+// SPDX-License-Identifier: GPL-2.0
/*
* Assorted bcachefs debug code
*
#include "btree_cache.h"
#include "btree_io.h"
#include "btree_iter.h"
+#include "btree_locking.h"
#include "btree_update.h"
#include "buckets.h"
#include "debug.h"
#include "extents.h"
#include "fsck.h"
#include "inode.h"
-#include "io.h"
#include "super.h"
#include <linux/console.h>
static struct dentry *bch_debug;
-#ifdef CONFIG_BCACHEFS_DEBUG
-
-void __bch2_btree_verify(struct bch_fs *c, struct btree *b)
+static bool bch2_btree_verify_replica(struct bch_fs *c, struct btree *b,
+ struct extent_ptr_decoded pick)
{
struct btree *v = c->verify_data;
- struct btree_node *n_ondisk, *n_sorted, *n_inmemory;
- struct bset *sorted, *inmemory;
- struct extent_ptr_decoded pick;
- struct bch_dev *ca;
+ struct btree_node *n_ondisk = c->verify_ondisk;
+ struct btree_node *n_sorted = c->verify_data->data;
+ 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, saw_error = false;
- if (c->opts.nochanges)
- return;
-
- btree_node_io_lock(b);
- mutex_lock(&c->verify_lock);
-
- n_ondisk = c->verify_ondisk;
- n_sorted = c->verify_data->data;
- n_inmemory = b->data;
-
- bkey_copy(&v->key, &b->key);
- v->written = 0;
- v->level = b->level;
- v->btree_id = b->btree_id;
- bch2_btree_keys_init(v, &c->expensive_debug_checks);
-
- if (bch2_bkey_pick_read_device(c, bkey_i_to_s_c(&b->key),
- NULL, &pick) <= 0)
- return;
-
- ca = bch_dev_bkey_exists(c, pick.ptr.dev);
if (!bch2_dev_get_ioref(ca, READ))
- return;
+ return false;
- bio = bio_alloc_bioset(GFP_NOIO,
- buf_pages(n_sorted, btree_bytes(c)),
- &c->btree_bio);
- bio_set_dev(bio, ca->disk_sb.bdev);
- bio->bi_opf = REQ_OP_READ|REQ_META;
+ bio = bio_alloc_bioset(ca->disk_sb.bdev,
+ buf_pages(n_sorted, btree_buf_bytes(b)),
+ REQ_OP_READ|REQ_META,
+ GFP_NOFS,
+ &c->btree_bio);
bio->bi_iter.bi_sector = pick.ptr.offset;
- bio->bi_iter.bi_size = btree_bytes(c);
- bch2_bio_map(bio, n_sorted);
+ 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));
- if (bch2_btree_node_read_done(c, v, false))
- goto out;
+ v->written = 0;
+ if (bch2_btree_node_read_done(c, ca, v, false, &saw_error) || saw_error)
+ return false;
n_sorted = c->verify_data->data;
sorted = &n_sorted->keys;
- inmemory = &n_inmemory->keys;
if (inmemory->u64s != sorted->u64s ||
memcmp(inmemory->start,
console_lock();
printk(KERN_ERR "*** in memory:\n");
- bch2_dump_bset(b, inmemory, 0);
+ bch2_dump_bset(c, b, inmemory, 0);
printk(KERN_ERR "*** read back in:\n");
- bch2_dump_bset(v, sorted, 0);
+ bch2_dump_bset(c, v, sorted, 0);
- while (offset < b->written) {
- if (!offset ) {
+ while (offset < v->written) {
+ if (!offset) {
i = &n_ondisk->keys;
sectors = vstruct_blocks(n_ondisk, c->block_bits) <<
c->block_bits;
}
printk(KERN_ERR "*** on disk block %u:\n", offset);
- bch2_dump_bset(b, i, offset);
+ bch2_dump_bset(c, b, i, offset);
offset += sectors;
}
- printk(KERN_ERR "*** block %u/%u not written\n",
- offset >> c->block_bits, btree_blocks(c));
-
for (j = 0; j < le16_to_cpu(inmemory->u64s); j++)
if (inmemory->_data[j] != sorted->_data[j])
break;
- printk(KERN_ERR "b->written %u\n", b->written);
-
console_unlock();
- panic("verify failed at %u\n", j);
+ bch_err(c, "verify failed at key %u", j);
+
+ failed = true;
+ }
+
+ if (v->written != b->written) {
+ bch_err(c, "written wrong: expected %u, got %u",
+ b->written, v->written);
+ failed = true;
+ }
+
+ return failed;
+}
+
+void __bch2_btree_verify(struct bch_fs *c, struct btree *b)
+{
+ struct bkey_ptrs_c ptrs;
+ struct extent_ptr_decoded p;
+ const union bch_extent_entry *entry;
+ struct btree *v;
+ struct bset *inmemory = &b->data->keys;
+ struct bkey_packed *k;
+ bool failed = false;
+
+ if (c->opts.nochanges)
+ return;
+
+ bch2_btree_node_io_lock(b);
+ mutex_lock(&c->verify_lock);
+
+ if (!c->verify_ondisk) {
+ c->verify_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL);
+ if (!c->verify_ondisk)
+ goto out;
+ }
+
+ if (!c->verify_data) {
+ c->verify_data = __bch2_btree_node_mem_alloc(c);
+ if (!c->verify_data)
+ goto out;
+
+ list_del_init(&c->verify_data->list);
+ }
+
+ BUG_ON(b->nsets != 1);
+
+ 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);
+ v->c.level = b->c.level;
+ v->c.btree_id = b->c.btree_id;
+ bch2_btree_keys_init(v);
+
+ ptrs = bch2_bkey_ptrs_c(bkey_i_to_s_c(&b->key));
+ bkey_for_each_ptr_decode(&b->key.k, ptrs, p, entry)
+ failed |= bch2_btree_verify_replica(c, b, p);
+
+ if (failed) {
+ struct printbuf buf = PRINTBUF;
+
+ bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
+ bch2_fs_fatal_error(c, "btree node verify failed for : %s\n", buf.buf);
+ printbuf_exit(&buf);
}
out:
mutex_unlock(&c->verify_lock);
- btree_node_io_unlock(b);
+ bch2_btree_node_io_unlock(b);
}
-#endif
+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 */
struct dump_iter {
- struct bpos from;
- struct bch_fs *c;
+ struct bch_fs *c;
enum btree_id id;
+ struct bpos from;
+ struct bpos prev_node;
+ u64 iter;
- char buf[PAGE_SIZE];
- size_t bytes; /* what's currently in buf */
+ struct printbuf buf;
char __user *ubuf; /* destination user buffer */
size_t size; /* size of requested read */
ssize_t ret; /* bytes read so far */
};
-static int flush_buf(struct dump_iter *i)
+static ssize_t flush_buf(struct dump_iter *i)
{
- if (i->bytes) {
- size_t bytes = min(i->bytes, i->size);
- int err = copy_to_user(i->ubuf, i->buf, bytes);
-
- if (err)
- return err;
-
- i->ret += bytes;
- i->ubuf += bytes;
- i->size -= bytes;
- i->bytes -= bytes;
- memmove(i->buf, i->buf + bytes, i->bytes);
+ if (i->buf.pos) {
+ size_t bytes = min_t(size_t, i->buf.pos, i->size);
+ int copied = bytes - copy_to_user(i->ubuf, i->buf.buf, bytes);
+
+ i->ret += copied;
+ i->ubuf += copied;
+ i->size -= copied;
+ i->buf.pos -= copied;
+ memmove(i->buf.buf, i->buf.buf + copied, i->buf.pos);
+
+ if (copied != bytes)
+ return -EFAULT;
}
- return 0;
+ return i->size ? 0 : i->ret;
}
static int bch2_dump_open(struct inode *inode, struct file *file)
file->private_data = i;
i->from = POS_MIN;
+ i->iter = 0;
i->c = container_of(bd, struct bch_fs, btree_debug[bd->id]);
i->id = bd->id;
+ i->buf = PRINTBUF;
return 0;
}
static int bch2_dump_release(struct inode *inode, struct file *file)
{
- kfree(file->private_data);
+ struct dump_iter *i = file->private_data;
+
+ printbuf_exit(&i->buf);
+ kfree(i);
return 0;
}
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;
- int err;
i->ubuf = buf;
i->size = size;
i->ret = 0;
- err = flush_buf(i);
- if (err)
- return err;
-
- if (!i->size)
- return i->ret;
-
- bch2_trans_init(&trans, i->c);
-
- iter = bch2_trans_get_iter(&trans, i->id, i->from, BTREE_ITER_PREFETCH);
- k = bch2_btree_iter_peek(iter);
-
- while (k.k && !(err = bkey_err(k))) {
- bch2_bkey_val_to_text(&PBUF(i->buf), i->c, k);
- i->bytes = strlen(i->buf);
- BUG_ON(i->bytes >= PAGE_SIZE);
- i->buf[i->bytes] = '\n';
- i->bytes++;
-
- k = bch2_btree_iter_next(iter);
- i->from = iter->pos;
-
- err = flush_buf(i);
- if (err)
- break;
-
- if (!i->size)
- break;
- }
- bch2_trans_exit(&trans);
-
- return err < 0 ? err : 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_iter *iter;
+ struct btree_trans *trans;
+ struct btree_iter iter;
struct btree *b;
- int err;
+ ssize_t ret;
i->ubuf = buf;
i->size = size;
i->ret = 0;
- err = flush_buf(i);
- if (err)
- return err;
+ ret = flush_buf(i);
+ if (ret)
+ return ret;
- if (!i->size || !bkey_cmp(POS_MAX, i->from))
+ if (bpos_eq(SPOS_MAX, i->from))
return i->ret;
- bch2_trans_init(&trans, i->c);
+ trans = bch2_trans_get(i->c);
+retry:
+ bch2_trans_begin(trans);
- for_each_btree_node(&trans, iter, i->id, i->from, 0, b) {
- bch2_btree_node_to_text(&PBUF(i->buf), i->c, b);
- i->bytes = strlen(i->buf);
- err = flush_buf(i);
- if (err)
- break;
-
- /*
- * can't easily correctly restart a btree node traversal across
- * all nodes, meh
- */
- i->from = bkey_cmp(POS_MAX, b->key.k.p)
- ? bkey_successor(b->key.k.p)
+ 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_eq(SPOS_MAX, b->key.k.p)
+ ? bpos_successor(b->key.k.p)
: b->key.k.p;
- if (!i->size)
+ ret = drop_locks_do(trans, flush_buf(i));
+ if (ret)
break;
}
- bch2_trans_exit(&trans);
+ bch2_trans_iter_exit(trans, &iter);
- return err < 0 ? err : i->ret;
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
+ goto retry;
+
+ bch2_trans_put(trans);
+
+ if (!ret)
+ ret = flush_buf(i);
+
+ return ret ?: i->ret;
}
static const struct file_operations btree_format_debug_ops = {
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;
- struct btree *prev_node = NULL;
- int err;
i->ubuf = buf;
i->size = size;
i->ret = 0;
- err = flush_buf(i);
- if (err)
- return err;
+ 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;
+}
- if (!i->size)
- return i->ret;
+static const struct file_operations bfloat_failed_debug_ops = {
+ .owner = THIS_MODULE,
+ .open = bch2_dump_open,
+ .release = bch2_dump_release,
+ .read = bch2_read_bfloat_failed,
+};
- bch2_trans_init(&trans, i->c);
+static void bch2_cached_btree_node_to_text(struct printbuf *out, struct bch_fs *c,
+ struct btree *b)
+{
+ if (!out->nr_tabstops)
+ printbuf_tabstop_push(out, 32);
+
+ prt_printf(out, "%px btree=%s l=%u ",
+ b,
+ bch2_btree_id_str(b->c.btree_id),
+ b->c.level);
+ prt_newline(out);
+
+ printbuf_indent_add(out, 2);
+
+ bch2_bkey_val_to_text(out, c, bkey_i_to_s_c(&b->key));
+ prt_newline(out);
+
+ prt_printf(out, "flags: ");
+ prt_tab(out);
+ prt_bitflags(out, bch2_btree_node_flags, b->flags);
+ prt_newline(out);
+
+ prt_printf(out, "pcpu read locks: ");
+ prt_tab(out);
+ prt_printf(out, "%u", b->c.lock.readers != NULL);
+ prt_newline(out);
+
+ prt_printf(out, "written:");
+ prt_tab(out);
+ prt_printf(out, "%u", b->written);
+ prt_newline(out);
+
+ prt_printf(out, "writes blocked:");
+ prt_tab(out);
+ prt_printf(out, "%u", !list_empty_careful(&b->write_blocked));
+ prt_newline(out);
+
+ prt_printf(out, "will make reachable:");
+ prt_tab(out);
+ prt_printf(out, "%lx", b->will_make_reachable);
+ prt_newline(out);
+
+ prt_printf(out, "journal pin %px:", &b->writes[0].journal);
+ prt_tab(out);
+ prt_printf(out, "%llu", b->writes[0].journal.seq);
+ prt_newline(out);
+
+ prt_printf(out, "journal pin %px:", &b->writes[1].journal);
+ prt_tab(out);
+ prt_printf(out, "%llu", b->writes[1].journal.seq);
+ prt_newline(out);
+
+ printbuf_indent_sub(out, 2);
+}
- iter = bch2_trans_get_iter(&trans, i->id, i->from, BTREE_ITER_PREFETCH);
+static ssize_t bch2_cached_btree_nodes_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;
+ bool done = false;
+ ssize_t ret = 0;
- while ((k = bch2_btree_iter_peek(iter)).k &&
- !(err = bkey_err(k))) {
- struct btree_iter_level *l = &iter->l[0];
- struct bkey_packed *_k =
- bch2_btree_node_iter_peek(&l->iter, l->b);
+ i->ubuf = buf;
+ i->size = size;
+ i->ret = 0;
- if (l->b != prev_node) {
- bch2_btree_node_to_text(&PBUF(i->buf), i->c, l->b);
- i->bytes = strlen(i->buf);
- err = flush_buf(i);
- if (err)
- break;
+ do {
+ struct bucket_table *tbl;
+ struct rhash_head *pos;
+ struct btree *b;
+
+ ret = flush_buf(i);
+ if (ret)
+ return ret;
+
+ rcu_read_lock();
+ i->buf.atomic++;
+ tbl = rht_dereference_rcu(c->btree_cache.table.tbl,
+ &c->btree_cache.table);
+ 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++;
+ } else {
+ done = true;
}
- prev_node = l->b;
+ --i->buf.atomic;
+ rcu_read_unlock();
+ } while (!done);
+
+ if (i->buf.allocation_failure)
+ ret = -ENOMEM;
- bch2_bfloat_to_text(&PBUF(i->buf), l->b, _k);
- i->bytes = strlen(i->buf);
+ if (!ret)
+ ret = flush_buf(i);
+
+ return ret ?: i->ret;
+}
+
+static const struct file_operations cached_btree_nodes_ops = {
+ .owner = THIS_MODULE,
+ .open = bch2_dump_open,
+ .release = bch2_dump_release,
+ .read = bch2_cached_btree_nodes_read,
+};
+
+static ssize_t bch2_btree_transactions_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;
+ struct btree_trans *trans;
+ ssize_t ret = 0;
+ u32 seq;
+
+ i->ubuf = buf;
+ i->size = size;
+ i->ret = 0;
+restart:
+ seqmutex_lock(&c->btree_trans_lock);
+ list_for_each_entry(trans, &c->btree_trans_list, list) {
+ 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) {
+ 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_task_backtrace(&i->buf, task, 0, GFP_KERNEL);
+ printbuf_indent_sub(&i->buf, 2);
+ prt_newline(&i->buf);
+
+ 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;
+
+ if (!ret)
+ ret = flush_buf(i);
+
+ return ret ?: i->ret;
+}
+
+static const struct file_operations btree_transactions_ops = {
+ .owner = THIS_MODULE,
+ .open = bch2_dump_open,
+ .release = bch2_dump_release,
+ .read = bch2_btree_transactions_read,
+};
+
+static ssize_t bch2_journal_pins_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;
+ bool done = false;
+ int err;
+
+ i->ubuf = buf;
+ i->size = size;
+ i->ret = 0;
+
+ do {
err = flush_buf(i);
if (err)
+ return err;
+
+ if (!i->size)
break;
- bch2_btree_iter_next(iter);
- i->from = iter->pos;
+ done = bch2_journal_seq_pins_to_text(&i->buf, &c->journal, &i->iter);
+ i->iter++;
+ } while (!done);
+
+ if (i->buf.allocation_failure)
+ return -ENOMEM;
+
+ return i->ret;
+}
+
+static const struct file_operations journal_pins_ops = {
+ .owner = THIS_MODULE,
+ .open = bch2_dump_open,
+ .release = bch2_dump_release,
+ .read = bch2_journal_pins_read,
+};
+
+static int btree_transaction_stats_open(struct inode *inode, struct file *file)
+{
+ struct bch_fs *c = inode->i_private;
+ struct dump_iter *i;
+
+ i = kzalloc(sizeof(struct dump_iter), GFP_KERNEL);
+
+ if (!i)
+ return -ENOMEM;
+
+ i->iter = 1;
+ i->c = c;
+ i->buf = PRINTBUF;
+ file->private_data = i;
+
+ return 0;
+}
+
+static int btree_transaction_stats_release(struct inode *inode, struct file *file)
+{
+ struct dump_iter *i = file->private_data;
+
+ printbuf_exit(&i->buf);
+ kfree(i);
+
+ return 0;
+}
+
+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;
+ int err;
+
+ i->ubuf = buf;
+ i->size = size;
+ i->ret = 0;
+
+ while (1) {
+ struct btree_transaction_stats *s = &c->btree_transaction_stats[i->iter];
err = flush_buf(i);
if (err)
- break;
+ return err;
if (!i->size)
break;
+
+ if (i->iter == ARRAY_SIZE(bch2_btree_transaction_fns) ||
+ !bch2_btree_transaction_fns[i->iter])
+ break;
+
+ prt_printf(&i->buf, "%s: ", bch2_btree_transaction_fns[i->iter]);
+ prt_newline(&i->buf);
+ printbuf_indent_add(&i->buf, 2);
+
+ mutex_lock(&s->lock);
+
+ 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);
+
+ printbuf_indent_add(&i->buf, 2);
+ bch2_time_stats_to_text(&i->buf, &s->lock_hold_times);
+ printbuf_indent_sub(&i->buf, 2);
+ }
+
+ if (s->max_paths_text) {
+ prt_printf(&i->buf, "Maximum allocated btree paths (%u):", s->nr_max_paths);
+ prt_newline(&i->buf);
+
+ printbuf_indent_add(&i->buf, 2);
+ prt_str_indented(&i->buf, s->max_paths_text);
+ printbuf_indent_sub(&i->buf, 2);
+ }
+
+ mutex_unlock(&s->lock);
+
+ printbuf_indent_sub(&i->buf, 2);
+ prt_newline(&i->buf);
+ i->iter++;
+ }
+
+ if (i->buf.allocation_failure)
+ return -ENOMEM;
+
+ return i->ret;
+}
+
+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,
+ size_t size, loff_t *ppos)
+{
+ struct dump_iter *i = file->private_data;
+ 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;
+
+ if (i->iter)
+ goto out;
+restart:
+ seqmutex_lock(&c->btree_trans_lock);
+ list_for_each_entry(trans, &c->btree_trans_list, list) {
+ 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) {
+ closure_put(&trans->ref);
+ goto out;
+ }
+
+ bch2_check_for_deadlock(trans, &i->buf);
+
+ i->iter = task->pid;
+
+ closure_put(&trans->ref);
+
+ if (!seqmutex_relock(&c->btree_trans_lock, seq))
+ goto restart;
}
- bch2_trans_exit(&trans);
+ seqmutex_unlock(&c->btree_trans_lock);
+out:
+ if (i->buf.allocation_failure)
+ ret = -ENOMEM;
- return err < 0 ? err : i->ret;
+ if (!ret)
+ ret = flush_buf(i);
+
+ return ret ?: i->ret;
}
-static const struct file_operations bfloat_failed_debug_ops = {
+static const struct file_operations btree_deadlock_ops = {
.owner = THIS_MODULE,
.open = bch2_dump_open,
.release = bch2_dump_release,
- .read = bch2_read_bfloat_failed,
+ .read = bch2_btree_deadlock_read,
};
void bch2_fs_debug_exit(struct bch_fs *c)
{
- if (!IS_ERR_OR_NULL(c->debug))
- debugfs_remove_recursive(c->debug);
+ if (!IS_ERR_OR_NULL(c->fs_debug_dir))
+ debugfs_remove_recursive(c->fs_debug_dir);
}
void bch2_fs_debug_init(struct bch_fs *c)
return;
snprintf(name, sizeof(name), "%pU", c->sb.user_uuid.b);
- c->debug = debugfs_create_dir(name, bch_debug);
- if (IS_ERR_OR_NULL(c->debug))
+ c->fs_debug_dir = debugfs_create_dir(name, bch_debug);
+ if (IS_ERR_OR_NULL(c->fs_debug_dir))
+ return;
+
+ debugfs_create_file("cached_btree_nodes", 0400, c->fs_debug_dir,
+ c->btree_debug, &cached_btree_nodes_ops);
+
+ debugfs_create_file("btree_transactions", 0400, c->fs_debug_dir,
+ c->btree_debug, &btree_transactions_ops);
+
+ debugfs_create_file("journal_pins", 0400, c->fs_debug_dir,
+ c->btree_debug, &journal_pins_ops);
+
+ debugfs_create_file("btree_transaction_stats", 0400, c->fs_debug_dir,
+ c, &btree_transaction_stats_op);
+
+ debugfs_create_file("btree_deadlock", 0400, c->fs_debug_dir,
+ c->btree_debug, &btree_deadlock_ops);
+
+ c->btree_debug_dir = debugfs_create_dir("btrees", c->fs_debug_dir);
+ if (IS_ERR_OR_NULL(c->btree_debug_dir))
return;
for (bd = c->btree_debug;
bd < c->btree_debug + ARRAY_SIZE(c->btree_debug);
bd++) {
bd->id = bd - c->btree_debug;
- bd->btree = debugfs_create_file(bch2_btree_ids[bd->id],
- 0400, c->debug, bd,
- &btree_debug_ops);
+ 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));
- bd->btree_format = debugfs_create_file(name, 0400, c->debug, bd,
- &btree_format_debug_ops);
+ 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));
- bd->failed = debugfs_create_file(name, 0400, c->debug, bd,
- &bfloat_failed_debug_ops);
+ 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;
}