f2fs: fix to flush multiple device in checkpoint
If f2fs manages multiple devices, in checkpoint, we need to issue flush
in those devices which contain dirty data/node in their cache before
we write checkpoint region, otherwise, filesystem metadata could be
corrupted if hitting SPO after checkpoint.
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index 933c3d5..57e79cc 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -1965,6 +1965,9 @@ static void init_sb_info(struct f2fs_sb_info *sbi)
for (j = HOT; j < NR_TEMP_TYPE; j++)
mutex_init(&sbi->wio_mutex[i][j]);
spin_lock_init(&sbi->cp_lock);
+
+ sbi->dirty_device = 0;
+ spin_lock_init(&sbi->dev_lock);
}
static int init_percpu_info(struct f2fs_sb_info *sbi)