/* ARGSUSED */
static void
ready(zio_t *zio, arc_buf_t *abuf, void *arg)
{
objset_impl_t *os = arg;
blkptr_t *bp = os->os_rootbp;
dnode_phys_t *dnp = &os->os_phys->os_meta_dnode;
int i;
ASSERT(bp == zio->io_bp);
/*
* Update rootbp fill count.
*/
bp->blk_fill = 1; /* count the meta-dnode */
for (i = 0; i < dnp->dn_nblkptr; i++)
bp->blk_fill += dnp->dn_blkptr[i].blk_fill;
BP_SET_TYPE(bp, DMU_OT_OBJSET);
BP_SET_LEVEL(bp, 0);
/* We must do this after we've set the bp's type and level */
if (!DVA_EQUAL(BP_IDENTITY(bp),
BP_IDENTITY(&zio->io_bp_orig))) {
if (zio->io_bp_orig.blk_birth == os->os_synctx->tx_txg)
(void) dsl_dataset_block_kill(os->os_dsl_dataset,
&zio->io_bp_orig, NULL, os->os_synctx);
dsl_dataset_block_born(os->os_dsl_dataset, bp, os->os_synctx);
}
}
/* ARGSUSED */
static void
ready(zio_t *zio, arc_buf_t *abuf, void *arg)
{
blkptr_t *bp = zio->io_bp;
blkptr_t *bp_orig = &zio->io_bp_orig;
objset_impl_t *os = arg;
dnode_phys_t *dnp = &os->os_phys->os_meta_dnode;
ASSERT(bp == os->os_rootbp);
ASSERT(BP_GET_TYPE(bp) == DMU_OT_OBJSET);
ASSERT(BP_GET_LEVEL(bp) == 0);
/*
* Update rootbp fill count.
*/
bp->blk_fill = 1; /* count the meta-dnode */
for (int i = 0; i < dnp->dn_nblkptr; i++)
bp->blk_fill += dnp->dn_blkptr[i].blk_fill;
if (zio->io_flags & ZIO_FLAG_IO_REWRITE) {
ASSERT(DVA_EQUAL(BP_IDENTITY(bp), BP_IDENTITY(bp_orig)));
} else {
if (zio->io_bp_orig.blk_birth == os->os_synctx->tx_txg)
(void) dsl_dataset_block_kill(os->os_dsl_dataset,
&zio->io_bp_orig, zio, os->os_synctx);
dsl_dataset_block_born(os->os_dsl_dataset, bp, os->os_synctx);
}
}
/* ARGSUSED */
static void
ready(zio_t *zio, arc_buf_t *abuf, void *arg)
{
int i;
blkptr_t *bp = zio->io_bp;
objset_impl_t *os = arg;
dnode_phys_t *dnp = &os->os_phys->os_meta_dnode;
ASSERTV(blkptr_t *bp_orig = &zio->io_bp_orig);
ASSERT(bp == os->os_rootbp);
ASSERT(BP_GET_TYPE(bp) == DMU_OT_OBJSET);
ASSERT(BP_GET_LEVEL(bp) == 0);
/*
* Update rootbp fill count: it should be the number of objects
* allocated in the object set (not counting the "special"
* objects that are stored in the objset_phys_t -- the meta
* dnode and user/group accounting objects).
*/
bp->blk_fill = 0;
for (i = 0; i < dnp->dn_nblkptr; i++)
bp->blk_fill += dnp->dn_blkptr[i].blk_fill;
if (zio->io_flags & ZIO_FLAG_IO_REWRITE) {
ASSERT(DVA_EQUAL(BP_IDENTITY(bp), BP_IDENTITY(bp_orig)));
} else {
if (zio->io_bp_orig.blk_birth == os->os_synctx->tx_txg)
(void) dsl_dataset_block_kill(os->os_dsl_dataset,
&zio->io_bp_orig, zio, os->os_synctx);
dsl_dataset_block_born(os->os_dsl_dataset, bp, os->os_synctx);
}
}
/* ARGSUSED */
static void
dmu_objset_write_done(zio_t *zio, arc_buf_t *abuf, void *arg)
{
blkptr_t *bp = zio->io_bp;
blkptr_t *bp_orig = &zio->io_bp_orig;
objset_t *os = arg;
if (zio->io_flags & ZIO_FLAG_IO_REWRITE) {
ASSERT(BP_EQUAL(bp, bp_orig));
} else {
dsl_dataset_t *ds = os->os_dsl_dataset;
dmu_tx_t *tx = os->os_synctx;
(void) dsl_dataset_block_kill(ds, bp_orig, tx, B_TRUE);
dsl_dataset_block_born(ds, bp, tx);
}
}
static void
free_blocks(dnode_t *dn, blkptr_t *bp, int num, dmu_tx_t *tx)
{
dsl_dataset_t *ds = dn->dn_objset->os_dsl_dataset;
uint64_t bytesfreed = 0;
int i;
dprintf("ds=%p obj=%llx num=%d\n", ds, dn->dn_object, num);
for (i = 0; i < num; i++, bp++) {
uint64_t lsize, lvl;
dmu_object_type_t type;
if (BP_IS_HOLE(bp))
continue;
bytesfreed += dsl_dataset_block_kill(ds, bp, tx, B_FALSE);
ASSERT3U(bytesfreed, <=, DN_USED_BYTES(dn->dn_phys));
/*
* Save some useful information on the holes being
* punched, including logical size, type, and indirection
* level. Retaining birth time enables detection of when
* holes are punched for reducing the number of free
* records transmitted during a zfs send.
*/
lsize = BP_GET_LSIZE(bp);
type = BP_GET_TYPE(bp);
lvl = BP_GET_LEVEL(bp);
bzero(bp, sizeof (blkptr_t));
if (spa_feature_is_active(dn->dn_objset->os_spa,
SPA_FEATURE_HOLE_BIRTH)) {
BP_SET_LSIZE(bp, lsize);
BP_SET_TYPE(bp, type);
BP_SET_LEVEL(bp, lvl);
BP_SET_BIRTH(bp, dmu_tx_get_txg(tx), 0);
}
}
dnode_diduse_space(dn, -bytesfreed);
}
/* called from dsl */
void
dmu_objset_sync(objset_impl_t *os, zio_t *pio, dmu_tx_t *tx)
{
int txgoff;
zbookmark_t zb;
writeprops_t wp = { 0 };
zio_t *zio;
list_t *list;
dbuf_dirty_record_t *dr;
dprintf_ds(os->os_dsl_dataset, "txg=%llu\n", tx->tx_txg);
ASSERT(dmu_tx_is_syncing(tx));
/* XXX the write_done callback should really give us the tx... */
os->os_synctx = tx;
if (os->os_dsl_dataset == NULL) {
/*
* This is the MOS. If we have upgraded,
* spa_max_replication() could change, so reset
* os_copies here.
*/
os->os_copies = spa_max_replication(os->os_spa);
}
/*
* Create the root block IO
*/
zb.zb_objset = os->os_dsl_dataset ? os->os_dsl_dataset->ds_object : 0;
zb.zb_object = 0;
zb.zb_level = -1;
zb.zb_blkid = 0;
if (BP_IS_OLDER(os->os_rootbp, tx->tx_txg)) {
(void) dsl_dataset_block_kill(os->os_dsl_dataset,
os->os_rootbp, pio, tx);
}
wp.wp_type = DMU_OT_OBJSET;
wp.wp_copies = os->os_copies;
wp.wp_level = (uint8_t)-1;
wp.wp_oschecksum = os->os_checksum;
wp.wp_oscompress = os->os_compress;
arc_release(os->os_phys_buf, &os->os_phys_buf);
zio = arc_write(pio, os->os_spa, &wp,
tx->tx_txg, os->os_rootbp, os->os_phys_buf, ready, NULL, os,
ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_METADATA,
&zb);
/*
* Sync meta-dnode - the parent IO for the sync is the root block
*/
os->os_meta_dnode->dn_zio = zio;
dnode_sync(os->os_meta_dnode, tx);
txgoff = tx->tx_txg & TXG_MASK;
dmu_objset_sync_dnodes(&os->os_free_dnodes[txgoff], tx);
dmu_objset_sync_dnodes(&os->os_dirty_dnodes[txgoff], tx);
list = &os->os_meta_dnode->dn_dirty_records[txgoff];
while (dr = list_head(list)) {
ASSERT(dr->dr_dbuf->db_level == 0);
list_remove(list, dr);
if (dr->dr_zio)
zio_nowait(dr->dr_zio);
}
/*
* Free intent log blocks up to this tx.
*/
zil_sync(os->os_zil, tx);
os->os_phys->os_zil_header = os->os_zil_header;
zio_nowait(zio);
}
/* called from dsl */
void
dmu_objset_sync(objset_impl_t *os, zio_t *pio, dmu_tx_t *tx)
{
int txgoff;
zbookmark_t zb;
writeprops_t wp = { 0 };
zio_t *zio;
list_t *list;
list_t *newlist = NULL;
dbuf_dirty_record_t *dr;
dprintf_ds(os->os_dsl_dataset, "txg=%llu\n", tx->tx_txg);
ASSERT(dmu_tx_is_syncing(tx));
/* XXX the write_done callback should really give us the tx... */
os->os_synctx = tx;
if (os->os_dsl_dataset == NULL) {
/*
* This is the MOS. If we have upgraded,
* spa_max_replication() could change, so reset
* os_copies here.
*/
os->os_copies = spa_max_replication(os->os_spa);
}
/*
* Create the root block IO
*/
zb.zb_objset = os->os_dsl_dataset ? os->os_dsl_dataset->ds_object : 0;
zb.zb_object = 0;
zb.zb_level = -1; /* for block ordering; it's level 0 on disk */
zb.zb_blkid = 0;
wp.wp_type = DMU_OT_OBJSET;
wp.wp_level = 0; /* on-disk BP level; see above */
wp.wp_copies = os->os_copies;
wp.wp_oschecksum = os->os_checksum;
wp.wp_oscompress = os->os_compress;
if (BP_IS_OLDER(os->os_rootbp, tx->tx_txg)) {
(void) dsl_dataset_block_kill(os->os_dsl_dataset,
os->os_rootbp, pio, tx);
}
arc_release(os->os_phys_buf, &os->os_phys_buf);
zio = arc_write(pio, os->os_spa, &wp, DMU_OS_IS_L2CACHEABLE(os),
tx->tx_txg, os->os_rootbp, os->os_phys_buf, ready, NULL, os,
ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb);
/*
* Sync special dnodes - the parent IO for the sync is the root block
*/
os->os_meta_dnode->dn_zio = zio;
dnode_sync(os->os_meta_dnode, tx);
os->os_phys->os_flags = os->os_flags;
if (os->os_userused_dnode &&
os->os_userused_dnode->dn_type != DMU_OT_NONE) {
os->os_userused_dnode->dn_zio = zio;
dnode_sync(os->os_userused_dnode, tx);
os->os_groupused_dnode->dn_zio = zio;
dnode_sync(os->os_groupused_dnode, tx);
}
txgoff = tx->tx_txg & TXG_MASK;
if (dmu_objset_userused_enabled(os)) {
newlist = &os->os_synced_dnodes;
/*
* We must create the list here because it uses the
* dn_dirty_link[] of this txg.
*/
list_create(newlist, sizeof (dnode_t),
offsetof(dnode_t, dn_dirty_link[txgoff]));
}
dmu_objset_sync_dnodes(&os->os_free_dnodes[txgoff], newlist, tx);
dmu_objset_sync_dnodes(&os->os_dirty_dnodes[txgoff], newlist, tx);
list = &os->os_meta_dnode->dn_dirty_records[txgoff];
while ((dr = list_head(list))) {
ASSERT(dr->dr_dbuf->db_level == 0);
list_remove(list, dr);
if (dr->dr_zio)
zio_nowait(dr->dr_zio);
}
/*
* Free intent log blocks up to this tx.
*/
zil_sync(os->os_zil, tx);
os->os_phys->os_zil_header = os->os_zil_header;
zio_nowait(zio);
}
