f2fs: add bitmaps for empty or full NAT blocks

This patches adds bitmaps to represent empty or full NAT blocks containing
free nid entries.

If we can find valid crc|cp_ver in the last block of checkpoint pack, we'll
use these bitmaps when building free nids. In order to avoid checkpointing
burden, up-to-date bitmaps will be flushed only during umount time. So,
normally we can get this gain, but when power-cut happens, we rely on fsck.f2fs
which recovers this bitmap again.

After this patch, we build free nids from nid #0 at mount time to make more
full NAT blocks, but in runtime, we check empty NAT blocks to load free nids
without loading any NAT pages from disk.

Signed-off-by: Jaegeuk Kim <[email protected]>

fs/f2fs/checkpoint.c

@@ -1024,6 +1024,10 @@ static void update_ckpt_flags(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 
 	spin_lock(&sbi->cp_lock);
 
+	if (cpc->reason == CP_UMOUNT && ckpt->cp_pack_total_block_count >
+			sbi->blocks_per_seg - NM_I(sbi)->nat_bits_blocks)
+		disable_nat_bits(sbi, false);
+
 	if (cpc->reason == CP_UMOUNT)
 		__set_ckpt_flags(ckpt, CP_UMOUNT_FLAG);
 	else
@@ -1136,6 +1140,28 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 
 	start_blk = __start_cp_next_addr(sbi);
 
+	/* write nat bits */
+	if (enabled_nat_bits(sbi, cpc)) {
+		__u64 cp_ver = cur_cp_version(ckpt);
+		unsigned int i;
+		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++)
+			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)) {
+			sync_meta_pages(sbi, META, LONG_MAX);
+			if (unlikely(f2fs_cp_error(sbi)))
+				return -EIO;
+		}
+	}
+
 	/* need to wait for end_io results */
 	wait_on_all_pages_writeback(sbi);
 	if (unlikely(f2fs_cp_error(sbi)))
@@ -1272,7 +1298,7 @@ int write_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	ckpt->checkpoint_ver = cpu_to_le64(++ckpt_ver);
 
 	/* write cached NAT/SIT entries to NAT/SIT area */
-	flush_nat_entries(sbi);
+	flush_nat_entries(sbi, cpc);
 	flush_sit_entries(sbi, cpc);
 
 	/* unlock all the fs_lock[] in do_checkpoint() */

fs/f2fs/debug.c

@@ -193,6 +193,7 @@ static void update_mem_info(struct f2fs_sb_info *sbi)
 	/* build nm */
 	si->base_mem += sizeof(struct f2fs_nm_info);
 	si->base_mem += __bitmap_size(sbi, NAT_BITMAP);
+	si->base_mem += (NM_I(sbi)->nat_bits_blocks << F2FS_BLKSIZE_BITS);
 
 get_cache:
 	si->cache_mem = 0;

fs/f2fs/f2fs.h

@@ -554,6 +554,7 @@ struct f2fs_nm_info {
 	struct list_head nat_entries;	/* cached nat entry list (clean) */
 	unsigned int nat_cnt;		/* the # of cached nat entries */
 	unsigned int dirty_nat_cnt;	/* total num of nat entries in set */
+	unsigned int nat_blocks;	/* # of nat blocks */
 
 	/* free node ids management */
 	struct radix_tree_root free_nid_root;/* root of the free_nid cache */
@@ -564,6 +565,11 @@ struct f2fs_nm_info {
 
 	/* for checkpoint */
 	char *nat_bitmap;		/* NAT bitmap pointer */
+
+	unsigned int nat_bits_blocks;	/* # of nat bits blocks */
+	unsigned char *nat_bits;	/* NAT bits blocks */
+	unsigned char *full_nat_bits;	/* full NAT pages */
+	unsigned char *empty_nat_bits;	/* empty NAT pages */
 #ifdef CONFIG_F2FS_CHECK_FS
 	char *nat_bitmap_mir;		/* NAT bitmap mirror */
 #endif
@@ -1171,6 +1177,27 @@ static inline void clear_ckpt_flags(struct f2fs_sb_info *sbi, unsigned int f)
 	spin_unlock(&sbi->cp_lock);
 }
 
+static inline void disable_nat_bits(struct f2fs_sb_info *sbi, bool lock)
+{
+	set_sbi_flag(sbi, SBI_NEED_FSCK);
+
+	if (lock)
+		spin_lock(&sbi->cp_lock);
+	__clear_ckpt_flags(F2FS_CKPT(sbi), CP_NAT_BITS_FLAG);
+	kfree(NM_I(sbi)->nat_bits);
+	NM_I(sbi)->nat_bits = NULL;
+	if (lock)
+		spin_unlock(&sbi->cp_lock);
+}
+
+static inline bool enabled_nat_bits(struct f2fs_sb_info *sbi,
+					struct cp_control *cpc)
+{
+	bool set = is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG);
+
+	return (cpc) ? (cpc->reason == CP_UMOUNT) && set : set;
+}
+
 static inline void f2fs_lock_op(struct f2fs_sb_info *sbi)
 {
 	down_read(&sbi->cp_rwsem);
@@ -2131,7 +2158,7 @@ void move_node_page(struct page *node_page, int gc_type);
 int fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
 			struct writeback_control *wbc, bool atomic);
 int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc);
-void build_free_nids(struct f2fs_sb_info *sbi, bool sync);
+void build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount);
 bool alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid);
 void alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid);
 void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid);
@@ -2142,7 +2169,7 @@ int recover_xattr_data(struct inode *inode, struct page *page,
 int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page);
 int restore_node_summary(struct f2fs_sb_info *sbi,
 			unsigned int segno, struct f2fs_summary_block *sum);
-void flush_nat_entries(struct f2fs_sb_info *sbi);
+void flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc);
 int build_node_manager(struct f2fs_sb_info *sbi);
 void destroy_node_manager(struct f2fs_sb_info *sbi);
 int __init create_node_manager_caches(void);