+ if (fsck_err_on(!bch2_dev_bucket_exists(c, pos), c,
+ "entry in %s btree for nonexistant dev:bucket %llu:%llu",
+ bch2_btree_ids[iter->btree_id], pos.inode, pos.offset))
+ goto delete;
+
+ a = bch2_alloc_to_v4(alloc_k, &a_convert);
+
+ if (fsck_err_on(a->data_type != state ||
+ (state == BCH_DATA_free &&
+ genbits != alloc_freespace_genbits(*a)), c,
+ "%s\n incorrectly set at %s:%llu:%llu:0 (free %u, genbits %llu should be %llu)",
+ (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf),
+ bch2_btree_ids[iter->btree_id],
+ iter->pos.inode,
+ iter->pos.offset,
+ a->data_type == state,
+ genbits >> 56, alloc_freespace_genbits(*a) >> 56))
+ goto delete;
+out:
+fsck_err:
+ set_btree_iter_dontneed(&alloc_iter);
+ bch2_trans_iter_exit(trans, &alloc_iter);
+ printbuf_exit(&buf);
+ return ret;
+delete:
+ ret = bch2_btree_delete_extent_at(trans, iter,
+ iter->btree_id == BTREE_ID_freespace ? 1 : 0, 0) ?:
+ bch2_trans_commit(trans, NULL, NULL,
+ BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW);
+ goto out;
+}
+
+static int bch2_check_discard_freespace_key(struct btree_trans *trans,
+ struct btree_iter *iter,
+ struct bpos end)
+{
+ if (!btree_node_type_is_extents(iter->btree_id)) {
+ return __bch2_check_discard_freespace_key(trans, iter);
+ } else {
+ int ret;
+
+ while (!bkey_eq(iter->pos, end) &&
+ !(ret = btree_trans_too_many_iters(trans) ?:
+ __bch2_check_discard_freespace_key(trans, iter)))
+ bch2_btree_iter_set_pos(iter, bpos_nosnap_successor(iter->pos));
+
+ return ret;
+ }
+}
+
+/*
+ * We've already checked that generation numbers in the bucket_gens btree are
+ * valid for buckets that exist; this just checks for keys for nonexistent
+ * buckets.
+ */
+static int bch2_check_bucket_gens_key(struct btree_trans *trans,
+ struct btree_iter *iter,
+ struct bkey_s_c k)
+{
+ struct bch_fs *c = trans->c;
+ struct bkey_i_bucket_gens g;
+ struct bch_dev *ca;
+ u64 start = bucket_gens_pos_to_alloc(k.k->p, 0).offset;
+ u64 end = bucket_gens_pos_to_alloc(bpos_nosnap_successor(k.k->p), 0).offset;
+ u64 b;
+ bool need_update = false, dev_exists;
+ struct printbuf buf = PRINTBUF;
+ int ret = 0;
+
+ BUG_ON(k.k->type != KEY_TYPE_bucket_gens);
+ bkey_reassemble(&g.k_i, k);
+
+ /* if no bch_dev, skip out whether we repair or not */
+ dev_exists = bch2_dev_exists2(c, k.k->p.inode);
+ if (!dev_exists) {
+ if (fsck_err_on(!dev_exists, c,
+ "bucket_gens key for invalid device:\n %s",
+ (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+ ret = bch2_btree_delete_at(trans, iter, 0);
+ }
+ goto out;
+ }
+
+ ca = bch_dev_bkey_exists(c, k.k->p.inode);
+ if (fsck_err_on(end <= ca->mi.first_bucket ||
+ start >= ca->mi.nbuckets, c,
+ "bucket_gens key for invalid buckets:\n %s",
+ (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
+ ret = bch2_btree_delete_at(trans, iter, 0);
+ goto out;
+ }
+
+ for (b = start; b < ca->mi.first_bucket; b++)
+ if (fsck_err_on(g.v.gens[b & KEY_TYPE_BUCKET_GENS_MASK], c,
+ "bucket_gens key has nonzero gen for invalid bucket")) {
+ g.v.gens[b & KEY_TYPE_BUCKET_GENS_MASK] = 0;
+ need_update = true;
+ }
+
+ for (b = ca->mi.nbuckets; b < end; b++)
+ if (fsck_err_on(g.v.gens[b & KEY_TYPE_BUCKET_GENS_MASK], c,
+ "bucket_gens key has nonzero gen for invalid bucket")) {
+ g.v.gens[b & KEY_TYPE_BUCKET_GENS_MASK] = 0;
+ need_update = true;
+ }
+
+ if (need_update) {
+ struct bkey_i *k;
+
+ k = bch2_trans_kmalloc(trans, sizeof(g));
+ ret = PTR_ERR_OR_ZERO(k);
+ if (ret)
+ goto out;
+
+ memcpy(k, &g, sizeof(g));
+ ret = bch2_trans_update(trans, iter, k, 0);
+ }
+out:
+fsck_err:
+ printbuf_exit(&buf);
+ return ret;
+}
+
+int bch2_check_alloc_info(struct bch_fs *c)
+{
+ struct btree_trans trans;
+ struct btree_iter iter, discard_iter, freespace_iter, bucket_gens_iter;
+ struct bkey hole;
+ struct bkey_s_c k;
+ int ret = 0;
+
+ bch2_trans_init(&trans, c, 0, 0);
+
+ bch2_trans_iter_init(&trans, &iter, BTREE_ID_alloc, POS_MIN,
+ BTREE_ITER_PREFETCH);
+ bch2_trans_iter_init(&trans, &discard_iter, BTREE_ID_need_discard, POS_MIN,
+ BTREE_ITER_PREFETCH);
+ bch2_trans_iter_init(&trans, &freespace_iter, BTREE_ID_freespace, POS_MIN,
+ BTREE_ITER_PREFETCH);
+ bch2_trans_iter_init(&trans, &bucket_gens_iter, BTREE_ID_bucket_gens, POS_MIN,
+ BTREE_ITER_PREFETCH);
+
+ while (1) {
+ struct bpos next;
+
+ bch2_trans_begin(&trans);
+
+ k = bch2_get_key_or_real_bucket_hole(&iter, &hole);
+ ret = bkey_err(k);
+ if (ret)
+ goto bkey_err;
+
+ if (!k.k)
+ break;
+
+ if (k.k->type) {
+ next = bpos_nosnap_successor(k.k->p);
+
+ ret = bch2_check_alloc_key(&trans,
+ k, &iter,
+ &discard_iter,
+ &freespace_iter,
+ &bucket_gens_iter);
+ if (ret)
+ goto bkey_err;
+ } else {
+ next = k.k->p;
+
+ ret = bch2_check_alloc_hole_freespace(&trans,
+ bkey_start_pos(k.k),
+ &next,
+ &freespace_iter) ?:
+ bch2_check_alloc_hole_bucket_gens(&trans,
+ bkey_start_pos(k.k),
+ &next,
+ &bucket_gens_iter);
+ if (ret)
+ goto bkey_err;
+ }
+
+ ret = bch2_trans_commit(&trans, NULL, NULL,
+ BTREE_INSERT_NOFAIL|
+ BTREE_INSERT_LAZY_RW);
+ if (ret)
+ goto bkey_err;
+
+ bch2_btree_iter_set_pos(&iter, next);
+bkey_err:
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+ continue;
+ if (ret)
+ break;
+ }
+ bch2_trans_iter_exit(&trans, &bucket_gens_iter);
+ bch2_trans_iter_exit(&trans, &freespace_iter);
+ bch2_trans_iter_exit(&trans, &discard_iter);
+ bch2_trans_iter_exit(&trans, &iter);
+
+ if (ret < 0)
+ goto err;
+
+ ret = for_each_btree_key2(&trans, iter,
+ BTREE_ID_need_discard, POS_MIN,
+ BTREE_ITER_PREFETCH, k,
+ bch2_check_discard_freespace_key(&trans, &iter, k.k->p)) ?:
+ for_each_btree_key2(&trans, iter,
+ BTREE_ID_freespace, POS_MIN,
+ BTREE_ITER_PREFETCH, k,
+ bch2_check_discard_freespace_key(&trans, &iter, k.k->p)) ?:
+ for_each_btree_key_commit(&trans, iter,
+ BTREE_ID_bucket_gens, POS_MIN,
+ BTREE_ITER_PREFETCH, k,
+ NULL, NULL, BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW,
+ bch2_check_bucket_gens_key(&trans, &iter, k));
+err:
+ bch2_trans_exit(&trans);
+ return ret < 0 ? ret : 0;
+}
+
+static int bch2_check_alloc_to_lru_ref(struct btree_trans *trans,
+ struct btree_iter *alloc_iter)
+{
+ struct bch_fs *c = trans->c;
+ struct btree_iter lru_iter;
+ struct bch_alloc_v4 a_convert;
+ const struct bch_alloc_v4 *a;
+ struct bkey_s_c alloc_k, lru_k;
+ struct printbuf buf = PRINTBUF;
+ int ret;
+
+ alloc_k = bch2_btree_iter_peek(alloc_iter);
+ if (!alloc_k.k)
+ return 0;
+
+ ret = bkey_err(alloc_k);
+ if (ret)
+ return ret;
+
+ a = bch2_alloc_to_v4(alloc_k, &a_convert);
+
+ if (a->data_type != BCH_DATA_cached)
+ return 0;
+
+ lru_k = bch2_bkey_get_iter(trans, &lru_iter, BTREE_ID_lru,
+ lru_pos(alloc_k.k->p.inode,
+ bucket_to_u64(alloc_k.k->p),
+ a->io_time[READ]), 0);
+ ret = bkey_err(lru_k);
+ if (ret)
+ return ret;
+
+ if (fsck_err_on(!a->io_time[READ], c,
+ "cached bucket with read_time 0\n"
+ " %s",
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)) ||
+ fsck_err_on(lru_k.k->type != KEY_TYPE_set, c,
+ "missing lru entry\n"
+ " %s",
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
+ u64 read_time = a->io_time[READ] ?:
+ atomic64_read(&c->io_clock[READ].now);
+
+ ret = bch2_lru_set(trans,
+ alloc_k.k->p.inode,
+ bucket_to_u64(alloc_k.k->p),
+ read_time);
+ if (ret)
+ goto err;
+
+ if (a->io_time[READ] != read_time) {
+ struct bkey_i_alloc_v4 *a_mut =
+ bch2_alloc_to_v4_mut(trans, alloc_k);
+ ret = PTR_ERR_OR_ZERO(a_mut);
+ if (ret)
+ goto err;
+
+ a_mut->v.io_time[READ] = read_time;
+ ret = bch2_trans_update(trans, alloc_iter,
+ &a_mut->k_i, BTREE_TRIGGER_NORUN);
+ if (ret)
+ goto err;
+ }
+ }
+err:
+fsck_err:
+ bch2_trans_iter_exit(trans, &lru_iter);
+ printbuf_exit(&buf);
+ return ret;
+}
+
+int bch2_check_alloc_to_lru_refs(struct bch_fs *c)
+{
+ struct btree_trans trans;
+ struct btree_iter iter;
+ struct bkey_s_c k;
+ int ret = 0;
+
+ bch2_trans_init(&trans, c, 0, 0);
+
+ for_each_btree_key_commit(&trans, iter, BTREE_ID_alloc,
+ POS_MIN, BTREE_ITER_PREFETCH, k,
+ NULL, NULL, BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW,
+ bch2_check_alloc_to_lru_ref(&trans, &iter));
+
+ bch2_trans_exit(&trans);
+ return ret < 0 ? ret : 0;
+}
+
+static int bch2_discard_one_bucket(struct btree_trans *trans,
+ struct btree_iter *need_discard_iter,
+ struct bpos *discard_pos_done,
+ u64 *seen,
+ u64 *open,
+ u64 *need_journal_commit,
+ u64 *discarded)
+{
+ struct bch_fs *c = trans->c;
+ struct bpos pos = need_discard_iter->pos;
+ struct btree_iter iter = { NULL };
+ struct bkey_s_c k;
+ struct bch_dev *ca;
+ struct bkey_i_alloc_v4 *a;
+ struct printbuf buf = PRINTBUF;
+ int ret = 0;
+
+ ca = bch_dev_bkey_exists(c, pos.inode);
+ if (!percpu_ref_tryget(&ca->io_ref)) {
+ bch2_btree_iter_set_pos(need_discard_iter, POS(pos.inode + 1, 0));
+ return 0;
+ }
+
+ if (bch2_bucket_is_open_safe(c, pos.inode, pos.offset)) {
+ (*open)++;
+ goto out;
+ }
+
+ if (bch2_bucket_needs_journal_commit(&c->buckets_waiting_for_journal,
+ c->journal.flushed_seq_ondisk,
+ pos.inode, pos.offset)) {
+ (*need_journal_commit)++;
+ goto out;
+ }
+
+ k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_alloc,
+ need_discard_iter->pos,
+ BTREE_ITER_CACHED);
+ ret = bkey_err(k);
+ if (ret)
+ goto out;
+
+ a = bch2_alloc_to_v4_mut(trans, k);
+ ret = PTR_ERR_OR_ZERO(a);
+ if (ret)
+ goto out;
+
+ if (BCH_ALLOC_V4_NEED_INC_GEN(&a->v)) {
+ a->v.gen++;
+ SET_BCH_ALLOC_V4_NEED_INC_GEN(&a->v, false);
+ goto write;
+ }
+
+ if (a->v.journal_seq > c->journal.flushed_seq_ondisk) {
+ if (test_bit(BCH_FS_CHECK_ALLOC_DONE, &c->flags)) {
+ bch2_trans_inconsistent(trans,
+ "clearing need_discard but journal_seq %llu > flushed_seq %llu\n"
+ "%s",
+ a->v.journal_seq,
+ c->journal.flushed_seq_ondisk,
+ (bch2_bkey_val_to_text(&buf, c, k), buf.buf));
+ ret = -EIO;
+ }
+ goto out;
+ }
+
+ if (a->v.data_type != BCH_DATA_need_discard) {
+ if (test_bit(BCH_FS_CHECK_ALLOC_DONE, &c->flags)) {
+ bch2_trans_inconsistent(trans,
+ "bucket incorrectly set in need_discard btree\n"
+ "%s",
+ (bch2_bkey_val_to_text(&buf, c, k), buf.buf));
+ ret = -EIO;
+ }
+
+ goto out;
+ }
+
+ if (!bkey_eq(*discard_pos_done, iter.pos) &&
+ ca->mi.discard && !c->opts.nochanges) {
+ /*
+ * This works without any other locks because this is the only
+ * thread that removes items from the need_discard tree
+ */
+ bch2_trans_unlock(trans);
+ blkdev_issue_discard(ca->disk_sb.bdev,
+ k.k->p.offset * ca->mi.bucket_size,
+ ca->mi.bucket_size,
+ GFP_KERNEL);
+ *discard_pos_done = iter.pos;
+
+ ret = bch2_trans_relock_notrace(trans);
+ if (ret)
+ goto out;
+ }