Security Advisory

CVE-2020-0556: BlueZ HID and HOGP Bonding Bypass

Last updated:
  • CVE-2020-0556

CVE-2020-0556 is an access-control flaw in BlueZ before 5.54: the input plugins accepted HID connections from Bluetooth devices that were never bonded, so a nearby attacker could act as a keyboard or mouse. NVD scores it 7.1. BlueZ 5.54 enforces bonding for HID over GATT, but the Classic HID check (ClassicBondedOnly) shipped switched off.

Vulnerability summary

CVSS base score

7.1

High severity

EPSS score (FIRST)

1.0%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2020-0556 · Source: NVD

Affected components

ComponentCPEVersion range
BlueZ (bluetoothd input and HID over GATT plugins)cpe:2.3:a:bluez:bluez:*:*:*:*:*:*:*:*before 5.54 per NVD; fixed in 5.54 (commits 8cdbd3b and 3cccdba), Classic HID part only with ClassicBondedOnly=true
OpenWrt packages feed, openwrt-19.07 branch: bluez-daemoncpe:/a:bluez:bluez5.50 (PKG_RELEASE 5), HID and HoG profiles not disabled, no HID bonding patch (package Makefile read 2026-11-16)

What does this mean for affected devices?

CVE-2020-0556 is an improper access control flaw in the BlueZ userspace Bluetooth stack. NVD's record describes it as allowing "an unauthenticated user to potentially enable escalation of privilege and denial of service via adjacent access" in BlueZ before 5.54, and NVD scores it 7.1 (high) with vector CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:L, weakness NVD-CWE-noinfo (NVD). The attacker must be within Bluetooth radio range; no pairing or user action is needed.

The fix is two BlueZ commits by Alain Michaud of the Chromium project, dated 2020-03-10 and both referencing Intel's advisory INTEL-SA-00352. The first, "HOGP must only accept data from bonded devices", makes the HID over GATT (Bluetooth Low Energy) plugin refuse unbonded devices outright. The second, "HID accepts bonded device connections only", adds a ClassicBondedOnly option to input.conf for Classic (BR/EDR) HID, but sets its default to false so older mice that cannot pair keep working. Both reached BlueZ 5.54, whose tarball on kernel.org is dated 2020-03-15. Ubuntu's notes record that Intel first named 5.53 as the fixed version and changed it to 5.54 on 2020-03-16.

That default is why this CVE returned in 2023. With ClassicBondedOnly false, the Classic HID path still accepted unbonded keyboards, and the keystroke-injection flaw CVE-2023-45866 was fixed by turning the 2020 option on by default in BlueZ 5.71. A device that only took the 5.54 code change is protected on the LE path and still exposed on the Classic path unless its input.conf sets the option.

The firmware angle is the build-system package. On 2026-11-16 the OpenWrt packages Makefile for the openwrt-19.07 branch builds BlueZ 5.50 (PKG_RELEASE 5), does not pass --disable-hid or --disable-hog, so both input plugins are built, and its patches directory carries board and build fixes but no HID bonding patch. An OpenWrt 19.07 image with bluez-daemon therefore runs a version below 5.54. The Debian tracker shows how distributions handled it: stretch fixed in 5.43-2+deb9u2 and buster in 5.50-1.2~deb10u1 (DSA-4647-1), both backports below 5.54.

The CVE is not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13), and FIRST EPSS gives it 0.0104, the 65th percentile, in the 2026-11-15 dataset (CISA KEV, FIRST EPSS).

Device makers should move to BlueZ 5.54 or later, preferably a release from 5.71 on, and set ClassicBondedOnly=true explicitly in the shipped input.conf. Products that accept no keyboards or mice can drop the input and HoG plugins altogether. Our BlueZ and Linux Bluetooth vulnerabilities in embedded devices guide shows how to find the BlueZ version and settings in an image, and our authentication bypass fix guide covers enforcing authentication on every input path.

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

  • Improper access control in BlueZ before 5.54; adjacent, unauthenticated; NVD CVSS v3.1 7.1 (CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:L), NVD-CWE-noinfo. — NVD
  • Commit 8cdbd3b (2020-03-10): HOGP must only accept data from bonded devices. — BlueZ commit 8cdbd3b
  • Commit 3cccdba (2020-03-10): adds ClassicBondedOnly for Classic HID, default false. — BlueZ commit 3cccdba
  • bluez-5.54.tar.xz dated 2020-03-15 on kernel.org. — kernel.org BlueZ release directory
  • Intel's advisory was updated on 2020-03-16 to change the fixed version to 5.54. — Ubuntu CVE-2020-0556
  • openwrt-19.07 bluez package: PKG_VERSION 5.50, PKG_RELEASE 5, no --disable-hid or --disable-hog, no HID bonding patch. — OpenWrt packages (openwrt-19.07)
  • Debian fixed stretch in 5.43-2+deb9u2 and buster in 5.50-1.2~deb10u1 (DSA-4647-1); ClassicBondedOnly defaults to false. — Debian Security Tracker
  • Not in the CISA KEV catalog as of 2026-11-16 (catalog version 2026.11.13). — CISA KEV
  • EPSS 0.01042 (65th percentile), dataset 2026-11-15, read 2026-11-16. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. BlueZ commit 8cdbd3b: HOGP must only accept data from bonded devices
  3. BlueZ commit 3cccdba: HID accepts bonded device connections only
  4. kernel.org: BlueZ release tarballs
  5. OpenWrt packages: bluez Makefile (openwrt-19.07)
  6. Debian Security Tracker
  7. Ubuntu Security: CVE-2020-0556
  8. CISA KEV
  9. FIRST EPSS

Frequently asked questions

Does upgrading to BlueZ 5.54 fully fix CVE-2020-0556?

Only on the Bluetooth Low Energy path. BlueZ 5.54 makes HID over GATT require bonding, but the Classic HID check is the ClassicBondedOnly option in input.conf, which defaulted to false until BlueZ 5.71 (per BlueZ commits 3cccdba and 25a471a). Set it to true explicitly.

How is CVE-2020-0556 related to CVE-2023-45866?

CVE-2023-45866 is the same Classic HID weakness left open by the false default. BlueZ fixed it in 5.71 by changing the ClassicBondedOnly default to true, and the Debian tracker notes the 2020 fix allowed the option to be set manually but defaulted to false (per the Debian Security Tracker).

Which embedded firmware builds ship an affected BlueZ?

Any image with BlueZ below 5.54 and no backport. On 2026-11-16 the OpenWrt packages feed for the openwrt-19.07 branch builds BlueZ 5.50 with the HID and HoG plugins and no bonding patch (per the OpenWrt packages repository).

Is CVE-2020-0556 in the CISA KEV catalog?

No, not as of 2026-11-16 (catalog version 2026.11.13). FIRST EPSS gives it 0.0104, the 65th percentile, in the 2026-11-15 dataset (per CISA KEV and FIRST EPSS).