Re: OpenBMC and https Vulnerable issue.

[Joseph Reynolds <jrey@linux.ibm.com>]

On 11/6/19 4:43 PM, James Feist wrote:
> On 11/6/19 2:38 PM, Bruce Mitchell wrote:
>>
>>
>>> -----Original Message-----
>>> From: openbmc [mailto:openbmc-
>>> bounces+bruce_mitchell=phoenix.com@lists.ozlabs.org] On Behalf Of Bruce
>>> Mitchell
>>> Sent: Wednesday, November 6, 2019 14:19
>>> To: James Feist; OpenBMC Maillist
>>> Subject: RE: OpenBMC and https Vulnerable issue.
>>>
>>>
>>>
>>>> -----Original Message-----
>>>> From: openbmc [mailto:openbmc-
>>>> bounces+bruce_mitchell=phoenix.com@lists.ozlabs.org] On Behalf Of
>>>> James Feist
>>>> Sent: Wednesday, November 6, 2019 13:52
>>>> To: Bruce Mitchell; OpenBMC Maillist
>>>> Subject: Re: OpenBMC and https Vulnerable issue.
>>>>
>>>> On 11/6/19 11:31 AM, Bruce Mitchell wrote:
>>>>>   From my investigations on TLS there seems to be 2 issues that 
>>>>> could be
>>>> corrected with OpenBMC's https:
>>>>>     1  Secure Client-Initiated Renegotiation     VULNERABLE (NOT 
>>>>> ok), DoS
>>>> threat
>>>>
>>>> This CVE is disputed https://www.cvedetails.com/cve/CVE-2011-1473/ due
>>>> to CPU consumption issues that might make it easier to cause a DOS
>>>> (which is arguably already not that difficult on a BMC). That being 
>>>> said
>>>> the fix is a 1 liner, so I implemented it and it seems to work, but I
>>>> need to see if there are any consequences.
>>>>
>>>> https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/26992

Thanks for looking at this.  Sorry about my delayed response.  The fix 
looks good to me, and it has merged already anyway.
I don't see any negative consequences.

FWIW, to address the ongoing issue of what ciphers to support, an 
OpenBMC network security considerations document was created to discuss 
relevant standards and the OpenBMC implementation.  Feel free to improve 
it with additional information.  It is here:
https://github.com/openbmc/docs/blob/master/security/network-security-considerations.md

- Joseph

>>>>
>>>>
>>>>
>>>>>     2  LUCKY13 (CVE-2013-0169), experimental     potentially 
>>>>> VULNERABLE,
>>>> uses cipher block chaining (CBC) ciphers with TLS
>>>>>        and xc023 ECDHE-ECDSA-AES128-SHA256         ECDH 521   AES 128
>>>> TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256
>>>>
>>>> Based on this https://wiki.crashtest-security.com/prevent-ssl-lucky13,
>>>> we are using the recommended ciphers,
>>>>
>>> https://github.com/openbmc/bmcweb/blob/1f8c7b5d6a679a38b8226106031
>>>> 0b876079d0f8b/include/ssl_key_handler.hpp#L330.
>>>> And based on this comment from the maintainer of test ssl, no tool can
>>>> determine this externally, and it's just a warning:
>>>> https://github.com/drwetter/testssl.sh/issues/1011#issuecomment-
>>>> 372953654.
>>>> Do you have any suggestions on if there is anything to change for 
>>>> this one?
>>>>
>>>> Thanks
>>>>
>>>> -James
>>>>
>>>
>>> Thanks James, I accept your assessment.
>>>
>>> -Bruce
>>>
>>
>> There are Mozilla Recommended configurations as well.
>> https://wiki.mozilla.org/Security/Server_Side_TLS#Recommended_configurations 
>>
>>
>
> I believe that's what was originally copied based on the variable name 
> in ssl_key_handler.hpp.
>
>> - Bruce
>>
>>>>
>>>>>
>>>>> Present standard of practice seems to be to not allow Secure Client-
>>>> Initiated Renegotiation and to not allow CBC ciphers.
>>>>>
>>>>> Is this your understanding as well?
>>>>>
>>>>> Thank you!
>>>>>
>>