OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: Jerome Etienne (jmeoff.net)
Date: Wed Jan 02 2002 - 13:57:54 CST

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    Hello,

    The following text describes a security hole in the encrypted loop
    device for linux. Because of it, an attacker is able to modify the
    content of the encrypted device without being detected. This text
    proposes to fix the hole by authenticating the device.

    comments are welcome

    ps: version in html, pdf and ps can be found in http://www.off.net/~jme

                    Vulnerability in encrypted loop device for Linux

                               Jerome Etienne jmeoff.net

    Abstract

       This text describes a security hole i found in encrypted loop device for
       Linux. An attacker is able to modify the content of the encrypted device
       without being detected (see section 2). This text proposes to fix the hole
       by authenticating the device (see section 3).

    1 Threat model

       Encrypting a disk device aims to protect against a off-line attacker who
       would be able to access the disk between 2 legitimate mounts.

       It isn't against an attacker who has access to the running computer when
       the encrypted device is mounted as either (i) the attacker is root and it
       can access the encrypted device anyway or (ii) he is an unprivileged user
       and can be stopped with Unix's right management (i.e. user/group).

    2 Attack description

       The vulnerability of encrypted loop device is due to its lack of
       authentication. The aim of encryption is to make the data unreadable for
       anybody who doesn't know the key. It doesn't prevent an attacker from
       modifying the data. People assume that an attacker won't do it because the
       attacker wouldn't be able to choose the resulting clear text. But this
       section shows that the attacker can choose the resulting clear text to
       some extends and that modifying the cypher text data may be interesting
       even if the attacker ignores the result.

       This attack is only applicable to device storing data which are reused
       across mounts: most file-system (e.g. ext2, reiserfs, ext3) but not swap.
       In some systems, encrypted devices are stored in the same location than
       the encrypted disk containing the operating system. For those systems the
       attacker who can access the encrypted device, can easily modify the OS to
       gain access (e.g. kernel) independtly of the encrypted device.

      2.1 To insert random data

       If the attacker modifies the cipher text without choosing the resulting
       clear text, it will likely produce random data. The legitimate user won't
       detect the modification and will use them as if they were valid. As they
       likely appears random, it will result of a Denial of Service (aka DoS).

      2.2 To insert chosen data

       The encryption mode used by encrypted loop device is CBC[oST81,sec 5.3].
       CBC allows cut/past attacks i.e. the attacker can cut encrypted data from
       one part of the device and paste them in another location. As both data
       sections have been encrypted by the same key, the clear text won't be
       completely random data.

       This lack of authentication isn't a CBC flaw. Authentication isn't
       considered a aim of the encryption mode, so most modes (e.g. ECB, CFB,
       OFB) doesn't authenticate the data. To use another mode would be flawed in
       the same way except if they explicitly protect against forgery. Recently
       some modes including authentication popped up to speed up the encryption /
       authentication couple but as far as i know they are all patented.

       In very short, encrypting with CBC is Cn=Enc(Cn-1 xor Pn) where Enc(x) is
       encrypting x, Pn is the nth block of plain text and Cn the nth block of
       cipher text. For the first block, Cn-1 is an Initial vector (aka IV) which
       may be public and must be unique for a given key. The decryption is Pn =
       Dec(Cn) xor Cn-1. See [oST81,sec 5.3] for a longer description of CBC.

       If the attacker copies s blocks from the location m to n (aka
       [Cn,...,Cn+s-1] == [Cm,...,Cm+s-1]), Pn+1 up to Pn+s-1 will the same as
       Pm+1 to Pm+s-1 and Pn will likely appears random. Cn (i.e. Cm) will be
       decrypted as Pn = Dec(Cm) xor Cn-1 but Cm-1 and Cn-1 are different so Pn
       will likely appears random. Nevertheless Pn+1 = Dec(Cn+1) xor Cn =
       Dec(Cm+1) xor Cm = Pm+1, so Pn+1=Pm+1. So if the attacker has an idea of
       the content of a group of blocks in the device, he can copy them to the
       Nth block, thus it can choose the content of it without being detected.

       As an file-system isn't designed to appears random, its content may be
       predictable to some extents (e.g. common directories and files, inode,
       superblock). The attacker may use such informations to guess the contents
       and do a knowledgeable cut/past. For example, an attacker knowing the
       location of a password file may replace a password by another one which is
       already known.

    3 Proposed fixes

       We propose 2 types of fixes: one which authenticate at mount time (see
       section 3.1) and the other which authenticates at the cluster level (see
       section 3.2). The choice between the two (see section 3.4) is a user
       matter as it mostly depends on the access pattern on the encrypted device.

       In the proposed fixes, the authentication is a MAC computed over the
       encrypted device. The MAC is HMAC[KBC97] combined with a configured hash
       function, preferably a well studied one such as SHA1[oST95] or MD5[Riv92].
       The MAC secret key is derived from the pass-phrase via PKCS-5 key
       derivations ([Kal00,sec 5.1]).

      3.1 Authenticating at mount time

       As we need to authenticate the device across mounts and not while it is
       mounted (see section 1), it is sufficient to authenticate the whole device
       during mount operations. It slows down mount operations but they are
       rather infrequent so we consider the trade-off delay/security acceptable.
       The MAC is verified during mount operations and generated during unmount
       operations. It isn't supposed to be valid while the device is mounted.

       The MAC generation is done when unmounting the device. The MAC is computed
       over all the sectors of the device and the result is appended in the
       device file after all the sectors.

       The MAC verification is done when mounting the device. The MAC is computed
       over all the sectors of the device. If the result is equal to the MAC
       appended to the block device, the verification succeed, else it failed.
       The verification may fail (i) if an attacker attempted to modify the
       device during 2 legitimates mounts or (ii) if the device hasn't been
       cleanly unmounted (e.g. computer crash). It is impossible to automatically
       distinguish both cases with certainty. So if the verification fails, the
       user is notified and the mount operation may be stopped depending on
       configuration.

      3.2 Authenticating at cluster level

       To authenticate the whole device at mount time, may be considered
       prohibitive by some users, so this section describe an alternative which
       authenticate the device at the cluster level. A cluster is a group of one
       or more sectors, the exact number depends on configuration. In this case,
       the MAC is verified each time a cluster is read from the disk and
       generated at each write.

       If the device isn't cleanly unmounted, the authentication of one or more
       cluster may fail (e.g. the super block). This case will be detected at
       mount time. But if an attacker forges data in the device, it will be
       detected only when the user read the modified data. The kernel will read
       the forged cluster and the authentication will fail. It may report it with
       a printk with a rate limitor, it isn't clean but i don't see any better
       way.

      3.3 MAC location

       Currently the encrypted loop file-system is stored in a regular file of a
       hosting file-system. Its size is a multiple of a sector size (i.e. 512
       byte). The MAC could be stored in a separate file or included in the
       regular file. To store the MAC in a separate file, generates problems
       while managing the loop device file (e.g. copy, backup). The administrator
       must not forget to copy the MAC file when he copies the device file, else
       the copied device won't be usable anymore. To store the MAC in the same
       file as the clusters doesn't has this disadvantage.

      3.4 Comparison

       To authenticate at the cluster level will increase the access time of each
       cluster but won't affect mount operation. The exact increase depends on
       the MAC and encryption algorithms. As a rule of thumb, MAC algorithms are
       typically 3 times faster than encryption ones so the time dedicated to
       cryptography for each block will increase by around 30%. To authenticate
       at mount time will largely slow down the mount operations but won't affect
       every access once mounted.

       The authentication at mount time will detect forgery at mount time,
       whereas the alternative detects it only when the forged cluster is read,
       possibly a long time after the modification. Users may consider that it is
       easier to diagnose who forged it if they have a better idea of when the
       attack occurred.

       To authenticate the whole device at mount time requires a single MAC per
       device, so the space overhead (typically 16 byte) is negligible compared
       to the device's size. To authenticate at the cluster level requires a MAC
       per cluster, it is significantly more but some people may consider it
       still negligible, especially with cheap disks.

       The choice between the two mostly depends on the access pattern on the
       encrypted device. If the device is used for interactive purpose, the
       increased access latency may be unsuitable. If the access latency is
       important or if every block is frequently modified, to authenticate only
       once at mount time may be more interesting. If the user can't stand long
       mount operations, to authenticate at cluster level will be more suitable.
       As only the final user knows the type access made on his encrypted device,
       he should be the one able to choose between the two.

    4 Acknowledgments

       Thanks to Andy Kleen and Phil Swan for their useful comments.

    5 Conclusion

       This text described an vulnerability in encrypted loop device which allows
       an attacker to modify the encrypted device without being detected (see
       section 2). We propose a fix which authenticate the whole device during
       mount operation (see section 3.1). This fix slows down mount operations
       but we consider the trade-off longer delay vs additional security very
       reasonable as mount operations are rather infrequent. We propose another
       fix which authenticate at cluster level for people who can't stand long
       mount operation. The choice between the two is a final user matter.

       The authentication may be optionally disabled thus if an user considers
       the trade-off delay/security not in favor of security, he may choose to be
       vulnerable to this attack and disable it. Nevertheless the author thinks
       encrypted loop device must be secure by default.

    References

       [Kal00]
               B. Kaliski. Pkcs 5: Password-based cryptography specification
               version 2.0. Request For Comment (Informational) RFC2898,
               September 2000.

       [KBC97]
               H. Krawczyk, M. Bellare, and R. Canetti. Hmac: Keyed-hashing for
               message authentication. Request For Comment (Informational)
               RFC2104, February 1997.

       [oST81]
               National Institute of Standards and Technology. implementing and
               using the nbs data encryption standard. Federal information
               processing standards fips74, April 1981.

       [oST95]
               National Institute of Standards and Technology. Secure hash
               standard. Federal information processing standards fips180-1,
               April 1995.

       [Riv92]
               R. Rivest. The md5 message-digest algorithm. Request For Comment
               (Informational) RFC1321, April 1992.