/*
* should _only_ called from journal_res_get() - when we actually want a
* journal reservation - journal entry is open means journal is dirty:
- *
- * returns:
- * 0: success
- * -ENOSPC: journal currently full, must invoke reclaim
- * -EAGAIN: journal blocked, must wait
- * -EROFS: insufficient rw devices or journal error
*/
static int journal_entry_open(struct journal *j)
{
if (!fifo_free(&j->pin))
return JOURNAL_ERR_journal_pin_full;
- if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf) - 1)
+ if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf))
return JOURNAL_ERR_max_in_flight;
BUG_ON(!j->cur_entry_sectors);
journal_entry_overhead(j);
u64s = clamp_t(int, u64s, 0, JOURNAL_ENTRY_CLOSED_VAL - 1);
- if (u64s <= 0)
+ if (u64s <= (ssize_t) j->early_journal_entries.nr)
return JOURNAL_ERR_journal_full;
if (fifo_empty(&j->pin) && j->reclaim_thread)
buf->data->seq = cpu_to_le64(journal_cur_seq(j));
buf->data->u64s = 0;
+ if (j->early_journal_entries.nr) {
+ memcpy(buf->data->_data, j->early_journal_entries.data,
+ j->early_journal_entries.nr * sizeof(u64));
+ le32_add_cpu(&buf->data->u64s, j->early_journal_entries.nr);
+ }
+
/*
* Must be set before marking the journal entry as open:
*/
BUG_ON(new.idx != (journal_cur_seq(j) & JOURNAL_BUF_MASK));
journal_state_inc(&new);
- new.cur_entry_offset = 0;
+
+ /* Handle any already added entries */
+ new.cur_entry_offset = le32_to_cpu(buf->data->u64s);
} while ((v = atomic64_cmpxchg(&j->reservations.counter,
old.v, new.v)) != old.v);
&j->write_work,
msecs_to_jiffies(c->opts.journal_flush_delay));
journal_wake(j);
+
+ if (j->early_journal_entries.nr)
+ darray_exit(&j->early_journal_entries);
return 0;
}
return 0;
if (bch2_journal_error(j))
- return -EROFS;
+ return -BCH_ERR_erofs_journal_err;
spin_lock(&j->lock);
ret = journal_entry_open(j);
if (ret == JOURNAL_ERR_max_in_flight)
- trace_journal_entry_full(c);
+ trace_and_count(c, journal_entry_full, c);
unlock:
if ((ret && ret != JOURNAL_ERR_insufficient_devices) &&
!j->res_get_blocked_start) {
j->res_get_blocked_start = local_clock() ?: 1;
- trace_journal_full(c);
+ trace_and_count(c, journal_full, c);
}
can_discard = j->can_discard;
}
}
- return ret == JOURNAL_ERR_insufficient_devices ? -EROFS : -EAGAIN;
+ return ret == JOURNAL_ERR_insufficient_devices
+ ? -EROFS
+ : -BCH_ERR_journal_res_get_blocked;
}
/*
int ret;
closure_wait_event(&j->async_wait,
- (ret = __journal_res_get(j, res, flags)) != -EAGAIN ||
+ (ret = __journal_res_get(j, res, flags)) !=
+ -BCH_ERR_journal_res_get_blocked||
(flags & JOURNAL_RES_GET_NONBLOCK));
return ret;
}
return bch2_journal_flush_seq(j, res.seq);
}
-int bch2_journal_log_msg(struct journal *j, const char *fmt, ...)
-{
- struct jset_entry_log *entry;
- struct journal_res res = { 0 };
- unsigned msglen, u64s;
- va_list args;
- int ret;
-
- va_start(args, fmt);
- msglen = vsnprintf(NULL, 0, fmt, args) + 1;
- va_end(args);
-
- u64s = jset_u64s(DIV_ROUND_UP(msglen, sizeof(u64)));
-
- ret = bch2_journal_res_get(j, &res, u64s, 0);
- if (ret)
- return ret;
-
- entry = container_of(journal_res_entry(j, &res),
- struct jset_entry_log, entry);;
- memset(entry, 0, u64s * sizeof(u64));
- entry->entry.type = BCH_JSET_ENTRY_log;
- entry->entry.u64s = u64s - 1;
-
- va_start(args, fmt);
- vsnprintf(entry->d, INT_MAX, fmt, args);
- va_end(args);
-
- bch2_journal_res_put(j, &res);
-
- return bch2_journal_flush_seq(j, res.seq);
-}
-
/* block/unlock the journal: */
void bch2_journal_unblock(struct journal *j)
bch2_journal_block(&c->journal);
}
- bu = kzalloc(nr_want * sizeof(*bu), GFP_KERNEL);
- ob = kzalloc(nr_want * sizeof(*ob), GFP_KERNEL);
- new_buckets = kzalloc(nr * sizeof(u64), GFP_KERNEL);
- new_bucket_seq = kzalloc(nr * sizeof(u64), GFP_KERNEL);
+ bu = kcalloc(nr_want, sizeof(*bu), GFP_KERNEL);
+ ob = kcalloc(nr_want, sizeof(*ob), GFP_KERNEL);
+ new_buckets = kcalloc(nr, sizeof(u64), GFP_KERNEL);
+ new_bucket_seq = kcalloc(nr, sizeof(u64), GFP_KERNEL);
if (!bu || !ob || !new_buckets || !new_bucket_seq) {
ret = -ENOMEM;
goto err_unblock;
if (new_fs) {
bu[nr_got] = bch2_bucket_alloc_new_fs(ca);
if (bu[nr_got] < 0) {
- ret = -ENOSPC;
+ ret = -BCH_ERR_ENOSPC_bucket_alloc;
break;
}
} else {
ob[nr_got] = bch2_bucket_alloc(c, ca, RESERVE_none,
false, cl);
- if (IS_ERR(ob[nr_got])) {
- ret = cl ? -EAGAIN : -ENOSPC;
+ ret = PTR_ERR_OR_ZERO(ob[nr_got]);
+ if (ret)
break;
- }
bu[nr_got] = ob[nr_got]->bucket;
}
if (!new_fs) {
for (i = 0; i < nr_got; i++) {
- ret = bch2_trans_do(c, NULL, NULL, BTREE_INSERT_NOFAIL,
+ ret = bch2_trans_run(c,
bch2_trans_mark_metadata_bucket(&trans, ca,
bu[i], BCH_DATA_journal,
ca->mi.bucket_size));
closure_init_stack(&cl);
- while (ja->nr != nr && (ret == 0 || ret == -EAGAIN)) {
+ while (ja->nr != nr && (ret == 0 || ret == -BCH_ERR_bucket_alloc_blocked)) {
struct disk_reservation disk_res = { 0, 0 };
closure_sync(&cl);
* reservation to ensure we'll actually be able to allocate:
*/
- if (bch2_disk_reservation_get(c, &disk_res,
- bucket_to_sector(ca, nr - ja->nr), 1, 0)) {
+ ret = bch2_disk_reservation_get(c, &disk_res,
+ bucket_to_sector(ca, nr - ja->nr), 1, 0);
+ if (ret) {
mutex_unlock(&c->sb_lock);
- return -ENOSPC;
+ return ret;
}
ret = __bch2_set_nr_journal_buckets(ca, nr, false, &cl);
bch2_sb_get_journal(sb);
struct bch_sb_field_journal_v2 *journal_buckets_v2 =
bch2_sb_get_journal_v2(sb);
- unsigned i;
+ unsigned i, nr_bvecs;
ja->nr = 0;
if (!ja->bucket_seq)
return -ENOMEM;
- ca->journal.bio = bio_kmalloc(GFP_KERNEL,
- DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE));
+ nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE);
+
+ ca->journal.bio = bio_kmalloc(nr_bvecs, GFP_KERNEL);
if (!ca->journal.bio)
return -ENOMEM;
+ bio_init(ca->journal.bio, NULL, ca->journal.bio->bi_inline_vecs, nr_bvecs, 0);
+
ja->buckets = kcalloc(ja->nr, sizeof(u64), GFP_KERNEL);
if (!ja->buckets)
return -ENOMEM;
{
unsigned i;
+ darray_exit(&j->early_journal_entries);
+
for (i = 0; i < ARRAY_SIZE(j->buf); i++)
kvpfree(j->buf[i].data, j->buf[i].buf_size);
free_fifo(&j->pin);
u64 seq;
unsigned i;
+ if (!out->nr_tabstops)
+ printbuf_tabstop_push(out, 24);
out->atomic++;
- out->tabstops[0] = 24;
rcu_read_lock();
s = READ_ONCE(j->reservations);
- pr_buf(out, "dirty journal entries:\t%llu/%llu\n",fifo_used(&j->pin), j->pin.size);
- pr_buf(out, "seq:\t\t\t%llu\n", journal_cur_seq(j));
- pr_buf(out, "seq_ondisk:\t\t%llu\n", j->seq_ondisk);
- pr_buf(out, "last_seq:\t\t%llu\n", journal_last_seq(j));
- pr_buf(out, "last_seq_ondisk:\t%llu\n", j->last_seq_ondisk);
- pr_buf(out, "flushed_seq_ondisk:\t%llu\n", j->flushed_seq_ondisk);
- pr_buf(out, "prereserved:\t\t%u/%u\n", j->prereserved.reserved, j->prereserved.remaining);
- pr_buf(out, "watermark:\t\t%s\n", bch2_journal_watermarks[j->watermark]);
- pr_buf(out, "each entry reserved:\t%u\n", j->entry_u64s_reserved);
- pr_buf(out, "nr flush writes:\t%llu\n", j->nr_flush_writes);
- pr_buf(out, "nr noflush writes:\t%llu\n", j->nr_noflush_writes);
- pr_buf(out, "nr direct reclaim:\t%llu\n", j->nr_direct_reclaim);
- pr_buf(out, "nr background reclaim:\t%llu\n", j->nr_background_reclaim);
- pr_buf(out, "reclaim kicked:\t\t%u\n", j->reclaim_kicked);
- pr_buf(out, "reclaim runs in:\t%u ms\n", time_after(j->next_reclaim, now)
+ prt_printf(out, "dirty journal entries:\t%llu/%llu\n", fifo_used(&j->pin), j->pin.size);
+ prt_printf(out, "seq:\t\t\t%llu\n", journal_cur_seq(j));
+ prt_printf(out, "seq_ondisk:\t\t%llu\n", j->seq_ondisk);
+ prt_printf(out, "last_seq:\t\t%llu\n", journal_last_seq(j));
+ prt_printf(out, "last_seq_ondisk:\t%llu\n", j->last_seq_ondisk);
+ prt_printf(out, "flushed_seq_ondisk:\t%llu\n", j->flushed_seq_ondisk);
+ prt_printf(out, "prereserved:\t\t%u/%u\n", j->prereserved.reserved, j->prereserved.remaining);
+ prt_printf(out, "watermark:\t\t%s\n", bch2_journal_watermarks[j->watermark]);
+ prt_printf(out, "each entry reserved:\t%u\n", j->entry_u64s_reserved);
+ prt_printf(out, "nr flush writes:\t%llu\n", j->nr_flush_writes);
+ prt_printf(out, "nr noflush writes:\t%llu\n", j->nr_noflush_writes);
+ prt_printf(out, "nr direct reclaim:\t%llu\n", j->nr_direct_reclaim);
+ prt_printf(out, "nr background reclaim:\t%llu\n", j->nr_background_reclaim);
+ prt_printf(out, "reclaim kicked:\t\t%u\n", j->reclaim_kicked);
+ prt_printf(out, "reclaim runs in:\t%u ms\n", time_after(j->next_reclaim, now)
? jiffies_to_msecs(j->next_reclaim - jiffies) : 0);
- pr_buf(out, "current entry sectors:\t%u\n", j->cur_entry_sectors);
- pr_buf(out, "current entry error:\t%s\n", bch2_journal_errors[j->cur_entry_error]);
- pr_buf(out, "current entry:\t\t");
+ prt_printf(out, "current entry sectors:\t%u\n", j->cur_entry_sectors);
+ prt_printf(out, "current entry error:\t%s\n", bch2_journal_errors[j->cur_entry_error]);
+ prt_printf(out, "current entry:\t\t");
switch (s.cur_entry_offset) {
case JOURNAL_ENTRY_ERROR_VAL:
- pr_buf(out, "error");
+ prt_printf(out, "error");
break;
case JOURNAL_ENTRY_CLOSED_VAL:
- pr_buf(out, "closed");
+ prt_printf(out, "closed");
break;
default:
- pr_buf(out, "%u/%u", s.cur_entry_offset, j->cur_entry_u64s);
+ prt_printf(out, "%u/%u", s.cur_entry_offset, j->cur_entry_u64s);
break;
}
- pr_newline(out);
+ prt_newline(out);
for (seq = journal_cur_seq(j);
seq >= journal_last_unwritten_seq(j);
--seq) {
i = seq & JOURNAL_BUF_MASK;
- pr_buf(out, "unwritten entry:");
- pr_tab(out);
- pr_buf(out, "%llu", seq);
- pr_newline(out);
- pr_indent_push(out, 2);
+ prt_printf(out, "unwritten entry:");
+ prt_tab(out);
+ prt_printf(out, "%llu", seq);
+ prt_newline(out);
+ printbuf_indent_add(out, 2);
- pr_buf(out, "refcount:");
- pr_tab(out);
- pr_buf(out, "%u", journal_state_count(s, i));
- pr_newline(out);
+ prt_printf(out, "refcount:");
+ prt_tab(out);
+ prt_printf(out, "%u", journal_state_count(s, i));
+ prt_newline(out);
- pr_buf(out, "sectors:");
- pr_tab(out);
- pr_buf(out, "%u", j->buf[i].sectors);
- pr_newline(out);
+ prt_printf(out, "sectors:");
+ prt_tab(out);
+ prt_printf(out, "%u", j->buf[i].sectors);
+ prt_newline(out);
- pr_buf(out, "expires");
- pr_tab(out);
- pr_buf(out, "%li jiffies", j->buf[i].expires - jiffies);
- pr_newline(out);
+ prt_printf(out, "expires");
+ prt_tab(out);
+ prt_printf(out, "%li jiffies", j->buf[i].expires - jiffies);
+ prt_newline(out);
- pr_indent_pop(out, 2);
+ printbuf_indent_sub(out, 2);
}
- pr_buf(out,
+ prt_printf(out,
"replay done:\t\t%i\n",
test_bit(JOURNAL_REPLAY_DONE, &j->flags));
- pr_buf(out, "space:\n");
- pr_buf(out, "\tdiscarded\t%u:%u\n",
+ prt_printf(out, "space:\n");
+ prt_printf(out, "\tdiscarded\t%u:%u\n",
j->space[journal_space_discarded].next_entry,
j->space[journal_space_discarded].total);
- pr_buf(out, "\tclean ondisk\t%u:%u\n",
+ prt_printf(out, "\tclean ondisk\t%u:%u\n",
j->space[journal_space_clean_ondisk].next_entry,
j->space[journal_space_clean_ondisk].total);
- pr_buf(out, "\tclean\t\t%u:%u\n",
+ prt_printf(out, "\tclean\t\t%u:%u\n",
j->space[journal_space_clean].next_entry,
j->space[journal_space_clean].total);
- pr_buf(out, "\ttotal\t\t%u:%u\n",
+ prt_printf(out, "\ttotal\t\t%u:%u\n",
j->space[journal_space_total].next_entry,
j->space[journal_space_total].total);
if (!ja->nr)
continue;
- pr_buf(out, "dev %u:\n", i);
- pr_buf(out, "\tnr\t\t%u\n", ja->nr);
- pr_buf(out, "\tbucket size\t%u\n", ca->mi.bucket_size);
- pr_buf(out, "\tavailable\t%u:%u\n", bch2_journal_dev_buckets_available(j, ja, journal_space_discarded), ja->sectors_free);
- pr_buf(out, "\tdiscard_idx\t%u\n", ja->discard_idx);
- pr_buf(out, "\tdirty_ondisk\t%u (seq %llu)\n", ja->dirty_idx_ondisk, ja->bucket_seq[ja->dirty_idx_ondisk]);
- pr_buf(out, "\tdirty_idx\t%u (seq %llu)\n", ja->dirty_idx, ja->bucket_seq[ja->dirty_idx]);
- pr_buf(out, "\tcur_idx\t\t%u (seq %llu)\n", ja->cur_idx, ja->bucket_seq[ja->cur_idx]);
+ prt_printf(out, "dev %u:\n", i);
+ prt_printf(out, "\tnr\t\t%u\n", ja->nr);
+ prt_printf(out, "\tbucket size\t%u\n", ca->mi.bucket_size);
+ prt_printf(out, "\tavailable\t%u:%u\n", bch2_journal_dev_buckets_available(j, ja, journal_space_discarded), ja->sectors_free);
+ prt_printf(out, "\tdiscard_idx\t%u\n", ja->discard_idx);
+ prt_printf(out, "\tdirty_ondisk\t%u (seq %llu)\n", ja->dirty_idx_ondisk, ja->bucket_seq[ja->dirty_idx_ondisk]);
+ prt_printf(out, "\tdirty_idx\t%u (seq %llu)\n", ja->dirty_idx, ja->bucket_seq[ja->dirty_idx]);
+ prt_printf(out, "\tcur_idx\t\t%u (seq %llu)\n", ja->cur_idx, ja->bucket_seq[ja->cur_idx]);
}
rcu_read_unlock();
pin_list = journal_seq_pin(j, *seq);
- pr_buf(out, "%llu: count %u", *seq, atomic_read(&pin_list->count));
- pr_newline(out);
- pr_indent_push(out, 2);
+ prt_printf(out, "%llu: count %u", *seq, atomic_read(&pin_list->count));
+ prt_newline(out);
+ printbuf_indent_add(out, 2);
list_for_each_entry(pin, &pin_list->list, list) {
- pr_buf(out, "\t%px %ps", pin, pin->flush);
- pr_newline(out);
+ prt_printf(out, "\t%px %ps", pin, pin->flush);
+ prt_newline(out);
}
list_for_each_entry(pin, &pin_list->key_cache_list, list) {
- pr_buf(out, "\t%px %ps", pin, pin->flush);
- pr_newline(out);
+ prt_printf(out, "\t%px %ps", pin, pin->flush);
+ prt_newline(out);
}
if (!list_empty(&pin_list->flushed)) {
- pr_buf(out, "flushed:");
- pr_newline(out);
+ prt_printf(out, "flushed:");
+ prt_newline(out);
}
list_for_each_entry(pin, &pin_list->flushed, list) {
- pr_buf(out, "\t%px %ps", pin, pin->flush);
- pr_newline(out);
+ prt_printf(out, "\t%px %ps", pin, pin->flush);
+ prt_newline(out);
}
- pr_indent_pop(out, 2);
+ printbuf_indent_sub(out, 2);
--out->atomic;
spin_unlock(&j->lock);