int
dmu_buf_hold_noread(objset_t *os, uint64_t object, uint64_t offset,
void *tag, dmu_buf_t **dbp)
{
dnode_t *dn;
uint64_t blkid;
dmu_buf_impl_t *db;
int err;
err = dnode_hold(os, object, FTAG, &dn);
if (err)
return (err);
blkid = dbuf_whichblock(dn, offset);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold(dn, blkid, tag);
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
if (db == NULL) {
*dbp = NULL;
return (SET_ERROR(EIO));
}
*dbp = &db->db;
return (err);
}
int
dmu_buf_hold(objset_t *os, uint64_t object, uint64_t offset,
void *tag, dmu_buf_t **dbp, int flags)
{
dnode_t *dn;
uint64_t blkid;
dmu_buf_impl_t *db;
int err;
int db_flags = DB_RF_CANFAIL;
if (flags & DMU_READ_NO_PREFETCH)
db_flags |= DB_RF_NOPREFETCH;
err = dnode_hold(os, object, FTAG, &dn);
if (err)
return (err);
blkid = dbuf_whichblock(dn, offset);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold(dn, blkid, tag);
rw_exit(&dn->dn_struct_rwlock);
if (db == NULL) {
err = EIO;
} else {
err = dbuf_read(db, NULL, db_flags);
if (err) {
dbuf_rele(db, tag);
db = NULL;
}
}
dnode_rele(dn, FTAG);
*dbp = &db->db; /* NULL db plus first field offset is NULL */
return (err);
}
int
dmu_buf_hold(objset_t *os, uint64_t object, uint64_t offset,
void *tag, dmu_buf_t **dbp)
{
dnode_t *dn;
uint64_t blkid;
dmu_buf_impl_t *db;
int err;
err = dnode_hold(os->os, object, FTAG, &dn);
if (err)
return (err);
blkid = dbuf_whichblock(dn, offset);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold(dn, blkid, tag);
rw_exit(&dn->dn_struct_rwlock);
if (db == NULL) {
err = EIO;
} else {
err = dbuf_read(db, NULL, DB_RF_CANFAIL);
if (err) {
dbuf_rele(db, tag);
db = NULL;
}
}
dnode_rele(dn, FTAG);
*dbp = &db->db;
return (err);
}
/*
* returns ENOENT, EIO, or 0.
*/
int
dmu_bonus_hold(objset_t *os, uint64_t object, void *tag, dmu_buf_t **dbp)
{
dnode_t *dn;
dmu_buf_impl_t *db;
int error;
error = dnode_hold(os->os, object, FTAG, &dn);
if (error)
return (error);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
if (dn->dn_bonus == NULL) {
rw_exit(&dn->dn_struct_rwlock);
rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
if (dn->dn_bonus == NULL)
dbuf_create_bonus(dn);
}
db = dn->dn_bonus;
rw_exit(&dn->dn_struct_rwlock);
/* as long as the bonus buf is held, the dnode will be held */
if (refcount_add(&db->db_holds, tag) == 1)
VERIFY(dnode_add_ref(dn, db));
dnode_rele(dn, FTAG);
VERIFY(0 == dbuf_read(db, NULL, DB_RF_MUST_SUCCEED));
*dbp = &db->db;
return (0);
}
int
dmu_rm_spill(objset_t *os, uint64_t object, dmu_tx_t *tx)
{
dnode_t *dn;
int error;
error = dnode_hold(os, object, FTAG, &dn);
dbuf_rm_spill(dn, tx);
rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
dnode_rm_spill(dn, tx);
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
return (error);
}
static dmu_tx_hold_t *
dmu_tx_hold_object_impl(dmu_tx_t *tx, objset_t *os, uint64_t object,
enum dmu_tx_hold_type type, uint64_t arg1, uint64_t arg2)
{
dnode_t *dn = NULL;
dmu_tx_hold_t *txh;
int err;
if (object != DMU_NEW_OBJECT) {
err = dnode_hold(os, object, FTAG, &dn);
if (err != 0) {
tx->tx_err = err;
return (NULL);
}
}
txh = dmu_tx_hold_dnode_impl(tx, dn, type, arg1, arg2);
if (dn != NULL)
dnode_rele(dn, FTAG);
return (txh);
}
/*
* returns ENOENT, EIO, or 0.
*/
int
dmu_bonus_hold(objset_t *os, uint64_t object, void *tag, dmu_buf_t **dbp)
{
dnode_t *dn;
dmu_buf_impl_t *db;
int error;
error = dnode_hold(os, object, FTAG, &dn);
if (error)
return (error);
rw_enter(&dn->dn_struct_rwlock, RW_READER);
if (dn->dn_bonus == NULL) {
rw_exit(&dn->dn_struct_rwlock);
rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
if (dn->dn_bonus == NULL)
dbuf_create_bonus(dn);
}
db = dn->dn_bonus;
/* as long as the bonus buf is held, the dnode will be held */
if (refcount_add(&db->db_holds, tag) == 1) {
VERIFY(dnode_add_ref(dn, db));
atomic_inc_32(&dn->dn_dbufs_count);
}
/*
* Wait to drop dn_struct_rwlock until after adding the bonus dbuf's
* hold and incrementing the dbuf count to ensure that dnode_move() sees
* a dnode hold for every dbuf.
*/
rw_exit(&dn->dn_struct_rwlock);
dnode_rele(dn, FTAG);
VERIFY(0 == dbuf_read(db, NULL, DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH));
*dbp = &db->db;
return (0);
}
int
dmu_objset_evict_dbufs(objset_t *os)
{
dnode_t *dn;
mutex_enter(&os->os_lock);
/* process the mdn last, since the other dnodes have holds on it */
list_remove(&os->os_dnodes, DMU_META_DNODE(os));
list_insert_tail(&os->os_dnodes, DMU_META_DNODE(os));
/*
* Find the first dnode with holds. We have to do this dance
* because dnode_add_ref() only works if you already have a
* hold. If there are no holds then it has no dbufs so OK to
* skip.
*/
for (dn = list_head(&os->os_dnodes);
dn && !dnode_add_ref(dn, FTAG);
dn = list_next(&os->os_dnodes, dn))
continue;
while (dn) {
dnode_t *next_dn = dn;
do {
next_dn = list_next(&os->os_dnodes, next_dn);
} while (next_dn && !dnode_add_ref(next_dn, FTAG));
mutex_exit(&os->os_lock);
dnode_evict_dbufs(dn);
dnode_rele(dn, FTAG);
mutex_enter(&os->os_lock);
dn = next_dn;
}
dn = list_head(&os->os_dnodes);
mutex_exit(&os->os_lock);
return (dn != DMU_META_DNODE(os));
}
void
dmu_objset_evict_dbufs(objset_t *os)
{
dnode_t dn_marker;
dnode_t *dn;
mutex_enter(&os->os_lock);
dn = list_head(&os->os_dnodes);
while (dn != NULL) {
/*
* Skip dnodes without holds. We have to do this dance
* because dnode_add_ref() only works if there is already a
* hold. If the dnode has no holds, then it has no dbufs.
*/
if (dnode_add_ref(dn, FTAG)) {
list_insert_after(&os->os_dnodes, dn, &dn_marker);
mutex_exit(&os->os_lock);
dnode_evict_dbufs(dn);
dnode_rele(dn, FTAG);
mutex_enter(&os->os_lock);
dn = list_next(&os->os_dnodes, &dn_marker);
list_remove(&os->os_dnodes, &dn_marker);
} else {
dn = list_next(&os->os_dnodes, dn);
}
}
mutex_exit(&os->os_lock);
if (DMU_USERUSED_DNODE(os) != NULL) {
dnode_evict_dbufs(DMU_GROUPUSED_DNODE(os));
dnode_evict_dbufs(DMU_USERUSED_DNODE(os));
}
dnode_evict_dbufs(DMU_META_DNODE(os));
}
void
dmu_objset_do_userquota_callbacks(objset_impl_t *os, dmu_tx_t *tx)
{
dnode_t *dn;
list_t *list = &os->os_synced_dnodes;
ASSERTV(static const char zerobuf[DN_MAX_BONUSLEN] = {0});
ASSERT(list_head(list) == NULL || dmu_objset_userused_enabled(os));
while ((dn = list_head(list))) {
dmu_object_type_t bonustype;
ASSERT(!DMU_OBJECT_IS_SPECIAL(dn->dn_object));
ASSERT(dn->dn_oldphys);
ASSERT(dn->dn_phys->dn_type == DMU_OT_NONE ||
dn->dn_phys->dn_flags &
DNODE_FLAG_USERUSED_ACCOUNTED);
/* Allocate the user/groupused objects if necessary. */
if (os->os_userused_dnode->dn_type == DMU_OT_NONE) {
VERIFY(0 == zap_create_claim(&os->os,
DMU_USERUSED_OBJECT,
DMU_OT_USERGROUP_USED, DMU_OT_NONE, 0, tx));
VERIFY(0 == zap_create_claim(&os->os,
DMU_GROUPUSED_OBJECT,
DMU_OT_USERGROUP_USED, DMU_OT_NONE, 0, tx));
}
/*
* If the object was not previously
* accounted, pretend that it was free.
*/
if (!(dn->dn_oldphys->dn_flags &
DNODE_FLAG_USERUSED_ACCOUNTED)) {
bzero(dn->dn_oldphys, sizeof (dnode_phys_t));
}
/*
* If the object was freed, use the previous bonustype.
*/
bonustype = dn->dn_phys->dn_bonustype ?
dn->dn_phys->dn_bonustype : dn->dn_oldphys->dn_bonustype;
ASSERT(dn->dn_phys->dn_type != 0 ||
(bcmp(DN_BONUS(dn->dn_phys), zerobuf,
DN_MAX_BONUSLEN) == 0 &&
DN_USED_BYTES(dn->dn_phys) == 0));
ASSERT(dn->dn_oldphys->dn_type != 0 ||
(bcmp(DN_BONUS(dn->dn_oldphys), zerobuf,
DN_MAX_BONUSLEN) == 0 &&
DN_USED_BYTES(dn->dn_oldphys) == 0));
used_cbs[os->os_phys->os_type](&os->os, bonustype,
DN_BONUS(dn->dn_oldphys), DN_BONUS(dn->dn_phys),
DN_USED_BYTES(dn->dn_oldphys),
DN_USED_BYTES(dn->dn_phys), tx);
/*
* The mutex is needed here for interlock with dnode_allocate.
*/
mutex_enter(&dn->dn_mtx);
zio_buf_free(dn->dn_oldphys, sizeof (dnode_phys_t));
dn->dn_oldphys = NULL;
mutex_exit(&dn->dn_mtx);
list_remove(list, dn);
dnode_rele(dn, list);
}
}
