#include "bbpos.h"
#include "alloc_background.h"
#include "backpointers.h"
+#include "bkey_buf.h"
#include "btree_cache.h"
#include "btree_update.h"
+#include "btree_update_interior.h"
#include "btree_write_buffer.h"
#include "error.h"
return false;
}
-int bch2_backpointer_invalid(const struct bch_fs *c, struct bkey_s_c k,
- unsigned flags, struct printbuf *err)
+int bch2_backpointer_invalid(struct bch_fs *c, struct bkey_s_c k,
+ enum bkey_invalid_flags flags,
+ struct printbuf *err)
{
struct bkey_s_c_backpointer bp = bkey_s_c_to_backpointer(k);
struct bpos bucket = bp_pos_to_bucket(c, bp.k->p);
+ int ret = 0;
- if (bkey_val_bytes(bp.k) < sizeof(*bp.v)) {
- prt_str(err, "incorrect value size");
- return -BCH_ERR_invalid_bkey;
- }
-
- if (!bpos_eq(bp.k->p, bucket_pos_to_bp(c, bucket, bp.v->bucket_offset))) {
- prt_str(err, "backpointer at wrong pos");
- return -BCH_ERR_invalid_bkey;
- }
-
- return 0;
+ bkey_fsck_err_on(!bpos_eq(bp.k->p, bucket_pos_to_bp(c, bucket, bp.v->bucket_offset)),
+ c, err,
+ backpointer_pos_wrong,
+ "backpointer at wrong pos");
+fsck_err:
+ return ret;
}
void bch2_backpointer_to_text(struct printbuf *out, const struct bch_backpointer *bp)
{
prt_printf(out, "btree=%s l=%u offset=%llu:%u len=%u pos=",
- bch2_btree_ids[bp->btree_id],
+ bch2_btree_id_str(bp->btree_id),
bp->level,
(u64) (bp->bucket_offset >> MAX_EXTENT_COMPRESS_RATIO_SHIFT),
(u32) bp->bucket_offset & ~(~0U << MAX_EXTENT_COMPRESS_RATIO_SHIFT),
void bch2_backpointer_k_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c k)
{
+ prt_str(out, "bucket=");
+ bch2_bpos_to_text(out, bp_pos_to_bucket(c, k.k->p));
+ prt_str(out, " ");
+
bch2_backpointer_to_text(out, bkey_s_c_to_backpointer(k).v);
}
{
struct bkey_s_backpointer bp = bkey_s_to_backpointer(k);
- bp.v->bucket_offset = swab32(bp.v->bucket_offset);
+ bp.v->bucket_offset = swab40(bp.v->bucket_offset);
bp.v->bucket_len = swab32(bp.v->bucket_len);
bch2_bpos_swab(&bp.v->pos);
}
-#define BACKPOINTER_OFFSET_MAX ((1ULL << 40) - 1)
-
-static inline int backpointer_cmp(struct bch_backpointer l, struct bch_backpointer r)
-{
- return cmp_int(l.bucket_offset, r.bucket_offset);
-}
-
-static int bch2_backpointer_del_by_offset(struct btree_trans *trans,
- struct bpos bucket,
- u64 bp_offset,
- struct bch_backpointer bp)
-{
- struct bch_fs *c = trans->c;
- struct btree_iter iter;
- struct bkey_s_c k;
- int ret;
-
- if (bp_offset < BACKPOINTER_OFFSET_MAX) {
- struct bch_backpointer *bps;
- struct bkey_i_alloc_v4 *a;
- unsigned i, nr;
-
- bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
- bucket,
- BTREE_ITER_INTENT|
- BTREE_ITER_SLOTS|
- 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_alloc_v4) {
- ret = -ENOENT;
- goto err;
- }
-
- a = bch2_alloc_to_v4_mut(trans, k);
- ret = PTR_ERR_OR_ZERO(a);
- if (ret)
- goto err;
- bps = alloc_v4_backpointers(&a->v);
- nr = BCH_ALLOC_V4_NR_BACKPOINTERS(&a->v);
-
- for (i = 0; i < nr; i++) {
- if (bps[i].bucket_offset == bp_offset)
- goto found;
- if (bps[i].bucket_offset > bp_offset)
- break;
- }
-
- ret = -ENOENT;
- goto err;
-found:
- if (memcmp(&bps[i], &bp, sizeof(bp))) {
- ret = -ENOENT;
- goto err;
- }
- array_remove_item(bps, nr, i);
- SET_BCH_ALLOC_V4_NR_BACKPOINTERS(&a->v, nr);
- set_alloc_v4_u64s(a);
- ret = bch2_trans_update(trans, &iter, &a->k_i, 0);
- } else {
- bp_offset -= BACKPOINTER_OFFSET_MAX;
-
- bch2_trans_iter_init(trans, &iter, BTREE_ID_backpointers,
- bucket_pos_to_bp(c, bucket, bp_offset),
- BTREE_ITER_INTENT|
- BTREE_ITER_SLOTS|
- 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_backpointer ||
- memcmp(bkey_s_c_to_backpointer(k).v, &bp, sizeof(bp))) {
- ret = -ENOENT;
- goto err;
- }
-
- ret = bch2_btree_delete_at(trans, &iter, 0);
- }
-err:
- bch2_trans_iter_exit(trans, &iter);
- return ret;
-}
-
-bool bch2_bucket_backpointer_del(struct btree_trans *trans,
- struct bkey_i_alloc_v4 *a,
- struct bch_backpointer bp)
-{
- struct bch_backpointer *bps = alloc_v4_backpointers(&a->v);
- unsigned i, nr = BCH_ALLOC_V4_NR_BACKPOINTERS(&a->v);
-
- for (i = 0; i < nr; i++) {
- int cmp = backpointer_cmp(bps[i], bp) ?:
- memcmp(&bps[i], &bp, sizeof(bp));
- if (!cmp) {
- array_remove_item(bps, nr, i);
- SET_BCH_ALLOC_V4_NR_BACKPOINTERS(&a->v, nr);
- set_alloc_v4_u64s(a);
- return true;
- }
- if (cmp >= 0)
- break;
- }
-
- return false;
-}
-
static noinline int backpointer_mod_err(struct btree_trans *trans,
struct bch_backpointer bp,
struct bkey_s_c bp_k,
bch2_bkey_val_to_text(&buf, c, orig_k);
bch_err(c, "%s", buf.buf);
- } else if (test_bit(BCH_FS_CHECK_BACKPOINTERS_DONE, &c->flags)) {
+ } else if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_extents_to_backpointers) {
prt_printf(&buf, "backpointer not found when deleting");
prt_newline(&buf);
printbuf_indent_add(&buf, 2);
printbuf_exit(&buf);
- if (test_bit(BCH_FS_CHECK_BACKPOINTERS_DONE, &c->flags)) {
- bch2_inconsistent_error(c);
- return -EIO;
+ if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_extents_to_backpointers) {
+ return bch2_inconsistent_error(c) ? BCH_ERR_erofs_unfixed_errors : 0;
} else {
return 0;
}
}
int bch2_bucket_backpointer_mod_nowritebuffer(struct btree_trans *trans,
- struct bkey_i_alloc_v4 *a,
+ struct bpos bucket,
struct bch_backpointer bp,
struct bkey_s_c orig_k,
bool insert)
{
- struct bch_fs *c = trans->c;
- struct bkey_i_backpointer *bp_k;
struct btree_iter bp_iter;
struct bkey_s_c k;
+ struct bkey_i_backpointer *bp_k;
int ret;
bp_k = bch2_trans_kmalloc_nomemzero(trans, sizeof(struct bkey_i_backpointer));
return ret;
bkey_backpointer_init(&bp_k->k_i);
- bp_k->k.p = bucket_pos_to_bp(c, a->k.p, bp.bucket_offset);
+ bp_k->k.p = bucket_pos_to_bp(trans->c, bucket, bp.bucket_offset);
bp_k->v = bp;
if (!insert) {
set_bkey_val_u64s(&bp_k->k, 0);
}
- bch2_trans_iter_init(trans, &bp_iter, BTREE_ID_backpointers,
- bucket_pos_to_bp(c, a->k.p, bp.bucket_offset),
- BTREE_ITER_INTENT|
- BTREE_ITER_SLOTS|
- BTREE_ITER_WITH_UPDATES);
- k = bch2_btree_iter_peek_slot(&bp_iter);
+ k = bch2_bkey_get_iter(trans, &bp_iter, BTREE_ID_backpointers,
+ bp_k->k.p,
+ BTREE_ITER_INTENT|
+ BTREE_ITER_SLOTS|
+ BTREE_ITER_WITH_UPDATES);
ret = bkey_err(k);
if (ret)
goto err;
*/
int bch2_get_next_backpointer(struct btree_trans *trans,
struct bpos bucket, int gen,
- u64 *bp_offset,
- struct bch_backpointer *dst,
+ struct bpos *bp_pos,
+ struct bch_backpointer *bp,
unsigned iter_flags)
{
struct bch_fs *c = trans->c;
- struct bpos bp_pos, bp_end_pos;
- struct btree_iter alloc_iter, bp_iter = { NULL };
+ struct bpos bp_end_pos = bucket_pos_to_bp(c, bpos_nosnap_successor(bucket), 0);
+ struct btree_iter alloc_iter = { NULL }, bp_iter = { NULL };
struct bkey_s_c k;
- struct bkey_s_c_alloc_v4 a;
- size_t i;
- int ret;
-
- if (*bp_offset == U64_MAX)
- return 0;
-
- bp_pos = bucket_pos_to_bp(c, bucket,
- max(*bp_offset, BACKPOINTER_OFFSET_MAX) - BACKPOINTER_OFFSET_MAX);
- bp_end_pos = bucket_pos_to_bp(c, bpos_nosnap_successor(bucket), 0);
-
- bch2_trans_iter_init(trans, &alloc_iter, BTREE_ID_alloc,
- bucket, BTREE_ITER_CACHED);
- k = bch2_btree_iter_peek_slot(&alloc_iter);
- ret = bkey_err(k);
- if (ret)
- goto out;
-
- if (k.k->type != KEY_TYPE_alloc_v4)
- goto done;
+ int ret = 0;
- a = bkey_s_c_to_alloc_v4(k);
- if (gen >= 0 && a.v->gen != gen)
+ if (bpos_ge(*bp_pos, bp_end_pos))
goto done;
- for (i = 0; i < BCH_ALLOC_V4_NR_BACKPOINTERS(a.v); i++) {
- if (alloc_v4_backpointers_c(a.v)[i].bucket_offset < *bp_offset)
- continue;
+ if (gen >= 0) {
+ k = bch2_bkey_get_iter(trans, &alloc_iter, BTREE_ID_alloc,
+ bucket, BTREE_ITER_CACHED|iter_flags);
+ ret = bkey_err(k);
+ if (ret)
+ goto out;
- *dst = alloc_v4_backpointers_c(a.v)[i];
- *bp_offset = dst->bucket_offset;
- goto out;
+ if (k.k->type != KEY_TYPE_alloc_v4 ||
+ bkey_s_c_to_alloc_v4(k).v->gen != gen)
+ goto done;
}
+ *bp_pos = bpos_max(*bp_pos, bucket_pos_to_bp(c, bucket, 0));
+
for_each_btree_key_norestart(trans, bp_iter, BTREE_ID_backpointers,
- bp_pos, 0, k, ret) {
+ *bp_pos, iter_flags, k, ret) {
if (bpos_ge(k.k->p, bp_end_pos))
break;
- if (k.k->type != KEY_TYPE_backpointer)
- continue;
-
- *dst = *bkey_s_c_to_backpointer(k).v;
- *bp_offset = dst->bucket_offset + BACKPOINTER_OFFSET_MAX;
+ *bp_pos = k.k->p;
+ *bp = *bkey_s_c_to_backpointer(k).v;
goto out;
}
done:
- *bp_offset = U64_MAX;
+ *bp_pos = SPOS_MAX;
out:
bch2_trans_iter_exit(trans, &bp_iter);
bch2_trans_iter_exit(trans, &alloc_iter);
}
static void backpointer_not_found(struct btree_trans *trans,
- struct bpos bucket,
- u64 bp_offset,
+ struct bpos bp_pos,
struct bch_backpointer bp,
- struct bkey_s_c k,
- const char *thing_it_points_to)
+ struct bkey_s_c k)
{
struct bch_fs *c = trans->c;
struct printbuf buf = PRINTBUF;
+ struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
+ /*
+ * If we're using the btree write buffer, the backpointer we were
+ * looking at may have already been deleted - failure to find what it
+ * pointed to is not an error:
+ */
if (likely(!bch2_backpointers_no_use_write_buffer))
return;
prt_printf(&buf, "backpointer doesn't match %s it points to:\n ",
- thing_it_points_to);
+ bp.level ? "btree node" : "extent");
prt_printf(&buf, "bucket: ");
bch2_bpos_to_text(&buf, bucket);
prt_printf(&buf, "\n ");
- if (bp_offset >= BACKPOINTER_OFFSET_MAX) {
- struct bpos bp_pos =
- bucket_pos_to_bp(c, bucket,
- bp_offset - BACKPOINTER_OFFSET_MAX);
- prt_printf(&buf, "backpointer pos: ");
- bch2_bpos_to_text(&buf, bp_pos);
- prt_printf(&buf, "\n ");
- }
+ prt_printf(&buf, "backpointer pos: ");
+ bch2_bpos_to_text(&buf, bp_pos);
+ prt_printf(&buf, "\n ");
bch2_backpointer_to_text(&buf, &bp);
prt_printf(&buf, "\n ");
bch2_bkey_val_to_text(&buf, c, k);
- if (!test_bit(BCH_FS_CHECK_BACKPOINTERS_DONE, &c->flags))
+ if (c->curr_recovery_pass >= BCH_RECOVERY_PASS_check_extents_to_backpointers)
bch_err_ratelimited(c, "%s", buf.buf);
else
bch2_trans_inconsistent(trans, "%s", buf.buf);
struct bkey_s_c bch2_backpointer_get_key(struct btree_trans *trans,
struct btree_iter *iter,
- struct bpos bucket,
- u64 bp_offset,
- struct bch_backpointer bp)
+ struct bpos bp_pos,
+ struct bch_backpointer bp,
+ unsigned iter_flags)
{
- struct bch_fs *c = trans->c;
- struct bkey_s_c k;
-
- bch2_trans_node_iter_init(trans, iter,
- bp.btree_id,
- bp.pos,
- 0,
- min(bp.level, c->btree_roots[bp.btree_id].level),
- 0);
- k = bch2_btree_iter_peek_slot(iter);
- if (bkey_err(k)) {
- bch2_trans_iter_exit(trans, iter);
- return k;
- }
-
- if (bp.level == c->btree_roots[bp.btree_id].level + 1)
- k = bkey_i_to_s_c(&c->btree_roots[bp.btree_id].key);
-
- if (k.k && extent_matches_bp(c, bp.btree_id, bp.level, k, bucket, bp))
- return k;
-
- bch2_trans_iter_exit(trans, iter);
+ if (likely(!bp.level)) {
+ struct bch_fs *c = trans->c;
+ struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
+ struct bkey_s_c k;
+
+ bch2_trans_node_iter_init(trans, iter,
+ bp.btree_id,
+ bp.pos,
+ 0, 0,
+ iter_flags);
+ k = bch2_btree_iter_peek_slot(iter);
+ if (bkey_err(k)) {
+ bch2_trans_iter_exit(trans, iter);
+ return k;
+ }
- if (unlikely(bch2_backpointers_no_use_write_buffer)) {
- if (bp.level) {
- struct btree *b;
+ if (k.k && extent_matches_bp(c, bp.btree_id, bp.level, k, bucket, bp))
+ return k;
- /*
- * If a backpointer for a btree node wasn't found, it may be
- * because it was overwritten by a new btree node that hasn't
- * been written out yet - backpointer_get_node() checks for
- * this:
- */
- b = bch2_backpointer_get_node(trans, iter, bucket, bp_offset, bp);
- if (!IS_ERR_OR_NULL(b))
- return bkey_i_to_s_c(&b->key);
+ bch2_trans_iter_exit(trans, iter);
+ backpointer_not_found(trans, bp_pos, bp, k);
+ return bkey_s_c_null;
+ } else {
+ struct btree *b = bch2_backpointer_get_node(trans, iter, bp_pos, bp);
+ if (IS_ERR_OR_NULL(b)) {
bch2_trans_iter_exit(trans, iter);
-
- if (IS_ERR(b))
- return bkey_s_c_err(PTR_ERR(b));
- return bkey_s_c_null;
+ return IS_ERR(b) ? bkey_s_c_err(PTR_ERR(b)) : bkey_s_c_null;
}
-
- backpointer_not_found(trans, bucket, bp_offset, bp, k, "extent");
+ return bkey_i_to_s_c(&b->key);
}
-
- return bkey_s_c_null;
}
struct btree *bch2_backpointer_get_node(struct btree_trans *trans,
struct btree_iter *iter,
- struct bpos bucket,
- u64 bp_offset,
+ struct bpos bp_pos,
struct bch_backpointer bp)
{
struct bch_fs *c = trans->c;
+ struct bpos bucket = bp_pos_to_bucket(c, bp_pos);
struct btree *b;
BUG_ON(!bp.level);
bp.level - 1,
0);
b = bch2_btree_iter_peek_node(iter);
- if (IS_ERR(b))
+ if (IS_ERR_OR_NULL(b))
goto err;
- if (b && extent_matches_bp(c, bp.btree_id, bp.level,
- bkey_i_to_s_c(&b->key),
- bucket, bp))
+ BUG_ON(b->c.level != bp.level - 1);
+
+ if (extent_matches_bp(c, bp.btree_id, bp.level,
+ bkey_i_to_s_c(&b->key),
+ bucket, bp))
return b;
- if (b && btree_node_will_make_reachable(b)) {
+ if (btree_node_will_make_reachable(b)) {
b = ERR_PTR(-BCH_ERR_backpointer_to_overwritten_btree_node);
} else {
- backpointer_not_found(trans, bucket, bp_offset, bp,
- bkey_i_to_s_c(&b->key), "btree node");
+ backpointer_not_found(trans, bp_pos, bp, bkey_i_to_s_c(&b->key));
b = NULL;
}
err:
{
struct bch_fs *c = trans->c;
struct btree_iter alloc_iter = { NULL };
- struct bch_dev *ca;
struct bkey_s_c alloc_k;
struct printbuf buf = PRINTBUF;
int ret = 0;
if (fsck_err_on(!bch2_dev_exists2(c, k.k->p.inode), c,
- "backpointer for mising device:\n%s",
+ backpointer_to_missing_device,
+ "backpointer for missing device:\n%s",
(bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
ret = bch2_btree_delete_at(trans, bp_iter, 0);
goto out;
}
- ca = bch_dev_bkey_exists(c, k.k->p.inode);
-
- bch2_trans_iter_init(trans, &alloc_iter, BTREE_ID_alloc,
- bp_pos_to_bucket(c, k.k->p), 0);
-
- alloc_k = bch2_btree_iter_peek_slot(&alloc_iter);
+ alloc_k = bch2_bkey_get_iter(trans, &alloc_iter, BTREE_ID_alloc,
+ bp_pos_to_bucket(c, k.k->p), 0);
ret = bkey_err(alloc_k);
if (ret)
goto out;
if (fsck_err_on(alloc_k.k->type != KEY_TYPE_alloc_v4, c,
+ backpointer_to_missing_alloc,
"backpointer for nonexistent alloc key: %llu:%llu:0\n%s",
alloc_iter.pos.inode, alloc_iter.pos.offset,
(bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
/* verify that every backpointer has a corresponding alloc key */
int bch2_check_btree_backpointers(struct bch_fs *c)
{
- struct btree_iter iter;
- struct bkey_s_c k;
-
- return bch2_trans_run(c,
- for_each_btree_key_commit(&trans, iter,
+ int ret = bch2_trans_run(c,
+ for_each_btree_key_commit(trans, iter,
BTREE_ID_backpointers, POS_MIN, 0, k,
- NULL, NULL, BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL,
- bch2_check_btree_backpointer(&trans, &iter, k)));
+ NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+ bch2_check_btree_backpointer(trans, &iter, k)));
+ bch_err_fn(c, ret);
+ return ret;
+}
+
+static inline bool bkey_and_val_eq(struct bkey_s_c l, struct bkey_s_c r)
+{
+ return bpos_eq(l.k->p, r.k->p) &&
+ bkey_bytes(l.k) == bkey_bytes(r.k) &&
+ !memcmp(l.v, r.v, bkey_val_bytes(l.k));
}
+struct extents_to_bp_state {
+ struct bpos bucket_start;
+ struct bpos bucket_end;
+ struct bkey_buf last_flushed;
+};
+
static int check_bp_exists(struct btree_trans *trans,
- struct bpos bucket_pos,
+ struct extents_to_bp_state *s,
+ struct bpos bucket,
struct bch_backpointer bp,
- struct bkey_s_c orig_k,
- struct bpos bucket_start,
- struct bpos bucket_end,
- struct bpos *last_flushed_pos)
+ struct bkey_s_c orig_k)
{
struct bch_fs *c = trans->c;
- struct btree_iter alloc_iter, bp_iter = { NULL };
+ struct btree_iter bp_iter = { NULL };
struct printbuf buf = PRINTBUF;
- struct bkey_s_c alloc_k, bp_k;
+ struct bkey_s_c bp_k;
+ struct bkey_buf tmp;
int ret;
- if (bpos_lt(bucket_pos, bucket_start) ||
- bpos_gt(bucket_pos, bucket_end))
- return 0;
+ bch2_bkey_buf_init(&tmp);
- bch2_trans_iter_init(trans, &alloc_iter, BTREE_ID_alloc, bucket_pos, 0);
- alloc_k = bch2_btree_iter_peek_slot(&alloc_iter);
- ret = bkey_err(alloc_k);
- if (ret)
- goto err;
+ if (bpos_lt(bucket, s->bucket_start) ||
+ bpos_gt(bucket, s->bucket_end))
+ return 0;
- if (alloc_k.k->type == KEY_TYPE_alloc_v4) {
- struct bkey_s_c_alloc_v4 a = bkey_s_c_to_alloc_v4(alloc_k);
- const struct bch_backpointer *bps = alloc_v4_backpointers_c(a.v);
- unsigned i, nr = BCH_ALLOC_V4_NR_BACKPOINTERS(a.v);
-
- for (i = 0; i < nr; i++) {
- int cmp = backpointer_cmp(bps[i], bp) ?:
- memcmp(&bps[i], &bp, sizeof(bp));
- if (!cmp)
- goto out;
- if (cmp >= 0)
- break;
- }
- } else {
+ if (!bch2_dev_bucket_exists(c, bucket))
goto missing;
- }
- bch2_trans_iter_init(trans, &bp_iter, BTREE_ID_backpointers,
- bucket_pos_to_bp(c, bucket_pos, bp.bucket_offset),
- 0);
- bp_k = bch2_btree_iter_peek_slot(&bp_iter);
+ bp_k = bch2_bkey_get_iter(trans, &bp_iter, BTREE_ID_backpointers,
+ bucket_pos_to_bp(c, bucket, bp.bucket_offset),
+ 0);
ret = bkey_err(bp_k);
if (ret)
goto err;
if (bp_k.k->type != KEY_TYPE_backpointer ||
memcmp(bkey_s_c_to_backpointer(bp_k).v, &bp, sizeof(bp))) {
- if (!bpos_eq(*last_flushed_pos, orig_k.k->p)) {
- *last_flushed_pos = orig_k.k->p;
- ret = bch2_btree_write_buffer_flush_sync(trans) ?:
- -BCH_ERR_transaction_restart_write_buffer_flush;
+ bch2_bkey_buf_reassemble(&tmp, c, orig_k);
+
+ if (!bkey_and_val_eq(orig_k, bkey_i_to_s_c(s->last_flushed.k))) {
+ if (bp.level) {
+ bch2_trans_unlock(trans);
+ bch2_btree_interior_updates_flush(c);
+ }
+
+ ret = bch2_btree_write_buffer_flush_sync(trans);
+ if (ret)
+ goto err;
+
+ bch2_bkey_buf_copy(&s->last_flushed, c, tmp.k);
+ ret = -BCH_ERR_transaction_restart_write_buffer_flush;
goto out;
}
goto missing;
err:
fsck_err:
bch2_trans_iter_exit(trans, &bp_iter);
- bch2_trans_iter_exit(trans, &alloc_iter);
+ bch2_bkey_buf_exit(&tmp, c);
printbuf_exit(&buf);
return ret;
missing:
-
prt_printf(&buf, "missing backpointer for btree=%s l=%u ",
- bch2_btree_ids[bp.btree_id], bp.level);
+ bch2_btree_id_str(bp.btree_id), bp.level);
bch2_bkey_val_to_text(&buf, c, orig_k);
prt_printf(&buf, "\nbp pos ");
bch2_bpos_to_text(&buf, bp_iter.pos);
- if (c->sb.version < bcachefs_metadata_version_backpointers ||
- c->opts.reconstruct_alloc ||
- fsck_err(c, "%s", buf.buf)) {
- struct bkey_i_alloc_v4 *a = bch2_alloc_to_v4_mut(trans, alloc_k);
-
- ret = PTR_ERR_OR_ZERO(a) ?:
- bch2_bucket_backpointer_mod(trans, a, bp, orig_k, true);
- }
+ if (c->opts.reconstruct_alloc ||
+ fsck_err(c, ptr_to_missing_backpointer, "%s", buf.buf))
+ ret = bch2_bucket_backpointer_mod(trans, bucket, bp, orig_k, true);
goto out;
}
static int check_extent_to_backpointers(struct btree_trans *trans,
- struct btree_iter *iter,
- struct bpos bucket_start,
- struct bpos bucket_end,
- struct bpos *last_flushed_pos)
+ struct extents_to_bp_state *s,
+ enum btree_id btree, unsigned level,
+ struct bkey_s_c k)
{
struct bch_fs *c = trans->c;
struct bkey_ptrs_c ptrs;
const union bch_extent_entry *entry;
struct extent_ptr_decoded p;
- struct bkey_s_c k;
int ret;
- k = bch2_btree_iter_peek_all_levels(iter);
- ret = bkey_err(k);
- if (ret)
- return ret;
- if (!k.k)
- return 0;
-
ptrs = bch2_bkey_ptrs_c(k);
bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
struct bpos bucket_pos;
if (p.ptr.cached)
continue;
- bch2_extent_ptr_to_bp(c, iter->btree_id, iter->path->level,
+ bch2_extent_ptr_to_bp(c, btree, level,
k, p, &bucket_pos, &bp);
- ret = check_bp_exists(trans, bucket_pos, bp, k,
- bucket_start, bucket_end,
- last_flushed_pos);
+ ret = check_bp_exists(trans, s, bucket_pos, bp, k);
if (ret)
return ret;
}
}
static int check_btree_root_to_backpointers(struct btree_trans *trans,
+ struct extents_to_bp_state *s,
enum btree_id btree_id,
- struct bpos bucket_start,
- struct bpos bucket_end,
- struct bpos *last_flushed_pos)
+ int *level)
{
struct bch_fs *c = trans->c;
struct btree_iter iter;
struct btree *b;
struct bkey_s_c k;
- struct bkey_ptrs_c ptrs;
- struct extent_ptr_decoded p;
- const union bch_extent_entry *entry;
int ret;
-
- bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN, 0,
- c->btree_roots[btree_id].level, 0);
+retry:
+ bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN,
+ 0, bch2_btree_id_root(c, btree_id)->b->c.level, 0);
b = bch2_btree_iter_peek_node(&iter);
ret = PTR_ERR_OR_ZERO(b);
if (ret)
goto err;
- BUG_ON(b != btree_node_root(c, b));
-
- k = bkey_i_to_s_c(&b->key);
- ptrs = bch2_bkey_ptrs_c(k);
- bkey_for_each_ptr_decode(k.k, ptrs, p, entry) {
- struct bpos bucket_pos;
- struct bch_backpointer bp;
-
- if (p.ptr.cached)
- continue;
+ if (b != btree_node_root(c, b)) {
+ bch2_trans_iter_exit(trans, &iter);
+ goto retry;
+ }
- bch2_extent_ptr_to_bp(c, iter.btree_id, iter.path->level + 1,
- k, p, &bucket_pos, &bp);
+ *level = b->c.level;
- ret = check_bp_exists(trans, bucket_pos, bp, k,
- bucket_start, bucket_end,
- last_flushed_pos);
- if (ret)
- goto err;
- }
+ k = bkey_i_to_s_c(&b->key);
+ ret = check_extent_to_backpointers(trans, s, btree_id, b->c.level + 1, k);
err:
bch2_trans_iter_exit(trans, &iter);
return ret;
};
}
-static size_t btree_nodes_fit_in_ram(struct bch_fs *c)
+static u64 mem_may_pin_bytes(struct bch_fs *c)
{
struct sysinfo i;
- u64 mem_bytes;
-
si_meminfo(&i);
- mem_bytes = i.totalram * i.mem_unit;
- return div_u64(mem_bytes >> 1, btree_bytes(c));
+
+ u64 mem_bytes = i.totalram * i.mem_unit;
+ return div_u64(mem_bytes * c->opts.fsck_memory_usage_percent, 100);
}
-int bch2_get_btree_in_memory_pos(struct btree_trans *trans,
- unsigned btree_leaf_mask,
- unsigned btree_interior_mask,
- struct bbpos start, struct bbpos *end)
+static size_t btree_nodes_fit_in_ram(struct bch_fs *c)
{
- struct btree_iter iter;
- struct bkey_s_c k;
- size_t btree_nodes = btree_nodes_fit_in_ram(trans->c);
- enum btree_id btree;
+ return div_u64(mem_may_pin_bytes(c), c->opts.btree_node_size);
+}
+
+static int bch2_get_btree_in_memory_pos(struct btree_trans *trans,
+ u64 btree_leaf_mask,
+ u64 btree_interior_mask,
+ struct bbpos start, struct bbpos *end)
+{
+ struct bch_fs *c = trans->c;
+ s64 mem_may_pin = mem_may_pin_bytes(c);
int ret = 0;
- for (btree = start.btree; btree < BTREE_ID_NR && !ret; btree++) {
- unsigned depth = ((1U << btree) & btree_leaf_mask) ? 1 : 2;
+ btree_interior_mask |= btree_leaf_mask;
+
+ c->btree_cache.pinned_nodes_leaf_mask = btree_leaf_mask;
+ c->btree_cache.pinned_nodes_interior_mask = btree_interior_mask;
+ c->btree_cache.pinned_nodes_start = start;
+ c->btree_cache.pinned_nodes_end = *end = BBPOS_MAX;
+
+ for (enum btree_id btree = start.btree;
+ btree < BTREE_ID_NR && !ret;
+ btree++) {
+ unsigned depth = ((1U << btree) & btree_leaf_mask) ? 0 : 1;
+ struct btree_iter iter;
+ struct btree *b;
if (!((1U << btree) & btree_leaf_mask) &&
!((1U << btree) & btree_interior_mask))
continue;
- bch2_trans_node_iter_init(trans, &iter, btree,
- btree == start.btree ? start.pos : POS_MIN,
- 0, depth, 0);
- /*
- * for_each_btree_key_contineu() doesn't check the return value
- * from bch2_btree_iter_advance(), which is needed when
- * iterating over interior nodes where we'll see keys at
- * SPOS_MAX:
- */
- do {
- k = __bch2_btree_iter_peek_and_restart(trans, &iter, 0);
- ret = bkey_err(k);
- if (!k.k || ret)
- break;
-
- --btree_nodes;
- if (!btree_nodes) {
- *end = BBPOS(btree, k.k->p);
+ __for_each_btree_node(trans, iter, btree,
+ btree == start.btree ? start.pos : POS_MIN,
+ 0, depth, BTREE_ITER_PREFETCH, b, ret) {
+ mem_may_pin -= btree_buf_bytes(b);
+ if (mem_may_pin <= 0) {
+ c->btree_cache.pinned_nodes_end = *end =
+ BBPOS(btree, b->key.k.p);
bch2_trans_iter_exit(trans, &iter);
return 0;
}
- } while (bch2_btree_iter_advance(&iter));
+ }
bch2_trans_iter_exit(trans, &iter);
}
- *end = BBPOS_MAX;
return ret;
}
static int bch2_check_extents_to_backpointers_pass(struct btree_trans *trans,
- struct bpos bucket_start,
- struct bpos bucket_end)
+ struct extents_to_bp_state *s)
{
- struct btree_iter iter;
- enum btree_id btree_id;
- struct bpos last_flushed_pos = SPOS_MAX;
+ struct bch_fs *c = trans->c;
int ret = 0;
- for (btree_id = 0; btree_id < BTREE_ID_NR; btree_id++) {
- unsigned depth = btree_type_has_ptrs(btree_id) ? 0 : 1;
-
- bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN, 0,
- depth,
- BTREE_ITER_ALL_LEVELS|
- BTREE_ITER_PREFETCH);
-
- do {
- ret = commit_do(trans, NULL, NULL,
- BTREE_INSERT_LAZY_RW|
- BTREE_INSERT_NOFAIL,
- check_extent_to_backpointers(trans, &iter,
- bucket_start, bucket_end,
- &last_flushed_pos));
- if (ret)
- break;
- } while (!bch2_btree_iter_advance(&iter));
-
- bch2_trans_iter_exit(trans, &iter);
-
- if (ret)
- break;
+ for (enum btree_id btree_id = 0;
+ btree_id < btree_id_nr_alive(c);
+ btree_id++) {
+ int level, depth = btree_type_has_ptrs(btree_id) ? 0 : 1;
ret = commit_do(trans, NULL, NULL,
- BTREE_INSERT_LAZY_RW|
- BTREE_INSERT_NOFAIL,
- check_btree_root_to_backpointers(trans, btree_id,
- bucket_start, bucket_end,
- &last_flushed_pos));
+ BCH_TRANS_COMMIT_no_enospc,
+ check_btree_root_to_backpointers(trans, s, btree_id, &level));
if (ret)
- break;
- }
- return ret;
-}
-
-static struct bpos bucket_pos_to_bp_safe(const struct bch_fs *c,
- struct bpos bucket)
-{
- return bch2_dev_exists2(c, bucket.inode)
- ? bucket_pos_to_bp(c, bucket, 0)
- : bucket;
-}
-
-int bch2_get_alloc_in_memory_pos(struct btree_trans *trans,
- struct bpos start, struct bpos *end)
-{
- struct btree_iter alloc_iter;
- struct btree_iter bp_iter;
- struct bkey_s_c alloc_k, bp_k;
- size_t btree_nodes = btree_nodes_fit_in_ram(trans->c);
- bool alloc_end = false, bp_end = false;
- int ret = 0;
+ return ret;
- bch2_trans_node_iter_init(trans, &alloc_iter, BTREE_ID_alloc,
- start, 0, 1, 0);
- bch2_trans_node_iter_init(trans, &bp_iter, BTREE_ID_backpointers,
- bucket_pos_to_bp_safe(trans->c, start), 0, 1, 0);
- while (1) {
- alloc_k = !alloc_end
- ? __bch2_btree_iter_peek_and_restart(trans, &alloc_iter, 0)
- : bkey_s_c_null;
- bp_k = !bp_end
- ? __bch2_btree_iter_peek_and_restart(trans, &bp_iter, 0)
- : bkey_s_c_null;
-
- ret = bkey_err(alloc_k) ?: bkey_err(bp_k);
- if ((!alloc_k.k && !bp_k.k) || ret) {
- *end = SPOS_MAX;
- break;
- }
+ while (level >= depth) {
+ struct btree_iter iter;
+ bch2_trans_node_iter_init(trans, &iter, btree_id, POS_MIN, 0,
+ level,
+ BTREE_ITER_PREFETCH);
+ while (1) {
+ bch2_trans_begin(trans);
+
+ struct bkey_s_c k = bch2_btree_iter_peek(&iter);
+ if (!k.k)
+ break;
+ ret = bkey_err(k) ?:
+ check_extent_to_backpointers(trans, s, btree_id, level, k) ?:
+ bch2_trans_commit(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc);
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart)) {
+ ret = 0;
+ continue;
+ }
+ if (ret)
+ break;
+ if (bpos_eq(iter.pos, SPOS_MAX))
+ break;
+ bch2_btree_iter_advance(&iter);
+ }
+ bch2_trans_iter_exit(trans, &iter);
- --btree_nodes;
- if (!btree_nodes) {
- *end = alloc_k.k->p;
- break;
- }
+ if (ret)
+ return ret;
- if (bpos_lt(alloc_iter.pos, SPOS_MAX) &&
- bpos_lt(bucket_pos_to_bp_safe(trans->c, alloc_iter.pos), bp_iter.pos)) {
- if (!bch2_btree_iter_advance(&alloc_iter))
- alloc_end = true;
- } else {
- if (!bch2_btree_iter_advance(&bp_iter))
- bp_end = true;
+ --level;
}
}
- bch2_trans_iter_exit(trans, &bp_iter);
- bch2_trans_iter_exit(trans, &alloc_iter);
- return ret;
+
+ return 0;
}
int bch2_check_extents_to_backpointers(struct bch_fs *c)
{
- struct btree_trans trans;
- struct bpos start = POS_MIN, end;
+ struct btree_trans *trans = bch2_trans_get(c);
+ struct extents_to_bp_state s = { .bucket_start = POS_MIN };
int ret;
- bch2_trans_init(&trans, c, 0, 0);
+ bch2_bkey_buf_init(&s.last_flushed);
+ bkey_init(&s.last_flushed.k->k);
+
while (1) {
- ret = bch2_get_alloc_in_memory_pos(&trans, start, &end);
+ struct bbpos end;
+ ret = bch2_get_btree_in_memory_pos(trans,
+ BIT_ULL(BTREE_ID_backpointers),
+ BIT_ULL(BTREE_ID_backpointers),
+ BBPOS(BTREE_ID_backpointers, s.bucket_start), &end);
if (ret)
break;
- if (bpos_eq(start, POS_MIN) && !bpos_eq(end, SPOS_MAX))
+ s.bucket_end = end.pos;
+
+ if ( bpos_eq(s.bucket_start, POS_MIN) &&
+ !bpos_eq(s.bucket_end, SPOS_MAX))
bch_verbose(c, "%s(): alloc info does not fit in ram, running in multiple passes with %zu nodes per pass",
__func__, btree_nodes_fit_in_ram(c));
- if (!bpos_eq(start, POS_MIN) || !bpos_eq(end, SPOS_MAX)) {
+ if (!bpos_eq(s.bucket_start, POS_MIN) ||
+ !bpos_eq(s.bucket_end, SPOS_MAX)) {
struct printbuf buf = PRINTBUF;
prt_str(&buf, "check_extents_to_backpointers(): ");
- bch2_bpos_to_text(&buf, start);
+ bch2_bpos_to_text(&buf, s.bucket_start);
prt_str(&buf, "-");
- bch2_bpos_to_text(&buf, end);
+ bch2_bpos_to_text(&buf, s.bucket_end);
bch_verbose(c, "%s", buf.buf);
printbuf_exit(&buf);
}
- ret = bch2_check_extents_to_backpointers_pass(&trans, start, end);
- if (ret || bpos_eq(end, SPOS_MAX))
+ ret = bch2_check_extents_to_backpointers_pass(trans, &s);
+ if (ret || bpos_eq(s.bucket_end, SPOS_MAX))
break;
- start = bpos_successor(end);
+ s.bucket_start = bpos_successor(s.bucket_end);
}
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
+ bch2_bkey_buf_exit(&s.last_flushed, c);
+ c->btree_cache.pinned_nodes_leaf_mask = 0;
+ c->btree_cache.pinned_nodes_interior_mask = 0;
+
+ bch_err_fn(c, ret);
return ret;
}
static int check_one_backpointer(struct btree_trans *trans,
- struct bpos bucket,
- u64 *bp_offset,
struct bbpos start,
- struct bbpos end)
+ struct bbpos end,
+ struct bkey_s_c_backpointer bp,
+ struct bpos *last_flushed_pos)
{
+ struct bch_fs *c = trans->c;
struct btree_iter iter;
- struct bch_backpointer bp;
- struct bbpos pos;
+ struct bbpos pos = bp_to_bbpos(*bp.v);
struct bkey_s_c k;
struct printbuf buf = PRINTBUF;
int ret;
- ret = bch2_get_next_backpointer(trans, bucket, -1, bp_offset, &bp, 0);
- if (ret || *bp_offset == U64_MAX)
- return ret;
-
- pos = bp_to_bbpos(bp);
if (bbpos_cmp(pos, start) < 0 ||
bbpos_cmp(pos, end) > 0)
return 0;
- k = bch2_backpointer_get_key(trans, &iter, bucket, *bp_offset, bp);
+ k = bch2_backpointer_get_key(trans, &iter, bp.k->p, *bp.v, 0);
ret = bkey_err(k);
if (ret == -BCH_ERR_backpointer_to_overwritten_btree_node)
return 0;
if (ret)
return ret;
- if (fsck_err_on(!k.k, trans->c,
- "%s backpointer points to missing extent\n%s",
- *bp_offset < BACKPOINTER_OFFSET_MAX ? "alloc" : "btree",
- (bch2_backpointer_to_text(&buf, &bp), buf.buf))) {
- ret = bch2_backpointer_del_by_offset(trans, bucket, *bp_offset, bp);
- if (ret == -ENOENT)
- bch_err(trans->c, "backpointer at %llu not found", *bp_offset);
+ if (!k.k && !bpos_eq(*last_flushed_pos, bp.k->p)) {
+ *last_flushed_pos = bp.k->p;
+ ret = bch2_btree_write_buffer_flush_sync(trans) ?:
+ -BCH_ERR_transaction_restart_write_buffer_flush;
+ goto out;
}
- bch2_trans_iter_exit(trans, &iter);
+ if (fsck_err_on(!k.k, c,
+ backpointer_to_missing_ptr,
+ "backpointer for missing %s\n %s",
+ bp.v->level ? "btree node" : "extent",
+ (bch2_bkey_val_to_text(&buf, c, bp.s_c), buf.buf))) {
+ ret = bch2_btree_delete_at_buffered(trans, BTREE_ID_backpointers, bp.k->p);
+ goto out;
+ }
+out:
fsck_err:
+ bch2_trans_iter_exit(trans, &iter);
printbuf_exit(&buf);
return ret;
}
struct bbpos start,
struct bbpos end)
{
- struct btree_iter iter;
- struct bkey_s_c k;
- int ret = 0;
-
- for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
- BTREE_ITER_PREFETCH, k, ret) {
- u64 bp_offset = 0;
-
- while (!(ret = commit_do(trans, NULL, NULL,
- BTREE_INSERT_LAZY_RW|
- BTREE_INSERT_NOFAIL,
- check_one_backpointer(trans, iter.pos, &bp_offset, start, end))) &&
- bp_offset < U64_MAX)
- bp_offset++;
+ struct bpos last_flushed_pos = SPOS_MAX;
- if (ret)
- break;
- }
- bch2_trans_iter_exit(trans, &iter);
- return ret < 0 ? ret : 0;
+ return for_each_btree_key_commit(trans, iter, BTREE_ID_backpointers,
+ POS_MIN, BTREE_ITER_PREFETCH, k,
+ NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+ check_one_backpointer(trans, start, end,
+ bkey_s_c_to_backpointer(k),
+ &last_flushed_pos));
}
int bch2_check_backpointers_to_extents(struct bch_fs *c)
{
- struct btree_trans trans;
+ struct btree_trans *trans = bch2_trans_get(c);
struct bbpos start = (struct bbpos) { .btree = 0, .pos = POS_MIN, }, end;
int ret;
- bch2_trans_init(&trans, c, 0, 0);
while (1) {
- ret = bch2_get_btree_in_memory_pos(&trans,
+ ret = bch2_get_btree_in_memory_pos(trans,
(1U << BTREE_ID_extents)|
(1U << BTREE_ID_reflink),
~0,
printbuf_exit(&buf);
}
- ret = bch2_check_backpointers_to_extents_pass(&trans, start, end);
+ ret = bch2_check_backpointers_to_extents_pass(trans, start, end);
if (ret || !bbpos_cmp(end, BBPOS_MAX))
break;
start = bbpos_successor(end);
}
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
+
+ c->btree_cache.pinned_nodes_leaf_mask = 0;
+ c->btree_cache.pinned_nodes_interior_mask = 0;
+ bch_err_fn(c, ret);
return ret;
}
projects / bcachefs-tools-debian / blobdiff