super.c

来自「基于组件方式开发操作系统的OSKIT源代码」· C语言 代码 · 共 1,342 行 · 第 1/3 页

 */
void sync_supers(kdev_t dev)
{
	struct super_block * sb;

	for (sb = sb_entry(super_blocks.next);
	     sb != sb_entry(&super_blocks); 
	     sb = sb_entry(sb->s_list.next)) {
		if (!sb->s_dev)
			continue;
		if (dev && sb->s_dev != dev)
			continue;
		if (!sb->s_dirt)
			continue;
		/* N.B. Should lock the superblock while writing */
		wait_on_super(sb);
		if (!sb->s_dev || !sb->s_dirt)
			continue;
		if (dev && (dev != sb->s_dev))
			continue;
		if (sb->s_op && sb->s_op->write_super)
			sb->s_op->write_super(sb);
	}
}

struct super_block * get_super(kdev_t dev)
{
	struct super_block * s;

	if (!dev)
		return NULL;
restart:
	s = sb_entry(super_blocks.next);
	while (s != sb_entry(&super_blocks))
		if (s->s_dev == dev) {
			wait_on_super(s);
			if (s->s_dev == dev)
				return s;
			goto restart;
		} else
			s = sb_entry(s->s_list.next);
	return NULL;
}

#ifndef OSKIT
asmlinkage int sys_ustat(dev_t dev, struct ustat * ubuf)
{
        struct super_block *s;
        struct ustat tmp;
        struct statfs sbuf;
        mm_segment_t old_fs;
	int err = -EINVAL;

	lock_kernel();
        s = get_super(to_kdev_t(dev));
        if (s == NULL)
                goto out;
	err = -ENOSYS;
        if (!(s->s_op->statfs))
                goto out;

        old_fs = get_fs();
        set_fs(get_ds());
        s->s_op->statfs(s,&sbuf,sizeof(struct statfs));
        set_fs(old_fs);

        memset(&tmp,0,sizeof(struct ustat));
        tmp.f_tfree = sbuf.f_bfree;
        tmp.f_tinode = sbuf.f_ffree;

        err = copy_to_user(ubuf,&tmp,sizeof(struct ustat)) ? -EFAULT : 0;
out:
	unlock_kernel();
	return err;
}
#endif /* OSKIT */

/*
 * Find a super_block with no device assigned.
 */
static struct super_block *get_empty_super(void)
{
	struct super_block *s;

	for (s  = sb_entry(super_blocks.next);
	     s != sb_entry(&super_blocks); 
	     s  = sb_entry(s->s_list.next)) {
		if (s->s_dev)
			continue;
		if (!s->s_lock)
			return s;
		printk("VFS: empty superblock %p locked!\n", s);
	}
	/* Need a new one... */
	if (nr_super_blocks >= max_super_blocks)
		return NULL;
	s = kmalloc(sizeof(struct super_block),  GFP_USER);
	if (s) {
		nr_super_blocks++;
		memset(s, 0, sizeof(struct super_block));
		INIT_LIST_HEAD(&s->s_dirty);
		list_add (&s->s_list, super_blocks.prev);
	}
	return s;
}

static struct super_block * read_super(kdev_t dev,const char *name,int flags,
				       void *data, int silent)
{
	struct super_block * s;
	struct file_system_type *type;

	if (!dev)
		goto out_null;
	check_disk_change(dev);
	s = get_super(dev);
	if (s)
		goto out;

	type = get_fs_type(name);
	if (!type) {
		printk("VFS: on device %s: get_fs_type(%s) failed\n",
		       kdevname(dev), name);
		goto out;
	}
	s = get_empty_super();
	if (!s)
		goto out;
	s->s_dev = dev;
	s->s_flags = flags;
	s->s_dirt = 0;
	sema_init(&s->s_vfs_rename_sem,1);
	/* N.B. Should lock superblock now ... */
	if (!type->read_super(s, data, silent))
		goto out_fail;
	s->s_dev = dev; /* N.B. why do this again?? */
	s->s_rd_only = 0;
	s->s_type = type;
out:
	return s;

	/* N.B. s_dev should be cleared in type->read_super */
out_fail:
	s->s_dev = 0;
out_null:
	s = NULL;
	goto out;
}

#ifndef OSKIT
/*
 * Unnamed block devices are dummy devices used by virtual
 * filesystems which don't use real block-devices.  -- jrs
 */

static unsigned int unnamed_dev_in_use[256/(8*sizeof(unsigned int))] = { 0, };

kdev_t get_unnamed_dev(void)
{
	int i;

	for (i = 1; i < 256; i++) {
		if (!test_and_set_bit(i,unnamed_dev_in_use))
			return MKDEV(UNNAMED_MAJOR, i);
	}
	return 0;
}

void put_unnamed_dev(kdev_t dev)
{
	if (!dev || MAJOR(dev) != UNNAMED_MAJOR)
		return;
	if (test_and_clear_bit(MINOR(dev), unnamed_dev_in_use))
		return;
	printk("VFS: put_unnamed_dev: freeing unused device %s\n",
			kdevname(dev));
}
#endif /* OSKIT */

#ifdef OSKIT
       int d_umount(struct super_block * sb)
#else
static int d_umount(struct super_block * sb)
#endif
{
	struct dentry * root = sb->s_root;
	struct dentry * covered = root->d_covers;

	if (root->d_count != 1)
		return -EBUSY;

	if (root->d_inode->i_state)
		return -EBUSY;

	sb->s_root = NULL;

	if (covered != root) {
		root->d_covers = root;
		covered->d_mounts = covered;
		dput(covered);
	}
	dput(root);
	return 0;
}

#ifndef OSKIT
static void d_mount(struct dentry *covered, struct dentry *dentry)
{
	if (covered->d_mounts != covered) {
		printk("VFS: mount - already mounted\n");
		return;
	}
	covered->d_mounts = dentry;
	dentry->d_covers = covered;
}

static int do_umount(kdev_t dev, int unmount_root, int flags)
{
	struct super_block * sb;
	int retval;
	
	retval = -ENOENT;
	sb = get_super(dev);
	if (!sb || !sb->s_root)
		goto out;

	/*
	 * Before checking whether the filesystem is still busy,
	 * make sure the kernel doesn't hold any quota files open
	 * on the device. If the umount fails, too bad -- there
	 * are no quotas running any more. Just turn them on again.
	 */
	DQUOT_OFF(dev);
	acct_auto_close(dev);

	/*
	 * If we may have to abort operations to get out of this
	 * mount, and they will themselves hold resources we must
	 * allow the fs to do things. In the Unix tradition of
	 * 'Gee thats tricky lets do it in userspace' the umount_begin
	 * might fail to complete on the first run through as other tasks
	 * must return, and the like. Thats for the mount program to worry
	 * about for the moment.
	 */
	 
	if( (flags&MNT_FORCE) && sb->s_op->umount_begin)
		sb->s_op->umount_begin(sb);

	/*
	 * Shrink dcache, then fsync. This guarantees that if the
	 * filesystem is quiescent at this point, then (a) only the
	 * root entry should be in use and (b) that root entry is
	 * clean.
	 */
	shrink_dcache_sb(sb);
	fsync_dev(dev);

	if (dev==ROOT_DEV && !unmount_root) {
		/*
		 * Special case for "unmounting" root ...
		 * we just try to remount it readonly.
		 */
		retval = 0;
		if (!(sb->s_flags & MS_RDONLY))
			retval = do_remount_sb(sb, MS_RDONLY, 0);
		return retval;
	}

	retval = d_umount(sb);
	if (retval)
		goto out;

	if (sb->s_op) {
		if (sb->s_op->write_super && sb->s_dirt)
			sb->s_op->write_super(sb);
	}

	lock_super(sb);
	if (sb->s_op) {
		if (sb->s_op->put_super)
			sb->s_op->put_super(sb);
	}

	/* Forget any remaining inodes */
	if (invalidate_inodes(sb)) {
		printk("VFS: Busy inodes after unmount. "
			"Self-destruct in 5 seconds.  Have a nice day...\n");
	}

	sb->s_dev = 0;		/* Free the superblock */
	unlock_super(sb);

	remove_vfsmnt(dev);
out:
	return retval;
}

static int umount_dev(kdev_t dev, int flags)
{
	int retval;
	struct inode * inode = get_empty_inode();

	retval = -ENOMEM;
	if (!inode)
		goto out;

	inode->i_rdev = dev;
	retval = -ENXIO;
	if (MAJOR(dev) >= MAX_BLKDEV)
		goto out_iput;

	fsync_dev(dev);

	down(&mount_sem);

	retval = do_umount(dev, 0, flags);
	if (!retval) {
		fsync_dev(dev);
		if (dev != ROOT_DEV) {
			blkdev_release(inode);
			put_unnamed_dev(dev);
		}
	}

	up(&mount_sem);
out_iput:
	iput(inode);
out:
	return retval;
}

/*
 * Now umount can handle mount points as well as block devices.
 * This is important for filesystems which use unnamed block devices.
 *
 * There is a little kludge here with the dummy_inode.  The current
 * vfs release functions only use the r_dev field in the inode so
 * we give them the info they need without using a real inode.
 * If any other fields are ever needed by any block device release
 * functions, they should be faked here.  -- jrs
 *
 * We now support a flag for forced unmount like the other 'big iron'
 * unixes. Our API is identical to OSF/1 to avoid making a mess of AMD
 */

asmlinkage int sys_umount(char * name, int flags)
{
	struct dentry * dentry;
	int retval;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	lock_kernel();
	dentry = namei(name);
	retval = PTR_ERR(dentry);
	if (!IS_ERR(dentry)) {
		struct inode * inode = dentry->d_inode;
		kdev_t dev = inode->i_rdev;

		retval = 0;		
		if (S_ISBLK(inode->i_mode)) {
			if (IS_NODEV(inode))
				retval = -EACCES;
		} else {
			struct super_block *sb = inode->i_sb;
			retval = -EINVAL;
			if (sb && inode == sb->s_root->d_inode) {
				dev = sb->s_dev;
				retval = 0;
			}
		}
		dput(dentry);

		if (!retval)
			retval = umount_dev(dev, flags);
	}
	unlock_kernel();
	return retval;
}

/*
 *	The 2.0 compatible umount. No flags. 
 */
 
asmlinkage int sys_oldumount(char * name)
{
	return sys_umount(name,0);
}

/*
 * Check whether we can mount the specified device.
 */
int fs_may_mount(kdev_t dev)
{
	struct super_block * sb = get_super(dev);
	int busy;

	busy = sb && sb->s_root &&
	       (sb->s_root->d_count != 1 || sb->s_root->d_covers != sb->s_root);
	return !busy;
}

/*
 * do_mount() does the actual mounting after sys_mount has done the ugly
 * parameter parsing. When enough time has gone by, and everything uses the
 * new mount() parameters, sys_mount() can then be cleaned up.
 *
 * We cannot mount a filesystem if it has active, used, or dirty inodes.
 * We also have to flush all inode-data for this device, as the new mount
 * might need new info.
 *
 * [21-Mar-97] T.Schoebel-Theuer: Now this can be overridden when
 * supplying a leading "!" before the dir_name, allowing "stacks" of
 * mounted filesystems. The stacking will only influence any pathname lookups
 * _after_ the mount, but open file descriptors or working directories that
 * are now covered remain valid. For example, when you overmount /home, any
 * process with old cwd /home/joe will continue to use the old versions,
 * as long as relative paths are used, but absolute paths like /home/joe/xxx
 * will go to the new "top of stack" version. In general, crossing a
 * mount point will always go to the top of stack element.
 * Anyone using this new feature must know what he/she is doing.
 */

int do_mount(kdev_t dev, const char * dev_name, const char * dir_name, const char * type, int flags, void * data)
{
	struct dentry * dir_d;
	struct super_block * sb;
	struct vfsmount *vfsmnt;
	int error;

	error = -EACCES;
	if (!(flags & MS_RDONLY) && dev && is_read_only(dev))
		goto out;

	/*
	 * Do the lookup first to force automounting.
	 */
	dir_d = namei(dir_name);
	error = PTR_ERR(dir_d);
	if (IS_ERR(dir_d))
		goto out;

	down(&mount_sem);
	error = -ENOTDIR;
	if (!S_ISDIR(dir_d->d_inode->i_mode))
		goto dput_and_out;