f2fs 的checkpoint 维护data、node和meta data(SIT,NAT)的数据一致性，把一起写到SSA区域的数据分别写回到SIT/NAT区域。
checkpoint 相关数据结构
super block区域里记录了checkpoint （CP)的起始block address，以及checkpoint 区域segment 的数量：
struct f2fs_super_block {
__le32 magic; /* Magic Number */
__le16 major_ver; /* Major Version */
__le16 minor_ver; /* Minor Version */
__le32 log_sectorsize; /* log2 sector size in bytes */
__le32 log_sectors_per_block; /* log2 # of sectors per block */
__le32 log_blocksize; /* log2 block size in bytes */
__le32 log_blocks_per_seg; /* log2 # of blocks per segment */
__le32 segs_per_sec; /* # of segments per section */
__le32 secs_per_zone; /* # of sections per zone */
__le32 checksum_offset; /* checksum offset inside super block */
__le64 block_count; /* total # of user blocks */
__le32 section_count; /* total # of sections */
__le32 segment_count; /* total # of segments */
__le32 segment_count_ckpt; /* # of segments for checkpoint */
__le32 segment_count_sit; /* # of segments for SIT */
__le32 segment_count_nat; /* # of segments for NAT */
__le32 segment_count_ssa; /* # of segments for SSA */
__le32 segment_count_main; /* # of segments for main area */
__le32 segment0_blkaddr; /* start block address of segment 0 */
__le32 cp_blkaddr; /* start block address of checkpoint */
.............check point 包含的信息如下：
#define F2FS_CP_PACKS 2 /* # of checkpoint packs */
struct f2fs_checkpoint {
__le64 checkpoint_ver; /* checkpoint block version number */
__le64 user_block_count; /* # of user blocks */
__le64 valid_block_count; /* # of valid blocks in main area */
__le32 rsvd_segment_count; /* # of reserved segments for gc */
__le32 overprov_segment_count; /* # of overprovision segments */
__le32 free_segment_count; /* # of free segments in main area */
/* information of current node segments */
__le32 cur_node_segno[MAX_ACTIVE_NODE_LOGS];
__le16 cur_node_blkoff[MAX_ACTIVE_NODE_LOGS];
/* information of current data segments */
__le32 cur_data_segno[MAX_ACTIVE_DATA_LOGS];
__le16 cur_data_blkoff[MAX_ACTIVE_DATA_LOGS];
__le32 ckpt_flags; /* Flags : umount and journal_present */
__le32 cp_pack_total_block_count; /* total # of one cp pack */
__le32 cp_pack_start_sum; /* start block number of data summary */
__le32 valid_node_count; /* Total number of valid nodes */
__le32 valid_inode_count; /* Total number of valid inodes */
__le32 next_free_nid; /* Next free node number */
__le32 sit_ver_bitmap_bytesize; /* Default value 64 */
__le32 nat_ver_bitmap_bytesize; /* Default value 256 */
__le32 checksum_offset; /* checksum offset inside cp block */
__le64 elapsed_time; /* mounted time */
/* allocation type of current segment */
unsigned char alloc_type[MAX_ACTIVE_LOGS];
/* SIT and NAT version bitmap */
unsigned char sit_nat_version_bitmap[1];
} __packed;checkpoint 主要工作
把cached 的SIT/NAT 信息写回到相应的SIT / NAT 区域。
主要的工作在 f2fs_write_checkpoint 中完成：
int f2fs_write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
{
struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
unsigned long long ckpt_ver;
int err = 0;
............
err = block_operations(sbi);
............
/*
* update checkpoint pack index
* Increase the version number so that
* SIT entries and seg summaries are written at correct place
*/
ckpt_ver = cur_cp_version(ckpt);
ckpt->checkpoint_ver = cpu_to_le64(++ckpt_ver);
/* write cached NAT/SIT entries to NAT/SIT area */
f2fs_flush_nat_entries(sbi, cpc);
f2fs_flush_sit_entries(sbi, cpc);
/* unlock all the fs_lock[] in do_checkpoint() */
err = do_checkpoint(sbi, cpc);
if (err)
f2fs_release_discard_addrs(sbi);
else
f2fs_clear_prefree_segments(sbi, cpc);
unblock_operations(sbi);
stat_inc_cp_count(sbi->stat_info);
......
}可以看到上面首先用block_operations()来阻塞一下IO operations ：
/*
* Freeze all the FS-operations for checkpoint.
*/
static int block_operations(struct f2fs_sb_info *sbi)
{
......
/* write all the dirty dentry pages */ //flush dirty dentry pages
if (get_pages(sbi, F2FS_DIRTY_DENTS)) {
f2fs_unlock_all(sbi);
err = f2fs_sync_dirty_inodes(sbi, DIR_INODE);
if (err)
goto out;
cond_resched();
goto retry_flush_dents;
}
.....
if (get_pages(sbi, F2FS_DIRTY_IMETA)) {
up_write(&sbi->node_change);
f2fs_unlock_all(sbi);
err = f2fs_sync_inode_meta(sbi);
.........
...... // flush dirty nodes
if (get_pages(sbi, F2FS_DIRTY_NODES)) {
up_write(&sbi->node_write);
atomic_inc(&sbi->wb_sync_req[NODE]);
err = f2fs_sync_node_pages(sbi, &wbc, false, FS_CP_NODE_IO);
......
接着刷回缓存的 NAT entry到 NAT 区域，这是通过f2fs_flush_nat_entries来实现的：
void f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
{
struct f2fs_nm_info *nm_i = NM_I(sbi);
struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
struct f2fs_journal *journal = curseg->journal;
struct nat_entry_set *setvec[SETVEC_SIZE];
struct nat_entry_set *set, *tmp;
unsigned int found;
nid_t set_idx = 0;
LIST_HEAD(sets);
/* during unmount, let's flush nat_bits before checking dirty_nat_cnt */
if (enabled_nat_bits(sbi, cpc)) {
down_write(&nm_i->nat_tree_lock);
remove_nats_in_journal(sbi);
up_write(&nm_i->nat_tree_lock);
}
.....
/*
* if there are no enough space in journal to store dirty nat
* entries, remove all entries from journal and merge them
* into nat entry set.
*/
if (enabled_nat_bits(sbi, cpc) ||
!__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL))
remove_nats_in_journal(sbi);
while ((found = __gang_lookup_nat_set(nm_i,
set_idx, SETVEC_SIZE, setvec))) {
unsigned idx;
set_idx = setvec[found - 1]->set + 1;
for (idx = 0; idx < found; idx++)
__adjust_nat_entry_set(setvec[idx], &sets,
MAX_NAT_JENTRIES(journal));
}
/* flush dirty nats in nat entry set */
list_for_each_entry_safe(set, tmp, &sets, set_list)
__flush_nat_entry_set(sbi, set, cpc);
上面步骤完成之后，还刷回缓存的SIT entry到 SIT 区域，这是通过下面的函数去实现：
void f2fs_flush_sit_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
{
struct sit_info *sit_i = SIT_I(sbi);
unsigned long *bitmap = sit_i->dirty_sentries_bitmap;
struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_COLD_DATA);
struct f2fs_journal *journal = curseg->journal;
struct sit_entry_set *ses, *tmp;
struct list_head *head = &SM_I(sbi)->sit_entry_set;
bool to_journal = true;
struct seg_entry *se;
......
/*
* if there are no enough space in journal to store dirty sit
* entries, remove all entries from journal and add and account
* them in sit entry set.
*/
if (!__has_cursum_space(journal, sit_i->dirty_sentries, SIT_JOURNAL))
remove_sits_in_journal(sbi);
/*
* there are two steps to flush sit entries:
* #1, flush sit entries to journal in current cold data summary block.
* #2, flush sit entries to sit page.
*/
list_for_each_entry_safe(ses, tmp, head, set_list) {
......
/* flush dirty sit entries in region of current sit set */
for_each_set_bit_from(segno, bitmap, end) {
int offset, sit_offset;上面把 NAT／SIT 都刷完了之后，会执行最关键的操作， do_checkpoint, 它主要的操作如下：
/* Flush all the NAT/SIT pages */
while (get_pages(sbi, F2FS_DIRTY_META)) {
f2fs_sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO);
if (unlikely(f2fs_cp_error(sbi)))
break;
}
.....
/* write nat bits */
if (enabled_nat_bits(sbi, cpc)) {
__u64 cp_ver = cur_cp_version(ckpt);
block_t blk;
cp_ver |= ((__u64)crc32 << 32);
*(__le64 *)nm_i->nat_bits = cpu_to_le64(cp_ver);
blk = start_blk + sbi->blocks_per_seg - nm_i->nat_bits_blocks;
for (i = 0; i < nm_i->nat_bits_blocks; i++)
f2fs_update_meta_page(sbi, nm_i->nat_bits +
(i << F2FS_BLKSIZE_BITS), blk + i);
/* Flush all the NAT BITS pages */
while (get_pages(sbi, F2FS_DIRTY_META)) {
f2fs_sync_meta_pages(sbi, META, LONG_MAX,
FS_CP_META_IO);
if (unlikely(f2fs_cp_error(sbi)))
break;
}
}
/* write out checkpoint buffer at block 0 */
f2fs_update_meta_page(sbi, ckpt, start_blk++);
for (i = 1; i < 1 + cp_payload_blks; i++)
f2fs_update_meta_page(sbi, (char *)ckpt + i * F2FS_BLKSIZE,
start_blk++);
.....
f2fs_write_data_summaries(sbi, start_blk);
start_blk += data_sum_blocks;
.....
/* update user_block_counts */
sbi->last_valid_block_count = sbi->total_valid_block_count;
percpu_counter_set(&sbi->alloc_valid_block_count, 0);
/* Here, we have one bio having CP pack except cp pack 2 page */
f2fs_sync_meta_pages(sbi, META, LONG_MAX, FS_CP_META_IO);
/* wait for previous submitted meta pages writeback */
f2fs_wait_on_all_pages_writeback(sbi);
/* flush all device cache */
err = f2fs_flush_device_cache(sbi);
if (err)
return err;
/* barrier and flush checkpoint cp pack 2 page if it can */
commit_checkpoint(sbi, ckpt, start_blk);
f2fs_wait_on_all_pages_writeback(sbi);
......触发 check point的时机
- gc: 当空闲segment 不足的时候，可以通过check point 释放。
if (gc_type == BG_GC && has_not_enough_free_secs(sbi, 0, 0)) { /* * For example, if there are many prefree_segments below given * threshold, we can make them free by checkpoint. Then, we * secure free segments which doesn't need fggc any more. */ if (prefree_segments(sbi)) { ret = f2fs_write_checkpoint(sbi, &cpc); if (ret) goto stop; } if (has_not_enough_free_secs(sbi, 0, 0)) gc_type = FG_GC; } -
recover:
当 f2fs_recover_fsync_data()找到有需要fsync的segment的时候，会调用f2fs_write_checkpoint: -
trim:
f2fs_trim 通过设置CP的reason 是CP_DISCARD， 然后走f2fs_write_checkpoint 实现：cpc.reason = CP_DISCARD; cpc.trim_minlen = max_t(__u64, 1, F2FS_BYTES_TO_BLK(range->minlen)); cpc.trim_start = start_segno; cpc.trim_end = end_segno; if (sbi->discard_blks == 0) goto out; mutex_lock(&sbi->gc_mutex); err = f2fs_write_checkpoint(sbi, &cpc); mutex_unlock(&sbi->gc_mutex); -
super block
super block 的put_super 接口需要通过f2fs_write_checkpoint() 来把没有来得及umount导致的可能不一致的数据重新做一次check point。/* ....... * We don't need to do checkpoint when superblock is clean. * But, the previous checkpoint was not done by umount, it needs to do * clean checkpoint again. */ if (is_sbi_flag_set(sbi, SBI_IS_DIRTY) || !is_set_ckpt_flags(sbi, CP_UMOUNT_FLAG)) { struct cp_control cpc = { .reason = CP_UMOUNT, }; f2fs_write_checkpoint(sbi, &cpc); } /* be sure to wait for any on-going discard commands */ dropped = f2fs_wait_discard_bios(sbi); if (f2fs_discard_en(sbi) && !sbi->discard_blks && !dropped) { struct cp_control cpc = { .reason = CP_UMOUNT | CP_TRIMMED, }; .....