#include "extents.h"
#include "fsck.h"
#include "inode.h"
-#include "io.h"
#include "super.h"
#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)),
+ buf_pages(n_sorted, btree_buf_bytes(b)),
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));
+ bch2_bio_map(bio, n_sorted, btree_buf_bytes(b));
submit_bio_wait(bio);
bio_put(bio);
percpu_ref_put(&ca->io_ref);
- memcpy(n_ondisk, n_sorted, btree_bytes(c));
+ memcpy(n_ondisk, n_sorted, btree_buf_bytes(b));
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;
mutex_lock(&c->verify_lock);
if (!c->verify_ondisk) {
- c->verify_ondisk = kvpmalloc(btree_bytes(c), GFP_KERNEL);
+ c->verify_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL);
if (!c->verify_ondisk)
goto out;
}
BUG_ON(b->nsets != 1);
- for (k = inmemory->start; k != vstruct_last(inmemory); k = bkey_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;
- }
+ 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 *) bkeyp_val(&b->format, k))->mem_ptr = 0;
v = c->verify_data;
bkey_copy(&v->key, &b->key);
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 = kvmalloc(btree_buf_bytes(b), 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_buf_bytes(b)),
+ 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_buf_bytes(b));
+
+ 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);
+ kvfree(n_ondisk);
+ percpu_ref_put(&ca->io_ref);
+}
+
#ifdef CONFIG_DEBUG_FS
/* XXX: bch_fs refcounting */
{
if (i->buf.pos) {
size_t bytes = min_t(size_t, i->buf.pos, i->size);
- int err = copy_to_user(i->ubuf, i->buf.buf, bytes);
+ int copied = bytes - copy_to_user(i->ubuf, i->buf.buf, bytes);
- if (err)
- return err;
+ i->ret += copied;
+ i->ubuf += copied;
+ i->size -= copied;
+ i->buf.pos -= copied;
+ memmove(i->buf.buf, i->buf.buf + copied, i->buf.pos);
- i->ret += bytes;
- i->ubuf += bytes;
- i->size -= bytes;
- i->buf.pos -= bytes;
- memmove(i->buf.buf, i->buf.buf + bytes, i->buf.pos);
+ if (copied != bytes)
+ return -EFAULT;
}
return i->size ? 0 : i->ret;
size_t size, loff_t *ppos)
{
struct dump_iter *i = file->private_data;
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- ssize_t ret;
i->ubuf = buf;
i->size = size;
i->ret = 0;
- 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;
- }));
- i->from = iter.pos;
-
- if (!ret)
- ret = flush_buf(i);
-
- bch2_trans_exit(&trans);
-
- return ret ?: i->ret;
+ return flush_buf(i) ?:
+ bch2_trans_run(i->c,
+ for_each_btree_key(trans, iter, i->id, i->from,
+ BTREE_ITER_PREFETCH|
+ BTREE_ITER_ALL_SNAPSHOTS, k, ({
+ bch2_bkey_val_to_text(&i->buf, i->c, k);
+ prt_newline(&i->buf);
+ bch2_trans_unlock(trans);
+ i->from = bpos_successor(iter.pos);
+ flush_buf(i);
+ }))) ?:
+ i->ret;
}
static const struct file_operations btree_debug_ops = {
size_t size, loff_t *ppos)
{
struct dump_iter *i = file->private_data;
- struct btree_trans trans;
+ struct btree_trans *trans;
struct btree_iter iter;
struct btree *b;
ssize_t 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);
-
- for_each_btree_node(&trans, iter, i->id, i->from, 0, b, ret) {
- ret = flush_buf(i);
- if (ret)
- break;
+ trans = bch2_trans_get(i->c);
+retry:
+ bch2_trans_begin(trans);
+ for_each_btree_node(trans, iter, i->id, i->from, 0, b, ret) {
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);
+ bch2_trans_iter_exit(trans, &iter);
+
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+ goto retry;
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
if (!ret)
ret = flush_buf(i);
size_t size, loff_t *ppos)
{
struct dump_iter *i = file->private_data;
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- ssize_t ret;
i->ubuf = buf;
i->size = size;
i->ret = 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, ({
- struct btree_path_level *l = &iter.path->l[0];
- 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) {
- 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;
- }));
- i->from = iter.pos;
-
- bch2_trans_exit(&trans);
-
- if (!ret)
- ret = flush_buf(i);
-
- return ret ?: i->ret;
+ return flush_buf(i) ?:
+ bch2_trans_run(i->c,
+ for_each_btree_key(trans, iter, i->id, i->from,
+ BTREE_ITER_PREFETCH|
+ BTREE_ITER_ALL_SNAPSHOTS, k, ({
+ struct btree_path_level *l =
+ &btree_iter_path(trans, &iter)->l[0];
+ struct bkey_packed *_k =
+ bch2_btree_node_iter_peek(&l->iter, l->b);
+
+ 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);
+ bch2_trans_unlock(trans);
+ i->from = bpos_successor(iter.pos);
+ flush_buf(i);
+ }))) ?:
+ i->ret;
}
static const struct file_operations bfloat_failed_debug_ops = {
prt_printf(out, "%px btree=%s l=%u ",
b,
- bch2_btree_ids[b->c.btree_id],
+ bch2_btree_id_str(b->c.btree_id),
b->c.level);
prt_newline(out);
if (i->iter < tbl->size) {
rht_for_each_entry_rcu(b, pos, tbl, i->iter, hash)
bch2_cached_btree_node_to_text(&i->buf, c, b);
- i->iter++;;
+ i->iter++;
} else {
done = true;
}
.read = bch2_cached_btree_nodes_read,
};
-static int prt_backtrace(struct printbuf *out, struct task_struct *task)
-{
- unsigned long entries[32];
- unsigned i, nr_entries;
- int ret;
-
- ret = down_read_killable(&task->signal->exec_update_lock);
- if (ret)
- return ret;
-
- nr_entries = stack_trace_save_tsk(task, entries, ARRAY_SIZE(entries), 0);
- for (i = 0; i < nr_entries; i++) {
- prt_printf(out, "[<0>] %pB", (void *)entries[i]);
- prt_newline(out);
- }
-
- up_read(&task->signal->exec_update_lock);
- return 0;
-}
-
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)
+ struct task_struct *task = READ_ONCE(trans->locking_wait.task);
+
+ if (!task || 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);
- prt_backtrace(&i->buf, trans->locking_wait.task);
+ bch2_prt_task_backtrace(&i->buf, task, 0, GFP_KERNEL);
printbuf_indent_sub(&i->buf, 2);
prt_newline(&i->buf);
- i->iter = trans->locking_wait.task->pid;
- }
- mutex_unlock(&c->btree_trans_lock);
+ i->iter = task->pid;
+ 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;
.read = bch2_journal_pins_read,
};
-static int lock_held_stats_open(struct inode *inode, struct file *file)
+static int btree_transaction_stats_open(struct inode *inode, struct file *file)
{
struct bch_fs *c = inode->i_private;
struct dump_iter *i;
if (!i)
return -ENOMEM;
- i->iter = 0;
+ i->iter = 1;
i->c = c;
i->buf = PRINTBUF;
file->private_data = i;
return 0;
}
-static int lock_held_stats_release(struct inode *inode, struct file *file)
+static int btree_transaction_stats_release(struct inode *inode, struct file *file)
{
struct dump_iter *i = file->private_data;
return 0;
}
-static ssize_t lock_held_stats_read(struct file *file, char __user *buf,
- size_t size, loff_t *ppos)
+static ssize_t btree_transaction_stats_read(struct file *file, char __user *buf,
+ size_t size, loff_t *ppos)
{
struct dump_iter *i = file->private_data;
struct bch_fs *c = i->c;
if (!i->size)
break;
- if (i->iter == ARRAY_SIZE(c->btree_transaction_fns) ||
- !c->btree_transaction_fns[i->iter])
+ if (i->iter == ARRAY_SIZE(bch2_btree_transaction_fns) ||
+ !bch2_btree_transaction_fns[i->iter])
break;
- prt_printf(&i->buf, "%s: ", c->btree_transaction_fns[i->iter]);
+ prt_printf(&i->buf, "%s: ", bch2_btree_transaction_fns[i->iter]);
prt_newline(&i->buf);
printbuf_indent_add(&i->buf, 2);
prt_printf(&i->buf, "Max mem used: %u", s->max_mem);
prt_newline(&i->buf);
+ prt_printf(&i->buf, "Transaction duration:");
+ prt_newline(&i->buf);
+
+ printbuf_indent_add(&i->buf, 2);
+ bch2_time_stats_to_text(&i->buf, &s->duration);
+ printbuf_indent_sub(&i->buf, 2);
+
if (IS_ENABLED(CONFIG_BCACHEFS_LOCK_TIME_STATS)) {
prt_printf(&i->buf, "Lock hold times:");
prt_newline(&i->buf);
return i->ret;
}
-static const struct file_operations lock_held_stats_op = {
- .owner = THIS_MODULE,
- .open = lock_held_stats_open,
- .release = lock_held_stats_release,
- .read = lock_held_stats_read,
+static const struct file_operations btree_transaction_stats_op = {
+ .owner = THIS_MODULE,
+ .open = btree_transaction_stats_open,
+ .release = btree_transaction_stats_release,
+ .read = btree_transaction_stats_read,
};
static ssize_t bch2_btree_deadlock_read(struct file *file, char __user *buf,
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)
+ struct task_struct *task = READ_ONCE(trans->locking_wait.task);
+
+ if (!task || 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;
+ i->iter = 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;
c->btree_debug, &journal_pins_ops);
debugfs_create_file("btree_transaction_stats", 0400, c->fs_debug_dir,
- c, &lock_held_stats_op);
+ c, &btree_transaction_stats_op);
debugfs_create_file("btree_deadlock", 0400, c->fs_debug_dir,
c->btree_debug, &btree_deadlock_ops);
bd < c->btree_debug + ARRAY_SIZE(c->btree_debug);
bd++) {
bd->id = bd - c->btree_debug;
- debugfs_create_file(bch2_btree_ids[bd->id],
+ debugfs_create_file(bch2_btree_id_str(bd->id),
0400, c->btree_debug_dir, bd,
&btree_debug_ops);
snprintf(name, sizeof(name), "%s-formats",
- bch2_btree_ids[bd->id]);
+ bch2_btree_id_str(bd->id));
debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
&btree_format_debug_ops);
snprintf(name, sizeof(name), "%s-bfloat-failed",
- bch2_btree_ids[bd->id]);
+ bch2_btree_id_str(bd->id));
debugfs_create_file(name, 0400, c->btree_debug_dir, bd,
&bfloat_failed_debug_ops);
int __init bch2_debug_init(void)
{
- int ret = 0;
-
bch_debug = debugfs_create_dir("bcachefs", NULL);
- return ret;
+ return 0;
}