lander/lsm/src/store/lsm_store_disk_write.c

173 lines
4.4 KiB
C

#include "lsm/store_internal.h"
static lsm_error lsm_fwrite(uint64_t *sum, FILE *f, uint64_t size,
uint64_t count, const void *val) {
size_t res = fwrite(val, size, count, f);
if (res < count) {
return lsm_error_failed_io;
}
if (sum != NULL) {
*sum += size * count;
}
return lsm_error_ok;
}
static lsm_error lsm_write_str(uint64_t *sum, FILE *f, const lsm_str *s) {
uint64_t len = lsm_str_len(s);
LSM_RES(lsm_fwrite(sum, f, sizeof(uint64_t), 1, &len));
uint64_t written = 0;
do {
written += fwrite(lsm_str_ptr(s), sizeof(char), len - written, f);
} while (written < len);
if (sum != NULL) {
*sum += len * sizeof(char);
}
return lsm_error_ok;
}
static lsm_error lsm_fseek(FILE *f, uint64_t pos) {
if (fseek(f, pos, SEEK_SET) != 0) {
return lsm_error_failed_io;
}
return lsm_error_ok;
}
lsm_error lsm_write_db_entry(uint64_t *size, FILE *db_file, lsm_entry *entry,
uint64_t pos) {
*size = 0;
LSM_RES(lsm_fseek(db_file, pos));
LSM_RES(lsm_write_str(size, db_file, entry->key));
LSM_RES(lsm_fwrite(size, db_file, sizeof(uint64_t), 1, &entry->data_len));
LSM_RES(lsm_fwrite(size, db_file, sizeof(uint8_t), 1, &entry->attrs.count));
for (uint8_t i = 0; i < entry->attrs.count; i++) {
LSM_RES(lsm_fwrite(size, db_file, sizeof(uint8_t), 1,
&entry->attrs.items[i].type));
LSM_RES(lsm_write_str(size, db_file, entry->attrs.items[i].str));
}
return lsm_error_ok;
}
lsm_error lsm_write_idx_entry(uint64_t *size, FILE *idx_file, uint64_t offset,
uint64_t len, uint64_t pos) {
*size = 0;
LSM_RES(lsm_fseek(idx_file, pos));
LSM_RES(lsm_fwrite(size, idx_file, sizeof(uint64_t), 1, &offset));
LSM_RES(lsm_fwrite(size, idx_file, sizeof(uint64_t), 1, &len));
return lsm_error_ok;
}
lsm_error lsm_entry_disk_insert(lsm_entry_handle *handle) {
lsm_store *store = handle->store;
pthread_mutex_lock(&store->db.lock);
uint64_t db_entry_index = store->db.size;
uint64_t db_entry_size;
lsm_error res = lsm_write_db_entry(&db_entry_size, store->db.f,
handle->wrapper->entry, store->db.size);
fflush(store->db.f);
pthread_mutex_unlock(&store->db.lock);
if (res != lsm_error_ok) {
return res;
}
// Append entry to index file
pthread_mutex_lock(&store->idx.lock);
uint64_t idx_entry_index = store->idx.size;
uint64_t idx_entry_size;
res = lsm_write_idx_entry(&idx_entry_size, store->idx.f, db_entry_index,
db_entry_size, store->idx.size);
if (res == lsm_error_ok) {
// Update the counter at the beginning of the file
rewind(store->idx.f);
uint64_t new_block_count = store->idx.block_count + 1;
res = lsm_fwrite(NULL, store->idx.f, sizeof(uint64_t), 1, &new_block_count);
if (res == lsm_error_ok) {
// Only if we successfully updated the on-disk counter do we make the code
// aware that the files' sizes have increased. This way, if a write to the
// counter fails, the code will simply reuse the already written content.
store->idx.size += idx_entry_size;
store->idx.block_count = new_block_count;
store->db.size += db_entry_size;
handle->wrapper->entry->idx_file_offset = idx_entry_index;
}
}
fflush(store->idx.f);
pthread_mutex_unlock(&store->idx.lock);
return res;
}
// Marking an entry as removed in the idx file is simply setting the length of
// its entry to zero
lsm_error lsm_entry_disk_remove(lsm_entry_handle *handle) {
lsm_store *store = handle->store;
const lsm_entry *entry = handle->wrapper->entry;
pthread_mutex_lock(&store->idx.lock);
lsm_error res =
lsm_fseek(store->idx.f, entry->idx_file_offset + sizeof(uint64_t));
if (res != lsm_error_ok) {
pthread_mutex_unlock(&store->idx.lock);
return res;
}
uint64_t val = 0;
res = lsm_fwrite(NULL, store->idx.f, sizeof(uint64_t), 1, &val);
pthread_mutex_unlock(&store->idx.lock);
if (res != lsm_error_ok) {
return res;
}
fflush(store->idx.f);
// Remove data file if present
if (entry->data_len > 0) {
if (handle->f != NULL) {
fclose(handle->f);
handle->f = NULL;
}
char data_path[lsm_entry_data_path_len(handle) + 1];
lsm_entry_data_path(data_path, handle);
if (remove(data_path) != 0) {
return lsm_error_failed_io;
}
}
return lsm_error_ok;
}