Security Advisory

CVE-2017-6519: Avahi Off-Link mDNS Amplification

Last updated:
  • CVE-2017-6519

CVE-2017-6519 lets anyone who can reach UDP port 5353 from outside the local link make avahi-daemon answer unicast mDNS queries, turning the device into a traffic amplifier and leaking its service names. It affects Avahi through 0.6.32 and 0.7, NVD scores it 9.1, and Avahi 0.8 fixes it by dropping such queries.

Vulnerability summary

CVSS base score

9.1

Critical severity

EPSS score (FIRST)

3.2%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2017-6519 · Source: NVD

Affected components

ComponentCPEVersion range
Avahi (avahi-daemon)cpe:2.3:a:avahi:avahi:*:*:*:*:*:*:*:*up to and including 0.6.32, and 0.7 (NVD); fixed in 0.8 (2020-02-18) by commit e111def
Buildroot package/avahi—2019.02 release ships 0.7 without the fix; 2019.02.11, 2019.11 and 2020.02 ship 0.7 with e111def backported; 2020.05 and later ship 0.8
OpenWrt packages feed: avahi—0.8 on openwrt-19.07, openwrt-21.02 and openwrt-23.05 (OpenWrt packages repository, as of 2026-11-19)
Debian and Ubuntu avahi packagespkg:deb/debian/avahiDebian fixed in 0.7-5 (unstable), rated unimportant; Ubuntu fixed in 0.7-3.1ubuntu1.2 (18.04), 0.6.32~rc+dfsg-1ubuntu2.3 (16.04), 0.6.31-4ubuntu1.3 (14.04), 0.7-4ubuntu2.1 (18.10) (USN-3876-1)

What does this mean for affected devices?

CVE-2017-6519 is a missing origin check in avahi-daemon, the mDNS/DNS-SD responder found on many embedded Linux devices. The CVE record says avahi-daemon in Avahi through 0.6.32 and 0.7 "inadvertently responds to IPv6 unicast queries with source addresses that are not on-link", which lets remote attackers cause traffic amplification and may leak information from the responding device via UDP port 5353. NVD's Primary score is 9.1, CVSS:3.0/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H, with CWE-346 (origin validation error); CISA's ADP entry gives the same 9.1 in CVSS 3.1 (NVD, CVE record). MITRE is the CNA and published no score.

Why does an mDNS responder become an amplifier?

Multicast DNS is meant for one link, but avahi-daemon also answers "legacy unicast" queries sent straight to port 5353. A 2018 report on Avahi issue 203 showed a spoofed query from an off-link source address being answered with a larger reply sent to the spoofed address, on 0.7 and on 0.6.32, so a device reachable from outside its link can be used to reflect traffic at a third party. The reply also names the device's services, host name and addresses. Our Avahi and mDNS/DNS-SD vulnerabilities in embedded Linux guide explains where avahi-daemon hides in a firmware image and why port 5353 belongs on the local link.

Which Avahi versions and embedded builds are affected?

NVD's CPE configuration covers Avahi up to and including 0.6.32 and 0.7. The fix is commit e111def of 2018-12-22, "Drop legacy unicast queries from address not on local link", which checks that the source address is inside a subnet on the local link before answering. It is part of Avahi 0.8, released on 2020-02-18, whose notes list it under security fixes.

  • Buildroot keeps Avahi 0.7 in 2019.11, 2020.02 and the 2019.02.11 point release but carries e111def as a backported patch (Buildroot 2020.02 avahi patch); the original 2019.02 release ships 0.7 without it, and 2020.05 moved to 0.8.
  • The OpenWrt packages feed carries Avahi 0.8 on the openwrt-19.07, 21.02 and 23.05 branches, as of 2026-11-19 (OpenWrt packages).
  • The Debian security tracker rates the issue unimportant, records the fix in 0.7-5 in unstable and names e111def; Ubuntu fixed 14.04 to 18.10 in USN-3876-1.

No device vendor or CERT advisory naming a product for this CVE appears among the NVD references as of 2026-11-19.

How do you fix CVE-2017-6519?

Ship Avahi 0.8 or later, or cherry-pick e111def into a 0.6 or 0.7 tree as Buildroot did. Whatever the version, bind avahi-daemon only to the interfaces that need discovery and block UDP 5353 on WAN and other untrusted interfaces. Our input handling fix guide covers checking where a packet came from before acting on it. The same builds often also need the local fixes for CVE-2021-3468 and CVE-2023-38469, and the on-link CNAME crash CVE-2026-24401.

The CVE is not in the CISA KEV catalog as of 2026-11-19 (catalog version 2026.11.18, 1,786 entries). CISA's SSVC entry records exploitation as "none". FIRST EPSS gives it 0.03238, the 88.5th percentile, in the 2026-11-18 dataset (CISA KEV, FIRST EPSS).

How to fix this class of issueNo dedicated SoftWSP guide covers this weakness class; the fix is usually the vendor patch or firmware update referenced above. Our related guide explains how to plan and verify it: NIST SP 800-40 Rev. 4: enterprise patch management planning, explained.

Evidence

  • NVD: avahi-daemon in Avahi through 0.6.32 and 0.7 responds to IPv6 unicast queries with source addresses that are not on-link (traffic amplification, information leakage via UDP 5353); NVD Primary CVSS v3.0 9.1 AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:H; CWE-346; CPE avahi up to 0.6.32 and 0.7; published 2017-05-01. Read 2026-11-19. — NVD
  • CVE.org: state PUBLISHED, CNA MITRE, no CNA score; CISA-ADP CVSS 3.1 9.1 same vector; SSVC exploitation none, automatable yes, technical impact partial. Read 2026-11-19. — CVE.org
  • Avahi 0.8 release (2020-02-18) security fixes: drop legacy unicast queries from address not on local link which can lead to UDP traffic amplification attacks (CVE-2017-6519). Read 2026-11-19. — Avahi GitHub releases
  • Commit e111def (Trent Lloyd, 2018-12-22) 'Drop legacy unicast queries from address not on local link' changes avahi-core/server.c, references CVE-2017-6519 and issue #203; contained in tag v0.8, not in v0.7. Read 2026-11-19. — Avahi GitHub
  • Avahi issue #203 (2018-11-08): spoofed query from off-link source 1.1.1.1 answered with a larger reply to the spoofed address on 0.7 and git master; 0.6.32 also responded. Read 2026-11-19. — Avahi GitHub issue 203
  • Debian: avahi fixed in 0.7-5 (unstable), urgency unimportant, bug 917047; notes list commit e111def. Read 2026-11-19. — Debian security tracker
  • USN-3876-1: CVE-2017-6519 and CVE-2018-1000845 fixed in avahi-daemon 0.7-4ubuntu2.1 (18.10), 0.7-3.1ubuntu1.2 (18.04), 0.6.32~rc+dfsg-1ubuntu2.3 (16.04), 0.6.31-4ubuntu1.3 (14.04). Read 2026-11-19. — Ubuntu
  • Buildroot avahi.mk: AVAHI_VERSION 0.7 in 2019.02 (no patches), 0.7 with patch 0001 backported from e111def in 2019.02.11, 2019.11 and 2020.02, 0.8 in 2020.05. Read 2026-11-19. — Buildroot
  • OpenWrt packages libs/avahi Makefile: PKG_VERSION 0.8 on openwrt-19.07, openwrt-21.02 and openwrt-23.05; master 0.9_rc5. Read 2026-11-19. — OpenWrt packages
  • Not in the CISA KEV catalog as of 2026-11-19 (catalog version 2026.11.18, 1,786 entries); EPSS 0.03238, percentile 0.88539, dataset 2026-11-18, read 2026-11-19. — CISA KEV / FIRST EPSS

Related advisories

Sources

  1. NVD
  2. CVE.org record CVE-2017-6519 (CNA: MITRE)
  3. Avahi 0.8 release notes (2020-02-18)
  4. Avahi commit e111def: drop legacy unicast queries from address not on local link
  5. Avahi issue 203: missing link-local checks allow mDNS reflection
  6. Debian security tracker: CVE-2017-6519
  7. Ubuntu USN-3876-1: Avahi vulnerabilities
  8. Buildroot 2020.02: avahi patch backported from e111def
  9. OpenWrt packages: libs/avahi Makefile (openwrt-23.05)
  10. CISA KEV
  11. FIRST EPSS

Frequently asked questions

Which Avahi version fixes CVE-2017-6519?

Avahi 0.8, released on 2020-02-18, which includes commit e111def to drop legacy unicast queries from addresses not on the local link. Debian fixed it in 0.7-5 and Buildroot 2020.02 backported the commit onto 0.7 (per the Avahi 0.8 release notes, the Debian security tracker and Buildroot).

How can I reduce exposure to CVE-2017-6519 without upgrading?

Cherry-pick commit e111def, which only touches avahi-core/server.c, and keep UDP port 5353 off WAN and other untrusted interfaces so off-link hosts cannot reach the responder; the attack depends on queries from addresses that are not on-link (per the Avahi commit and NVD).

Why does NVD rate a traffic amplification bug 9.1?

NVD's CVSS 3.0 vector treats it as network-reachable with no privileges or user interaction, with high confidentiality impact for the leaked device information and high availability impact for the amplification; CISA's ADP gives the same 9.1. Debian rates the issue unimportant for its packages (per NVD, CVE.org and the Debian security tracker).

Does any device vendor list CVE-2017-6519?

No device vendor or CERT advisory naming a product for this CVE is among the NVD references as of 2026-11-19; the references are the Avahi issue, Red Hat and Ubuntu entries such as USN-3876-1, and a mailing-list thread. Check any product that ships avahi-daemon from Buildroot, OpenWrt or Yocto by version and patch list (per NVD and Ubuntu).

Is CVE-2017-6519 exploited?

It is not in the CISA KEV catalog as of 2026-11-19 (catalog version 2026.11.18), and CISA's SSVC entry records exploitation as none. FIRST EPSS gives it 0.03238, the 88.5th percentile, in the 2026-11-18 dataset (per CISA KEV, CVE.org and FIRST EPSS).