In this document:
- CVE Process
- Threat Modeling
- Use Case: Integrate NVMe Driver
- Use Case: NVMe over Fabrics Target
- Use Case: Vhost integration
Common Vulnerabilities and Exposures (CVE) Process
The SPDK CVE process is designed to be simple yet effective. The process is here to help make sure that security vulnerabilities are dealt with efficiently, and with the least amount of advertising possible. An SPDK CVE sub-team exists to handle this process. If you are interested in joining please contact one of the SPDK maintainers. Issues can come from anywhere but when one comes to the attention of the community the sub-team leader should be notified via private communications ASAP. This is to ensure that we don’t advertise possible vulnerabilities before they are fully understood.
If you believe you have identified a potential security issue, please contact the sub-team directly via private email (see below), do not discuss on Slack or report the issue in GitHub.
Sub-team members are responsible for the following process:
- Upon receiving a notification of a security vulnerability, the sub-team leader will call for a sub-team meeting to discuss the potential issue.
- The sub-team will identify the resources required to investigate and determine the scope of the vulnerability, including a fix or workaround.
- When the fix or workaround has been identified, the sub-team will approve it or ask for further investigation.
- Once approved, the patch will be submitted via normal channels without verbose information on the vulnerability itself. It will simply state what the patch is doing (not why).
- Once the patch is merged, the most recent official release will be used as a baseline for a maintenance release and will include only the fix(es) identified for the issue at hand.
- Once the release is tagged, the sub-team leader will file an issue with the CVE Org
- After the issue has been filed, an announcement will be made on the mailing list with more information about the patch and the fix so that community members can decide for themselves what their exposure is and when, if at all, they should move to the new release.
CVE Sub-Team Members:
- Felipe Francosi, email@example.com
- Jim Harris, firstname.lastname@example.org
- Changpeng Liu, email@example.com
- Paul Luse, firstname.lastname@example.org
- Shuhei Matsumoto, email@example.com
- Darek Stojaczyk, firstname.lastname@example.org
- Ben Walker, email@example.com
- Anna Wan, firstname.lastname@example.org
- Tomek Zawadzki, email@example.com
Threat modeling is a common industry practice for identifying security vulnerabilities. SPDK will leverage threat modeling in an effort to proactively identify vulnerabilities and address them. Threat modeling involves identifying the most common use cases, mapping out what components are involved, and identifying possible attack surfaces and attack motives.
The use case below, NVMe Driver Integration, can serve as a model for future use cases. Everyone is encouraged to participate.
The community will hold hack-fests periodically, either at summits or developer meetups, where developers are given a use case and asked to find vulnerabilities based on the threat modeling. It will be a fun and interactive way of helping make SPDK more secure.