[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <874loh4uao.fsf@mail.parknet.co.jp>
Date: Sat, 23 Dec 2017 19:23:27 +0900
From: OGAWA Hirofumi <hirofumi@...l.parknet.co.jp>
To: Chen Guanqiao <chen.chenchacha@...mail.com>
Cc: linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/2] fs: fat: add ioctl method in fat filesystem driver
Chen Guanqiao <chen.chenchacha@...mail.com> writes:
> +static int fat_check_volume_label(const char *label)
> +{
> + int i;
> +
> + for (i=0; i<11; ++i) {
> + switch (label[i]) {
> + case 0x20:
> + case 'A' ... 'Z':
> + case '0' ... '9':
> + continue;
> + case 0:
> + return 0;
> + default:
> + return -EINVAL;
> + }
> + }
> + return -EINVAL;
> +}
This check is really work? Especially what Windows show if '\0' is
included in label?
> +static int fat_ioctl_get_volume_label(struct inode *inode,
> + u8 __user *vol_label)
> +{
> + int ret = 0;
> + struct msdos_sb_info *sbi = MSDOS_SB(inode->i_sb);
> +
> + if (copy_to_user(vol_label, sbi->vol_label, sizeof(sbi->vol_label)))
> + ret = -EFAULT;
> +
> + return ret;
> +}
This should handle the label in root dir too.
> + inode_lock(inode);
Lock is not enough.
> + if (sb_rdonly(inode->i_sb)) {
> + err = -EROFS;
> + goto out_unlock_inode;
> + }
mnt_want_write_file() checks it already.
> + /* handling sector's vol_label */
> + bh = sb_bread(inode->i_sb, 0);
> + if (bh == NULL) {
> + fat_msg(inode->i_sb, KERN_ERR,
> + "unable to read boot sector to write volume label");
> + err = -EFAULT;
-EIO
> + if (b->fat16.signature == 0x28 || b->fat32.signature == 0x28) {
> + fat_msg(inode->i_sb, KERN_ERR,
> + "volume label supported since OS/2 1.2 and MS-DOS 4.0 "
> + "and higher");
> + err = -EFAULT;
> + goto out_unlock_inode;
> + }
I don't know though, is this check necessary?
> + /* handling root directory's vol_label */
> + bh = sb_bread(sb, sbi->dir_start);
> + entry = &((struct msdos_dir_entry *)(bh->b_data))[0];
> +
> + lock_buffer(bh);
> + memcpy(entry->name, label, sizeof(entry->name));
> + mark_buffer_dirty(bh);
> + unlock_buffer(bh);
> + brelse(bh);
This totally doesn't work. Please check other implementations how to
handle label.
> + out_unlock_inode:
> + inode_unlock(inode);
Missing release mnt_want_write_file().
Thanks.
Powered by blists - more mailing lists