#include <linux/console.h>
#include <linux/debugfs.h>
#include <linux/module.h>
-#include <linux/pretty-printers.h>
#include <linux/random.h>
#include <linux/seq_file.h>
struct bset *sorted, *inmemory = &b->data->keys;
struct bch_dev *ca = bch_dev_bkey_exists(c, pick.ptr.dev);
struct bio *bio;
- bool failed = false;
+ bool failed = false, saw_error = false;
if (!bch2_dev_get_ioref(ca, READ))
return false;
bio = bio_alloc_bioset(ca->disk_sb.bdev,
buf_pages(n_sorted, btree_bytes(c)),
REQ_OP_READ|REQ_META,
- GFP_NOIO,
+ GFP_NOFS,
&c->btree_bio);
bio->bi_iter.bi_sector = pick.ptr.offset;
bch2_bio_map(bio, n_sorted, btree_bytes(c));
memcpy(n_ondisk, n_sorted, btree_bytes(c));
v->written = 0;
- if (bch2_btree_node_read_done(c, ca, v, false))
+ if (bch2_btree_node_read_done(c, ca, v, false, &saw_error) || saw_error)
return false;
n_sorted = c->verify_data->data;
BUG_ON(b->nsets != 1);
- for (k = inmemory->start; k != vstruct_last(inmemory); k = bkey_next(k))
+ for (k = inmemory->start; k != vstruct_last(inmemory); k = bkey_p_next(k))
if (k->type == KEY_TYPE_btree_ptr_v2) {
struct bch_btree_ptr_v2 *v = (void *) bkeyp_val(&b->format, k);
v->mem_ptr = 0;
bch2_btree_node_io_unlock(b);
}
+void bch2_btree_node_ondisk_to_text(struct printbuf *out, struct bch_fs *c,
+ const struct btree *b)
+{
+ struct btree_node *n_ondisk = NULL;
+ struct extent_ptr_decoded pick;
+ struct bch_dev *ca;
+ struct bio *bio = NULL;
+ unsigned offset = 0;
+ int ret;
+
+ if (bch2_bkey_pick_read_device(c, bkey_i_to_s_c(&b->key), NULL, &pick) <= 0) {
+ prt_printf(out, "error getting device to read from: invalid device\n");
+ return;
+ }
+
+ ca = bch_dev_bkey_exists(c, pick.ptr.dev);
+ if (!bch2_dev_get_ioref(ca, READ)) {
+ prt_printf(out, "error getting device to read from: not online\n");
+ return;
+ }
+
+ n_ondisk = kvpmalloc(btree_bytes(c), GFP_KERNEL);
+ if (!n_ondisk) {
+ prt_printf(out, "memory allocation failure\n");
+ goto out;
+ }
+
+ bio = bio_alloc_bioset(ca->disk_sb.bdev,
+ buf_pages(n_ondisk, btree_bytes(c)),
+ REQ_OP_READ|REQ_META,
+ GFP_NOFS,
+ &c->btree_bio);
+ bio->bi_iter.bi_sector = pick.ptr.offset;
+ bch2_bio_map(bio, n_ondisk, btree_bytes(c));
+
+ ret = submit_bio_wait(bio);
+ if (ret) {
+ prt_printf(out, "IO error reading btree node: %s\n", bch2_err_str(ret));
+ goto out;
+ }
+
+ while (offset < btree_sectors(c)) {
+ struct bset *i;
+ struct nonce nonce;
+ struct bch_csum csum;
+ struct bkey_packed *k;
+ unsigned sectors;
+
+ if (!offset) {
+ i = &n_ondisk->keys;
+
+ if (!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i))) {
+ prt_printf(out, "unknown checksum type at offset %u: %llu\n",
+ offset, BSET_CSUM_TYPE(i));
+ goto out;
+ }
+
+ nonce = btree_nonce(i, offset << 9);
+ csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, n_ondisk);
+
+ if (bch2_crc_cmp(csum, n_ondisk->csum)) {
+ prt_printf(out, "invalid checksum\n");
+ goto out;
+ }
+
+ bset_encrypt(c, i, offset << 9);
+
+ sectors = vstruct_sectors(n_ondisk, c->block_bits);
+ } else {
+ struct btree_node_entry *bne = (void *) n_ondisk + (offset << 9);
+
+ i = &bne->keys;
+
+ if (i->seq != n_ondisk->keys.seq)
+ break;
+
+ if (!bch2_checksum_type_valid(c, BSET_CSUM_TYPE(i))) {
+ prt_printf(out, "unknown checksum type at offset %u: %llu\n",
+ offset, BSET_CSUM_TYPE(i));
+ goto out;
+ }
+
+ nonce = btree_nonce(i, offset << 9);
+ csum = csum_vstruct(c, BSET_CSUM_TYPE(i), nonce, bne);
+
+ if (bch2_crc_cmp(csum, bne->csum)) {
+ prt_printf(out, "invalid checksum");
+ goto out;
+ }
+
+ bset_encrypt(c, i, offset << 9);
+
+ sectors = vstruct_sectors(bne, c->block_bits);
+ }
+
+ prt_printf(out, " offset %u version %u, journal seq %llu\n",
+ offset,
+ le16_to_cpu(i->version),
+ le64_to_cpu(i->journal_seq));
+ offset += sectors;
+
+ printbuf_indent_add(out, 4);
+
+ for (k = i->start; k != vstruct_last(i); k = bkey_p_next(k)) {
+ struct bkey u;
+
+ bch2_bkey_val_to_text(out, c, bkey_disassemble(b, k, &u));
+ prt_newline(out);
+ }
+
+ printbuf_indent_sub(out, 4);
+ }
+out:
+ if (bio)
+ bio_put(bio);
+ kvpfree(n_ondisk, btree_bytes(c));
+ percpu_ref_put(&ca->io_ref);
+}
+
#ifdef CONFIG_DEBUG_FS
/* XXX: bch_fs refcounting */
i->size = size;
i->ret = 0;
- bch2_trans_init(&trans, i->c, 0, 0);
+ ret = flush_buf(i);
+ if (ret)
+ return ret;
+ bch2_trans_init(&trans, i->c, 0, 0);
ret = for_each_btree_key2(&trans, iter, i->id, i->from,
BTREE_ITER_PREFETCH|
BTREE_ITER_ALL_SNAPSHOTS, k, ({
- ret = flush_buf(i);
- if (ret)
- break;
-
bch2_bkey_val_to_text(&i->buf, i->c, k);
prt_newline(&i->buf);
- 0;
+ drop_locks_do(&trans, flush_buf(i));
}));
i->from = iter.pos;
+ bch2_trans_exit(&trans);
+
if (!ret)
ret = flush_buf(i);
- bch2_trans_exit(&trans);
-
return ret ?: i->ret;
}
if (ret)
return ret;
- if (!bpos_cmp(SPOS_MAX, i->from))
+ if (bpos_eq(SPOS_MAX, i->from))
return i->ret;
bch2_trans_init(&trans, i->c, 0, 0);
+retry:
+ bch2_trans_begin(&trans);
for_each_btree_node(&trans, iter, i->id, i->from, 0, b, ret) {
- ret = flush_buf(i);
- if (ret)
- break;
-
bch2_btree_node_to_text(&i->buf, i->c, b);
- i->from = bpos_cmp(SPOS_MAX, b->key.k.p)
+ i->from = !bpos_eq(SPOS_MAX, b->key.k.p)
? bpos_successor(b->key.k.p)
: b->key.k.p;
+
+ ret = drop_locks_do(&trans, flush_buf(i));
+ if (ret)
+ break;
}
bch2_trans_iter_exit(&trans, &iter);
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+ goto retry;
+
bch2_trans_exit(&trans);
if (!ret)
struct bkey_packed *_k =
bch2_btree_node_iter_peek(&l->iter, l->b);
- ret = flush_buf(i);
- if (ret)
- break;
-
- if (bpos_cmp(l->b->key.k.p, i->prev_node) > 0) {
+ if (bpos_gt(l->b->key.k.p, i->prev_node)) {
bch2_btree_node_to_text(&i->buf, i->c, l->b);
i->prev_node = l->b->key.k.p;
}
bch2_bfloat_to_text(&i->buf, l->b, _k);
- 0;
+ drop_locks_do(&trans, flush_buf(i));
}));
i->from = iter.pos;
.read = bch2_cached_btree_nodes_read,
};
+#ifdef CONFIG_BCACHEFS_DEBUG_TRANSACTIONS
static ssize_t bch2_btree_transactions_read(struct file *file, char __user *buf,
size_t size, loff_t *ppos)
{
struct bch_fs *c = i->c;
struct btree_trans *trans;
ssize_t ret = 0;
+ u32 seq;
i->ubuf = buf;
i->size = size;
i->ret = 0;
-
- mutex_lock(&c->btree_trans_lock);
+restart:
+ seqmutex_lock(&c->btree_trans_lock);
list_for_each_entry(trans, &c->btree_trans_list, list) {
if (trans->locking_wait.task->pid <= i->iter)
continue;
+ closure_get(&trans->ref);
+ seq = seqmutex_seq(&c->btree_trans_lock);
+ seqmutex_unlock(&c->btree_trans_lock);
+
ret = flush_buf(i);
- if (ret)
- return ret;
+ if (ret) {
+ closure_put(&trans->ref);
+ goto unlocked;
+ }
bch2_btree_trans_to_text(&i->buf, trans);
prt_printf(&i->buf, "backtrace:");
prt_newline(&i->buf);
printbuf_indent_add(&i->buf, 2);
- bch2_prt_backtrace(&i->buf, trans->locking_wait.task);
+ bch2_prt_task_backtrace(&i->buf, trans->locking_wait.task);
printbuf_indent_sub(&i->buf, 2);
prt_newline(&i->buf);
i->iter = trans->locking_wait.task->pid;
- }
- mutex_unlock(&c->btree_trans_lock);
+ closure_put(&trans->ref);
+
+ if (!seqmutex_relock(&c->btree_trans_lock, seq))
+ goto restart;
+ }
+ seqmutex_unlock(&c->btree_trans_lock);
+unlocked:
if (i->buf.allocation_failure)
ret = -ENOMEM;
.release = bch2_dump_release,
.read = bch2_btree_transactions_read,
};
+#endif /* CONFIG_BCACHEFS_DEBUG_TRANSACTIONS */
static ssize_t bch2_journal_pins_read(struct file *file, char __user *buf,
size_t size, loff_t *ppos)
struct bch_fs *c = i->c;
struct btree_trans *trans;
ssize_t ret = 0;
+ u32 seq;
i->ubuf = buf;
i->size = size;
if (i->iter)
goto out;
-
- mutex_lock(&c->btree_trans_lock);
+restart:
+ seqmutex_lock(&c->btree_trans_lock);
list_for_each_entry(trans, &c->btree_trans_list, list) {
if (trans->locking_wait.task->pid <= i->iter)
continue;
+ closure_get(&trans->ref);
+ seq = seqmutex_seq(&c->btree_trans_lock);
+ seqmutex_unlock(&c->btree_trans_lock);
+
ret = flush_buf(i);
- if (ret)
- return ret;
+ if (ret) {
+ closure_put(&trans->ref);
+ goto out;
+ }
bch2_check_for_deadlock(trans, &i->buf);
i->iter = trans->locking_wait.task->pid;
+
+ closure_put(&trans->ref);
+
+ if (!seqmutex_relock(&c->btree_trans_lock, seq))
+ goto restart;
}
- mutex_unlock(&c->btree_trans_lock);
+ seqmutex_unlock(&c->btree_trans_lock);
out:
if (i->buf.allocation_failure)
ret = -ENOMEM;
debugfs_create_file("cached_btree_nodes", 0400, c->fs_debug_dir,
c->btree_debug, &cached_btree_nodes_ops);
+#ifdef CONFIG_BCACHEFS_DEBUG_TRANSACTIONS
debugfs_create_file("btree_transactions", 0400, c->fs_debug_dir,
c->btree_debug, &btree_transactions_ops);
+#endif
debugfs_create_file("journal_pins", 0400, c->fs_debug_dir,
c->btree_debug, &journal_pins_ops);
projects / bcachefs-tools-debian / blobdiff