一个文件系统获得了内核的支持，仅仅是功能上的可能性，如果要使用，还需要将其mount到rootfs下，才能可见并进而被使用，本文主要讨论mount的实现过程。

SYSCALL_DEFINE5(mount...)	//namespace.c
	do_mount()
		user_path()
		do_remount()
		do_loopback()
		do_change_type()
		do_move_mount()
		do_new_mount()
			struct file_system_type *type
			struct vfs_mount *mnt
			mnt = vfs_kern_mount()
				struct mount * mnt
				mnt = alloc_vfsmnt()
				root = mount_fs()
					struct super_block *sb
					root = type->mount()		//回调file_system_type的mount方法
					sb = root ->d_sb
					security_sb_kern_mount()
					up_write()
				init mnt
				list_add_tail(&mnt->mnt_instance, &root->d_sb->s_mounts)
			do_add_mount(real_mount(mnt),path,mnt_flags)
ext4_mnt()		//ext4
	mount_bdev(...ext4_fill_super)
		struct block_device *bdev
		struct super_block *s
		bdev = blkdev_get_by_path()
		s = sget(...test_bdev_super,set_bdev_super...)		//find or create a superblock
			alloc_super()
				kzalloc()
				init_waitqueue_head()
				s->s_bdi = &noop_backing_dev_info
				...other initialization of s
			set()							//set_bdev_super()
				s->s_bdev = data;
				s->s_dev = s->s_bdev->bd_dev;
				s->s_bdi = &bdev_get_queue(s->s_bdev)->backing_dev_info;
					bdev->bd_disk->queue;
			list_add_tail(...&super_blocks)
			hlist_add_head()
			get_filesystem()
				__module_get()
		sb_set_blocksize()
		fill_super()						//ext4_fill_super()
			struct ext4_sb_info *sbi
			sbi = kzalloc()
			... init sbi...
			ext4_msg()
			setup_timer()
			sb->s_op = &ext4_sops
			sb->s_export_op = &ext4_export_ops
			sb->s_xattr = ext4_xattr_handlers
			sb->s_root = d_make_root()
			ext4_setup_super()
			ext4_ext_init()
			ext4_mb_init()
			sbi->s_kobj.kset = ext4_kset
			init_completion()
			kobject_init_and_add(&sbi->s_kobj, &ext4_ktype...)
		s->s_flags|=MS_ACTIVE
		bdev->bd_super = s
		dget()

bdev文件系统在其初始化时就已经挂载到内核。

bdev_mount()		//bdev	
	mount_pseudo(...&bdev_sops...)
sysfs_mount()		//sysfs
	struct dentry* root
	void * ns
	ns = kobj_ns_grab_current(KOBJ_NS_TYPE_NET);
	root = kernfs_mount_ns