On Mon, Feb 4, 2019 at 3:01 PM Michel Thebeau <michel.thebeau@windriver.com> wrote:
On Mon, 2019-02-04 at 14:41 -0600, Victor Rodriguez wrote:
On Mon, Feb 4, 2019 at 2:25 PM Michel Thebeau <michel.thebeau@windriver.com> wrote:
Hi Victor,
My recommendation will be to use the highest one despite if the score came from V2 or V3
The two sets of metrics are not comparable. StarlingX policy should refer to either CVSS v2 or v3.
But slide in the wiki say : Criticality >= 7 , if V2 = 5 adn V3 =8 shoudl that CVE be consider as Criticality >= 7 ? My observation is that we shoudl specify Criticality >= 7 ( for either V2 or V3 )
I believe that Ken has a particular metric in mind, which was not documented. The value of '7' is specific to the metric not documentated. There is a hint in the document that the value '7' is for 'critical' CVEs issues.
... refer to this document: https://nvd.nist.gov/vuln-metrics/cvss
'7' is the boundary for the highest rating of CVSS v2. Where as CVSS v3 lists critical as >9.0
Ok , then Based on NVD Vulnerability Severity Ratings 7 is ok for both V2 or V3 We cond then said in the wiki that if CVE critilarity is >= High ...
One more question, when you said critical issues are fixed if corrections are available upstream it means that...
The text of one of the slides has written "StarlingX depends on the upstream OS community to fix CVEs". Is this reference to upstream you intend? In that text 'OS community' would refer to CentOS community rather than the example MariaDB community.
Ok, thanks for the clarification. I think that there might be some special cases where we might prompt to apply a CVE patch before CentOS, but that might have to be desided in the security meeeting when they analyze the CVEs
I expect that it is unlikely a team in StarlingX will match the proficiency of the teams behind CentOS. But I agree that the StarlingX community will discuss these things.
I have seen cases before, as CVE mantianer of an OS for some time there were times were the severity of the CVE make us really urgent to update the package and release a new version of the OS, our mindset was security first. As an open comunity I woudl like to hear more feedabck from multiple users and developers Regards Victor
M
M
On Mon, 2019-02-04 at 13:26 -0600, Victor Rodriguez wrote:
On Mon, Feb 4, 2019 at 12:49 PM Young, Ken <Ken.Young@windriver.c om> wrote:
Team,
A “Lights On” feature for the 2019.05 release is “CVE Upgrades”. This feature will enable ongoing security updates for the master branch and selectively provide CVE corrective content to supported releases. The first step for this feature is to define a policy. With the help of the Starling X security team, a draft of this policy has been provided below:
https://wiki.openstack.org/wiki/StarlingX/Security/CVE_Support_ Poli cy
Please review and provide comments. I plan to reserve a spot on the Community call for any discussion on Wednesday and start the discussion with the build team to identify tools to support this policy on Thursday.
Question on the last slide you propose a formula as
Criticality >= 7
What standard are you plan to use? CVSS v3.0 or v2.0
For example, taking this MariaDB
https://nvd.nist.gov/vuln/detail/CVE-2017-15365
Base Score: 8.8 HIGH in V3 and 6.5 MEDIUM in V2
My recommendation will be to use the highest one despite if the score came from V2 or V3 However, I think we should specify that somewhere
One more question, when you said critical issues are fixed if corrections are available upstream it means that ( taking the previous example ) if MariaDB provides a Patch, is merged in master and released in the latest release, like in the previous example : https://github.com/MariaDB/server/commit/0b5a5258abbeaf8a0c3a18c7 e753 699787fdf46e
But CentOS has not taken it yet, are we OK to apply this patch in STX until CentOS apply in incoming future?
Regards
Victor R
Regards,
Ken Y
_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-di scus s
_______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-disc uss