linux-kernel - RE: [PATCH v3] hfs: remove BUG() from hfs_release_folio()/hfs_test_inode()/hfs_write

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <c99951ae12dc1f5a51b1f6c82bbf7b61b2f12e02.camel@ibm.com>
Date: Wed, 23 Jul 2025 18:19:38 +0000
From: Viacheslav Dubeyko <Slava.Dubeyko@....com>
To: "penguin-kernel@...ove.SAKURA.ne.jp" <penguin-kernel@...ove.SAKURA.ne.jp>,
        "willy@...radead.org" <willy@...radead.org>
CC: "glaubitz@...sik.fu-berlin.de" <glaubitz@...sik.fu-berlin.de>,
        "frank.li@...o.com" <frank.li@...o.com>,
        "slava@...eyko.com"
	<slava@...eyko.com>,
        "linux-fsdevel@...r.kernel.org"
	<linux-fsdevel@...r.kernel.org>,
        "linux-kernel@...r.kernel.org"
	<linux-kernel@...r.kernel.org>,
        "akpm@...ux-foundation.org"
	<akpm@...ux-foundation.org>
Subject: RE: [PATCH v3] hfs: remove BUG() from
 hfs_release_folio()/hfs_test_inode()/hfs_write_inode()

On Wed, 2025-07-23 at 11:16 +0900, Tetsuo Handa wrote:
> With below change, legitimate HFS filesystem images can be mounted.
> 
> But is crafted HFS filesystem images can not be mounted expected result?
> 
>   # losetup -a
>   /dev/loop0: [0001]:7185 (/memfd:syzkaller (deleted))
>   # mount -t hfs /dev/loop0 /mnt/
>   mount: /mnt: filesystem was mounted, but any subsequent operation failed: Operation not permitted.
>   # fsck.hfs /dev/loop0
>   ** /dev/loop0
>      Executing fsck_hfs (version 540.1-Linux).
>   ** Checking HFS volume.
>      Invalid extent entry
>   (3, 0)
>   ** The volume   could not be verified completely.
>   # mount -t hfs /dev/loop0 /mnt/
>   mount: /mnt: filesystem was mounted, but any subsequent operation failed: Operation not permitted.
> 
> Also, are IDs which should be excluded from make_bad_inode() conditions
> same for HFS_CDR_FIL and HFS_CDR_DIR ?
> 
> 
> --- a/fs/hfs/inode.c
> +++ b/fs/hfs/inode.c
> @@ -358,6 +358,8 @@ static int hfs_read_inode(struct inode *inode, void *data)
>                 inode->i_op = &hfs_file_inode_operations;
>                 inode->i_fop = &hfs_file_operations;
>                 inode->i_mapping->a_ops = &hfs_aops;
> +               if (inode->i_ino < HFS_FIRSTUSER_CNID)
> +                       goto check_reserved_ino;
>                 break;
>         case HFS_CDR_DIR:
>                 inode->i_ino = be32_to_cpu(rec->dir.DirID);
> @@ -368,6 +370,24 @@ static int hfs_read_inode(struct inode *inode, void *data)
>                                       inode_set_atime_to_ts(inode, inode_set_ctime_to_ts(inode, hfs_m_to_utime(rec->dir.MdDat))));
>                 inode->i_op = &hfs_dir_inode_operations;
>                 inode->i_fop = &hfs_dir_operations;
> +               if (inode->i_ino < HFS_FIRSTUSER_CNID)
> +                       goto check_reserved_ino;
> +               break;
> +       default:
> +               make_bad_inode(inode);
> +       }
> +       return 0;
> +check_reserved_ino:
> +       switch (inode->i_ino) {
> +       case HFS_POR_CNID:
> +       case HFS_ROOT_CNID:
> +       case HFS_EXT_CNID:
> +       case HFS_CAT_CNID:
> +       case HFS_BAD_CNID:
> +       case HFS_ALLOC_CNID:
> +       case HFS_START_CNID:
> +       case HFS_ATTR_CNID:
> +       case HFS_EXCH_CNID:
>                 break;
>         default:
>                 make_bad_inode(inode);

I have missed that this list contains [1]:

#define HFS_POR_CNID		1	/* Parent Of the Root */
#define HFS_ROOT_CNID		2	/* ROOT directory */

Of course, hfs_read_inode() can be called for the root directory and parent of
the root cases. So, HFS_POR_CNID and HFS_ROOT_CNID are legitimate values.
However, the other constants cannot be used because they should be described in
superblock (MDB) and Catalog File cannot have the records for them.

Thanks,
Slava.

[1] https://elixir.bootlin.com/linux/v6.16-rc7/source/fs/hfs/hfs.h#L41