1. 概述
Linux 虚拟文件系统是建立在具体文件系统之上，其包括几种主要的对象，分别是超级块对象，目录项对象，索引节点对象，与进程相关的文件对象，安装点对象，文件系统类型对象。在VFS中，多个文件系统可以被安装在同一个目录，例如/dev/sda和/dev/sdb先后被安装在/project目录下，则/dev/sdb会隐藏/dev/sda文件系统。同一个文件系统也可以安装在不同的目录下，此时只有一个超级块对象。因此，超级块对象代表了文件系统。目录项对象主要用于路径查找过程，例如，将文件系统/dev/sda安装在/project目录下，首先需要找到/project/目录的目录项对象，另外，目录项对象和目录不同，每个目录和文件都具有目录项对象，存在于内存。索引节点对象描述了文件或目录的元数据，包括文件的访问时间等，在创建索引节点的时候，将文件的相关操作指针赋给索引节点对象i_fpos。每个打开的文件对象对应一个文件描述符，常用于文件的open操作。每个文件系统都有对应的类型对象，如ext3,ext2,ext4和fuse文件系统，而这个结构体存储了相关的文件类型。安装点对象表示文件系统的挂载位置以及不同文件系统之间的关系，新挂载的文件系统都需要加入mount tree.
2. 对象
(1) superblock对象<linux/fs.h>
struct super_block{
struct list_head s_list /*Pointers for superblock list所有的superblock链表*/
unsigned long s_blocksize /*Block size in bytes每个块的字节数*/
unsigned long s_old_blocksize /*Block size in bytes as reported by the underlying block device drive*/
unsigned char s_blocksize_bits /*Block size in number of bits每个块占多少比特*/
unsigned char s_dirt /*Modified (dirty) flag超级块为脏，需要和磁盘上的超级块同步*/
unsigned long long s_maxbytes /*Maximum size of the files文件的最大长度*/
struct file_system_type * s_type /*Filesystem type所属的文件系统类型如ext3*/
struct super_operations * s_op /*Superblock methods超级块的相关操作，对象包括属性和操作*/
struct dquot_operations * dq_op /*Disk quota handling methods*/
struct quotactl_ops * s_qcop /*Disk quota administration methods*/
struct export_operations * s_export_op /*Export operations used by network filesystems网络文件系统相关操作*/
unsigned long s_flags /*Mount flags挂载标志*/
unsigned long s_magic /*Filesystem magic number文件系统魔数*/
struct dentry * s_root /*Dentry object of the filesystem's root directory文件系统的根目录项对象*/
struct rw_semaphore s_umount /*Semaphore used for unmounting umount时使用*/
struct semaphore s_lock /*Superblock semaphore*/
int s_count /*Reference counter计数器，一个文件系统可能被安装到多个目录*/
int s_syncing /*Flag indicating that inodes of the superblock are being synchronized */
int s_need_sync_fs /*Flag used when synchronizing the superblock's mounted filesystem*/
atomic_t s_active /*Secondary reference counter*/
void * s_security /* Pointer to superblock security structure*/
struct xattr_handler ** s_xattr /*Pointer to superblock extended attribute structure*/
struct list_head s_inodes /*List of all inodes所有的inode节点*/
struct list_head s_dirty /*List of modified inodes修改的inode节点链表*/
struct list_head s_io /*List of inodes waiting to be written to disk等待写到磁盘上的inode节点列表*/
struct hlist_head s_anon /*List of anonymous dentries for handling remote network filesystems*/
struct list_head s_files /*List of file objects文件对象链表*/
struct block_device *s_bdev /*Pointer to the block device driver descriptor块设备描述符*/
struct list_head s_instances /*Pointers for a list of superblock objects of a given filesystem type属于指定文件系统类型的链表指针*/
struct quota_info s_dquot /*Descriptor for disk quota*/
int s_frozen /*Flag used when freezing the filesystem (forcing it to a consistent state)*/
wait_queue_head_t s_wait_unfrozen /*Wait queue where processes sleep until the filesystem is unforzen*/
char[] s_id /*Name of the block device containing the superblock*/
void * s_fs_info /*Pointer to superblock information of a specific filesystem具体文件系统的超级块信息*/
struct semaphore s_vfs_rename_sem /*Semaphore used by VFS when renaming files across directories*/
u32 s_time_gran /*Timestamp's granularity (in nanoseconds)*/
}
/*
* NOTE: write_inode, delete_inode, clear_inode, put_inode can be called
* without the big kernel lock held in all filesystems.
*/
struct super_operations {
struct inode *(*alloc_inode)(struct super_block *sb); /*为索引节点分配空间*/
void (*destroy_inode)(struct inode *); /*destory 索引节点对象*/
void (*read_inode) (struct inode *); /*利用i_ino索引节点号从磁盘上读取相关的数据来填充索引节点*/
void (*dirty_inode) (struct inode *); /*当索引节点被标识为dirty调用，即修改过*/
int (*write_inode) (struct inode *, int flag); /*更新文件系统的索引节点，flag表示是否需要同步*/
void (*put_inode) (struct inode *); /*当索引节点被释放时*/
void (*drop_inode) (struct inode *); /*当索引节点即将被destory*/
void (*delete_inode) (struct inode *); /*当索引节点被destory时，删除VFS的索引节点和磁盘上的文件数据和元数据*/
void (*put_super) (struct super_block *); /*释放超级块对象，因为相应的文件系统已经umounted*/
void (*write_super) (struct super_block *); /*更新文件系统的超级块*/
int (*sync_fs)(struct super_block *sb, int wait); /*同步文件系统*/
void (*write_super_lockfs) (struct super_block *);
void (*unlockfs) (struct super_block *);
int (*statfs) (struct dentry *, struct kstatfs *buf); /*返回文件的信息放在buf*/
int (*remount_fs) (struct super_block *, int *, char *); /'*重新安装文件系统*/
void (*clear_inode) (struct inode *); /*当索引节点正在destory*/
void (*umount_begin) (struct vfsmount *, int); /*终止mount操作，因为相应的umount正在开始，仅用于网络文件系统*/
int (*show_options)(struct seq_file *, struct vfsmount *);
int (*show_stats)(struct seq_file *, struct vfsmount *);
#ifdef CONFIG_QUOTA
ssize_t (*quota_read)(struct super_block *, int, char *, size_t, loff_t);
ssize_t (*quota_write)(struct super_block *, int, const char *, size_t, loff_t);
#endif
};
(2)inode节点对象，每个文件拥有一个inode节点对象，包含了文件的元数据信息，在<linux/fs.h>头文件
struct inode {
struct hlist_node i_hash; /*Pointers for the hash list指向hash链表指针*/
struct list_head i_list; /*Pointers for the list that describes the inode's current state索引节点状态链表，分为没有使用，正在使用和脏链表*/
struct list_head i_sb_list; /*指向超级块链表的指针*/
struct list_head i_dentry; /*与这个inode节点相关联的目录项对象*/
unsigned long i_ino; /*索引节点号*/
atomic_t i_count; /*索引节点的使用计数*/
unsigned int i_nlink; /*硬连接数，多个硬连接共享同一个索引节点，但目录项对象不同(在不同的目录下)*/
uid_t i_uid; /*owner identifier*/
gid_t i_gid; /*group identifier*/
dev_t i_rdev; /*real device identifier实际设备标识*/
unsigned long i_version; /*version number,automatically increased after each use*/
loff_t i_size; /*文件长度*/
#ifdef __NEED_I_SIZE_ORDERED
seqcount_t i_size_seqcount;
#endif
struct timespec i_atime; /*文件的最后访问时间*/
struct timespec i_mtime; /*文件最后修改时间*/
struct timespec i_ctime; /*索引节点最后修改时间*/
unsigned int i_blkbits; /*块大小*/
blkcnt_t i_blocks; /*文件的块数量*/
unsigned short i_bytes; /*文件最后一个块的字节数*/
umode_t i_mode; /*文件类型和存取权限*/
spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
struct mutex i_mutex;
struct rw_semaphore i_alloc_sem;
struct inode_operations *i_op; /*索引节点相关操作*/
const struct file_operations *i_fop; /* former ->i_op->default_file_ops 在索引节点创建时赋值，在打开文件时赋值给文件对象file*/
struct super_block *i_sb; /*超级块对象指针*/
struct file_lock *i_flock;
struct address_space *i_mapping; /*地址空间对象指针*/
struct address_space i_data; /*地址空间对象*/
#ifdef CONFIG_QUOTA
struct dquot *i_dquot[MAXQUOTAS];
#endif
struct list_head i_devices;
union {
struct pipe_inode_info *i_pipe;
struct block_device *i_bdev;
struct cdev *i_cdev;
};
int i_cindex;
__u32 i_generation;
#ifdef CONFIG_DNOTIFY
unsigned long i_dnotify_mask; /* Directory notify events */
struct dnotify_struct *i_dnotify; /* for directory notifications */
#endif
#ifdef CONFIG_INOTIFY
struct list_head inotify_watches; /* watches on this inode */
struct mutex inotify_mutex; /* protects the watches list */
#endif
unsigned long i_state; /*索引节点状态，是否为脏*/
unsigned long dirtied_when; /* jiffies of first dirtying */
unsigned int i_flags; /*文件系统安装标志*/
atomic_t i_writecount;
#ifdef CONFIG_SECURITY
void *i_security;
#endif
void *i_private; /* fs or device private pointer */
};
struct inode_operations {
int (*create) (struct inode *,struct dentry *,int, struct nameidata *); /*创建一个和目录项对象相关联的磁盘索引节点*/
struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *); /*在目录项高速缓存查找和索引节点相关的目录项对象*/
int (*link) (struct dentry *old_dentry,struct inode *,struct dentry *new_dentry); /*创建一个硬连接,关联新的目录项对象*/
int (*unlink) (struct inode *,struct dentry *); /*移除硬连接*/
int (*symlink) (struct inode *,struct dentry *,const char *); /*创建符号链接*/
int (*mkdir) (struct inode *,struct dentry *,int); /*创建一个新的索引节点*/
int (*rmdir) (struct inode *,struct dentry *);
int (*mknod) (struct inode *,struct dentry *,int,dev_t);
int (*rename) (struct inode *, struct dentry *,
struct inode *, struct dentry *);
int (*readlink) (struct dentry *, char __user *,int);
void * (*follow_link) (struct dentry *, struct nameidata *); /*转化符号链接*/
void (*put_link) (struct dentry *, struct nameidata *, void *);
void (*truncate) (struct inode *);
int (*permission) (struct inode *, int, struct nameidata *);
int (*setattr) (struct dentry *, struct iattr *);
int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
ssize_t (*listxattr) (struct dentry *, char *, size_t);
int (*removexattr) (struct dentry *, const char *);
void (*truncate_range)(struct inode *, loff_t, loff_t);
};
(3)文件对象struct file,当进程打开文件时，在内存创建一个文件对象，在磁盘上没有相对应的映像，所以没有dirty标志，<linux/fs.h>
struct file {
/*
* fu_list becomes invalid after file_free is called and queued via
* fu_rcuhead for RCU freeing
*/
union {
struct list_head fu_list; /*指向文件对象的链表指针*/
struct rcu_head fu_rcuhead;
} f_u;
struct path f_path;
#define f_dentry f_path.dentry /*和文件对象相关联的目录项对象*/
#define f_vfsmnt f_path.mnt /*包含文件对象的安装点对象*/
const struct file_operations *f_op; /*文件操作表指针*/
atomic_t f_count; /*文件对象索引次数，一个文件可以被多个进程同时打开，每个进程的struct file不相同*/
unsigned int f_flags; /*打开文件时的指定标志*/
mode_t f_mode; /*打开方式*/
loff_t f_pos; /*当前文件指针的位置，非常重要*/
struct fown_struct f_owner;
unsigned int f_uid, f_gid; /*用户ID，组ID*/
struct file_ra_state f_ra; /*文件预读状态*/
unsigned long f_version;
#ifdef CONFIG_SECURITY
void *f_security;
#endif
/* needed for tty driver, and maybe others */
void *private_data;
#ifdef CONFIG_EPOLL
/* Used by fs/eventpoll.c to link all the hooks to this file */
struct list_head f_ep_links;
spinlock_t f_ep_lock;
#endif /* #ifdef CONFIG_EPOLL */
struct address_space *f_mapping; /*地址空间指针*/
};
/*
* NOTE:
* read, write, poll, fsync, readv, writev, unlocked_ioctl and compat_ioctl
* can be called without the big kernel lock held in all filesystems.
*/
struct file_operations {
struct module *owner;
loff_t (*llseek) (struct file *, loff_t, int);
ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
int (*readdir) (struct file *, void *, filldir_t);
unsigned int (*poll) (struct file *, struct poll_table_struct *);
int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
int (*mmap) (struct file *, struct vm_area_struct *);
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *, fl_owner_t id);
int (*release) (struct inode *, struct file *);
int (*fsync) (struct file *, struct dentry *, int datasync);
int (*aio_fsync) (struct kiocb *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
ssize_t (*sendfile) (struct file *, loff_t *, size_t, read_actor_t, void *);
ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
int (*check_flags)(int);
int (*dir_notify)(struct file *filp, unsigned long arg);
int (*flock) (struct file *, int, struct file_lock *);
ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
};
(4)目录项对象，内核为路径的每一部分创建目录项对象，如/tmp/test，会为/,tmp和test分别创建目录项对象，用于定位文件或目录。为了加快查找，目录项对象通常存放在目录项高速缓存。目录项对象在磁盘上没有相对应的映像。当目录项对象不在dentry_cache中，就需要从磁盘读取相关信息，在内存创建一个目录项对象和相关联的索引节点对象。在<linux/dcache.h>
struct dentry {
atomic_t d_count; /*使用计数*/
unsigned int d_flags; /* protected by d_lock 目录项缓存标志 */
spinlock_t d_lock; /* per dentry lock */
struct inode *d_inode; /* Where the name belongs to - NULL is
* negative 和文件名相关联的inode节点*/
/*
* The next three fields are touched by __d_lookup. Place them here
* so they all fit in a cache line.
*/
struct hlist_node d_hash; /* lookup hash list */
struct dentry *d_parent; /* parent directory 父目录项对象*/
struct qstr d_name; /*文件名结构体，包括文件名长度，文件名，hash值*/
struct list_head d_lru; /* LRU list */
/*
* d_child and d_rcu can share memory
*/
union {
struct list_head d_child; /* child of parent list子目录项链表 */
struct rcu_head d_rcu;
} d_u;
struct list_head d_subdirs; /* our children */
struct list_head d_alias; /* inode alias list */
unsigned long d_time; /* used by d_revalidate */
struct dentry_operations *d_op; /*目录项操作方法*/
struct super_block *d_sb; /* The root of the dentry tree */
void *d_fsdata; /* fs-specific data */
#ifdef CONFIG_PROFILING
struct dcookie_struct *d_cookie; /* cookie, if any */
#endif
int d_mounted; /*文件系统的安装次数，同一目录可以安装不同的文件系统*/
unsigned char d_iname[DNAME_INLINE_LEN_MIN]; /* small names */
};
struct dentry_operations {
int (*d_revalidate)(struct dentry *, struct nameidata *); /*判定目录项是否还有效*/
int (*d_hash) (struct dentry *, struct qstr *); /*hash值*/
int (*d_compare) (struct dentry *, struct qstr *, struct qstr *); /*比较文件名*/
int (*d_delete)(struct dentry *); /*d_count为0时调用*/
void (*d_release)(struct dentry *);
void (*d_iput)(struct dentry *, struct inode *);
};
(5)进程相关结构体:每个进程拥有自己的当前工作目录和根目录，通过fs_struct结构体维护<linux/fs_struct.h>
struct fs_struct {
atomic_t count;
rwlock_t lock;
int umask; /*设置文件权限*/
struct dentry * root, * pwd, * altroot; /*根目录项对象，当前目录项对象，altroot通常为空*/
struct vfsmount * rootmnt, * pwdmnt, * altrootmnt; /*根目录的安装的文件系统安装点对象,当前工作目录安装的文件系统安装点对象*/
};
(6)与进程关联的files_struct对象，包括指向struct file对象的数组指针fd,一个进程打开的最多文件数max_fds等，在<linux/file.h>
/*
* Open file table structure
*/
struct files_struct {
/*
* read mostly part
*/
atomic_t count;
struct fdtable *fdt; /*文件表指针*/
struct fdtable fdtab;
/*
* written part on a separate cache line in SMP
*/
spinlock_t file_lock ____cacheline_aligned_in_smp;
int next_fd; /*下一个文件描述符*/
struct embedded_fd_set close_on_exec_init;
struct embedded_fd_set open_fds_init;
struct file * fd_array[NR_OPEN_DEFAULT]; /*文件对象数组指针*/
};
struct fdtable {
unsigned int max_fds; /*当前最大的文件描述符数*/
struct file ** fd; /* current fd array */
fd_set *close_on_exec;
fd_set *open_fds; /*已经打开的文件描述符集*/
struct rcu_head rcu;
struct fdtable *next;
};
(7)文件系统类型file_system_type ,每个文件系统类型对应多个具体的文件系统，在<linux/fs.h>
/*文件系统类型说明 */
struct file_system_type {
const char *name;/*文件系统名称*/
int fs_flags;/*文件系统类型标志*/
int (*get_sb) (struct file_system_type *, int,
const char *, void *, struct vfsmount *);/*读超级块*/
void (*kill_sb) (struct super_block *);/*移除超级块*/
struct module *owner;/*指向实现文件系统的module*/
struct file_system_type * next;/*指向文件系统类型表的下一个*/
struct list_head fs_supers;/*给定类型文件系统的超级块表头*/
struct lock_class_key s_lock_key;
struct lock_class_key s_umount_key;
};
其中，
(1) file_systems是所有的文件系统类型链表头，如ext2,ext3,ext4. 而next指针指向下一个文件系统类型。
(2)fs_supers是同类文件系统的链表头，下一个由超级块对象的s_instances指针指向.
(3)get_sb和kill_sb分配一个超级块对象和销毁一个超级块对象。
(8)mount point 对象，为了在内存中保存挂载点，挂载标志，以及和其它文件系统之间的关系，如父子关系。必须维护一个vfsmount对象，在<linux/mount.h>
struct vfsmount {
struct list_head mnt_hash; /*指向hash表指针*/
struct vfsmount *mnt_parent; /* fs we are mounted on 父安装点*/
struct dentry *mnt_mountpoint; /* dentry of mountpoint 安装点目录项对象*/
struct dentry *mnt_root; /* root of the mounted tree 安装点根的目录项对象*/
struct super_block *mnt_sb; /* pointer to superblock 指向超级块指针*/
struct list_head mnt_mounts; /* list of children, anchored here 子安装点头指针*/
struct list_head mnt_child; /* and going through their mnt_child 子安装点的下一个对象*/
atomic_t mnt_count;/*使用次数*/
int mnt_flags;/*安装标志*/
int mnt_expiry_mark; /* true if marked for expiry */
char *mnt_devname; /* Name of device e.g. /dev/dsk/hda1 设备名*/
struct list_head mnt_list; /*指向命名空间的下一个安装点对象，其中链表头存在命名空间的list域内*/
struct list_head mnt_expire; /* link in fs-specific expiry list */
struct list_head mnt_share; /* circular list of shared mounts */
struct list_head mnt_slave_list;/* list of slave mounts */
struct list_head mnt_slave; /* slave list entry */
struct vfsmount *mnt_master; /* slave is on master->mnt_slave_list */
struct mnt_namespace *mnt_ns; /* containing namespace 指向进程的命名空间 */
int mnt_pinned;
};
处理文件系统挂载的相关函数:
alloc_vfsmnt(name)
Allocates and initializes a mounted filesystem descriptor
free_vfsmnt(mnt)
Frees a mounted filesystem descriptor pointed by mnt
lookup_mnt(mnt,dentry)
Looks up a descriptor in the hash table and returns its address
3. 总结
super_block-关联每个具体的文件系统
inode-对应文件元数据信息
dentry-目录项对象，内存表示形式，文件和目录都拥有目录项对象
file-打开文件在内存的表示形式
fs_struct-进程相关的根目录，工作目录，根安装点和当前目录安装点
file_system_type-在启动时注册文件系统类型，如EXT3,或者是加载模块时注册
vfsmount-维护已挂载的文件系统之间的关系，如父子关系