bcachefs: bcachefs_metadata_version_fast_device_removal

Fast device removal, that uses backpointers to find pointers to the
device being removed instead of a full metadata scan.

This requires BCH_SB_MEMBER_DELETED_UUID, which is an incompatible
change - hence the version number bump. We don't fully trust
backpointers, so we don't want to reuse device indexes until after a
fsck has verified that there aren't any pointers to removed devices.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This commit is contained in:
Kent Overstreet 2025-05-03 21:26:04 -04:00
parent 09fa6c3039
commit a8539ad8fa
3 changed files with 28 additions and 6 deletions

View File

@ -696,7 +696,8 @@ struct bch_sb_field_ext {
x(stripe_lru, BCH_VERSION(1, 23)) \
x(casefolding, BCH_VERSION(1, 24)) \
x(extent_flags, BCH_VERSION(1, 25)) \
x(snapshot_deletion_v2, BCH_VERSION(1, 26))
x(snapshot_deletion_v2, BCH_VERSION(1, 26)) \
x(fast_device_removal, BCH_VERSION(1, 27))
enum bcachefs_metadata_version {
bcachefs_metadata_version_min = 9,

View File

@ -2197,13 +2197,15 @@ static int bch2_invalidate_stripe_to_dev_from_alloc(struct btree_trans *trans, s
int bch2_dev_remove_stripes(struct bch_fs *c, unsigned dev_idx, unsigned flags)
{
return bch2_trans_run(c,
int ret = bch2_trans_run(c,
for_each_btree_key_max_commit(trans, iter,
BTREE_ID_alloc, POS(dev_idx, 0), POS(dev_idx, U64_MAX),
BTREE_ITER_intent, k,
NULL, NULL, 0, ({
bch2_invalidate_stripe_to_dev_from_alloc(trans, k, flags);
})));
bch_err_fn(c, ret);
return ret;
}
/* startup/shutdown */

View File

@ -1726,6 +1726,8 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
{
struct bch_member *m;
unsigned dev_idx = ca->dev_idx, data;
bool fast_device_removal = !bch2_request_incompat_feature(c,
bcachefs_metadata_version_fast_device_removal);
int ret;
down_write(&c->state_lock);
@ -1744,12 +1746,25 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
__bch2_dev_read_only(c, ca);
ret = bch2_dev_data_drop(c, ca->dev_idx, flags) ?:
bch2_dev_remove_stripes(c, ca->dev_idx, flags);
bch_err_msg(ca, ret, "bch2_dev_data_drop()");
ret = fast_device_removal
? bch2_dev_data_drop_by_backpointers(c, ca->dev_idx, flags)
: (bch2_dev_data_drop(c, ca->dev_idx, flags) ?:
bch2_dev_remove_stripes(c, ca->dev_idx, flags));
if (ret)
goto err;
/* Check if device still has data before blowing away alloc info */
struct bch_dev_usage usage = bch2_dev_usage_read(ca);
for (unsigned i = 0; i < BCH_DATA_NR; i++)
if (!data_type_is_empty(i) &&
!data_type_is_hidden(i) &&
usage.buckets[i]) {
bch_err(ca, "Remove failed: still has data (%s, %llu buckets)",
__bch2_data_types[i], usage.buckets[i]);
ret = -EBUSY;
goto err;
}
ret = bch2_dev_remove_alloc(c, ca);
bch_err_msg(ca, ret, "bch2_dev_remove_alloc()");
if (ret)
@ -1813,7 +1828,11 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags)
*/
mutex_lock(&c->sb_lock);
m = bch2_members_v2_get_mut(c->disk_sb.sb, dev_idx);
memset(&m->uuid, 0, sizeof(m->uuid));
if (fast_device_removal)
m->uuid = BCH_SB_MEMBER_DELETED_UUID;
else
memset(&m->uuid, 0, sizeof(m->uuid));
bch2_write_super(c);