On Wed, Mar 11, 2020, at 11:36 AM, Matěj Týč wrote:
One of RHEL7 bugzillas [1] shows an interesting discrepancy between
our
content and STIG:
* We feature [2] a rule "Use Only FIPS 140-2 Validated Ciphers"
* STIG has its own [3] "A FIPS 140-2 approved cryptographic algorithm
must be used for SSH communications."
There is a discrepancy between the two - while we claim that the
following ciphers are FIPS 140-2 certified on Red Hat Enterprise Linux
7, only three of them are recognized as such by the STIG:
* aes128-ctr(STIG)
* aes192-ctr(STIG)
* aes256-ctr(STIG)
* aes128-cb
* aes192-cbc
* aes256-cbc
* 3des-cbc
* rijndael-cbc(a)lysator.liu.se
I have confirmed correctness of our description with our FIPS SME Tomas
Mraz (in CC), so this issue looks as a bug in STIG - either the
requirement is too strict, so it is incorrect, or it is supposed to be
strict, and it should therefore be reworded, and we need to create a
new rule in our content.
Indeed, the STIG allows fewer ciphers than FIPS allows; the STIG currently says "If
any ciphers other than "aes128-ctr", "aes192-ctr", or
"aes256-ctr" are listed, the "Ciphers" keyword is missing, or the
returned line is commented out, this is a finding." [4]
Similarly, for the MACs allowed, "If any ciphers other than "hmac-sha2-256"
or "hmac-sha2-512" are listed or the returned line is commented out, this is a
finding." [5]
More are permitted by FIPS than by SSH, but the STIG is FIPS compliant as-is, using a
subset of the FIPS ciphers. From the Security Policy document for the certification [6],
Only the following ciphers are allowed:
- aes128-ctr
- aes192-ctr
- aes256-ctr
- aes128-cbc
- aes192-cbc
- aes256-cbc
- 3des-cbc
- rijndael-cbc(a)lysator.liu.se
Only the following message authentication codes are allowed:
- hmac-sha1
- hmac-sha2-256
- hmac-sha2-512
- hmac-sha1-etm(a)openssh.com
- hmac-sha2-256-etm(a)openssh.com
- hmac-sha2-512-etm(a)openssh.com
What is the procedure in cases like this?
I'd just configure the STIG subset of FIPS, and maybe ask DISA to add a clarification
note to the STIG. Many folks are concerned about the FIPS-permitted 3DES algorithm [7],
"3DES is deprecated for all new applications and usage is disallowed after
2023"
V/r,
James Cassell
[4]
https://vaulted.io/library/disa-stigs-srgs/red_hat_enterprise_linux_7_sec...
[5]
https://vaulted.io/library/disa-stigs-srgs/red_hat_enterprise_linux_7_sec...
[6]
https://csrc.nist.gov/CSRC/media/projects/cryptographic-module-validation...
[7]
https://www.cryptomathic.com/news-events/blog/3des-is-officially-being-re...
> References:
>
>
> > [1]:
https://bugzilla.redhat.com/show_bug.cgi?id=1781244
> > [2]:
https://static.open-scap.org/ssg-guides/ssg-rhel7-guide-stig.html#xccdf_o...
> > [3]:
>
https://www.stigviewer.com/stig/red_hat_enterprise_linux_7/2017-12-14/fin...
>