#include <linux/sched/task.h>
#include <linux/sort.h>
+static void bch2_discard_one_bucket_fast(struct bch_fs *c, struct bpos bucket);
+
/* Persistent alloc info: */
static const unsigned BCH_ALLOC_V1_FIELD_BYTES[] = {
return DIV_ROUND_UP(bytes, sizeof(u64));
}
-int bch2_alloc_v1_invalid(const struct bch_fs *c, struct bkey_s_c k,
+int bch2_alloc_v1_invalid(struct bch_fs *c, struct bkey_s_c k,
enum bkey_invalid_flags flags,
struct printbuf *err)
{
struct bkey_s_c_alloc a = bkey_s_c_to_alloc(k);
+ int ret = 0;
/* allow for unknown fields */
- if (bkey_val_u64s(a.k) < bch_alloc_v1_val_u64s(a.v)) {
- prt_printf(err, "incorrect value size (%zu < %u)",
- bkey_val_u64s(a.k), bch_alloc_v1_val_u64s(a.v));
- return -BCH_ERR_invalid_bkey;
- }
-
- return 0;
+ bkey_fsck_err_on(bkey_val_u64s(a.k) < bch_alloc_v1_val_u64s(a.v), c, err,
+ alloc_v1_val_size_bad,
+ "incorrect value size (%zu < %u)",
+ bkey_val_u64s(a.k), bch_alloc_v1_val_u64s(a.v));
+fsck_err:
+ return ret;
}
-int bch2_alloc_v2_invalid(const struct bch_fs *c, struct bkey_s_c k,
+int bch2_alloc_v2_invalid(struct bch_fs *c, struct bkey_s_c k,
enum bkey_invalid_flags flags,
struct printbuf *err)
{
struct bkey_alloc_unpacked u;
+ int ret = 0;
- if (bch2_alloc_unpack_v2(&u, k)) {
- prt_printf(err, "unpack error");
- return -BCH_ERR_invalid_bkey;
- }
-
- return 0;
+ bkey_fsck_err_on(bch2_alloc_unpack_v2(&u, k), c, err,
+ alloc_v2_unpack_error,
+ "unpack error");
+fsck_err:
+ return ret;
}
-int bch2_alloc_v3_invalid(const struct bch_fs *c, struct bkey_s_c k,
+int bch2_alloc_v3_invalid(struct bch_fs *c, struct bkey_s_c k,
enum bkey_invalid_flags flags,
struct printbuf *err)
{
struct bkey_alloc_unpacked u;
+ int ret = 0;
- if (bch2_alloc_unpack_v3(&u, k)) {
- prt_printf(err, "unpack error");
- return -BCH_ERR_invalid_bkey;
- }
-
- return 0;
+ bkey_fsck_err_on(bch2_alloc_unpack_v3(&u, k), c, err,
+ alloc_v2_unpack_error,
+ "unpack error");
+fsck_err:
+ return ret;
}
-int bch2_alloc_v4_invalid(const struct bch_fs *c, struct bkey_s_c k,
- enum bkey_invalid_flags flags,
- struct printbuf *err)
+int bch2_alloc_v4_invalid(struct bch_fs *c, struct bkey_s_c k,
+ enum bkey_invalid_flags flags, struct printbuf *err)
{
struct bkey_s_c_alloc_v4 a = bkey_s_c_to_alloc_v4(k);
- int rw = flags & WRITE;
-
- if (alloc_v4_u64s(a.v) > bkey_val_u64s(k.k)) {
- prt_printf(err, "bad val size (%u > %lu)",
- alloc_v4_u64s(a.v), bkey_val_u64s(k.k));
- return -BCH_ERR_invalid_bkey;
- }
-
- if (!BCH_ALLOC_V4_BACKPOINTERS_START(a.v) &&
- BCH_ALLOC_V4_NR_BACKPOINTERS(a.v)) {
- prt_printf(err, "invalid backpointers_start");
- return -BCH_ERR_invalid_bkey;
- }
+ int ret = 0;
- if (rw == WRITE &&
- !(flags & BKEY_INVALID_JOURNAL) &&
- c->curr_recovery_pass > BCH_RECOVERY_PASS_check_btree_backpointers) {
- unsigned i, bp_len = 0;
+ bkey_fsck_err_on(alloc_v4_u64s(a.v) > bkey_val_u64s(k.k), c, err,
+ alloc_v4_val_size_bad,
+ "bad val size (%u > %zu)",
+ alloc_v4_u64s(a.v), bkey_val_u64s(k.k));
- for (i = 0; i < BCH_ALLOC_V4_NR_BACKPOINTERS(a.v); i++)
- bp_len += alloc_v4_backpointers_c(a.v)[i].bucket_len;
+ bkey_fsck_err_on(!BCH_ALLOC_V4_BACKPOINTERS_START(a.v) &&
+ BCH_ALLOC_V4_NR_BACKPOINTERS(a.v), c, err,
+ alloc_v4_backpointers_start_bad,
+ "invalid backpointers_start");
- if (bp_len > a.v->dirty_sectors) {
- prt_printf(err, "too many backpointers");
- return -BCH_ERR_invalid_bkey;
- }
- }
-
- if (rw == WRITE) {
- if (alloc_data_type(*a.v, a.v->data_type) != a.v->data_type) {
- prt_printf(err, "invalid data type (got %u should be %u)",
- a.v->data_type, alloc_data_type(*a.v, a.v->data_type));
- return -BCH_ERR_invalid_bkey;
- }
+ bkey_fsck_err_on(alloc_data_type(*a.v, a.v->data_type) != a.v->data_type, c, err,
+ alloc_key_data_type_bad,
+ "invalid data type (got %u should be %u)",
+ a.v->data_type, alloc_data_type(*a.v, a.v->data_type));
- switch (a.v->data_type) {
- case BCH_DATA_free:
- case BCH_DATA_need_gc_gens:
- case BCH_DATA_need_discard:
- if (a.v->dirty_sectors ||
- a.v->cached_sectors ||
- a.v->stripe) {
- prt_printf(err, "empty data type free but have data");
- return -BCH_ERR_invalid_bkey;
- }
- break;
- case BCH_DATA_sb:
- case BCH_DATA_journal:
- case BCH_DATA_btree:
- case BCH_DATA_user:
- case BCH_DATA_parity:
- if (!a.v->dirty_sectors) {
- prt_printf(err, "data_type %s but dirty_sectors==0",
- bch2_data_types[a.v->data_type]);
- return -BCH_ERR_invalid_bkey;
- }
- break;
- case BCH_DATA_cached:
- if (!a.v->cached_sectors ||
- a.v->dirty_sectors ||
- a.v->stripe) {
- prt_printf(err, "data type inconsistency");
- return -BCH_ERR_invalid_bkey;
- }
-
- if (!a.v->io_time[READ] &&
- c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_to_lru_refs) {
- prt_printf(err, "cached bucket with read_time == 0");
- return -BCH_ERR_invalid_bkey;
- }
- break;
- case BCH_DATA_stripe:
- if (!a.v->stripe) {
- prt_printf(err, "data_type %s but stripe==0",
- bch2_data_types[a.v->data_type]);
- return -BCH_ERR_invalid_bkey;
- }
- break;
- }
+ switch (a.v->data_type) {
+ case BCH_DATA_free:
+ case BCH_DATA_need_gc_gens:
+ case BCH_DATA_need_discard:
+ bkey_fsck_err_on(bch2_bucket_sectors(*a.v) || a.v->stripe,
+ c, err, alloc_key_empty_but_have_data,
+ "empty data type free but have data");
+ break;
+ case BCH_DATA_sb:
+ case BCH_DATA_journal:
+ case BCH_DATA_btree:
+ case BCH_DATA_user:
+ case BCH_DATA_parity:
+ bkey_fsck_err_on(!bch2_bucket_sectors_dirty(*a.v),
+ c, err, alloc_key_dirty_sectors_0,
+ "data_type %s but dirty_sectors==0",
+ bch2_data_type_str(a.v->data_type));
+ break;
+ case BCH_DATA_cached:
+ bkey_fsck_err_on(!a.v->cached_sectors ||
+ bch2_bucket_sectors_dirty(*a.v) ||
+ a.v->stripe,
+ c, err, alloc_key_cached_inconsistency,
+ "data type inconsistency");
+
+ bkey_fsck_err_on(!a.v->io_time[READ] &&
+ c->curr_recovery_pass > BCH_RECOVERY_PASS_check_alloc_to_lru_refs,
+ c, err, alloc_key_cached_but_read_time_zero,
+ "cached bucket with read_time == 0");
+ break;
+ case BCH_DATA_stripe:
+ break;
}
-
- return 0;
-}
-
-static inline u64 swab40(u64 x)
-{
- return (((x & 0x00000000ffULL) << 32)|
- ((x & 0x000000ff00ULL) << 16)|
- ((x & 0x0000ff0000ULL) >> 0)|
- ((x & 0x00ff000000ULL) >> 16)|
- ((x & 0xff00000000ULL) >> 32));
+fsck_err:
+ return ret;
}
void bch2_alloc_v4_swab(struct bkey_s k)
a->io_time[1] = swab64(a->io_time[1]);
a->stripe = swab32(a->stripe);
a->nr_external_backpointers = swab32(a->nr_external_backpointers);
+ a->fragmentation_lru = swab64(a->fragmentation_lru);
bps = alloc_v4_backpointers(a);
for (bp = bps; bp < bps + BCH_ALLOC_V4_NR_BACKPOINTERS(a); bp++) {
{
struct bch_alloc_v4 _a;
const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &_a);
- unsigned i;
prt_newline(out);
printbuf_indent_add(out, 2);
- prt_printf(out, "gen %u oldest_gen %u data_type %s",
- a->gen, a->oldest_gen,
- a->data_type < BCH_DATA_NR
- ? bch2_data_types[a->data_type]
- : "(invalid data type)");
+ prt_printf(out, "gen %u oldest_gen %u data_type ", a->gen, a->oldest_gen);
+ bch2_prt_data_type(out, a->data_type);
prt_newline(out);
prt_printf(out, "journal_seq %llu", a->journal_seq);
prt_newline(out);
prt_printf(out, "fragmentation %llu", a->fragmentation_lru);
prt_newline(out);
prt_printf(out, "bp_start %llu", BCH_ALLOC_V4_BACKPOINTERS_START(a));
- prt_newline(out);
-
- if (BCH_ALLOC_V4_NR_BACKPOINTERS(a)) {
- struct bkey_s_c_alloc_v4 a_raw = bkey_s_c_to_alloc_v4(k);
- const struct bch_backpointer *bps = alloc_v4_backpointers_c(a_raw.v);
-
- prt_printf(out, "backpointers: %llu", BCH_ALLOC_V4_NR_BACKPOINTERS(a_raw.v));
- printbuf_indent_add(out, 2);
-
- for (i = 0; i < BCH_ALLOC_V4_NR_BACKPOINTERS(a_raw.v); i++) {
- prt_newline(out);
- bch2_backpointer_to_text(out, &bps[i]);
- }
-
- printbuf_indent_sub(out, 2);
- }
-
printbuf_indent_sub(out, 2);
}
: 0;
}
-int bch2_bucket_gens_invalid(const struct bch_fs *c, struct bkey_s_c k,
+int bch2_bucket_gens_invalid(struct bch_fs *c, struct bkey_s_c k,
enum bkey_invalid_flags flags,
struct printbuf *err)
{
- if (bkey_val_bytes(k.k) != sizeof(struct bch_bucket_gens)) {
- prt_printf(err, "bad val size (%lu != %zu)",
- bkey_val_bytes(k.k), sizeof(struct bch_bucket_gens));
- return -BCH_ERR_invalid_bkey;
- }
+ int ret = 0;
- return 0;
+ bkey_fsck_err_on(bkey_val_bytes(k.k) != sizeof(struct bch_bucket_gens), c, err,
+ bucket_gens_val_size_bad,
+ "bad val size (%zu != %zu)",
+ bkey_val_bytes(k.k), sizeof(struct bch_bucket_gens));
+fsck_err:
+ return ret;
}
void bch2_bucket_gens_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c k)
int bch2_bucket_gens_init(struct bch_fs *c)
{
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- struct bch_alloc_v4 a;
+ struct btree_trans *trans = bch2_trans_get(c);
struct bkey_i_bucket_gens g;
bool have_bucket_gens_key = false;
- unsigned offset;
- struct bpos pos;
- u8 gen;
int ret;
- bch2_trans_init(&trans, c, 0, 0);
-
- for_each_btree_key(&trans, iter, BTREE_ID_alloc, POS_MIN,
- BTREE_ITER_PREFETCH, k, ret) {
+ ret = for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
+ BTREE_ITER_PREFETCH, k, ({
/*
* Not a fsck error because this is checked/repaired by
* bch2_check_alloc_key() which runs later:
if (!bch2_dev_bucket_exists(c, k.k->p))
continue;
- gen = bch2_alloc_to_v4(k, &a)->gen;
- pos = alloc_gens_pos(iter.pos, &offset);
+ struct bch_alloc_v4 a;
+ u8 gen = bch2_alloc_to_v4(k, &a)->gen;
+ unsigned offset;
+ struct bpos pos = alloc_gens_pos(iter.pos, &offset);
if (have_bucket_gens_key && bkey_cmp(iter.pos, pos)) {
- ret = commit_do(&trans, NULL, NULL,
- BTREE_INSERT_NOFAIL|
- BTREE_INSERT_LAZY_RW,
- __bch2_btree_insert(&trans, BTREE_ID_bucket_gens, &g.k_i, 0));
+ ret = commit_do(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc,
+ bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0));
if (ret)
break;
have_bucket_gens_key = false;
}
g.v.gens[offset] = gen;
- }
- bch2_trans_iter_exit(&trans, &iter);
+ 0;
+ }));
if (have_bucket_gens_key && !ret)
- ret = commit_do(&trans, NULL, NULL,
- BTREE_INSERT_NOFAIL|
- BTREE_INSERT_LAZY_RW,
- __bch2_btree_insert(&trans, BTREE_ID_bucket_gens, &g.k_i, 0));
+ ret = commit_do(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc,
+ bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0));
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
- if (ret)
- bch_err_fn(c, ret);
+ bch_err_fn(c, ret);
return ret;
}
int bch2_alloc_read(struct bch_fs *c)
{
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- struct bch_dev *ca;
+ struct btree_trans *trans = bch2_trans_get(c);
int ret;
down_read(&c->gc_lock);
- bch2_trans_init(&trans, c, 0, 0);
if (c->sb.version_upgrade_complete >= bcachefs_metadata_version_bucket_gens) {
- const struct bch_bucket_gens *g;
- u64 b;
-
- for_each_btree_key(&trans, iter, BTREE_ID_bucket_gens, POS_MIN,
- BTREE_ITER_PREFETCH, k, ret) {
+ ret = for_each_btree_key(trans, iter, BTREE_ID_bucket_gens, POS_MIN,
+ BTREE_ITER_PREFETCH, k, ({
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;
if (k.k->type != KEY_TYPE_bucket_gens)
continue;
- g = bkey_s_c_to_bucket_gens(k).v;
+ const struct bch_bucket_gens *g = bkey_s_c_to_bucket_gens(k).v;
/*
* Not a fsck error because this is checked/repaired by
if (!bch2_dev_exists2(c, k.k->p.inode))
continue;
- ca = bch_dev_bkey_exists(c, k.k->p.inode);
+ struct bch_dev *ca = bch_dev_bkey_exists(c, k.k->p.inode);
- for (b = max_t(u64, ca->mi.first_bucket, start);
+ for (u64 b = max_t(u64, ca->mi.first_bucket, start);
b < min_t(u64, ca->mi.nbuckets, end);
b++)
*bucket_gen(ca, b) = g->gens[b & KEY_TYPE_BUCKET_GENS_MASK];
- }
- bch2_trans_iter_exit(&trans, &iter);
+ 0;
+ }));
} else {
- struct bch_alloc_v4 a;
-
- for_each_btree_key(&trans, iter, BTREE_ID_alloc, POS_MIN,
- BTREE_ITER_PREFETCH, k, ret) {
+ ret = for_each_btree_key(trans, iter, BTREE_ID_alloc, POS_MIN,
+ BTREE_ITER_PREFETCH, k, ({
/*
* Not a fsck error because this is checked/repaired by
* bch2_check_alloc_key() which runs later:
if (!bch2_dev_bucket_exists(c, k.k->p))
continue;
- ca = bch_dev_bkey_exists(c, k.k->p.inode);
+ struct bch_dev *ca = bch_dev_bkey_exists(c, k.k->p.inode);
+ struct bch_alloc_v4 a;
*bucket_gen(ca, k.k->p.offset) = bch2_alloc_to_v4(k, &a)->gen;
- }
- bch2_trans_iter_exit(&trans, &iter);
+ 0;
+ }));
}
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
up_read(&c->gc_lock);
- if (ret)
- bch_err_fn(c, ret);
-
+ bch_err_fn(c, ret);
return ret;
}
"incorrect key when %s %s:%llu:%llu:0 (got %s should be %s)\n"
" for %s",
set ? "setting" : "clearing",
- bch2_btree_ids[btree],
+ bch2_btree_id_str(btree),
iter.pos.inode,
iter.pos.offset,
bch2_bkey_types[old.k->type],
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)
+int bch2_trigger_alloc(struct btree_trans *trans,
+ enum btree_id btree, unsigned level,
+ struct bkey_s_c old, struct bkey_s 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);
+ if (bch2_trans_inconsistent_on(!bch2_dev_bucket_exists(c, new.k->p), trans,
+ "alloc key for invalid device or bucket"))
+ return -EIO;
- old_a = bch2_alloc_to_v4(old, &old_a_convert);
- new_a = &bkey_i_to_alloc_v4(new)->v;
+ struct bch_dev *ca = bch_dev_bkey_exists(c, new.k->p.inode);
- new_a->data_type = alloc_data_type(*new_a, new_a->data_type);
+ struct bch_alloc_v4 old_a_convert;
+ const struct bch_alloc_v4 *old_a = bch2_alloc_to_v4(old, &old_a_convert);
- 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 (flags & BTREE_TRIGGER_TRANSACTIONAL) {
+ struct bch_alloc_v4 *new_a = bkey_s_to_alloc_v4(new).v;
- 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);
- }
+ new_a->data_type = alloc_data_type(*new_a, new_a->data_type);
- 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 (bch2_bucket_sectors(*new_a) > bch2_bucket_sectors(*old_a)) {
+ 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 (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));
+ 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);
+ }
- old_lru = alloc_lru_idx_read(*old_a);
- new_lru = alloc_lru_idx_read(*new_a);
+ 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, new.s_c, new_a, true);
+ if (ret)
+ return ret;
+ }
- if (old_lru != new_lru) {
- ret = bch2_lru_change(trans, new->k.p.inode,
- bucket_to_u64(new->k.p),
- old_lru, new_lru);
- 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));
+
+ u64 old_lru = alloc_lru_idx_read(*old_a);
+ u64 new_lru = alloc_lru_idx_read(*new_a);
+ if (old_lru != new_lru) {
+ ret = bch2_lru_change(trans, new.k->p.inode,
+ bucket_to_u64(new.k->p),
+ old_lru, new_lru);
+ if (ret)
+ return ret;
+ }
+
+ new_a->fragmentation_lru = alloc_lru_idx_fragmentation(*new_a,
+ bch_dev_bkey_exists(c, new.k->p.inode));
+ if (old_a->fragmentation_lru != new_a->fragmentation_lru) {
+ ret = bch2_lru_change(trans,
+ BCH_LRU_FRAGMENTATION_START,
+ bucket_to_u64(new.k->p),
+ old_a->fragmentation_lru, new_a->fragmentation_lru);
+ if (ret)
+ return ret;
+ }
+
+ if (old_a->gen != new_a->gen) {
+ ret = bch2_bucket_gen_update(trans, new.k->p, new_a->gen);
+ if (ret)
+ return ret;
+ }
+
+ /*
+ * need to know if we're getting called from the invalidate path or
+ * not:
+ */
+
+ if ((flags & BTREE_TRIGGER_BUCKET_INVALIDATE) &&
+ old_a->cached_sectors) {
+ ret = bch2_update_cached_sectors_list(trans, new.k->p.inode,
+ -((s64) old_a->cached_sectors));
+ if (ret)
+ return ret;
+ }
}
- new_a->fragmentation_lru = alloc_lru_idx_fragmentation(*new_a,
- bch_dev_bkey_exists(c, new->k.p.inode));
+ if ((flags & BTREE_TRIGGER_ATOMIC) && (flags & BTREE_TRIGGER_INSERT)) {
+ struct bch_alloc_v4 *new_a = bkey_s_to_alloc_v4(new).v;
+ u64 journal_seq = trans->journal_res.seq;
+ u64 bucket_journal_seq = new_a->journal_seq;
- if (old_a->fragmentation_lru != new_a->fragmentation_lru) {
- ret = bch2_lru_change(trans,
- BCH_LRU_FRAGMENTATION_START,
- bucket_to_u64(new->k.p),
- old_a->fragmentation_lru, new_a->fragmentation_lru);
- if (ret)
- return ret;
+ if ((flags & BTREE_TRIGGER_INSERT) &&
+ data_type_is_empty(old_a->data_type) !=
+ data_type_is_empty(new_a->data_type) &&
+ new.k->type == KEY_TYPE_alloc_v4) {
+ struct bch_alloc_v4 *v = bkey_s_to_alloc_v4(new).v;
+
+ /*
+ * If the btree updates referring to a bucket weren't flushed
+ * before the bucket became empty again, then the we don't have
+ * to wait on a journal flush before we can reuse the bucket:
+ */
+ v->journal_seq = bucket_journal_seq =
+ data_type_is_empty(new_a->data_type) &&
+ (journal_seq == v->journal_seq ||
+ bch2_journal_noflush_seq(&c->journal, v->journal_seq))
+ ? 0 : journal_seq;
+ }
+
+ if (!data_type_is_empty(old_a->data_type) &&
+ data_type_is_empty(new_a->data_type) &&
+ bucket_journal_seq) {
+ ret = bch2_set_bucket_needs_journal_commit(&c->buckets_waiting_for_journal,
+ c->journal.flushed_seq_ondisk,
+ new.k->p.inode, new.k->p.offset,
+ bucket_journal_seq);
+ if (ret) {
+ bch2_fs_fatal_error(c,
+ "error setting bucket_needs_journal_commit: %i", ret);
+ return ret;
+ }
+ }
+
+ percpu_down_read(&c->mark_lock);
+ if (new_a->gen != old_a->gen)
+ *bucket_gen(ca, new.k->p.offset) = new_a->gen;
+
+ bch2_dev_usage_update(c, ca, old_a, new_a, journal_seq, false);
+ percpu_up_read(&c->mark_lock);
+
+#define eval_state(_a, expr) ({ const struct bch_alloc_v4 *a = _a; expr; })
+#define statechange(expr) !eval_state(old_a, expr) && eval_state(new_a, expr)
+#define bucket_flushed(a) (!a->journal_seq || a->journal_seq <= c->journal.flushed_seq_ondisk)
+
+ if (statechange(a->data_type == BCH_DATA_free) &&
+ bucket_flushed(new_a))
+ closure_wake_up(&c->freelist_wait);
+
+ if (statechange(a->data_type == BCH_DATA_need_discard) &&
+ !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) &&
+ bucket_flushed(new_a))
+ bch2_discard_one_bucket_fast(c, new.k->p);
+
+ if (statechange(a->data_type == BCH_DATA_cached) &&
+ !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) &&
+ should_invalidate_buckets(ca, bch2_dev_usage_read(ca)))
+ bch2_do_invalidates(c);
+
+ if (statechange(a->data_type == BCH_DATA_need_gc_gens))
+ bch2_do_gc_gens(c);
}
- if (old_a->gen != new_a->gen) {
- ret = bch2_bucket_gen_update(trans, new->k.p, new_a->gen);
- if (ret)
- return ret;
+ if ((flags & BTREE_TRIGGER_GC) &&
+ (flags & BTREE_TRIGGER_BUCKET_INVALIDATE)) {
+ struct bch_alloc_v4 new_a_convert;
+ const struct bch_alloc_v4 *new_a = bch2_alloc_to_v4(new.s_c, &new_a_convert);
+
+ percpu_down_read(&c->mark_lock);
+ struct bucket *g = gc_bucket(ca, new.k->p.offset);
+
+ bucket_lock(g);
+
+ g->gen_valid = 1;
+ g->gen = new_a->gen;
+ g->data_type = new_a->data_type;
+ g->stripe = new_a->stripe;
+ g->stripe_redundancy = new_a->stripe_redundancy;
+ g->dirty_sectors = new_a->dirty_sectors;
+ g->cached_sectors = new_a->cached_sectors;
+
+ bucket_unlock(g);
+ percpu_up_read(&c->mark_lock);
}
return 0;
bch2_trans_copy_iter(&iter2, iter);
- if (!bpos_eq(iter->path->l[0].b->key.k.p, SPOS_MAX))
- end = bkey_min(end, bpos_nosnap_successor(iter->path->l[0].b->key.k.p));
+ struct btree_path *path = btree_iter_path(iter->trans, iter);
+ if (!bpos_eq(path->l[0].b->key.k.p, SPOS_MAX))
+ end = bkey_min(end, bpos_nosnap_successor(path->l[0].b->key.k.p));
end = bkey_min(end, POS(iter->pos.inode, iter->pos.offset + U32_MAX - 1));
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;
}
rcu_read_lock();
- iter = bucket->inode;
- ca = __bch2_next_dev(c, &iter, NULL);
+ ca = __bch2_next_dev_idx(c, bucket->inode, NULL);
if (ca)
*bucket = POS(ca->dev_idx, ca->mi.first_bucket);
rcu_read_unlock();
int ret;
if (fsck_err_on(!bch2_dev_bucket_exists(c, alloc_k.k->p), c,
+ alloc_key_to_missing_dev_bucket,
"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);
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"
+ fsck_err(c, need_discard_key_wrong,
+ "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],
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"
+ fsck_err(c, freespace_key_wrong,
+ "incorrect key in freespace btree (got %s should be %s)\n"
" %s",
bch2_bkey_types[k.k->type],
bch2_bkey_types[freespace_key_type],
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"
+ fsck_err(c, bucket_gens_key_wrong,
+ "incorrect gen in bucket_gens btree (got %u should be %u)\n"
" %s",
alloc_gen(k, gens_offset), a->gen,
(printbuf_reset(&buf),
if (k.k->type != KEY_TYPE_set &&
(c->opts.reconstruct_alloc ||
- fsck_err(c, "hole in alloc btree missing in freespace btree\n"
+ fsck_err(c, freespace_hole_missing,
+ "hole in alloc btree missing in freespace btree\n"
" device %llu buckets %llu-%llu",
freespace_iter->pos.inode,
freespace_iter->pos.offset,
unsigned i, gens_offset, gens_end_offset;
int ret;
- if (c->sb.version < bcachefs_metadata_version_bucket_gens)
- 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);
for (i = gens_offset; i < gens_end_offset; i++) {
if (fsck_err_on(g.v.gens[i], c,
+ bucket_gens_hole_wrong,
"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,
}
if (need_update) {
- struct bkey_i *k = bch2_trans_kmalloc(trans, sizeof(g));
+ struct bkey_i *u = bch2_trans_kmalloc(trans, sizeof(g));
- ret = PTR_ERR_OR_ZERO(k);
+ ret = PTR_ERR_OR_ZERO(u);
if (ret)
goto err;
- memcpy(k, &g, sizeof(g));
+ memcpy(u, &g, sizeof(g));
- ret = bch2_trans_update(trans, bucket_gens_iter, k, 0);
+ ret = bch2_trans_update(trans, bucket_gens_iter, u, 0);
if (ret)
goto err;
}
return ret;
}
-static noinline_for_stack int __bch2_check_discard_freespace_key(struct btree_trans *trans,
+static noinline_for_stack int bch2_check_discard_freespace_key(struct btree_trans *trans,
struct btree_iter *iter)
{
struct bch_fs *c = trans->c;
return ret;
if (fsck_err_on(!bch2_dev_bucket_exists(c, pos), c,
+ need_discard_freespace_key_to_invalid_dev_bucket,
"entry in %s btree for nonexistant dev:bucket %llu:%llu",
- bch2_btree_ids[iter->btree_id], pos.inode, pos.offset))
+ bch2_btree_id_str(iter->btree_id), pos.inode, pos.offset))
goto delete;
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,
+ need_discard_freespace_key_bad,
"%s\n incorrectly set at %s:%llu:%llu:0 (free %u, genbits %llu should be %llu)",
(bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf),
- bch2_btree_ids[iter->btree_id],
+ bch2_btree_id_str(iter->btree_id),
iter->pos.inode,
iter->pos.offset,
a->data_type == state,
ret = bch2_btree_delete_extent_at(trans, iter,
iter->btree_id == BTREE_ID_freespace ? 1 : 0, 0) ?:
bch2_trans_commit(trans, NULL, NULL,
- BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW);
+ BCH_TRANS_COMMIT_no_enospc);
goto out;
}
-static int bch2_check_discard_freespace_key(struct btree_trans *trans,
- struct btree_iter *iter,
- struct bpos end)
-{
- if (!btree_id_is_extents(iter->btree_id)) {
- return __bch2_check_discard_freespace_key(trans, iter);
- } else {
- int ret;
-
- while (!bkey_eq(iter->pos, end) &&
- !(ret = btree_trans_too_many_iters(trans) ?:
- __bch2_check_discard_freespace_key(trans, iter)))
- bch2_btree_iter_set_pos(iter, bpos_nosnap_successor(iter->pos));
-
- return ret;
- }
-}
-
/*
* 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
dev_exists = bch2_dev_exists2(c, k.k->p.inode);
if (!dev_exists) {
if (fsck_err_on(!dev_exists, c,
+ bucket_gens_to_invalid_dev,
"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);
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_to_invalid_buckets,
"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);
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_nonzero_for_invalid_buckets,
"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_nonzero_for_invalid_buckets,
"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;
+ struct bkey_i *u = bch2_trans_kmalloc(trans, sizeof(g));
- k = bch2_trans_kmalloc(trans, sizeof(g));
- ret = PTR_ERR_OR_ZERO(k);
+ ret = PTR_ERR_OR_ZERO(u);
if (ret)
goto out;
- memcpy(k, &g, sizeof(g));
- ret = bch2_trans_update(trans, iter, k, 0);
+ memcpy(u, &g, sizeof(g));
+ ret = bch2_trans_update(trans, iter, u, 0);
}
out:
fsck_err:
int bch2_check_alloc_info(struct bch_fs *c)
{
- struct btree_trans trans;
+ struct btree_trans *trans = bch2_trans_get(c);
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,
+ 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,
+ 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,
+ 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,
+ 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);
+ bch2_trans_begin(trans);
k = bch2_get_key_or_real_bucket_hole(&iter, &hole);
ret = bkey_err(k);
if (k.k->type) {
next = bpos_nosnap_successor(k.k->p);
- ret = bch2_check_alloc_key(&trans,
+ ret = bch2_check_alloc_key(trans,
k, &iter,
&discard_iter,
&freespace_iter,
} else {
next = k.k->p;
- ret = bch2_check_alloc_hole_freespace(&trans,
+ ret = bch2_check_alloc_hole_freespace(trans,
bkey_start_pos(k.k),
&next,
&freespace_iter) ?:
- bch2_check_alloc_hole_bucket_gens(&trans,
+ bch2_check_alloc_hole_bucket_gens(trans,
bkey_start_pos(k.k),
&next,
&bucket_gens_iter);
goto bkey_err;
}
- ret = bch2_trans_commit(&trans, NULL, NULL,
- BTREE_INSERT_NOFAIL|
- BTREE_INSERT_LAZY_RW);
+ ret = bch2_trans_commit(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc);
if (ret)
goto bkey_err;
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);
+ 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_key2(&trans, iter,
+ ret = for_each_btree_key(trans, iter,
BTREE_ID_need_discard, POS_MIN,
BTREE_ITER_PREFETCH, k,
- bch2_check_discard_freespace_key(&trans, &iter, k.k->p)) ?:
- for_each_btree_key2(&trans, iter,
- BTREE_ID_freespace, POS_MIN,
- BTREE_ITER_PREFETCH, k,
- bch2_check_discard_freespace_key(&trans, &iter, k.k->p)) ?:
- for_each_btree_key_commit(&trans, iter,
+ bch2_check_discard_freespace_key(trans, &iter));
+ if (ret)
+ goto err;
+
+ bch2_trans_iter_init(trans, &iter, BTREE_ID_freespace, POS_MIN,
+ BTREE_ITER_PREFETCH);
+ while (1) {
+ bch2_trans_begin(trans);
+ k = bch2_btree_iter_peek(&iter);
+ if (!k.k)
+ break;
+
+ ret = bkey_err(k) ?:
+ bch2_check_discard_freespace_key(trans, &iter);
+ if (bch2_err_matches(ret, BCH_ERR_transaction_restart)) {
+ ret = 0;
+ continue;
+ }
+ if (ret) {
+ struct printbuf buf = PRINTBUF;
+ bch2_bkey_val_to_text(&buf, c, k);
+
+ bch_err(c, "while checking %s", buf.buf);
+ printbuf_exit(&buf);
+ break;
+ }
+
+ bch2_btree_iter_set_pos(&iter, bpos_nosnap_successor(iter.pos));
+ }
+ bch2_trans_iter_exit(trans, &iter);
+ if (ret)
+ goto err;
+
+ ret = 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));
+ NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+ bch2_check_bucket_gens_key(trans, &iter, k));
err:
- bch2_trans_exit(&trans);
- if (ret)
- bch_err_fn(c, ret);
+ bch2_trans_put(trans);
+ bch_err_fn(c, ret);
return ret;
}
if (a->data_type != BCH_DATA_cached)
return 0;
+ if (fsck_err_on(!a->io_time[READ], c,
+ alloc_key_cached_but_read_time_zero,
+ "cached bucket with read_time 0\n"
+ " %s",
+ (printbuf_reset(&buf),
+ bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) {
+ 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] = atomic64_read(&c->io_clock[READ].now);
+ ret = bch2_trans_update(trans, alloc_iter,
+ &a_mut->k_i, BTREE_TRIGGER_NORUN);
+ if (ret)
+ goto err;
+
+ a = &a_mut->v;
+ }
+
lru_k = bch2_bkey_get_iter(trans, &lru_iter, BTREE_ID_lru,
lru_pos(alloc_k.k->p.inode,
bucket_to_u64(alloc_k.k->p),
if (ret)
return ret;
- 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(lru_k.k->type != KEY_TYPE_set, c,
+ if (fsck_err_on(lru_k.k->type != KEY_TYPE_set, c,
+ alloc_key_to_missing_lru_entry,
"missing lru entry\n"
" %s",
(printbuf_reset(&buf),
bch2_bkey_val_to_text(&buf, c, alloc_k), buf.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,
bucket_to_u64(alloc_k.k->p),
- read_time);
+ a->io_time[READ]);
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:
int bch2_check_alloc_to_lru_refs(struct bch_fs *c)
{
- struct btree_iter iter;
- struct bkey_s_c k;
- int ret = 0;
-
- ret = bch2_trans_run(c,
- for_each_btree_key_commit(&trans, iter, BTREE_ID_alloc,
+ int ret = bch2_trans_run(c,
+ 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)));
- if (ret)
- bch_err_fn(c, ret);
+ NULL, NULL, BCH_TRANS_COMMIT_no_enospc,
+ bch2_check_alloc_to_lru_ref(trans, &iter)));
+ bch_err_fn(c, ret);
+ return ret;
+}
+
+static int discard_in_flight_add(struct bch_fs *c, struct bpos bucket)
+{
+ int ret;
+
+ mutex_lock(&c->discard_buckets_in_flight_lock);
+ darray_for_each(c->discard_buckets_in_flight, i)
+ if (bkey_eq(*i, bucket)) {
+ ret = -EEXIST;
+ goto out;
+ }
+
+ ret = darray_push(&c->discard_buckets_in_flight, bucket);
+out:
+ mutex_unlock(&c->discard_buckets_in_flight_lock);
return ret;
}
+static void discard_in_flight_remove(struct bch_fs *c, struct bpos bucket)
+{
+ mutex_lock(&c->discard_buckets_in_flight_lock);
+ darray_for_each(c->discard_buckets_in_flight, i)
+ if (bkey_eq(*i, bucket)) {
+ darray_remove_item(&c->discard_buckets_in_flight, i);
+ goto found;
+ }
+ BUG();
+found:
+ mutex_unlock(&c->discard_buckets_in_flight_lock);
+}
+
+struct discard_buckets_state {
+ u64 seen;
+ u64 open;
+ u64 need_journal_commit;
+ u64 discarded;
+ struct bch_dev *ca;
+ u64 need_journal_commit_this_dev;
+};
+
+static void discard_buckets_next_dev(struct bch_fs *c, struct discard_buckets_state *s, struct bch_dev *ca)
+{
+ if (s->ca == ca)
+ return;
+
+ if (s->ca && s->need_journal_commit_this_dev >
+ bch2_dev_usage_read(s->ca).d[BCH_DATA_free].buckets)
+ bch2_journal_flush_async(&c->journal, NULL);
+
+ if (s->ca)
+ percpu_ref_put(&s->ca->ref);
+ if (ca)
+ percpu_ref_get(&ca->ref);
+ s->ca = ca;
+ s->need_journal_commit_this_dev = 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 discard_buckets_state *s)
{
struct bch_fs *c = trans->c;
struct bpos pos = need_discard_iter->pos;
struct bch_dev *ca;
struct bkey_i_alloc_v4 *a;
struct printbuf buf = PRINTBUF;
+ bool discard_locked = false;
int ret = 0;
ca = bch_dev_bkey_exists(c, pos.inode);
+
if (!percpu_ref_tryget(&ca->io_ref)) {
bch2_btree_iter_set_pos(need_discard_iter, POS(pos.inode + 1, 0));
return 0;
}
+ discard_buckets_next_dev(c, s, ca);
+
if (bch2_bucket_is_open_safe(c, pos.inode, pos.offset)) {
- (*open)++;
+ s->open++;
goto out;
}
if (bch2_bucket_needs_journal_commit(&c->buckets_waiting_for_journal,
c->journal.flushed_seq_ondisk,
pos.inode, pos.offset)) {
- (*need_journal_commit)++;
+ s->need_journal_commit++;
+ s->need_journal_commit_this_dev++;
goto out;
}
goto out;
}
+ if (discard_in_flight_add(c, SPOS(iter.pos.inode, iter.pos.offset, true)))
+ goto out;
+
+ discard_locked = true;
+
if (!bkey_eq(*discard_pos_done, iter.pos) &&
ca->mi.discard && !c->opts.nochanges) {
/*
* This works without any other locks because this is the only
* thread that removes items from the need_discard tree
*/
- bch2_trans_unlock(trans);
+ bch2_trans_unlock_long(trans);
blkdev_issue_discard(ca->disk_sb.bdev,
k.k->p.offset * ca->mi.bucket_size,
ca->mi.bucket_size,
ret = bch2_trans_update(trans, &iter, &a->k_i, 0) ?:
bch2_trans_commit(trans, NULL, NULL,
BCH_WATERMARK_btree|
- BTREE_INSERT_NOFAIL);
+ BCH_TRANS_COMMIT_no_enospc);
if (ret)
goto out;
- this_cpu_inc(c->counters[BCH_COUNTER_bucket_discard]);
- (*discarded)++;
+ count_event(c, bucket_discard);
+ s->discarded++;
out:
- (*seen)++;
+ if (discard_locked)
+ discard_in_flight_remove(c, iter.pos);
+ s->seen++;
bch2_trans_iter_exit(trans, &iter);
percpu_ref_put(&ca->io_ref);
printbuf_exit(&buf);
static void bch2_do_discards_work(struct work_struct *work)
{
struct bch_fs *c = container_of(work, struct bch_fs, discard_work);
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- u64 seen = 0, open = 0, need_journal_commit = 0, discarded = 0;
+ struct discard_buckets_state s = {};
struct bpos discard_pos_done = POS_MAX;
int ret;
- bch2_trans_init(&trans, c, 0, 0);
-
/*
* We're doing the commit in bch2_discard_one_bucket instead of using
* for_each_btree_key_commit() so that we can increment counters after
* successful commit:
*/
- ret = for_each_btree_key2(&trans, iter,
- BTREE_ID_need_discard, POS_MIN, 0, k,
- bch2_discard_one_bucket(&trans, &iter, &discard_pos_done,
- &seen,
- &open,
- &need_journal_commit,
- &discarded));
+ ret = bch2_trans_run(c,
+ for_each_btree_key(trans, iter,
+ BTREE_ID_need_discard, POS_MIN, 0, k,
+ bch2_discard_one_bucket(trans, &iter, &discard_pos_done, &s)));
- bch2_trans_exit(&trans);
+ discard_buckets_next_dev(c, &s, NULL);
- if (need_journal_commit * 2 > seen)
- bch2_journal_flush_async(&c->journal, NULL);
+ trace_discard_buckets(c, s.seen, s.open, s.need_journal_commit, s.discarded,
+ bch2_err_str(ret));
bch2_write_ref_put(c, BCH_WRITE_REF_discard);
-
- trace_discard_buckets(c, seen, open, need_journal_commit, discarded,
- bch2_err_str(ret));
}
void bch2_do_discards(struct bch_fs *c)
bch2_write_ref_put(c, BCH_WRITE_REF_discard);
}
+static int bch2_clear_bucket_needs_discard(struct btree_trans *trans, struct bpos bucket)
+{
+ struct btree_iter iter;
+ bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc, bucket, BTREE_ITER_INTENT);
+ struct bkey_s_c k = bch2_btree_iter_peek_slot(&iter);
+ int ret = bkey_err(k);
+ if (ret)
+ goto err;
+
+ struct bkey_i_alloc_v4 *a = bch2_alloc_to_v4_mut(trans, k);
+ ret = PTR_ERR_OR_ZERO(a);
+ if (ret)
+ goto err;
+
+ SET_BCH_ALLOC_V4_NEED_DISCARD(&a->v, false);
+ a->v.data_type = alloc_data_type(a->v, a->v.data_type);
+
+ ret = bch2_trans_update(trans, &iter, &a->k_i, 0);
+err:
+ bch2_trans_iter_exit(trans, &iter);
+ return ret;
+}
+
+static void bch2_do_discards_fast_work(struct work_struct *work)
+{
+ struct bch_fs *c = container_of(work, struct bch_fs, discard_fast_work);
+
+ while (1) {
+ bool got_bucket = false;
+ struct bpos bucket;
+ struct bch_dev *ca;
+
+ mutex_lock(&c->discard_buckets_in_flight_lock);
+ darray_for_each(c->discard_buckets_in_flight, i) {
+ if (i->snapshot)
+ continue;
+
+ ca = bch_dev_bkey_exists(c, i->inode);
+
+ if (!percpu_ref_tryget(&ca->io_ref)) {
+ darray_remove_item(&c->discard_buckets_in_flight, i);
+ continue;
+ }
+
+ got_bucket = true;
+ bucket = *i;
+ i->snapshot = true;
+ break;
+ }
+ mutex_unlock(&c->discard_buckets_in_flight_lock);
+
+ if (!got_bucket)
+ break;
+
+ blkdev_issue_discard(ca->disk_sb.bdev,
+ bucket.offset * ca->mi.bucket_size,
+ ca->mi.bucket_size,
+ GFP_KERNEL);
+
+ int ret = bch2_trans_do(c, NULL, NULL,
+ BCH_WATERMARK_btree|
+ BCH_TRANS_COMMIT_no_enospc,
+ bch2_clear_bucket_needs_discard(trans, bucket));
+ bch_err_fn(c, ret);
+
+ percpu_ref_put(&ca->io_ref);
+ discard_in_flight_remove(c, bucket);
+
+ if (ret)
+ break;
+ }
+
+ bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast);
+}
+
+static void bch2_discard_one_bucket_fast(struct bch_fs *c, struct bpos bucket)
+{
+ struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode);
+
+ if (!percpu_ref_is_dying(&ca->io_ref) &&
+ !discard_in_flight_add(c, bucket) &&
+ bch2_write_ref_tryget(c, BCH_WRITE_REF_discard_fast) &&
+ !queue_work(c->write_ref_wq, &c->discard_fast_work))
+ bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast);
+}
+
static int invalidate_one_bucket(struct btree_trans *trans,
struct btree_iter *lru_iter,
struct bkey_s_c lru_k,
BTREE_TRIGGER_BUCKET_INVALIDATE) ?:
bch2_trans_commit(trans, NULL, NULL,
BCH_WATERMARK_btree|
- BTREE_INSERT_NOFAIL);
+ BCH_TRANS_COMMIT_no_enospc);
if (ret)
goto out;
static void bch2_do_invalidates_work(struct work_struct *work)
{
struct bch_fs *c = container_of(work, struct bch_fs, invalidate_work);
- struct bch_dev *ca;
- struct btree_trans trans;
- struct btree_iter iter;
- struct bkey_s_c k;
- unsigned i;
+ struct btree_trans *trans = bch2_trans_get(c);
int ret = 0;
- bch2_trans_init(&trans, c, 0, 0);
-
- ret = bch2_btree_write_buffer_flush(&trans);
+ ret = bch2_btree_write_buffer_tryflush(trans);
if (ret)
goto err;
- for_each_member_device(ca, c, i) {
+ for_each_member_device(c, ca) {
s64 nr_to_invalidate =
should_invalidate_buckets(ca, bch2_dev_usage_read(ca));
- ret = for_each_btree_key2_upto(&trans, iter, BTREE_ID_lru,
+ ret = for_each_btree_key_upto(trans, iter, BTREE_ID_lru,
lru_pos(ca->dev_idx, 0, 0),
lru_pos(ca->dev_idx, U64_MAX, LRU_TIME_MAX),
BTREE_ITER_INTENT, k,
- invalidate_one_bucket(&trans, &iter, k, &nr_to_invalidate));
+ invalidate_one_bucket(trans, &iter, k, &nr_to_invalidate));
if (ret < 0) {
percpu_ref_put(&ca->ref);
}
}
err:
- bch2_trans_exit(&trans);
+ bch2_trans_put(trans);
bch2_write_ref_put(c, BCH_WRITE_REF_invalidate);
}
bch2_write_ref_put(c, BCH_WRITE_REF_invalidate);
}
-static int bch2_dev_freespace_init(struct bch_fs *c, struct bch_dev *ca,
- unsigned long *last_updated)
+int bch2_dev_freespace_init(struct bch_fs *c, struct bch_dev *ca,
+ u64 bucket_start, u64 bucket_end)
{
- struct btree_trans trans;
+ struct btree_trans *trans = bch2_trans_get(c);
struct btree_iter iter;
struct bkey_s_c k;
struct bkey hole;
- struct bpos end = POS(ca->dev_idx, ca->mi.nbuckets);
+ struct bpos end = POS(ca->dev_idx, bucket_end);
struct bch_member *m;
+ unsigned long last_updated = jiffies;
int ret;
- bch2_trans_init(&trans, c, 0, 0);
+ BUG_ON(bucket_start > bucket_end);
+ BUG_ON(bucket_end > ca->mi.nbuckets);
- bch2_trans_iter_init(&trans, &iter, BTREE_ID_alloc,
- POS(ca->dev_idx, ca->mi.first_bucket),
- BTREE_ITER_PREFETCH);
+ bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc,
+ POS(ca->dev_idx, max_t(u64, ca->mi.first_bucket, bucket_start)),
+ BTREE_ITER_PREFETCH);
/*
* Scan the alloc btree for every bucket on @ca, and add buckets to the
* freespace/need_discard/need_gc_gens btrees as needed:
*/
while (1) {
- if (*last_updated + HZ * 10 < jiffies) {
+ if (last_updated + HZ * 10 < jiffies) {
bch_info(ca, "%s: currently at %llu/%llu",
__func__, iter.pos.offset, ca->mi.nbuckets);
- *last_updated = jiffies;
+ last_updated = jiffies;
}
- bch2_trans_begin(&trans);
+ bch2_trans_begin(trans);
if (bkey_ge(iter.pos, end)) {
ret = 0;
struct bch_alloc_v4 a_convert;
const struct bch_alloc_v4 *a = bch2_alloc_to_v4(k, &a_convert);
- ret = bch2_bucket_do_index(&trans, k, a, true) ?:
- bch2_trans_commit(&trans, NULL, NULL,
- BTREE_INSERT_LAZY_RW|
- BTREE_INSERT_NOFAIL);
+ ret = bch2_bucket_do_index(trans, k, a, true) ?:
+ bch2_trans_commit(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc);
if (ret)
goto bkey_err;
} else {
struct bkey_i *freespace;
- freespace = bch2_trans_kmalloc(&trans, sizeof(*freespace));
+ freespace = bch2_trans_kmalloc(trans, sizeof(*freespace));
ret = PTR_ERR_OR_ZERO(freespace);
if (ret)
goto bkey_err;
freespace->k.p = k.k->p;
freespace->k.size = k.k->size;
- ret = __bch2_btree_insert(&trans, BTREE_ID_freespace, freespace, 0) ?:
- bch2_trans_commit(&trans, NULL, NULL,
- BTREE_INSERT_LAZY_RW|
- BTREE_INSERT_NOFAIL);
+ ret = bch2_btree_insert_trans(trans, BTREE_ID_freespace, freespace, 0) ?:
+ bch2_trans_commit(trans, NULL, NULL,
+ BCH_TRANS_COMMIT_no_enospc);
if (ret)
goto bkey_err;
break;
}
- bch2_trans_iter_exit(&trans, &iter);
- bch2_trans_exit(&trans);
+ bch2_trans_iter_exit(trans, &iter);
+ bch2_trans_put(trans);
if (ret < 0) {
- bch_err(ca, "error initializing free space: %s", bch2_err_str(ret));
+ bch_err_msg(ca, ret, "initializing free space");
return ret;
}
mutex_lock(&c->sb_lock);
- m = bch2_sb_get_members(c->disk_sb.sb)->members + ca->dev_idx;
+ m = bch2_members_v2_get_mut(c->disk_sb.sb, ca->dev_idx);
SET_BCH_MEMBER_FREESPACE_INITIALIZED(m, true);
mutex_unlock(&c->sb_lock);
int bch2_fs_freespace_init(struct bch_fs *c)
{
- struct bch_dev *ca;
- unsigned i;
int ret = 0;
bool doing_init = false;
- unsigned long last_updated = jiffies;
/*
* We can crash during the device add path, so we need to check this on
* every mount:
*/
- for_each_member_device(ca, c, i) {
+ for_each_member_device(c, ca) {
if (ca->mi.freespace_initialized)
continue;
doing_init = true;
}
- ret = bch2_dev_freespace_init(c, ca, &last_updated);
+ ret = bch2_dev_freespace_init(c, ca, 0, ca->mi.nbuckets);
if (ret) {
percpu_ref_put(&ca->ref);
bch_err_fn(c, ret);
void bch2_recalc_capacity(struct bch_fs *c)
{
- struct bch_dev *ca;
u64 capacity = 0, reserved_sectors = 0, gc_reserve;
unsigned bucket_size_max = 0;
unsigned long ra_pages = 0;
- unsigned i;
lockdep_assert_held(&c->state_lock);
- for_each_online_member(ca, c, i) {
+ for_each_online_member(c, ca) {
struct backing_dev_info *bdi = ca->disk_sb.bdev->bd_disk->bdi;
ra_pages += bdi->ra_pages;
bch2_set_ra_pages(c, ra_pages);
- for_each_rw_member(ca, c, i) {
+ for_each_rw_member(c, ca) {
u64 dev_reserve = 0;
/*
closure_wake_up(&c->freelist_wait);
}
+u64 bch2_min_rw_member_capacity(struct bch_fs *c)
+{
+ u64 ret = U64_MAX;
+
+ for_each_rw_member(c, ca)
+ ret = min(ret, ca->mi.nbuckets * ca->mi.bucket_size);
+ return ret;
+}
+
static bool bch2_dev_has_open_write_point(struct bch_fs *c, struct bch_dev *ca)
{
struct open_bucket *ob;
set_bit(ca->dev_idx, c->rw_devs[i].d);
}
+void bch2_fs_allocator_background_exit(struct bch_fs *c)
+{
+ darray_exit(&c->discard_buckets_in_flight);
+}
+
void bch2_fs_allocator_background_init(struct bch_fs *c)
{
spin_lock_init(&c->freelist_lock);
+ mutex_init(&c->discard_buckets_in_flight_lock);
INIT_WORK(&c->discard_work, bch2_do_discards_work);
+ INIT_WORK(&c->discard_fast_work, bch2_do_discards_fast_work);
INIT_WORK(&c->invalidate_work, bch2_do_invalidates_work);
}