2012年02月28日
情報科学類 オペレーティングシステム II
筑波大学 システム情報工学研究科
コンピュータサイエンス専攻, 電子・情報工学系
新城 靖
<yas@is.tsukuba.ac.jp>
このページは、次の URL にあります。
http://www.coins.tsukuba.ac.jp/~yas/coins/os2-2011/2012-02-28
あるいは、次のページから手繰っていくこともできます。
http://www.coins.tsukuba.ac.jp/~yas/
http://www.cs.tsukuba.ac.jp/~yas/
試験について
解決策:

図? ファイル・システム実現に置ける層構造
ディスク・キャッシュの層もある。解決策
$ ls -l /bin/{gunzip,gzip,zcat}
-rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/gunzip
-rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/gzip
-rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/zcat
$ ls -li /bin/{gunzip,gzip,zcat}
5636134 -rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/gunzip
5636134 -rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/gzip
5636134 -rwxr-xr-x 3 root root 62872 Jan 21 2010 /bin/zcat
$
$ cat /etc/fstab
LABEL=/ / ext3 defaults 1 1
tmpfs /dev/shm tmpfs defaults 0 0
devpts /dev/pts devpts gid=5,mode=620 0 0
sysfs /sys sysfs defaults 0 0
proc /proc proc defaults 0 0
LABEL=SWAP-sda3 swap swap defaults 0 0
130.158.86.2:/vol/vol1/home /home nfs rw,hard,bg,nfsvers=3,intr 0 0
130.158.86.2:/vol/vol5/local3 /usr/local3 nfs rw,hard,bg,nfsvers=3,intr 0 0
$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda2 222G 13G 198G 6% /
tmpfs 2.0G 0 2.0G 0% /dev/shm
130.158.86.2:/vol/vol1/home
4.0T 1.5T 2.5T 38% /home
130.158.86.2:/vol/vol5/local3
80G 4.9G 76G 7% /usr/local3
$

図? スーパーブロック、inode、dentry、file
linux-3.1.3/include/linux/fs.h
953: struct file {
...
962: struct path f_path;
963: #define f_dentry f_path.dentry
965: const struct file_operations *f_op;
...
970: atomic_long_t f_count;
972: fmode_t f_mode;
973: loff_t f_pos;
...
983: void *private_data;
...
993: };
struct fileの操作は、たとえば次のような形で行われる。 第1引数は、struct file *。
struct file *file;
file->f_op->read(file, buf, count, pos);
f_op には、次のような手続きがある。各ファイルシステム
(ext4,nfs,tmpfs,...) ごとに、手続きの実体は異なるが、インタフェースは同じ。
linux-3.1.3/include/linux/fs.h
1563: struct file_operations {
1564: struct module *owner;
1565: loff_t (*llseek) (struct file *, loff_t, int);
1566: ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);
1567: ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);
1568: ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1569: ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t);
1570: int (*readdir) (struct file *, void *, filldir_t);
1571: unsigned int (*poll) (struct file *, struct poll_table_struct *);
1572: long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long);
1573: long (*compat_ioctl) (struct file *, unsigned int, unsigned long);
1574: int (*mmap) (struct file *, struct vm_area_struct *);
1575: int (*open) (struct inode *, struct file *);
1576: int (*flush) (struct file *, fl_owner_t id);
1577: int (*release) (struct inode *, struct file *);
1578: int (*fsync) (struct file *, loff_t, loff_t, int datasync);
1579: int (*aio_fsync) (struct kiocb *, int datasync);
1580: int (*fasync) (int, struct file *, int);
1581: int (*lock) (struct file *, int, struct file_lock *);
1582: ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int);
1583: unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long);
1584: int (*check_flags)(int);
1585: int (*flock) (struct file *, int, struct file_lock *);
1586: ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int);
1587: ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int);
1588: int (*setlease)(struct file *, long, struct file_lock **);
1589: long (*fallocate)(struct file *file, int mode, loff_t offset,
1590: loff_t len);
1591: };
主な手続きの意味
linux-3.1.3/include/linux/dcache.h
116: struct dentry {
...
121: struct dentry *d_parent; /* parent directory */
122: struct qstr d_name;
123: struct inode *d_inode; /* Where the name belongs to - NULL is
124: * negative */
125: unsigned char d_iname[DNAME_INLINE_LEN]; /* small names */
...
128: unsigned int d_count; /* protected by d_lock */
130: const struct dentry_operations *d_op;
131: struct super_block *d_sb; /* The root of the dentry tree */
133: void *d_fsdata; /* fs-specific data */
...
143: struct list_head d_subdirs; /* our children */
...
145: };
...
35: struct qstr {
36: unsigned int hash;
37: unsigned int len;
38: const unsigned char *name;
39: };
...
112: # define DNAME_INLINE_LEN 40 /* 128 bytes */
linux-3.1.3/include/linux/fs.h
748: struct inode {
749: umode_t i_mode;
750: unsigned short i_opflags;
751: uid_t i_uid;
752: gid_t i_gid;
...
760: const struct inode_operations *i_op;
761: struct super_block *i_sb;
...
769: unsigned long i_ino;
770: unsigned int i_nlink;
771: dev_t i_rdev;
772: loff_t i_size;
773: struct timespec i_atime;
774: struct timespec i_mtime;
775: struct timespec i_ctime;
776: unsigned int i_blkbits;
777: blkcnt_t i_blocks;
...
784: unsigned long i_state;
790: struct hlist_node i_hash;
795: struct list_head i_dentry;
798: atomic_t i_count;
...
826: void *i_private; /* fs or device private pointer */
827: };
struct inode *inode;
...
inode->i_op->truncate(inode);
i_op には、次のような手続きがある。各ファイルシステム
(ext3,nfs,tmpfs,...) ごとに、手続きの実体は異なるが、インタフェースは同じ。
linux-3.1.3/include/linux/fs.h
1593: struct inode_operations {
1594: struct dentry * (*lookup) (struct inode *,struct dentry *, struct nameidata *);
1595: void * (*follow_link) (struct dentry *, struct nameidata *);
1596: int (*permission) (struct inode *, int);
1597: struct posix_acl * (*get_acl)(struct inode *, int);
1598:
1599: int (*readlink) (struct dentry *, char __user *,int);
1600: void (*put_link) (struct dentry *, struct nameidata *, void *);
1601:
1602: int (*create) (struct inode *,struct dentry *,int, struct nameidata *);
1603: int (*link) (struct dentry *,struct inode *,struct dentry *);
1604: int (*unlink) (struct inode *,struct dentry *);
1605: int (*symlink) (struct inode *,struct dentry *,const char *);
1606: int (*mkdir) (struct inode *,struct dentry *,int);
1607: int (*rmdir) (struct inode *,struct dentry *);
1608: int (*mknod) (struct inode *,struct dentry *,int,dev_t);
1609: int (*rename) (struct inode *, struct dentry *,
1610: struct inode *, struct dentry *);
1611: void (*truncate) (struct inode *);
1612: int (*setattr) (struct dentry *, struct iattr *);
1613: int (*getattr) (struct vfsmount *mnt, struct dentry *, struct kstat *);
1614: int (*setxattr) (struct dentry *, const char *,const void *,size_t,int);
1615: ssize_t (*getxattr) (struct dentry *, const char *, void *, size_t);
1616: ssize_t (*listxattr) (struct dentry *, char *, size_t);
1617: int (*removexattr) (struct dentry *, const char *);
1618: void (*truncate_range)(struct inode *, loff_t, loff_t);
1619: int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start,
1620: u64 len);
1621: } ____cacheline_aligned;
p->files->fd_array[fd] の struct file を表
す。
linux-3.1.3/include/linux/sched.h
1220: struct task_struct {
...
1386: struct files_struct *files;
...
1572: };
linux-3.1.3/include/linux/fdtable.h
44: struct files_struct {
...
58: struct file __rcu * fd_array[NR_OPEN_DEFAULT];
59: };
![図? p->files->fd_array[fd]](images/task-files_struct-fd-file.png)
図? task_struct、ファイル記述子、file構造体
linux-3.1.3/fs/read_write.c
444: SYSCALL_DEFINE3(read, unsigned int, fd, char __user *, buf, size_t, count)
445: {
446: struct file *file;
447: ssize_t ret = -EBADF;
448: int fput_needed;
449:
450: file = fget_light(fd, &fput_needed);
451: if (file) {
452: loff_t pos = file_pos_read(file);
453: ret = vfs_read(file, buf, count, &pos);
454: file_pos_write(file, pos);
455: fput_light(file, fput_needed);
456: }
457:
458: return ret;
459: }
348: ssize_t vfs_read(struct file *file, char __user *buf, size_t count, loff_t *pos)
349: {
350: ssize_t ret;
351:
352: if (!(file->f_mode & FMODE_READ))
353: return -EBADF;
354: if (!file->f_op || (!file->f_op->read && !file->f_op->aio_read))
355: return -EINVAL;
356: if (unlikely(!access_ok(VERIFY_WRITE, buf, count)))
357: return -EFAULT;
358:
359: ret = rw_verify_area(READ, file, pos, count);
360: if (ret >= 0) {
361: count = ret;
362: if (file->f_op->read)
363: ret = file->f_op->read(file, buf, count, pos);
364: else
365: ret = do_sync_read(file, buf, count, pos);
366: if (ret > 0) {
367: fsnotify_access(file);
368: add_rchar(current, ret);
369: }
370: inc_syscr(current);
371: }
372:
373: return ret;
374: }
linux-3.1.3/fs/ext4/file.c
276: const struct file_operations ext4_file_operations = {
277: .llseek = ext4_llseek,
278: .read = do_sync_read,
279: .write = do_sync_write,
280: .aio_read = generic_file_aio_read,
281: .aio_write = ext4_file_write,
282: .unlocked_ioctl = ext4_ioctl,
283: #ifdef CONFIG_COMPAT
284: .compat_ioctl = ext4_compat_ioctl,
285: #endif
286: .mmap = ext4_file_mmap,
287: .open = ext4_file_open,
288: .release = ext4_release_file,
289: .fsync = ext4_sync_file,
290: .splice_read = generic_file_splice_read,
291: .splice_write = generic_file_splice_write,
292: .fallocate = ext4_fallocate,
293: };
295: const struct inode_operations ext4_file_inode_operations = {
...
306: };
linux-3.1.3/fs/ext4/super.c
1256: static const struct super_operations ext4_sops = {
...
1275: };
linux-3.1.3/fs/ext4/ext4.h
761: struct ext4_inode_info {
762: __le32 i_data[15]; /* unconverted */
763: __u32 i_dtime;
764: ext4_fsblk_t i_file_acl;
...
821: struct inode vfs_inode;
...
869: };
1230: static inline struct ext4_inode_info *EXT4_I(struct inode *inode)
1231: {
1232: return container_of(inode, struct ext4_inode_info, vfs_inode);
1233: }
linux-3.1.3/include/linux/kernel.h
653: #define container_of(ptr, type, member) ({ \
654: const typeof( ((type *)0)->member ) *__mptr = (ptr); \
655: (type *)( (char *)__mptr - offsetof(type,member) );}
)

図? Ext4 ファイルシステムで使う構造体 ext4_inode_info での struct inode の保持
ssize_t vfs_write(struct file *file, const char __user *buf, size_t count, loff_t *pos)
{
...
ret = rw_verify_area(WRITE, file, pos, count);
if (ret >= 0) {
count = ret;
if (file->f_op->/*空欄(a)*/)
ret = /*空欄(b)*/(/*空欄(c)*/, buf, count, pos);
...
}
return ret;
}