static int
dmu_tx_check_ioerr(zio_t *zio, dnode_t *dn, int level, uint64_t blkid)
{
int err;
dmu_buf_impl_t *db;
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold_level(dn, level, blkid, FTAG);
rw_exit(&dn->dn_struct_rwlock);
if (db == NULL)
return (SET_ERROR(EIO));
err = dbuf_read(db, zio, DB_RF_CANFAIL | DB_RF_NOPREFETCH);
dbuf_rele(db, FTAG);
return (err);
}
static void
dnode_increase_indirection(dnode_t *dn, dmu_tx_t *tx)
{
dmu_buf_impl_t *db;
int txgoff = tx->tx_txg & TXG_MASK;
int nblkptr = dn->dn_phys->dn_nblkptr;
int old_toplvl = dn->dn_phys->dn_nlevels - 1;
int new_level = dn->dn_next_nlevels[txgoff];
int i;
rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
/* this dnode can't be paged out because it's dirty */
ASSERT(dn->dn_phys->dn_type != DMU_OT_NONE);
ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock));
ASSERT(new_level > 1 && dn->dn_phys->dn_nlevels > 0);
db = dbuf_hold_level(dn, dn->dn_phys->dn_nlevels, 0, FTAG);
ASSERT(db != NULL);
dn->dn_phys->dn_nlevels = new_level;
dprintf("os=%p obj=%llu, increase to %d\n", dn->dn_objset,
dn->dn_object, dn->dn_phys->dn_nlevels);
/* check for existing blkptrs in the dnode */
for (i = 0; i < nblkptr; i++)
if (!BP_IS_HOLE(&dn->dn_phys->dn_blkptr[i]))
break;
if (i != nblkptr) {
/* transfer dnode's block pointers to new indirect block */
(void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT);
ASSERT(db->db.db_data);
ASSERT(arc_released(db->db_buf));
ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size);
bcopy(dn->dn_phys->dn_blkptr, db->db.db_data,
sizeof (blkptr_t) * nblkptr);
arc_buf_freeze(db->db_buf);
}
/* ARGSUSED */
static void
dmu_tx_count_write(dmu_tx_hold_t *txh, uint64_t off, uint64_t len)
{
dnode_t *dn = txh->txh_dnode;
uint64_t start, end, i;
int min_bs, max_bs, min_ibs, max_ibs, epbs, bits;
int err = 0;
if (len == 0)
return;
min_bs = SPA_MINBLOCKSHIFT;
max_bs = SPA_MAXBLOCKSHIFT;
min_ibs = DN_MIN_INDBLKSHIFT;
max_ibs = DN_MAX_INDBLKSHIFT;
if (dn) {
uint64_t history[DN_MAX_LEVELS];
int nlvls = dn->dn_nlevels;
int delta;
/*
* For i/o error checking, read the first and last level-0
* blocks (if they are not aligned), and all the level-1 blocks.
*/
if (dn->dn_maxblkid == 0) {
delta = dn->dn_datablksz;
start = (off < dn->dn_datablksz) ? 0 : 1;
end = (off+len <= dn->dn_datablksz) ? 0 : 1;
if (start == 0 && (off > 0 || len < dn->dn_datablksz)) {
err = dmu_tx_check_ioerr(NULL, dn, 0, 0);
if (err)
goto out;
delta -= off;
}
} else {
zio_t *zio = zio_root(dn->dn_objset->os_spa,
NULL, NULL, ZIO_FLAG_CANFAIL);
/* first level-0 block */
start = off >> dn->dn_datablkshift;
if (P2PHASE(off, dn->dn_datablksz) ||
len < dn->dn_datablksz) {
err = dmu_tx_check_ioerr(zio, dn, 0, start);
if (err)
goto out;
}
/* last level-0 block */
end = (off+len-1) >> dn->dn_datablkshift;
if (end != start && end <= dn->dn_maxblkid &&
P2PHASE(off+len, dn->dn_datablksz)) {
err = dmu_tx_check_ioerr(zio, dn, 0, end);
if (err)
goto out;
}
/* level-1 blocks */
if (nlvls > 1) {
int shft = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
for (i = (start>>shft)+1; i < end>>shft; i++) {
err = dmu_tx_check_ioerr(zio, dn, 1, i);
if (err)
goto out;
}
}
err = zio_wait(zio);
if (err)
goto out;
delta = P2NPHASE(off, dn->dn_datablksz);
}
if (dn->dn_maxblkid > 0) {
/*
* The blocksize can't change,
* so we can make a more precise estimate.
*/
ASSERT(dn->dn_datablkshift != 0);
min_bs = max_bs = dn->dn_datablkshift;
min_ibs = max_ibs = dn->dn_indblkshift;
} else if (dn->dn_indblkshift > max_ibs) {
/*
* This ensures that if we reduce DN_MAX_INDBLKSHIFT,
* the code will still work correctly on older pools.
*/
min_ibs = max_ibs = dn->dn_indblkshift;
}
/*
* If this write is not off the end of the file
* we need to account for overwrites/unref.
*/
if (start <= dn->dn_maxblkid) {
for (int l = 0; l < DN_MAX_LEVELS; l++)
history[l] = -1ULL;
}
while (start <= dn->dn_maxblkid) {
dmu_buf_impl_t *db;
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold_level(dn, 0, start, FTAG);
rw_exit(&dn->dn_struct_rwlock);
dmu_tx_count_twig(txh, dn, db, 0, start, B_FALSE,
history);
dbuf_rele(db, FTAG);
if (++start > end) {
/*
* Account for new indirects appearing
* before this IO gets assigned into a txg.
*/
bits = 64 - min_bs;
epbs = min_ibs - SPA_BLKPTRSHIFT;
for (bits -= epbs * (nlvls - 1);
bits >= 0; bits -= epbs)
txh->txh_fudge += 1ULL << max_ibs;
goto out;
}
off += delta;
if (len >= delta)
len -= delta;
delta = dn->dn_datablksz;
}
}
static void
dnode_increase_indirection(dnode_t *dn, dmu_tx_t *tx)
{
dmu_buf_impl_t *db;
int txgoff = tx->tx_txg & TXG_MASK;
int nblkptr = dn->dn_phys->dn_nblkptr;
int old_toplvl = dn->dn_phys->dn_nlevels - 1;
int new_level = dn->dn_next_nlevels[txgoff];
int i;
rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
/* this dnode can't be paged out because it's dirty */
ASSERT(dn->dn_phys->dn_type != DMU_OT_NONE);
ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock));
ASSERT(new_level > 1 && dn->dn_phys->dn_nlevels > 0);
db = dbuf_hold_level(dn, dn->dn_phys->dn_nlevels, 0, FTAG);
ASSERT(db != NULL);
dn->dn_phys->dn_nlevels = new_level;
dprintf("os=%p obj=%llu, increase to %d\n", dn->dn_objset,
dn->dn_object, dn->dn_phys->dn_nlevels);
/* transfer dnode's block pointers to new indirect block */
(void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT);
ASSERT(db->db.db_data);
ASSERT(arc_released(db->db_buf));
ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size);
bcopy(dn->dn_phys->dn_blkptr, db->db.db_data,
sizeof (blkptr_t) * nblkptr);
arc_buf_freeze(db->db_buf);
/* set dbuf's parent pointers to new indirect buf */
for (i = 0; i < nblkptr; i++) {
dmu_buf_impl_t *child =
dbuf_find(dn->dn_objset, dn->dn_object, old_toplvl, i);
if (child == NULL)
continue;
#ifdef DEBUG
DB_DNODE_ENTER(child);
ASSERT3P(DB_DNODE(child), ==, dn);
DB_DNODE_EXIT(child);
#endif /* DEBUG */
if (child->db_parent && child->db_parent != dn->dn_dbuf) {
ASSERT(child->db_parent->db_level == db->db_level);
ASSERT(child->db_blkptr !=
&dn->dn_phys->dn_blkptr[child->db_blkid]);
mutex_exit(&child->db_mtx);
continue;
}
ASSERT(child->db_parent == NULL ||
child->db_parent == dn->dn_dbuf);
child->db_parent = db;
dbuf_add_ref(db, child);
if (db->db.db_data)
child->db_blkptr = (blkptr_t *)db->db.db_data + i;
else
child->db_blkptr = NULL;
dprintf_dbuf_bp(child, child->db_blkptr,
"changed db_blkptr to new indirect %s", "");
mutex_exit(&child->db_mtx);
}
bzero(dn->dn_phys->dn_blkptr, sizeof (blkptr_t) * nblkptr);
dbuf_rele(db, FTAG);
rw_exit(&dn->dn_struct_rwlock);
}
