Security Advisory

CVE-2019-13377: SAE/EAP-pwd Brainpool Timing Leak

Last updated:
  • CVE-2019-13377

CVE-2019-13377 is an observable timing and cache side channel (CWE-203) in the SAE (WPA3-Personal) and EAP-pwd handshakes of hostapd and wpa_supplicant through 2.8. It applies only when Brainpool groups 28-30 are enabled and can leak enough to recover the password offline. NVD scores it 5.9; 2.9 disables those groups.

Vulnerability summary

CVSS base score

5.9

Medium severity

EPSS score (FIRST)

2.5%

Probability of exploitation in the wild in the next 30 days

CISA KEV catalog

No

Not listed

Published

Vulnerability ID: CVE-2019-13377 · Source: NVD

Affected components

ComponentCPEVersion range
hostapd (SAE and EAP-pwd)cpe:2.3:a:w1.fi:hostapd:*:*:*:*:*:*:*:*2.0 to 2.8 (NVD CPE); only with Brainpool groups 28-30 configured in sae_groups or pwd_group; fixed in 2.9 (w1.fi 2019-6)
wpa_supplicant (SAE and EAP-pwd peer)cpe:2.3:a:w1.fi:wpa_supplicant:*:*:*:*:*:*:*:*through 2.8 per the NVD description and w1.fi 2019-6 (NVD CPE lists hostapd only); fixed in 2.9
OpenWrt hostapd / wpad packages—OpenWrt builds hostapd and wpa_supplicant from git snapshots dated 2026-08-07 on main, 2025-08-26 on openwrt-25.12 and 2024-09-15 on openwrt-24.10 (read 2026-11-18)

What does this mean for affected devices?

NVD's description: "The implementations of SAE and EAP-pwd in hostapd and wpa_supplicant 2.x through 2.8 are vulnerable to side-channel attacks as a result of observable timing differences and cache access patterns when Brainpool curves are used. An attacker may be able to gain leaked information from a side-channel attack that can be used for full password recovery." NVD's Primary score is CVSS v3.1 5.9 (CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:N/A:N) and CVSS v2 4.3, with CWE-203. NVD's CPE lists hostapd 2.0 through 2.8, though the description also names wpa_supplicant. The record was published on 2019-08-15 (NVD).

Where does the leak come from?

Mathy Vanhoef, who found the original Dragonblood flaws, reported this one after the first fixes. His Dragonblood page explains that the Dragonfly password encoding looks for a hash output smaller than the curve prime: with the default NIST curves it is "practically always found immediately", but with Brainpool curves "several iterations may have to be executed", and the number depends on the password and the client's MAC address. He confirmed the leak against the then-latest hostapd and "were able to brute-force the password using the leaked information." It follows CVE-2019-9494 and CVE-2019-9495, the first SAE and EAP-pwd side channels.

Which configurations are exposed?

The w1.fi 2019-6 advisory, published 2019-08-07, says the earlier fixes "did not remove all differences", especially for groups 28-30, and that cache attacks "can still be feasible when using these groups in SAE or EAP-pwd with wpa_supplicant/hostapd v2.8." It narrows the exposure: the groups "are not enabled by default in v2.8 (and in case of wpa_supplicant, in any version)" and must be added to sae_groups; the hostapd EAP-pwd server uses group 19 by default, so only a pwd_group of 28-30 is affected, while a wpa_supplicant EAP-pwd peer follows whatever group a rogue server picks if the crypto library supports Brainpool. OpenSSL 1.0.2 and newer do; BoringSSL does not. The Debian security tracker marks jessie and stretch not affected, since Brainpool support for SAE was "Introduced in 2.5", and lists 2:2.7+git20190128+0c1e29f-6+deb10u1 for buster (DSA-4538-1). We found no device vendor advisory naming this CVE; our hostapd and wpa_supplicant firmware guide shows how to check sae_groups and pwd_group in a device configuration.

How do you fix or mitigate it?

Ship hostapd and wpa_supplicant 2.9 or later, or merge the six w1.fi 2019-6 patches, the last of which is "dragonfly: Disable use of groups using Brainpool curves". On any version, keep sae_groups and pwd_group to the default NIST groups. w1.fi recommends disabling Brainpool for SAE and EAP-pwd because constant-time handling of those primes "may not be practical for all devices." The EAP-pwd crash fixed in 2.8 is CVE-2019-11555; for the Wi-Fi Direct bug see CVE-2021-27803.

CVE-2019-13377 is not in the CISA KEV catalog as of 2026-11-18 (catalog version 2026.11.17). FIRST EPSS gives it 0.02536, the 85.2nd percentile, in the 2026-11-17 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: SAE and EAP-pwd in hostapd and wpa_supplicant 2.x through 2.8, timing and cache side channels with Brainpool curves, full password recovery; Primary CVSS v3.1 5.9 AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:N/A:N, CVSS v2 4.3; CWE-203; CPE hostapd 2.0 to 2.8; published 2019-08-15. — NVD
  • Mathy Vanhoef, Dragonblood: Brainpool curves need several iterations depending on password and MAC address; leak confirmed against the latest hostapd and password brute-forced; CVE-2019-13377 allocated. — Dragonblood (Mathy Vanhoef)
  • w1.fi 2019-6 (2019-08-07): earlier fixes incomplete for groups 28-30; cache attacks still feasible with v2.8; groups not enabled by default in v2.8 hostapd or any wpa_supplicant; EAP-pwd server default group 19; OpenSSL 1.0.2+ supports Brainpool, BoringSSL does not; update to v2.9 or merge six patches including disabling Brainpool groups. — w1.fi
  • Debian: jessie and stretch not affected (introduced in 2.5, Brainpool support for SAE); buster fixed 2:2.7+git20190128+0c1e29f-6+deb10u1, DSA-4538-1; unstable 2:2.9-1. — Debian security tracker
  • OpenWrt hostapd Makefile: PKG_SOURCE_DATE 2026-08-07 (main), 2025-08-26 (openwrt-25.12), 2024-09-15 (openwrt-24.10). — OpenWrt
  • CVE-2019-13377 is not in the CISA KEV catalog, version 2026.11.17 (1,782 entries), checked 2026-11-18. — CISA KEV
  • EPSS 0.02536, percentile 0.85221, dataset 2026-11-17. — FIRST EPSS

Related advisories

Sources

  1. NVD
  2. Dragonblood: Analysing WPA3's Dragonfly Handshake (Mathy Vanhoef)
  3. w1.fi security advisory 2019-6
  4. Debian security tracker
  5. OpenWrt: hostapd package Makefile
  6. CISA KEV
  7. FIRST EPSS

Frequently asked questions

Is a WPA3 access point with default settings affected by CVE-2019-13377?

Usually not. w1.fi says the Brainpool groups 28-30 are not enabled by default in hostapd 2.8 or in any wpa_supplicant version and must be added to sae_groups; the EAP-pwd server defaults to group 19 (per w1.fi advisory 2019-6).

What can an attacker get from CVE-2019-13377?

Timing and cache-access differences that reveal how many iterations the password encoding took. Mathy Vanhoef used this leak against hostapd to brute-force the password offline (per the Dragonblood research page).

Which version fixes CVE-2019-13377?

hostapd and wpa_supplicant 2.9, which includes the w1.fi 2019-6 patches and disables Brainpool groups for SAE and EAP-pwd. Debian fixed buster in 2:2.7+git20190128+0c1e29f-6+deb10u1 (per w1.fi and the Debian security tracker).

Is CVE-2019-13377 in the CISA KEV catalog?

No, not as of 2026-11-18 (catalog version 2026.11.17). FIRST EPSS gives it 0.02536, the 85.2nd percentile, in the 2026-11-17 dataset (per CISA KEV and FIRST EPSS).