Device-os: Unpatched upstream vulnerabilities in mbedtls library

[chan0415]

submission notes

**Important:** Please sanitize/remove any confidential info like usernames, passwords, org names, product names/ids, access tokens, client ids/secrets, or anything else you don't wish to share.

Please choose one of the following templates **Bug Report** or **Feature/Enhancement Request** and delete the other.

You may also delete this submission notes header if you'd like. Thank you for contributing!

Bug Report

While scanning for unpatched upstream vulnerabilities in opensource projects. I found that there are several unpatched upstream vulnerabilities in mbedtls library.

• CVE-2018-9988
  Brief description: Buffer over-read in ssl_parse_server_key_exchange() that could cause a crash on invalid input.
  upstream patches: https://github.com/ARMmbed/mbedtls/commit/027f84c69f4ef30c0693832a6c396ef19e563ca1, https://github.com/ARMmbed/mbedtls/commit/a1098f81c252b317ad34ea978aea2bc47760b215
  NVD link: https://nvd.nist.gov/vuln/detail/CVE-2018-9988

• CVE-2018-9989
  Brief description: Buffer over-read in ssl_parse_server_psk_hint() that could cause a crash on invalid input.
  upstream patches: https://github.com/ARMmbed/mbedtls/commit/5224a7544c95552553e2e6be0b4a789956a6464e, https://github.com/ARMmbed/mbedtls/commit/740b218386083dc708ce98ccc94a63a95cd5629e
  NVD link: https://nvd.nist.gov/vuln/detail/CVE-2018-9989

• CVE-2018-0487
  Brief description: Allows remote attackers to execute arbitrary code or cause a denial of service (buffer overflow) via a crafted certificate chain that is mishandled during RSASSA-PSS signature verification within a TLS or DTLS session
  upstream patches: https://github.com/ARMmbed/mbedtls/commit/28a0c727957990ac655cbe40c7eb20b7ef01167d
  NVD link: https://nvd.nist.gov/vuln/detail/CVE-2018-0487
  Advisory: https://tls.mbed.org/tech-updates/security-advisories/mbedtls-security-advisory-2018-01

• CVE-2018-0488
  Brief description: When the truncated HMAC extension and CBC are used, allows remote attackers to execute arbitrary code or cause a denial of service (heap corruption) via a crafted application packet within a TLS or DTLS session
  upstream patches: https://github.com/ARMmbed/mbedtls/commit/992b6872f3ca717282ae367749a47f006d337a87, https://github.com/ARMmbed/mbedtls/commit/464147cadc694379b7717afb7b517fe05cdb323f
  NVD link: https://nvd.nist.gov/vuln/detail/CVE-2018-0488
  Advisory: https://tls.mbed.org/tech-updates/security-advisories/mbedtls-security-advisory-2018-01

• CVE-2017-18187
  Brief description: There is a bounds-check bypass through an integer overflow in PSK identity parsing in the ssl_parse_client_psk_identity() function in library/ssl_srv.c.
  upstream patches: https://github.com/ARMmbed/mbedtls/commit/83c9f495ffe70c7dd280b41fdfd4881485a3bc28
  NVD link: https://nvd.nist.gov/vuln/detail/CVE-2017-18187

If possible, upgrade to the latest mbedtls 2.8.0 version

[towynlin]

Thanks so much for the report. 🙇 We're discussing internally this morning.

[m-mcgowan]

@chan0415 - Many thanks for bringing this to our attention. We plan to go ahead and upgrade the Device OS (firmware) to the latest version of mbedtls to fix these CVEs. The work is scheduled to happen within the next week. Assuming the upgrade doesn't cause any regressions and still remains compatible with the DTLS implementation in the cloud, this will be part of the imminent 0.8.0-rc.4 release.

While we don't know of any active exploits of these vulnerabilities, we want to ensure the Device OS remains patched with the latest fixes regularly to close out known vulnerabilities.

I will post here with updates as the work proceeds!

[m-mcgowan]

Work has begun integrating the latest (2.8.0) mbedtls implementation into this repo. It's not trivial since our version includes a number of changes relating to TLS certificates, in particular allowing raw certificates, so it's far from simply clicking a button and pulling in the latest changes. This is also due in part to the code being pulled in wholesale into a subdirectory without git assistance.

Going forward, we have a fork of the repo with our changes and will use that as a submodule so that pulling in changes from upstream will be much simpler.

[avtolstoy]

mbedTLS has been upgraded since the posting of this issue.