f2fs: rebuild nat_bits during umount

If all free_nat_bitmap are available, we can rebuild nat_bits from
free_nat_bitmap entirely during umount, let's make another chance
to reenable nat_bits for image.

Signed-off-by: Chao Yu <chao@kernel.org>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>

1 files changed, 78 insertions, 23 deletions

diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 043cb831b289..e863136081b4 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -2213,6 +2213,24 @@ static void __move_free_nid(struct f2fs_sb_info *sbi, struct free_nid *i,
 	}
 }
 
+bool f2fs_nat_bitmap_enabled(struct f2fs_sb_info *sbi)
+{
+	struct f2fs_nm_info *nm_i = NM_I(sbi);
+	unsigned int i;
+	bool ret = true;
+
+	down_read(&nm_i->nat_tree_lock);
+	for (i = 0; i < nm_i->nat_blocks; i++) {
+		if (!test_bit_le(i, nm_i->nat_block_bitmap)) {
+			ret = false;
+			break;
+		}
+	}
+	up_read(&nm_i->nat_tree_lock);
+
+	return ret;
+}
+
 static void update_free_nid_bitmap(struct f2fs_sb_info *sbi, nid_t nid,
 							bool set, bool build)
 {
@@ -2884,7 +2902,23 @@ add_out:
 	list_add_tail(&nes->set_list, head);
 }
 
-static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
+static void __update_nat_bits(struct f2fs_nm_info *nm_i, unsigned int nat_ofs,
+							unsigned int valid)
+{
+	if (valid == 0) {
+		__set_bit_le(nat_ofs, nm_i->empty_nat_bits);
+		__clear_bit_le(nat_ofs, nm_i->full_nat_bits);
+		return;
+	}
+
+	__clear_bit_le(nat_ofs, nm_i->empty_nat_bits);
+	if (valid == NAT_ENTRY_PER_BLOCK)
+		__set_bit_le(nat_ofs, nm_i->full_nat_bits);
+	else
+		__clear_bit_le(nat_ofs, nm_i->full_nat_bits);
+}
+
+static void update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
 						struct page *page)
 {
 	struct f2fs_nm_info *nm_i = NM_I(sbi);
@@ -2893,7 +2927,7 @@ static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
 	int valid = 0;
 	int i = 0;
 
-	if (!enabled_nat_bits(sbi, NULL))
+	if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG))
 		return;
 
 	if (nat_index == 0) {
@@ -2904,17 +2938,36 @@ static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
 		if (le32_to_cpu(nat_blk->entries[i].block_addr) != NULL_ADDR)
 			valid++;
 	}
-	if (valid == 0) {
-		__set_bit_le(nat_index, nm_i->empty_nat_bits);
-		__clear_bit_le(nat_index, nm_i->full_nat_bits);
-		return;
+
+	__update_nat_bits(nm_i, nat_index, valid);
+}
+
+void f2fs_enable_nat_bits(struct f2fs_sb_info *sbi)
+{
+	struct f2fs_nm_info *nm_i = NM_I(sbi);
+	unsigned int nat_ofs;
+
+	down_read(&nm_i->nat_tree_lock);
+
+	for (nat_ofs = 0; nat_ofs < nm_i->nat_blocks; nat_ofs++) {
+		unsigned int valid = 0, nid_ofs = 0;
+
+		/* handle nid zero due to it should never be used */
+		if (unlikely(nat_ofs == 0)) {
+			valid = 1;
+			nid_ofs = 1;
+		}
+
+		for (; nid_ofs < NAT_ENTRY_PER_BLOCK; nid_ofs++) {
+			if (!test_bit_le(nid_ofs,
+					nm_i->free_nid_bitmap[nat_ofs]))
+				valid++;
+		}
+
+		__update_nat_bits(nm_i, nat_ofs, valid);
 	}
 
-	__clear_bit_le(nat_index, nm_i->empty_nat_bits);
-	if (valid == NAT_ENTRY_PER_BLOCK)
-		__set_bit_le(nat_index, nm_i->full_nat_bits);
-	else
-		__clear_bit_le(nat_index, nm_i->full_nat_bits);
+	up_read(&nm_i->nat_tree_lock);
 }
 
 static int __flush_nat_entry_set(struct f2fs_sb_info *sbi,
@@ -2933,7 +2986,7 @@ static int __flush_nat_entry_set(struct f2fs_sb_info *sbi,
 	 * #1, flush nat entries to journal in current hot data summary block.
 	 * #2, flush nat entries to nat page.
 	 */
-	if (enabled_nat_bits(sbi, cpc) ||
+	if ((cpc->reason & CP_UMOUNT) ||
 		!__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL))
 		to_journal = false;
 
@@ -2980,7 +3033,7 @@ static int __flush_nat_entry_set(struct f2fs_sb_info *sbi,
 	if (to_journal) {
 		up_write(&curseg->journal_rwsem);
 	} else {
-		__update_nat_bits(sbi, start_nid, page);
+		update_nat_bits(sbi, start_nid, page);
 		f2fs_put_page(page, 1);
 	}
 
@@ -3011,7 +3064,7 @@ int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	 * during unmount, let's flush nat_bits before checking
 	 * nat_cnt[DIRTY_NAT].
 	 */
-	if (enabled_nat_bits(sbi, cpc)) {
+	if (cpc->reason & CP_UMOUNT) {
 		down_write(&nm_i->nat_tree_lock);
 		remove_nats_in_journal(sbi);
 		up_write(&nm_i->nat_tree_lock);
@@ -3027,7 +3080,7 @@ int f2fs_flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
 	 * entries, remove all entries from journal and merge them
 	 * into nat entry set.
 	 */
-	if (enabled_nat_bits(sbi, cpc) ||
+	if (cpc->reason & CP_UMOUNT ||
 		!__has_cursum_space(journal,
 			nm_i->nat_cnt[DIRTY_NAT], NAT_JOURNAL))
 		remove_nats_in_journal(sbi);
@@ -3064,15 +3117,18 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
 	__u64 cp_ver = cur_cp_version(ckpt);
 	block_t nat_bits_addr;
 
-	if (!enabled_nat_bits(sbi, NULL))
-		return 0;
-
 	nm_i->nat_bits_blocks = F2FS_BLK_ALIGN((nat_bits_bytes << 1) + 8);
 	nm_i->nat_bits = f2fs_kvzalloc(sbi,
 			nm_i->nat_bits_blocks << F2FS_BLKSIZE_BITS, GFP_KERNEL);
 	if (!nm_i->nat_bits)
 		return -ENOMEM;
 
+	nm_i->full_nat_bits = nm_i->nat_bits + 8;
+	nm_i->empty_nat_bits = nm_i->full_nat_bits + nat_bits_bytes;
+
+	if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG))
+		return 0;
+
 	nat_bits_addr = __start_cp_addr(sbi) + sbi->blocks_per_seg -
 						nm_i->nat_bits_blocks;
 	for (i = 0; i < nm_i->nat_bits_blocks; i++) {
@@ -3089,13 +3145,12 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
 
 	cp_ver |= (cur_cp_crc(ckpt) << 32);
 	if (cpu_to_le64(cp_ver) != *(__le64 *)nm_i->nat_bits) {
-		disable_nat_bits(sbi, true);
+		clear_ckpt_flags(sbi, CP_NAT_BITS_FLAG);
+		f2fs_notice(sbi, "Disable nat_bits due to incorrect cp_ver (%llu, %llu)",
+			cp_ver, le64_to_cpu(*(__le64 *)nm_i->nat_bits));
 		return 0;
 	}
 
-	nm_i->full_nat_bits = nm_i->nat_bits + 8;
-	nm_i->empty_nat_bits = nm_i->full_nat_bits + nat_bits_bytes;
-
 	f2fs_notice(sbi, "Found nat_bits in checkpoint");
 	return 0;
 }
@@ -3106,7 +3161,7 @@ static inline void load_free_nid_bitmap(struct f2fs_sb_info *sbi)
 	unsigned int i = 0;
 	nid_t nid, last_nid;
 
-	if (!enabled_nat_bits(sbi, NULL))
+	if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG))
 		return;
 
 	for (i = 0; i < nm_i->nat_blocks; i++) {