+ bkey_init(&k->k);
+ k->k.type = new_type;
+
+ switch (a->data_type) {
+ case BCH_DATA_free:
+ btree = BTREE_ID_freespace;
+ k->k.p = alloc_freespace_pos(alloc_k.k->p, *a);
+ bch2_key_resize(&k->k, 1);
+ break;
+ case BCH_DATA_need_discard:
+ btree = BTREE_ID_need_discard;
+ k->k.p = alloc_k.k->p;
+ break;
+ default:
+ return 0;
+ }
+
+ bch2_trans_iter_init(trans, &iter, btree,
+ bkey_start_pos(&k->k),
+ BTREE_ITER_INTENT);
+ old = bch2_btree_iter_peek_slot(&iter);
+ ret = bkey_err(old);
+ if (ret)
+ goto err;
+
+ if (ca->mi.freespace_initialized &&
+ bch2_trans_inconsistent_on(old.k->type != old_type, trans,
+ "incorrect key when %s %s btree (got %s should be %s)\n"
+ " for %s",
+ set ? "setting" : "clearing",
+ bch2_btree_ids[btree],
+ bch2_bkey_types[old.k->type],
+ bch2_bkey_types[old_type],
+ (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
+ ret = -EIO;
+ goto err;
+ }
+
+ ret = bch2_trans_update(trans, &iter, k, 0);
+err:
+ bch2_trans_iter_exit(trans, &iter);
+ printbuf_exit(&buf);
+ return ret;
+}
+
+static noinline int bch2_bucket_gen_update(struct btree_trans *trans,
+ struct bpos bucket, u8 gen)
+{
+ struct btree_iter iter;
+ unsigned offset;
+ struct bpos pos = alloc_gens_pos(bucket, &offset);
+ struct bkey_i_bucket_gens *g;
+ struct bkey_s_c k;
+ int ret;
+
+ g = bch2_trans_kmalloc(trans, sizeof(*g));
+ ret = PTR_ERR_OR_ZERO(g);
+ if (ret)
+ return ret;
+
+ bch2_trans_iter_init(trans, &iter, BTREE_ID_bucket_gens, pos,
+ BTREE_ITER_INTENT|
+ BTREE_ITER_WITH_UPDATES);
+ k = bch2_btree_iter_peek_slot(&iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ if (k.k->type != KEY_TYPE_bucket_gens) {
+ bkey_bucket_gens_init(&g->k_i);
+ g->k.p = iter.pos;
+ } else {
+ bkey_reassemble(&g->k_i, k);
+ }
+
+ g->v.gens[offset] = gen;
+
+ ret = bch2_trans_update(trans, &iter, &g->k_i, 0);
+err:
+ bch2_trans_iter_exit(trans, &iter);
+ return ret;
+}
+
+int bch2_trans_mark_alloc(struct btree_trans *trans,
+ enum btree_id btree_id, unsigned level,
+ struct bkey_s_c old, struct bkey_i *new,
+ unsigned flags)
+{
+ struct bch_fs *c = trans->c;
+ struct bch_alloc_v4 old_a_convert, *new_a;
+ const struct bch_alloc_v4 *old_a;
+ u64 old_lru, new_lru;
+ int ret = 0;
+
+ /*
+ * Deletion only happens in the device removal path, with
+ * BTREE_TRIGGER_NORUN:
+ */
+ BUG_ON(new->k.type != KEY_TYPE_alloc_v4);
+
+ old_a = bch2_alloc_to_v4(old, &old_a_convert);
+ new_a = &bkey_i_to_alloc_v4(new)->v;
+
+ new_a->data_type = alloc_data_type(*new_a, new_a->data_type);
+
+ if (new_a->dirty_sectors > old_a->dirty_sectors ||
+ new_a->cached_sectors > old_a->cached_sectors) {
+ new_a->io_time[READ] = max_t(u64, 1, atomic64_read(&c->io_clock[READ].now));
+ new_a->io_time[WRITE]= max_t(u64, 1, atomic64_read(&c->io_clock[WRITE].now));
+ SET_BCH_ALLOC_V4_NEED_INC_GEN(new_a, true);
+ SET_BCH_ALLOC_V4_NEED_DISCARD(new_a, true);
+ }
+
+ if (data_type_is_empty(new_a->data_type) &&
+ BCH_ALLOC_V4_NEED_INC_GEN(new_a) &&
+ !bch2_bucket_is_open_safe(c, new->k.p.inode, new->k.p.offset)) {
+ new_a->gen++;
+ SET_BCH_ALLOC_V4_NEED_INC_GEN(new_a, false);
+ }
+
+ if (old_a->data_type != new_a->data_type ||
+ (new_a->data_type == BCH_DATA_free &&
+ alloc_freespace_genbits(*old_a) != alloc_freespace_genbits(*new_a))) {
+ ret = bch2_bucket_do_index(trans, old, old_a, false) ?:
+ bch2_bucket_do_index(trans, bkey_i_to_s_c(new), new_a, true);
+ if (ret)
+ return ret;
+ }
+
+ if (new_a->data_type == BCH_DATA_cached &&
+ !new_a->io_time[READ])
+ new_a->io_time[READ] = max_t(u64, 1, atomic64_read(&c->io_clock[READ].now));
+
+ old_lru = alloc_lru_idx(*old_a);
+ new_lru = alloc_lru_idx(*new_a);
+
+ if (old_lru != new_lru) {
+ ret = bch2_lru_change(trans, new->k.p.inode, new->k.p.offset,
+ old_lru, &new_lru, old);
+ if (ret)
+ return ret;
+
+ if (new_a->data_type == BCH_DATA_cached)
+ new_a->io_time[READ] = new_lru;
+ }
+
+ if (old_a->gen != new_a->gen) {
+ ret = bch2_bucket_gen_update(trans, new->k.p, new_a->gen);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+/*
+ * This synthesizes deleted extents for holes, similar to BTREE_ITER_SLOTS for
+ * extents style btrees, but works on non-extents btrees:
+ */
+struct bkey_s_c bch2_get_key_or_hole(struct btree_iter *iter, struct bpos end, struct bkey *hole)
+{
+ struct bkey_s_c k = bch2_btree_iter_peek_slot(iter);
+
+ if (bkey_err(k))
+ return k;
+
+ if (k.k->type) {
+ return k;
+ } else {
+ struct btree_iter iter2;
+ struct bpos next;
+
+ bch2_trans_copy_iter(&iter2, iter);
+ k = bch2_btree_iter_peek_upto(&iter2,
+ bkey_min(bkey_min(end,
+ iter->path->l[0].b->key.k.p),
+ POS(iter->pos.inode, iter->pos.offset + U32_MAX - 1)));
+ next = iter2.pos;
+ bch2_trans_iter_exit(iter->trans, &iter2);
+
+ BUG_ON(next.offset >= iter->pos.offset + U32_MAX);
+
+ if (bkey_err(k))
+ return k;
+
+ bkey_init(hole);
+ hole->p = iter->pos;
+
+ bch2_key_resize(hole, next.offset - iter->pos.offset);
+ return (struct bkey_s_c) { hole, NULL };
+ }
+}
+
+static bool next_bucket(struct bch_fs *c, struct bpos *bucket)
+{
+ struct bch_dev *ca;
+ unsigned iter;
+
+ if (bch2_dev_bucket_exists(c, *bucket))
+ return true;
+
+ if (bch2_dev_exists2(c, bucket->inode)) {
+ ca = bch_dev_bkey_exists(c, bucket->inode);
+
+ if (bucket->offset < ca->mi.first_bucket) {
+ bucket->offset = ca->mi.first_bucket;
+ return true;
+ }
+
+ bucket->inode++;
+ bucket->offset = 0;
+ }
+
+ rcu_read_lock();
+ iter = bucket->inode;
+ ca = __bch2_next_dev(c, &iter, NULL);
+ if (ca)
+ bucket->offset = ca->mi.first_bucket;
+ rcu_read_unlock();
+
+ return ca != NULL;
+}
+
+struct bkey_s_c bch2_get_key_or_real_bucket_hole(struct btree_iter *iter, struct bkey *hole)
+{
+ struct bch_fs *c = iter->trans->c;
+ struct bkey_s_c k;
+again:
+ k = bch2_get_key_or_hole(iter, POS_MAX, hole);
+ if (bkey_err(k))
+ return k;
+
+ if (!k.k->type) {
+ struct bpos bucket = bkey_start_pos(k.k);
+
+ if (!bch2_dev_bucket_exists(c, bucket)) {
+ if (!next_bucket(c, &bucket))
+ return bkey_s_c_null;
+
+ bch2_btree_iter_set_pos(iter, bucket);
+ goto again;
+ }
+
+ if (!bch2_dev_bucket_exists(c, k.k->p)) {
+ struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode);
+
+ bch2_key_resize(hole, ca->mi.nbuckets - bucket.offset);
+ }
+ }
+
+ return k;
+}
+
+static int bch2_check_alloc_key(struct btree_trans *trans,
+ struct bkey_s_c alloc_k,
+ struct btree_iter *alloc_iter,
+ struct btree_iter *discard_iter,
+ struct btree_iter *freespace_iter,
+ struct btree_iter *bucket_gens_iter)
+{
+ struct bch_fs *c = trans->c;
+ struct bch_dev *ca;
+ struct bch_alloc_v4 a_convert;
+ const struct bch_alloc_v4 *a;
+ unsigned discard_key_type, freespace_key_type;
+ unsigned gens_offset;
+ struct bkey_s_c k;
+ struct printbuf buf = PRINTBUF;
+ int ret;
+
+ if (fsck_err_on(!bch2_dev_bucket_exists(c, alloc_k.k->p), c,
+ "alloc key for invalid device:bucket %llu:%llu",
+ alloc_k.k->p.inode, alloc_k.k->p.offset))
+ return bch2_btree_delete_at(trans, alloc_iter, 0);
+
+ ca = bch_dev_bkey_exists(c, alloc_k.k->p.inode);
+ if (!ca->mi.freespace_initialized)
+ return 0;
+
+ a = bch2_alloc_to_v4(alloc_k, &a_convert);
+
+ discard_key_type = a->data_type == BCH_DATA_need_discard ? KEY_TYPE_set : 0;
+ bch2_btree_iter_set_pos(discard_iter, alloc_k.k->p);
+ k = bch2_btree_iter_peek_slot(discard_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ if (k.k->type != discard_key_type &&
+ (c->opts.reconstruct_alloc ||
+ fsck_err(c, "incorrect key in need_discard btree (got %s should be %s)\n"
+ " %s",
+ bch2_bkey_types[k.k->type],
+ bch2_bkey_types[discard_key_type],
+ (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+ struct bkey_i *update =
+ bch2_trans_kmalloc(trans, sizeof(*update));
+
+ ret = PTR_ERR_OR_ZERO(update);
+ if (ret)
+ goto err;
+
+ bkey_init(&update->k);
+ update->k.type = discard_key_type;
+ update->k.p = discard_iter->pos;
+
+ ret = bch2_trans_update(trans, discard_iter, update, 0);
+ if (ret)
+ goto err;
+ }
+
+ freespace_key_type = a->data_type == BCH_DATA_free ? KEY_TYPE_set : 0;
+ bch2_btree_iter_set_pos(freespace_iter, alloc_freespace_pos(alloc_k.k->p, *a));
+ k = bch2_btree_iter_peek_slot(freespace_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ if (k.k->type != freespace_key_type &&
+ (c->opts.reconstruct_alloc ||
+ fsck_err(c, "incorrect key in freespace btree (got %s should be %s)\n"
+ " %s",
+ bch2_bkey_types[k.k->type],
+ bch2_bkey_types[freespace_key_type],
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+ struct bkey_i *update =
+ bch2_trans_kmalloc(trans, sizeof(*update));
+
+ ret = PTR_ERR_OR_ZERO(update);
+ if (ret)
+ goto err;
+
+ bkey_init(&update->k);
+ update->k.type = freespace_key_type;
+ update->k.p = freespace_iter->pos;
+ bch2_key_resize(&update->k, 1);
+
+ ret = bch2_trans_update(trans, freespace_iter, update, 0);
+ if (ret)
+ goto err;
+ }
+
+ bch2_btree_iter_set_pos(bucket_gens_iter, alloc_gens_pos(alloc_k.k->p, &gens_offset));
+ k = bch2_btree_iter_peek_slot(bucket_gens_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ if (a->gen != alloc_gen(k, gens_offset) &&
+ (c->opts.reconstruct_alloc ||
+ fsck_err(c, "incorrect gen in bucket_gens btree (got %u should be %u)\n"
+ " %s",
+ alloc_gen(k, gens_offset), a->gen,
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) {
+ struct bkey_i_bucket_gens *g =
+ bch2_trans_kmalloc(trans, sizeof(*g));
+
+ ret = PTR_ERR_OR_ZERO(g);
+ if (ret)
+ goto err;
+
+ if (k.k->type == KEY_TYPE_bucket_gens) {
+ bkey_reassemble(&g->k_i, k);
+ } else {
+ bkey_bucket_gens_init(&g->k_i);
+ g->k.p = alloc_gens_pos(alloc_k.k->p, &gens_offset);
+ }
+
+ g->v.gens[gens_offset] = a->gen;
+
+ ret = bch2_trans_update(trans, bucket_gens_iter, &g->k_i, 0);
+ if (ret)
+ goto err;
+ }
+err:
+fsck_err:
+ printbuf_exit(&buf);
+ return ret;
+}
+
+static int bch2_check_alloc_hole_freespace(struct btree_trans *trans,
+ struct bpos start,
+ struct bpos *end,
+ struct btree_iter *freespace_iter)
+{
+ struct bch_fs *c = trans->c;
+ struct bch_dev *ca;
+ struct bkey_s_c k;
+ struct printbuf buf = PRINTBUF;
+ int ret;
+
+ ca = bch_dev_bkey_exists(c, start.inode);
+ if (!ca->mi.freespace_initialized)
+ return 0;
+
+ bch2_btree_iter_set_pos(freespace_iter, start);
+
+ k = bch2_btree_iter_peek_slot(freespace_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ *end = bkey_min(k.k->p, *end);
+
+ if (k.k->type != KEY_TYPE_set &&
+ (c->opts.reconstruct_alloc ||
+ fsck_err(c, "hole in alloc btree missing in freespace btree\n"
+ " device %llu buckets %llu-%llu",
+ freespace_iter->pos.inode,
+ freespace_iter->pos.offset,
+ end->offset))) {
+ struct bkey_i *update =
+ bch2_trans_kmalloc(trans, sizeof(*update));
+
+ ret = PTR_ERR_OR_ZERO(update);
+ if (ret)
+ goto err;
+
+ bkey_init(&update->k);
+ update->k.type = KEY_TYPE_set;
+ update->k.p = freespace_iter->pos;
+ bch2_key_resize(&update->k,
+ min_t(u64, U32_MAX, end->offset -
+ freespace_iter->pos.offset));
+
+ ret = bch2_trans_update(trans, freespace_iter, update, 0);
+ if (ret)
+ goto err;
+ }
+err:
+fsck_err:
+ printbuf_exit(&buf);
+ return ret;
+}
+
+static int bch2_check_alloc_hole_bucket_gens(struct btree_trans *trans,
+ struct bpos start,
+ struct bpos *end,
+ struct btree_iter *bucket_gens_iter)
+{
+ struct bch_fs *c = trans->c;
+ struct bkey_s_c k;
+ struct printbuf buf = PRINTBUF;
+ unsigned i, gens_offset, gens_end_offset;
+ int ret;
+
+ if (c->sb.version < bcachefs_metadata_version_bucket_gens &&
+ !c->opts.version_upgrade)
+ return 0;
+
+ bch2_btree_iter_set_pos(bucket_gens_iter, alloc_gens_pos(start, &gens_offset));
+
+ k = bch2_btree_iter_peek_slot(bucket_gens_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ if (bkey_cmp(alloc_gens_pos(start, &gens_offset),
+ alloc_gens_pos(*end, &gens_end_offset)))
+ gens_end_offset = KEY_TYPE_BUCKET_GENS_NR;
+
+ if (k.k->type == KEY_TYPE_bucket_gens) {
+ struct bkey_i_bucket_gens g;
+ bool need_update = false;
+
+ bkey_reassemble(&g.k_i, k);
+
+ for (i = gens_offset; i < gens_end_offset; i++) {
+ if (fsck_err_on(g.v.gens[i], c,
+ "hole in alloc btree at %llu:%llu with nonzero gen in bucket_gens btree (%u)",
+ bucket_gens_pos_to_alloc(k.k->p, i).inode,
+ bucket_gens_pos_to_alloc(k.k->p, i).offset,
+ g.v.gens[i])) {
+ g.v.gens[i] = 0;
+ need_update = true;
+ }
+ }
+
+ if (need_update) {
+ struct bkey_i *k = bch2_trans_kmalloc(trans, sizeof(g));
+
+ ret = PTR_ERR_OR_ZERO(k);
+ if (ret)
+ goto err;
+
+ memcpy(k, &g, sizeof(g));
+
+ ret = bch2_trans_update(trans, bucket_gens_iter, k, 0);
+ if (ret)
+ goto err;
+ }
+ }
+
+ *end = bkey_min(*end, bucket_gens_pos_to_alloc(bpos_nosnap_successor(k.k->p), 0));
+err:
+fsck_err:
+ printbuf_exit(&buf);
+ return ret;
+}
+
+static int bch2_check_discard_freespace_key(struct btree_trans *trans,
+ struct btree_iter *iter)
+{
+ struct bch_fs *c = trans->c;
+ struct btree_iter alloc_iter;
+ struct bkey_s_c alloc_k;
+ struct bch_alloc_v4 a_convert;
+ const struct bch_alloc_v4 *a;
+ u64 genbits;
+ struct bpos pos;
+ enum bch_data_type state = iter->btree_id == BTREE_ID_need_discard
+ ? BCH_DATA_need_discard
+ : BCH_DATA_free;
+ struct printbuf buf = PRINTBUF;
+ int ret;
+
+ pos = iter->pos;
+ pos.offset &= ~(~0ULL << 56);
+ genbits = iter->pos.offset & (~0ULL << 56);
+
+ bch2_trans_iter_init(trans, &alloc_iter, BTREE_ID_alloc, pos, 0);
+
+ 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;
+
+ alloc_k = bch2_btree_iter_peek_slot(&alloc_iter);
+ ret = bkey_err(alloc_k);
+ if (ret)
+ goto err;
+
+ 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 in %s index (free %u, genbits %llu should be %llu)",
+ (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf),
+ bch2_btree_ids[iter->btree_id],
+ a->data_type == state,
+ genbits >> 56, alloc_freespace_genbits(*a) >> 56))
+ goto delete;
+out:
+err:
+fsck_err:
+ 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);
+ goto out;
+}
+
+/*
+ * 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;
+ struct printbuf buf = PRINTBUF;
+ int ret = 0;
+
+ BUG_ON(k.k->type != KEY_TYPE_bucket_gens);
+ bkey_reassemble(&g.k_i, k);
+
+ if (fsck_err_on(!bch2_dev_exists2(c, k.k->p.inode), 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_key_commit(&trans, iter,
+ BTREE_ID_need_discard, POS_MIN,
+ BTREE_ITER_PREFETCH, k,
+ NULL, NULL, BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW,
+ bch2_check_discard_freespace_key(&trans, &iter)) ?:
+ for_each_btree_key_commit(&trans, iter,
+ BTREE_ID_freespace, POS_MIN,
+ BTREE_ITER_PREFETCH, k,
+ NULL, NULL, BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW,
+ bch2_check_discard_freespace_key(&trans, &iter)) ?:
+ 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, k;
+ struct printbuf buf = PRINTBUF;
+ struct printbuf buf2 = 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;
+
+ bch2_trans_iter_init(trans, &lru_iter, BTREE_ID_lru,
+ POS(alloc_k.k->p.inode, a->io_time[READ]), 0);
+
+ k = bch2_btree_iter_peek_slot(&lru_iter);
+ ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ 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(k.k->type != KEY_TYPE_lru ||
+ le64_to_cpu(bkey_s_c_to_lru(k).v->idx) != alloc_k.k->p.offset, c,
+ "incorrect/missing lru entry\n"
+ " %s\n"
+ " %s",
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf),
+ (bch2_bkey_val_to_text(&buf2, c, k), buf2.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,
+ alloc_k.k->p.offset,
+ &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(&buf2);
+ 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;
+ bool did_discard = false;
+ int ret = 0;