+ return (err);
}
- err = dmu_object_set_blocksize(os, obj, newsz, 0, tx);
- ASSERT0(err);
+ VERIFY0(dmu_object_set_blocksize(os, obj, newsz, 0, tx));
zap->zap_m.zap_num_chunks =
db->db_size / MZAP_ENT_LEN - 1;
}
return (0);
}
+static int
+zap_lockdir_by_dnode(dnode_t *dn, dmu_tx_t *tx,
+ krw_t lti, boolean_t fatreader, boolean_t adding, void *tag, zap_t **zapp)
+{
+ dmu_buf_t *db;
+
+ int err = dmu_buf_hold_by_dnode(dn, 0, tag, &db, DMU_READ_NO_PREFETCH);
+ if (err != 0) {
+ return (err);
+ }
+#ifdef ZFS_DEBUG
+ {
+ dmu_object_info_t doi;
+ dmu_object_info_from_db(db, &doi);
+ ASSERT3U(DMU_OT_BYTESWAP(doi.doi_type), ==, DMU_BSWAP_ZAP);
+ }
+#endif
+
+ err = zap_lockdir_impl(db, tag, tx, lti, fatreader, adding, zapp);
+ if (err != 0) {
+ dmu_buf_rele(db, tag);
+ }
+ return (err);
+}
+
+int
+zap_lockdir(objset_t *os, uint64_t obj, dmu_tx_t *tx,
+ krw_t lti, boolean_t fatreader, boolean_t adding, void *tag, zap_t **zapp)
+{
+ dmu_buf_t *db;
+
+ int err = dmu_buf_hold(os, obj, 0, tag, &db, DMU_READ_NO_PREFETCH);
+ if (err != 0)
+ return (err);
+#ifdef ZFS_DEBUG
+ {
+ dmu_object_info_t doi;
+ dmu_object_info_from_db(db, &doi);
+ ASSERT3U(DMU_OT_BYTESWAP(doi.doi_type), ==, DMU_BSWAP_ZAP);
+ }
+#endif
+ err = zap_lockdir_impl(db, tag, tx, lti, fatreader, adding, zapp);
+ if (err != 0)
+ dmu_buf_rele(db, tag);
+ return (err);
+}
+
void
-zap_unlockdir(zap_t *zap)
+zap_unlockdir(zap_t *zap, void *tag)
{
rw_exit(&zap->zap_rwlock);
- dmu_buf_rele(zap->zap_dbuf, NULL);
+ dmu_buf_rele(zap->zap_dbuf, tag);
}
static int
-mzap_upgrade(zap_t **zapp, dmu_tx_t *tx, zap_flags_t flags)
+mzap_upgrade(zap_t **zapp, void *tag, dmu_tx_t *tx, zap_flags_t flags)
{
- mzap_phys_t *mzp;
- int i, sz, nchunks;
int err = 0;
zap_t *zap = *zapp;
ASSERT(RW_WRITE_HELD(&zap->zap_rwlock));
- sz = zap->zap_dbuf->db_size;
- mzp = kmem_alloc(sz, KM_PUSHPAGE | KM_NODEBUG);
+ int sz = zap->zap_dbuf->db_size;
+ mzap_phys_t *mzp = vmem_alloc(sz, KM_SLEEP);
bcopy(zap->zap_dbuf->db_data, mzp, sz);
- nchunks = zap->zap_m.zap_num_chunks;
+ int nchunks = zap->zap_m.zap_num_chunks;
if (!flags) {
err = dmu_object_set_blocksize(zap->zap_objset, zap->zap_object,
1ULL << fzap_default_block_shift, 0, tx);
- if (err) {
- kmem_free(mzp, sz);
+ if (err != 0) {
+ vmem_free(mzp, sz);
return (err);
}
}
fzap_upgrade(zap, tx, flags);
- for (i = 0; i < nchunks; i++) {
+ for (int i = 0; i < nchunks; i++) {
mzap_ent_phys_t *mze = &mzp->mz_chunk[i];
- zap_name_t *zn;
if (mze->mze_name[0] == 0)
continue;
dprintf("adding %s=%llu\n",
mze->mze_name, mze->mze_value);
- zn = zap_name_alloc(zap, mze->mze_name, MT_EXACT);
- err = fzap_add_cd(zn, 8, 1, &mze->mze_value, mze->mze_cd, tx);
+ zap_name_t *zn = zap_name_alloc(zap, mze->mze_name, 0);
+ /* If we fail here, we would end up losing entries */
+ VERIFY0(fzap_add_cd(zn, 8, 1, &mze->mze_value, mze->mze_cd,
+ tag, tx));
zap = zn->zn_zap; /* fzap_add_cd() may change zap */
zap_name_free(zn);
- if (err)
- break;
}
- kmem_free(mzp, sz);
+ vmem_free(mzp, sz);
*zapp = zap;
- return (err);
+ return (0);
}
+/*
+ * The "normflags" determine the behavior of the matchtype_t which is
+ * passed to zap_lookup_norm(). Names which have the same normalized
+ * version will be stored with the same hash value, and therefore we can
+ * perform normalization-insensitive lookups. We can be Unicode form-
+ * insensitive and/or case-insensitive. The following flags are valid for
+ * "normflags":
+ *
+ * U8_TEXTPREP_NFC
+ * U8_TEXTPREP_NFD
+ * U8_TEXTPREP_NFKC
+ * U8_TEXTPREP_NFKD
+ * U8_TEXTPREP_TOUPPER
+ *
+ * The *_NF* (Normalization Form) flags are mutually exclusive; at most one
+ * of them may be supplied.
+ */
void
-mzap_create_impl(objset_t *os, uint64_t obj, int normflags, zap_flags_t flags,
- dmu_tx_t *tx)
+mzap_create_impl(dnode_t *dn, int normflags, zap_flags_t flags, dmu_tx_t *tx)
{
dmu_buf_t *db;
- mzap_phys_t *zp;
- VERIFY(0 == dmu_buf_hold(os, obj, 0, FTAG, &db, DMU_READ_NO_PREFETCH));
-
-#ifdef ZFS_DEBUG
- {
- dmu_object_info_t doi;
- dmu_object_info_from_db(db, &doi);
- ASSERT3U(DMU_OT_BYTESWAP(doi.doi_type), ==, DMU_BSWAP_ZAP);
- }
-#endif
+ VERIFY0(dmu_buf_hold_by_dnode(dn, 0, FTAG, &db, DMU_READ_NO_PREFETCH));
dmu_buf_will_dirty(db, tx);
- zp = db->db_data;
+ mzap_phys_t *zp = db->db_data;
zp->mz_block_type = ZBT_MICRO;
- zp->mz_salt = ((uintptr_t)db ^ (uintptr_t)tx ^ (obj << 1)) | 1ULL;
+ zp->mz_salt =
+ ((uintptr_t)db ^ (uintptr_t)tx ^ (dn->dn_object << 1)) | 1ULL;
zp->mz_normflags = normflags;
- dmu_buf_rele(db, FTAG);
if (flags != 0) {
zap_t *zap;
/* Only fat zap supports flags; upgrade immediately. */
- VERIFY(0 == zap_lockdir(os, obj, tx, RW_WRITER,
+ VERIFY0(zap_lockdir_impl(db, FTAG, tx, RW_WRITER,
B_FALSE, B_FALSE, &zap));
- VERIFY3U(0, ==, mzap_upgrade(&zap, tx, flags));
- zap_unlockdir(zap);
+ VERIFY0(mzap_upgrade(&zap, FTAG, tx, flags));
+ zap_unlockdir(zap, FTAG);
+ } else {
+ dmu_buf_rele(db, FTAG);
+ }
+}
+
+static uint64_t
+zap_create_impl(objset_t *os, int normflags, zap_flags_t flags,
+ dmu_object_type_t ot, int leaf_blockshift, int indirect_blockshift,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize,
+ dnode_t **allocated_dnode, void *tag, dmu_tx_t *tx)
+{
+ uint64_t obj;
+
+ ASSERT3U(DMU_OT_BYTESWAP(ot), ==, DMU_BSWAP_ZAP);
+
+ if (allocated_dnode == NULL) {
+ dnode_t *dn;
+ obj = dmu_object_alloc_hold(os, ot, 1ULL << leaf_blockshift,
+ indirect_blockshift, bonustype, bonuslen, dnodesize,
+ &dn, FTAG, tx);
+ mzap_create_impl(dn, normflags, flags, tx);
+ dnode_rele(dn, FTAG);
+ } else {
+ obj = dmu_object_alloc_hold(os, ot, 1ULL << leaf_blockshift,
+ indirect_blockshift, bonustype, bonuslen, dnodesize,
+ allocated_dnode, tag, tx);
+ mzap_create_impl(*allocated_dnode, normflags, flags, tx);
}
+
+ return (obj);
}
int
zap_create_claim(objset_t *os, uint64_t obj, dmu_object_type_t ot,
dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx)
{
- return (zap_create_claim_norm(os, obj,
- 0, ot, bonustype, bonuslen, tx));
+ return (zap_create_claim_dnsize(os, obj, ot, bonustype, bonuslen,
+ 0, tx));
+}
+
+int
+zap_create_claim_dnsize(objset_t *os, uint64_t obj, dmu_object_type_t ot,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx)
+{
+ return (zap_create_claim_norm_dnsize(os, obj,
+ 0, ot, bonustype, bonuslen, dnodesize, tx));
}
int
dmu_object_type_t ot,
dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx)
{
- int err;
+ return (zap_create_claim_norm_dnsize(os, obj, normflags, ot, bonustype,
+ bonuslen, 0, tx));
+}
+
+int
+zap_create_claim_norm_dnsize(objset_t *os, uint64_t obj, int normflags,
+ dmu_object_type_t ot, dmu_object_type_t bonustype, int bonuslen,
+ int dnodesize, dmu_tx_t *tx)
+{
+ dnode_t *dn;
+ int error;
+
+ ASSERT3U(DMU_OT_BYTESWAP(ot), ==, DMU_BSWAP_ZAP);
+ error = dmu_object_claim_dnsize(os, obj, ot, 0, bonustype, bonuslen,
+ dnodesize, tx);
+ if (error != 0)
+ return (error);
+
+ error = dnode_hold(os, obj, FTAG, &dn);
+ if (error != 0)
+ return (error);
+
+ mzap_create_impl(dn, normflags, 0, tx);
+
+ dnode_rele(dn, FTAG);
- err = dmu_object_claim(os, obj, ot, 0, bonustype, bonuslen, tx);
- if (err != 0)
- return (err);
- mzap_create_impl(os, obj, normflags, 0, tx);
return (0);
}
return (zap_create_norm(os, 0, ot, bonustype, bonuslen, tx));
}
+uint64_t
+zap_create_dnsize(objset_t *os, dmu_object_type_t ot,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx)
+{
+ return (zap_create_norm_dnsize(os, 0, ot, bonustype, bonuslen,
+ dnodesize, tx));
+}
+
uint64_t
zap_create_norm(objset_t *os, int normflags, dmu_object_type_t ot,
dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx)
{
- uint64_t obj = dmu_object_alloc(os, ot, 0, bonustype, bonuslen, tx);
+ return (zap_create_norm_dnsize(os, normflags, ot, bonustype, bonuslen,
+ 0, tx));
+}
- mzap_create_impl(os, obj, normflags, 0, tx);
- return (obj);
+uint64_t
+zap_create_norm_dnsize(objset_t *os, int normflags, dmu_object_type_t ot,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx)
+{
+ return (zap_create_impl(os, normflags, 0, ot, 0, 0,
+ bonustype, bonuslen, dnodesize, NULL, NULL, tx));
}
uint64_t
dmu_object_type_t ot, int leaf_blockshift, int indirect_blockshift,
dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx)
{
- uint64_t obj = dmu_object_alloc(os, ot, 0, bonustype, bonuslen, tx);
-
- ASSERT(leaf_blockshift >= SPA_MINBLOCKSHIFT &&
- leaf_blockshift <= SPA_MAXBLOCKSHIFT &&
- indirect_blockshift >= SPA_MINBLOCKSHIFT &&
- indirect_blockshift <= SPA_MAXBLOCKSHIFT);
+ return (zap_create_flags_dnsize(os, normflags, flags, ot,
+ leaf_blockshift, indirect_blockshift, bonustype, bonuslen, 0, tx));
+}
- VERIFY(dmu_object_set_blocksize(os, obj,
- 1ULL << leaf_blockshift, indirect_blockshift, tx) == 0);
+uint64_t
+zap_create_flags_dnsize(objset_t *os, int normflags, zap_flags_t flags,
+ dmu_object_type_t ot, int leaf_blockshift, int indirect_blockshift,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx)
+{
+ return (zap_create_impl(os, normflags, flags, ot, leaf_blockshift,
+ indirect_blockshift, bonustype, bonuslen, dnodesize, NULL, NULL,
+ tx));
+}
- mzap_create_impl(os, obj, normflags, flags, tx);
- return (obj);
+/*
+ * Create a zap object and return a pointer to the newly allocated dnode via
+ * the allocated_dnode argument. The returned dnode will be held and the
+ * caller is responsible for releasing the hold by calling dnode_rele().
+ */
+uint64_t
+zap_create_hold(objset_t *os, int normflags, zap_flags_t flags,
+ dmu_object_type_t ot, int leaf_blockshift, int indirect_blockshift,
+ dmu_object_type_t bonustype, int bonuslen, int dnodesize,
+ dnode_t **allocated_dnode, void *tag, dmu_tx_t *tx)
+{
+ return (zap_create_impl(os, normflags, flags, ot, leaf_blockshift,
+ indirect_blockshift, bonustype, bonuslen, dnodesize,
+ allocated_dnode, tag, tx));
}
int
return (dmu_object_free(os, zapobj, tx));
}
-_NOTE(ARGSUSED(0))
void
-zap_evict(dmu_buf_t *db, void *vzap)
+zap_evict_sync(void *dbu)
{
- zap_t *zap = vzap;
+ zap_t *zap = dbu;
rw_destroy(&zap->zap_rwlock);
zap_count(objset_t *os, uint64_t zapobj, uint64_t *count)
{
zap_t *zap;
- int err;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
return (err);
if (!zap->zap_ismicro) {
err = fzap_count(zap, count);
} else {
*count = zap->zap_m.zap_num_entries;
}
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
static boolean_t
mzap_normalization_conflict(zap_t *zap, zap_name_t *zn, mzap_ent_t *mze)
{
- mzap_ent_t *other;
int direction = AVL_BEFORE;
boolean_t allocdzn = B_FALSE;
return (B_FALSE);
again:
- for (other = avl_walk(&zap->zap_m.zap_avl, mze, direction);
+ for (mzap_ent_t *other = avl_walk(&zap->zap_m.zap_avl, mze, direction);
other && other->mze_hash == mze->mze_hash;
other = avl_walk(&zap->zap_m.zap_avl, other, direction)) {
if (zn == NULL) {
zn = zap_name_alloc(zap, MZE_PHYS(zap, mze)->mze_name,
- MT_FIRST);
+ MT_NORMALIZE);
allocdzn = B_TRUE;
}
if (zap_match(zn, MZE_PHYS(zap, other)->mze_name)) {
uint64_t integer_size, uint64_t num_integers, void *buf)
{
return (zap_lookup_norm(os, zapobj, name, integer_size,
- num_integers, buf, MT_EXACT, NULL, 0, NULL));
+ num_integers, buf, 0, NULL, 0, NULL));
}
-int
-zap_lookup_norm(objset_t *os, uint64_t zapobj, const char *name,
+static int
+zap_lookup_impl(zap_t *zap, const char *name,
uint64_t integer_size, uint64_t num_integers, void *buf,
matchtype_t mt, char *realname, int rn_len,
boolean_t *ncp)
{
- zap_t *zap;
- int err;
- mzap_ent_t *mze;
- zap_name_t *zn;
+ int err = 0;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
- return (err);
- zn = zap_name_alloc(zap, name, mt);
- if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_name_t *zn = zap_name_alloc(zap, name, mt);
+ if (zn == NULL)
return (SET_ERROR(ENOTSUP));
- }
if (!zap->zap_ismicro) {
err = fzap_lookup(zn, integer_size, num_integers, buf,
realname, rn_len, ncp);
} else {
- mze = mze_find(zn);
+ mzap_ent_t *mze = mze_find(zn);
if (mze == NULL) {
err = SET_ERROR(ENOENT);
} else {
}
}
zap_name_free(zn);
- zap_unlockdir(zap);
return (err);
}
int
-zap_prefetch_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
- int key_numints)
+zap_lookup_norm(objset_t *os, uint64_t zapobj, const char *name,
+ uint64_t integer_size, uint64_t num_integers, void *buf,
+ matchtype_t mt, char *realname, int rn_len,
+ boolean_t *ncp)
+{
+ zap_t *zap;
+
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
+ return (err);
+ err = zap_lookup_impl(zap, name, integer_size,
+ num_integers, buf, mt, realname, rn_len, ncp);
+ zap_unlockdir(zap, FTAG);
+ return (err);
+}
+
+int
+zap_prefetch(objset_t *os, uint64_t zapobj, const char *name)
{
zap_t *zap;
int err;
zap_name_t *zn;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
+ err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
if (err)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ zn = zap_name_alloc(zap, name, 0);
+ if (zn == NULL) {
+ zap_unlockdir(zap, FTAG);
+ return (SET_ERROR(ENOTSUP));
+ }
+
+ fzap_prefetch(zn);
+ zap_name_free(zn);
+ zap_unlockdir(zap, FTAG);
+ return (err);
+}
+
+int
+zap_lookup_by_dnode(dnode_t *dn, const char *name,
+ uint64_t integer_size, uint64_t num_integers, void *buf)
+{
+ return (zap_lookup_norm_by_dnode(dn, name, integer_size,
+ num_integers, buf, 0, NULL, 0, NULL));
+}
+
+int
+zap_lookup_norm_by_dnode(dnode_t *dn, const char *name,
+ uint64_t integer_size, uint64_t num_integers, void *buf,
+ matchtype_t mt, char *realname, int rn_len,
+ boolean_t *ncp)
+{
+ zap_t *zap;
+
+ int err = zap_lockdir_by_dnode(dn, NULL, RW_READER, TRUE, FALSE,
+ FTAG, &zap);
+ if (err != 0)
+ return (err);
+ err = zap_lookup_impl(zap, name, integer_size,
+ num_integers, buf, mt, realname, rn_len, ncp);
+ zap_unlockdir(zap, FTAG);
+ return (err);
+}
+
+int
+zap_prefetch_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
+ int key_numints)
+{
+ zap_t *zap;
+
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
+ return (err);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
fzap_prefetch(zn);
zap_name_free(zn);
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
int key_numints, uint64_t integer_size, uint64_t num_integers, void *buf)
{
zap_t *zap;
- int err;
- zap_name_t *zn;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
err = fzap_lookup(zn, integer_size, num_integers, buf,
NULL, 0, NULL);
zap_name_free(zn);
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
zap_contains(objset_t *os, uint64_t zapobj, const char *name)
{
int err = zap_lookup_norm(os, zapobj, name, 0,
- 0, NULL, MT_EXACT, NULL, 0, NULL);
+ 0, NULL, 0, NULL, 0, NULL);
if (err == EOVERFLOW || err == EINVAL)
err = 0; /* found, but skipped reading the value */
return (err);
uint64_t *integer_size, uint64_t *num_integers)
{
zap_t *zap;
- int err;
- mzap_ent_t *mze;
- zap_name_t *zn;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc(zap, name, MT_EXACT);
+ zap_name_t *zn = zap_name_alloc(zap, name, 0);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
if (!zap->zap_ismicro) {
err = fzap_length(zn, integer_size, num_integers);
} else {
- mze = mze_find(zn);
+ mzap_ent_t *mze = mze_find(zn);
if (mze == NULL) {
err = SET_ERROR(ENOENT);
} else {
}
}
zap_name_free(zn);
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
int key_numints, uint64_t *integer_size, uint64_t *num_integers)
{
zap_t *zap;
- int err;
- zap_name_t *zn;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
err = fzap_length(zn, integer_size, num_integers);
zap_name_free(zn);
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
static void
mzap_addent(zap_name_t *zn, uint64_t value)
{
- int i;
zap_t *zap = zn->zn_zap;
int start = zap->zap_m.zap_alloc_next;
- uint32_t cd;
ASSERT(RW_WRITE_HELD(&zap->zap_rwlock));
#ifdef ZFS_DEBUG
- for (i = 0; i < zap->zap_m.zap_num_chunks; i++) {
- ASSERTV(mzap_ent_phys_t *mze);
- ASSERT(mze = &zap->zap_m.zap_phys->mz_chunk[i]);
+ for (int i = 0; i < zap->zap_m.zap_num_chunks; i++) {
+ mzap_ent_phys_t *mze = &zap_m_phys(zap)->mz_chunk[i];
ASSERT(strcmp(zn->zn_key_orig, mze->mze_name) != 0);
}
#endif
- cd = mze_find_unused_cd(zap, zn->zn_hash);
+ uint32_t cd = mze_find_unused_cd(zap, zn->zn_hash);
/* given the limited size of the microzap, this can't happen */
ASSERT(cd < zap_maxcd(zap));
again:
- for (i = start; i < zap->zap_m.zap_num_chunks; i++) {
- mzap_ent_phys_t *mze = &zap->zap_m.zap_phys->mz_chunk[i];
+ for (int i = start; i < zap->zap_m.zap_num_chunks; i++) {
+ mzap_ent_phys_t *mze = &zap_m_phys(zap)->mz_chunk[i];
if (mze->mze_name[0] == 0) {
mze->mze_value = value;
mze->mze_cd = cd;
- (void) strcpy(mze->mze_name, zn->zn_key_orig);
+ (void) strlcpy(mze->mze_name, zn->zn_key_orig,
+ sizeof (mze->mze_name));
zap->zap_m.zap_num_entries++;
zap->zap_m.zap_alloc_next = i+1;
if (zap->zap_m.zap_alloc_next ==
start = 0;
goto again;
}
- ASSERT(!"out of entries!");
+ cmn_err(CE_PANIC, "out of entries!");
}
-int
-zap_add(objset_t *os, uint64_t zapobj, const char *key,
+static int
+zap_add_impl(zap_t *zap, const char *key,
int integer_size, uint64_t num_integers,
- const void *val, dmu_tx_t *tx)
+ const void *val, dmu_tx_t *tx, void *tag)
{
- zap_t *zap;
- int err;
- mzap_ent_t *mze;
const uint64_t *intval = val;
- zap_name_t *zn;
+ int err = 0;
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, &zap);
- if (err)
- return (err);
- zn = zap_name_alloc(zap, key, MT_EXACT);
+ zap_name_t *zn = zap_name_alloc(zap, key, 0);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, tag);
return (SET_ERROR(ENOTSUP));
}
if (!zap->zap_ismicro) {
- err = fzap_add(zn, integer_size, num_integers, val, tx);
+ err = fzap_add(zn, integer_size, num_integers, val, tag, tx);
zap = zn->zn_zap; /* fzap_add() may change zap */
} else if (integer_size != 8 || num_integers != 1 ||
- strlen(key) >= MZAP_NAME_LEN) {
- err = mzap_upgrade(&zn->zn_zap, tx, 0);
- if (err == 0)
- err = fzap_add(zn, integer_size, num_integers, val, tx);
+ strlen(key) >= MZAP_NAME_LEN ||
+ !mze_canfit_fzap_leaf(zn, zn->zn_hash)) {
+ err = mzap_upgrade(&zn->zn_zap, tag, tx, 0);
+ if (err == 0) {
+ err = fzap_add(zn, integer_size, num_integers, val,
+ tag, tx);
+ }
zap = zn->zn_zap; /* fzap_add() may change zap */
} else {
- mze = mze_find(zn);
- if (mze != NULL) {
+ if (mze_find(zn) != NULL) {
err = SET_ERROR(EEXIST);
} else {
mzap_addent(zn, *intval);
ASSERT(zap == zn->zn_zap);
zap_name_free(zn);
if (zap != NULL) /* may be NULL if fzap_add() failed */
- zap_unlockdir(zap);
+ zap_unlockdir(zap, tag);
+ return (err);
+}
+
+int
+zap_add(objset_t *os, uint64_t zapobj, const char *key,
+ int integer_size, uint64_t num_integers,
+ const void *val, dmu_tx_t *tx)
+{
+ zap_t *zap;
+ int err;
+
+ err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, FTAG, &zap);
+ if (err != 0)
+ return (err);
+ err = zap_add_impl(zap, key, integer_size, num_integers, val, tx, FTAG);
+ /* zap_add_impl() calls zap_unlockdir() */
+ return (err);
+}
+
+int
+zap_add_by_dnode(dnode_t *dn, const char *key,
+ int integer_size, uint64_t num_integers,
+ const void *val, dmu_tx_t *tx)
+{
+ zap_t *zap;
+ int err;
+
+ err = zap_lockdir_by_dnode(dn, tx, RW_WRITER, TRUE, TRUE, FTAG, &zap);
+ if (err != 0)
+ return (err);
+ err = zap_add_impl(zap, key, integer_size, num_integers, val, tx, FTAG);
+ /* zap_add_impl() calls zap_unlockdir() */
return (err);
}
const void *val, dmu_tx_t *tx)
{
zap_t *zap;
- int err;
- zap_name_t *zn;
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
- err = fzap_add(zn, integer_size, num_integers, val, tx);
+ err = fzap_add(zn, integer_size, num_integers, val, FTAG, tx);
zap = zn->zn_zap; /* fzap_add() may change zap */
zap_name_free(zn);
if (zap != NULL) /* may be NULL if fzap_add() failed */
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
int integer_size, uint64_t num_integers, const void *val, dmu_tx_t *tx)
{
zap_t *zap;
- mzap_ent_t *mze;
const uint64_t *intval = val;
- zap_name_t *zn;
- int err;
-
-#ifdef ZFS_DEBUG
- uint64_t oldval;
- /*
- * If there is an old value, it shouldn't change across the
- * lockdir (eg, due to bprewrite's xlation).
- */
- if (integer_size == 8 && num_integers == 1)
- (void) zap_lookup(os, zapobj, name, 8, 1, &oldval);
-#endif
-
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc(zap, name, MT_EXACT);
+ zap_name_t *zn = zap_name_alloc(zap, name, 0);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
if (!zap->zap_ismicro) {
- err = fzap_update(zn, integer_size, num_integers, val, tx);
+ err = fzap_update(zn, integer_size, num_integers, val,
+ FTAG, tx);
zap = zn->zn_zap; /* fzap_update() may change zap */
} else if (integer_size != 8 || num_integers != 1 ||
strlen(name) >= MZAP_NAME_LEN) {
dprintf("upgrading obj %llu: intsz=%u numint=%llu name=%s\n",
zapobj, integer_size, num_integers, name);
- err = mzap_upgrade(&zn->zn_zap, tx, 0);
- if (err == 0)
+ err = mzap_upgrade(&zn->zn_zap, FTAG, tx, 0);
+ if (err == 0) {
err = fzap_update(zn, integer_size, num_integers,
- val, tx);
+ val, FTAG, tx);
+ }
zap = zn->zn_zap; /* fzap_update() may change zap */
} else {
- mze = mze_find(zn);
+ mzap_ent_t *mze = mze_find(zn);
if (mze != NULL) {
- ASSERT3U(MZE_PHYS(zap, mze)->mze_value, ==, oldval);
MZE_PHYS(zap, mze)->mze_value = *intval;
} else {
mzap_addent(zn, *intval);
ASSERT(zap == zn->zn_zap);
zap_name_free(zn);
if (zap != NULL) /* may be NULL if fzap_upgrade() failed */
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
int integer_size, uint64_t num_integers, const void *val, dmu_tx_t *tx)
{
zap_t *zap;
- zap_name_t *zn;
- int err;
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, TRUE, FTAG, &zap);
+ if (err != 0)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
- err = fzap_update(zn, integer_size, num_integers, val, tx);
+ err = fzap_update(zn, integer_size, num_integers, val, FTAG, tx);
zap = zn->zn_zap; /* fzap_update() may change zap */
zap_name_free(zn);
if (zap != NULL) /* may be NULL if fzap_upgrade() failed */
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
int
zap_remove(objset_t *os, uint64_t zapobj, const char *name, dmu_tx_t *tx)
{
- return (zap_remove_norm(os, zapobj, name, MT_EXACT, tx));
+ return (zap_remove_norm(os, zapobj, name, 0, tx));
}
-int
-zap_remove_norm(objset_t *os, uint64_t zapobj, const char *name,
+static int
+zap_remove_impl(zap_t *zap, const char *name,
matchtype_t mt, dmu_tx_t *tx)
{
- zap_t *zap;
- int err;
- mzap_ent_t *mze;
- zap_name_t *zn;
+ int err = 0;
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, FALSE, &zap);
- if (err)
- return (err);
- zn = zap_name_alloc(zap, name, mt);
- if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_name_t *zn = zap_name_alloc(zap, name, mt);
+ if (zn == NULL)
return (SET_ERROR(ENOTSUP));
- }
if (!zap->zap_ismicro) {
err = fzap_remove(zn, tx);
} else {
- mze = mze_find(zn);
+ mzap_ent_t *mze = mze_find(zn);
if (mze == NULL) {
err = SET_ERROR(ENOENT);
} else {
zap->zap_m.zap_num_entries--;
- bzero(&zap->zap_m.zap_phys->mz_chunk[mze->mze_chunkid],
+ bzero(&zap_m_phys(zap)->mz_chunk[mze->mze_chunkid],
sizeof (mzap_ent_phys_t));
mze_remove(zap, mze);
}
}
zap_name_free(zn);
- zap_unlockdir(zap);
return (err);
}
int
-zap_remove_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
- int key_numints, dmu_tx_t *tx)
+zap_remove_norm(objset_t *os, uint64_t zapobj, const char *name,
+ matchtype_t mt, dmu_tx_t *tx)
{
zap_t *zap;
int err;
- zap_name_t *zn;
- err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, FALSE, &zap);
+ err = zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, FALSE, FTAG, &zap);
if (err)
return (err);
- zn = zap_name_alloc_uint64(zap, key, key_numints);
+ err = zap_remove_impl(zap, name, mt, tx);
+ zap_unlockdir(zap, FTAG);
+ return (err);
+}
+
+int
+zap_remove_by_dnode(dnode_t *dn, const char *name, dmu_tx_t *tx)
+{
+ zap_t *zap;
+ int err;
+
+ err = zap_lockdir_by_dnode(dn, tx, RW_WRITER, TRUE, FALSE, FTAG, &zap);
+ if (err)
+ return (err);
+ err = zap_remove_impl(zap, name, 0, tx);
+ zap_unlockdir(zap, FTAG);
+ return (err);
+}
+
+int
+zap_remove_uint64(objset_t *os, uint64_t zapobj, const uint64_t *key,
+ int key_numints, dmu_tx_t *tx)
+{
+ zap_t *zap;
+
+ int err =
+ zap_lockdir(os, zapobj, tx, RW_WRITER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
+ return (err);
+ zap_name_t *zn = zap_name_alloc_uint64(zap, key, key_numints);
if (zn == NULL) {
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (SET_ERROR(ENOTSUP));
}
err = fzap_remove(zn, tx);
zap_name_free(zn);
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (err);
}
{
if (zc->zc_zap) {
rw_enter(&zc->zc_zap->zap_rwlock, RW_READER);
- zap_unlockdir(zc->zc_zap);
+ zap_unlockdir(zc->zc_zap, NULL);
zc->zc_zap = NULL;
}
if (zc->zc_leaf) {
zap_cursor_retrieve(zap_cursor_t *zc, zap_attribute_t *za)
{
int err;
- avl_index_t idx;
- mzap_ent_t mze_tofind;
- mzap_ent_t *mze;
if (zc->zc_hash == -1ULL)
return (SET_ERROR(ENOENT));
if (zc->zc_zap == NULL) {
int hb;
err = zap_lockdir(zc->zc_objset, zc->zc_zapobj, NULL,
- RW_READER, TRUE, FALSE, &zc->zc_zap);
- if (err)
+ RW_READER, TRUE, FALSE, NULL, &zc->zc_zap);
+ if (err != 0)
return (err);
/*
if (!zc->zc_zap->zap_ismicro) {
err = fzap_cursor_retrieve(zc->zc_zap, zc, za);
} else {
+ avl_index_t idx;
+ mzap_ent_t mze_tofind;
+
mze_tofind.mze_hash = zc->zc_hash;
mze_tofind.mze_cd = zc->zc_cd;
- mze = avl_find(&zc->zc_zap->zap_m.zap_avl, &mze_tofind, &idx);
+ mzap_ent_t *mze =
+ avl_find(&zc->zc_zap->zap_m.zap_avl, &mze_tofind, &idx);
if (mze == NULL) {
mze = avl_nearest(&zc->zc_zap->zap_m.zap_avl,
idx, AVL_AFTER);
zc->zc_cd++;
}
-int
-zap_cursor_move_to_key(zap_cursor_t *zc, const char *name, matchtype_t mt)
-{
- int err = 0;
- mzap_ent_t *mze;
- zap_name_t *zn;
-
- if (zc->zc_zap == NULL) {
- err = zap_lockdir(zc->zc_objset, zc->zc_zapobj, NULL,
- RW_READER, TRUE, FALSE, &zc->zc_zap);
- if (err)
- return (err);
- } else {
- rw_enter(&zc->zc_zap->zap_rwlock, RW_READER);
- }
-
- zn = zap_name_alloc(zc->zc_zap, name, mt);
- if (zn == NULL) {
- rw_exit(&zc->zc_zap->zap_rwlock);
- return (SET_ERROR(ENOTSUP));
- }
-
- if (!zc->zc_zap->zap_ismicro) {
- err = fzap_cursor_move_to_key(zc, zn);
- } else {
- mze = mze_find(zn);
- if (mze == NULL) {
- err = SET_ERROR(ENOENT);
- goto out;
- }
- zc->zc_hash = mze->mze_hash;
- zc->zc_cd = mze->mze_cd;
- }
-
-out:
- zap_name_free(zn);
- rw_exit(&zc->zc_zap->zap_rwlock);
- return (err);
-}
-
int
zap_get_stats(objset_t *os, uint64_t zapobj, zap_stats_t *zs)
{
- int err;
zap_t *zap;
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
+ int err =
+ zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, FTAG, &zap);
+ if (err != 0)
return (err);
bzero(zs, sizeof (zap_stats_t));
} else {
fzap_get_stats(zap, zs);
}
- zap_unlockdir(zap);
+ zap_unlockdir(zap, FTAG);
return (0);
}
-int
-zap_count_write(objset_t *os, uint64_t zapobj, const char *name, int add,
- uint64_t *towrite, uint64_t *tooverwrite)
-{
- zap_t *zap;
- int err = 0;
-
-
- /*
- * Since, we don't have a name, we cannot figure out which blocks will
- * be affected in this operation. So, account for the worst case :
- * - 3 blocks overwritten: target leaf, ptrtbl block, header block
- * - 4 new blocks written if adding:
- * - 2 blocks for possibly split leaves,
- * - 2 grown ptrtbl blocks
- *
- * This also accomodates the case where an add operation to a fairly
- * large microzap results in a promotion to fatzap.
- */
- if (name == NULL) {
- *towrite += (3 + (add ? 4 : 0)) * SPA_MAXBLOCKSIZE;
- return (err);
- }
-
- /*
- * We lock the zap with adding == FALSE. Because, if we pass
- * the actual value of add, it could trigger a mzap_upgrade().
- * At present we are just evaluating the possibility of this operation
- * and hence we donot want to trigger an upgrade.
- */
- err = zap_lockdir(os, zapobj, NULL, RW_READER, TRUE, FALSE, &zap);
- if (err)
- return (err);
-
- if (!zap->zap_ismicro) {
- zap_name_t *zn = zap_name_alloc(zap, name, MT_EXACT);
- if (zn) {
- err = fzap_count_write(zn, add, towrite,
- tooverwrite);
- zap_name_free(zn);
- } else {
- /*
- * We treat this case as similar to (name == NULL)
- */
- *towrite += (3 + (add ? 4 : 0)) * SPA_MAXBLOCKSIZE;
- }
- } else {
- /*
- * We are here if (name != NULL) and this is a micro-zap.
- * We account for the header block depending on whether it
- * is freeable.
- *
- * Incase of an add-operation it is hard to find out
- * if this add will promote this microzap to fatzap.
- * Hence, we consider the worst case and account for the
- * blocks assuming this microzap would be promoted to a
- * fatzap.
- *
- * 1 block overwritten : header block
- * 4 new blocks written : 2 new split leaf, 2 grown
- * ptrtbl blocks
- */
- if (dmu_buf_freeable(zap->zap_dbuf))
- *tooverwrite += SPA_MAXBLOCKSIZE;
- else
- *towrite += SPA_MAXBLOCKSIZE;
-
- if (add) {
- *towrite += 4 * SPA_MAXBLOCKSIZE;
- }
- }
-
- zap_unlockdir(zap);
- return (err);
-}
-
-#if defined(_KERNEL) && defined(HAVE_SPL)
+#if defined(_KERNEL)
EXPORT_SYMBOL(zap_create);
+EXPORT_SYMBOL(zap_create_dnsize);
EXPORT_SYMBOL(zap_create_norm);
+EXPORT_SYMBOL(zap_create_norm_dnsize);
EXPORT_SYMBOL(zap_create_flags);
+EXPORT_SYMBOL(zap_create_flags_dnsize);
EXPORT_SYMBOL(zap_create_claim);
EXPORT_SYMBOL(zap_create_claim_norm);
+EXPORT_SYMBOL(zap_create_claim_norm_dnsize);
+EXPORT_SYMBOL(zap_create_hold);
EXPORT_SYMBOL(zap_destroy);
EXPORT_SYMBOL(zap_lookup);
+EXPORT_SYMBOL(zap_lookup_by_dnode);
EXPORT_SYMBOL(zap_lookup_norm);
EXPORT_SYMBOL(zap_lookup_uint64);
EXPORT_SYMBOL(zap_contains);
+EXPORT_SYMBOL(zap_prefetch);
EXPORT_SYMBOL(zap_prefetch_uint64);
-EXPORT_SYMBOL(zap_count_write);
EXPORT_SYMBOL(zap_add);
+EXPORT_SYMBOL(zap_add_by_dnode);
EXPORT_SYMBOL(zap_add_uint64);
EXPORT_SYMBOL(zap_update);
EXPORT_SYMBOL(zap_update_uint64);
EXPORT_SYMBOL(zap_length);
EXPORT_SYMBOL(zap_length_uint64);
EXPORT_SYMBOL(zap_remove);
+EXPORT_SYMBOL(zap_remove_by_dnode);
EXPORT_SYMBOL(zap_remove_norm);
EXPORT_SYMBOL(zap_remove_uint64);
EXPORT_SYMBOL(zap_count);
EXPORT_SYMBOL(zap_cursor_retrieve);
EXPORT_SYMBOL(zap_cursor_advance);
EXPORT_SYMBOL(zap_cursor_serialize);
-EXPORT_SYMBOL(zap_cursor_move_to_key);
EXPORT_SYMBOL(zap_cursor_init_serialized);
EXPORT_SYMBOL(zap_get_stats);
#endif
projects / mirror_zfs.git / blobdiff