CVE-2020-24336 is an AMNESIA:33 buffer overflow in the DNS64 code of the uIP stack in Contiki-OS and Contiki-NG: one crafted DNS Type A answer can overflow memory on a device with NAT64 enabled. NVD scores it 9.8. Contiki-NG merged its fix to the develop branch on 2026-09-07, after its newest tagged release, v5.2.
Vulnerability summary
CVSS base score
9.8
Critical severity
EPSS score (FIRST)
59.1%
Probability of exploitation in the wild in the next 30 days
CISA KEV catalog
No
Not listed
Published
Vulnerability ID: CVE-2020-24336 · Source: NVD
Affected components
| Component | CPE | Version range |
|---|---|---|
| Contiki-NG (uIP stack, NAT64/DNS64 in ip64-dns64.c) | cpe:2.3:o:contiki-ng:contiki-ng:*:*:*:*:*:*:*:* | NVD and CISA ICSA-20-343-01: 4.5 and earlier; the Contiki-NG fix (PR #3254) was merged to develop on 2026-09-07, after the newest tag release/v5.2 (2026-08-14) |
| Contiki-OS (uIP stack, NAT64/DNS64) | cpe:2.3:o:contiki-os:contiki:*:*:*:*:*:*:*:* | 3.0 and earlier; end of life per CISA ICSA-20-343-01 |
What does this mean for affected devices?
CVE-2020-24336 is one of the 33 AMNESIA:33 vulnerabilities that Forescout researchers reported in open-source embedded TCP/IP stacks, disclosed on 2020-12-08. The CVE text says the code "for parsing Type A domain name answers in ip64-dns64.c doesn't verify whether the address in the answer's length is sane", so "when copying an address of an arbitrary length, a buffer overflow can occur", and that "this bug can be exploited whenever NAT64 is enabled". NVD classes it as CWE-120 and scores it 9.8 with vector CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H for Contiki through 3.0 and Contiki-NG through 4.5 (NVD). CERT/CC's VU#815128 names the function: ip64_dns64_4to6 "accepts DNS A record response of arbitrary length which can result in RCE or DoS on the target device". CISA's ICSA-20-343-01 (CSAF) gives the same 9.8 score and lists uIP-Contiki-NG 4.5 and earlier and uIP-Contiki-OS 3.0 and earlier, marking Contiki-OS end of life.
Is there a fixed Contiki-NG release?
The fix landed years after disclosure. Contiki-NG pull request #3254, merged to the develop branch on 2026-09-07, says ip64_dns64_4to6() "copied every answer record using the length field from the wire without checking it against the message, so a single UDP packet from port 53 can overrun ip64_packet_buffer on any build with NAT64 enabled, and the Type A branch of that copy is CVE-2020-24336". It bounds the DNS64 parser in both directions and drops a reply it cannot translate. The newest tagged release on 2026-11-16 is release/v5.2, published 2026-08-14, before that merge. NVD's range stops at 4.5, but the pull request describes the unchecked copy in the code it replaced, so a firmware image built with NAT64 from any tagged Contiki-NG release should be treated as affected until it carries the #3254 changes. Contiki-OS receives no fixes.
Which devices does it reach?
Contiki-NG is an operating system for constrained IoT devices, and the bug is reachable only where the image enables NAT64, typically a border router or gateway that lets an IPv6 mesh reach IPv4 hosts. Leaf sensors built without NAT64 do not compile the vulnerable path, so the build configuration is the evidence. CERT/CC notes that projects such as open-iscsi inherited uIP code, but open-iscsi's own advisory says only CVE-2020-17437, CVE-2020-13988 and CVE-2020-13987 had matching code in its iscsiuio copy, so this CVE does not apply there.
Is CVE-2020-24336 exploited?
It is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.5909, the 99.1st percentile, in the 2026-11-16 dataset (CISA KEV, FIRST EPSS).
What should device makers do?
Check whether the build enables NAT64; if it does, backport the #3254 commits or build from a develop revision that contains them, then rebuild and test DNS64 translation. Contiki-OS users should plan a move to Contiki-NG or another maintained stack. Our embedded TCP/IP stack vulnerabilities guide covers identifying the stack in an image and recording a not-affected decision, our memory corruption fix guide covers bounds checks against the received length, and the picoTCP DNS flaw from the same disclosure is CVE-2020-24338. For DNS parsing bugs in other embedded stacks, see CVE-2020-27736 in Nucleus NET.
Evidence
- Contiki through 3.0 and Contiki-NG through 4.5: ip64-dns64.c does not check the Type A answer address length; buffer overflow whenever NAT64 is enabled; NVD CVSS v3.1 9.8, CWE-120. — NVD
- ip64_dns64_4to6 accepts DNS A record responses of arbitrary length, RCE or DoS (CWE-787); open-iscsi and Apache NuttX inherited some AMNESIA:33 code. — CERT/CC VU#815128
- CISA: CVSS 9.8; uIP-Contiki-NG 4.5 and prior, uIP-Contiki-OS (end of life) 3.0 and prior. — CISA ICSA-20-343-01 (CSAF)
- PR #3254, merged to develop 2026-09-07, bounds the DNS64 parser; states the Type A branch of the unchecked copy is CVE-2020-24336. — Contiki-NG pull request #3254
- Newest Contiki-NG tagged release is release/v5.2, published 2026-08-14. — Contiki-NG release/v5.2
- Only CVE-2020-17437, CVE-2020-13988 and CVE-2020-13987 had matching code in open-iscsi's iscsiuio uIP copy. — open-iscsi GHSA-r278-fm99-8rgp
- Not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). — CISA KEV
- EPSS 0.59089 (99.1st percentile), dataset 2026-11-16. — FIRST EPSS

Related advisories
Sources
- NVD
- CERT/CC VU#815128: Embedded TCP/IP stacks have memory corruption vulnerabilities
- CISA ICSA-20-343-01: Multiple Embedded TCP/IP Stacks (CSAF)
- Contiki-NG pull request #3254: ip64: bound the DNS64 parser and correct its answer rewriting
- Contiki-NG release/v5.2
- open-iscsi security advisory GHSA-r278-fm99-8rgp
- Forescout Vedere Labs: AMNESIA:33
- CISA KEV
- FIRST EPSS
Frequently asked questions
Which Contiki-NG release fixes CVE-2020-24336?
No tagged release carries the fix as of 2026-11-16. Contiki-NG merged pull request #3254, which names this CVE, to the develop branch on 2026-09-07; the newest tag, release/v5.2, was published on 2026-08-14 (per Contiki-NG on GitHub).
Is every Contiki device affected by CVE-2020-24336?
No. The CVE text says the bug can be exploited whenever NAT64 is enabled, and the vulnerable code is the DNS64 translation in ip64-dns64.c, so images built without NAT64 do not include it (per NVD and Contiki-NG PR #3254).
Does CVE-2020-24336 affect open-iscsi?
No. open-iscsi's advisory says only CVE-2020-17437, CVE-2020-13988 and CVE-2020-13987 had matching code in the uIP copy inside iscsiuio (per open-iscsi GHSA-r278-fm99-8rgp).
Is CVE-2020-24336 in the CISA KEV catalog?
No, not as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.5909, the 99.1st percentile, in the 2026-11-16 dataset (per CISA KEV and FIRST EPSS).
Questions about this page?