CVE-2015-4141 is a buffer overflow (CWE-119) in the WPS UPnP HTTP parser shared by hostapd and wpa_supplicant 0.7.0 to 2.4: a very large chunk length becomes negative and memcpy() overruns, ending the process. It needs WPS UPnP enabled and an attacker on the IP network. NVD scores it 4.3.
Vulnerability summary
CVSS base score
4.3
Medium severity
EPSS score (FIRST)
3.0%
Probability of exploitation in the wild in the next 30 days
CISA KEV catalog
No
Not listed
Published
Vulnerability ID: CVE-2015-4141 · Source: NVD
Affected components
| Component | CPE | Version range |
|---|---|---|
| hostapd (WPS UPnP) | cpe:2.3:a:w1.fi:hostapd:*:*:*:*:*:*:*:* | 0.7.0 to 2.4 with CONFIG_WPS_UPNP=y and upnp_iface set; fixed in 2.5 (w1.fi 2015-2) |
| wpa_supplicant (WPS external registrar) | cpe:2.3:a:w1.fi:wpa_supplicant:*:*:*:*:*:*:*:* | 0.7.0 to 2.4 with CONFIG_WPS_ER=y and WPS_ER_START issued; fixed in 2.5 (w1.fi 2015-2) |
| 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); WPS UPnP is commented out in the full build configuration |
What does this mean for affected devices?
NVD's description: "The WPS UPnP function in hostapd, when using WPS AP, and wpa_supplicant, when using WPS external registrar (ER), 0.7.0 through 2.4 allows remote attackers to cause a denial of service (crash) via a negative chunk length, which triggers an out-of-bounds read or heap-based buffer overflow." NVD's Primary score is CVSS v2 4.3 (AV:N/AC:M/Au:N/C:N/I:N/A:P) with CWE-119; the record has no CVSS v3 score. The CPE list names every hostapd and wpa_supplicant release from 0.7.0 to 2.4, and the record was published on 2015-06-15 (NVD).
What goes wrong in the parser?
The w1.fi 2015-2 advisory, published 2015-05-04, says the HTTP code used for WPS UPnP "uses a signed integer for storing the length of a HTTP chunk" when chunked transfer encoding is used, so a length of 0x8000000 or more can go negative. The length checks then pass "a negative value to the memcpy", causing "stack buffer read overflow and heap buffer write overflow." Because the negative value is read as a huge size, the advisory says "process execution terminates in practice" before the corrupted heap is used, so the result is a denial of service. Kostya Kortchinsky of Google Security Team reported it, and the CVE was requested on oss-security on 2015-05-09.
Which builds can be hit?
Only builds with the WPS UPnP code compiled in and switched on. On the access point side that means hostapd with CONFIG_WPS_UPNP=y and a upnp_iface line in the runtime configuration; w1.fi notes that for a "WPS AP (hostapd) with a wired network connectivity, this is commonly enabled." On the client side it means wpa_supplicant with CONFIG_WPS_ER=y and an external registrar started with WPS_ER_START. The attacker must reach the device over IP, normally from the LAN. OpenWrt's full hostapd build configuration leaves CONFIG_WPS_UPNP commented out, and its wpa_supplicant full configuration does the same for CONFIG_WPS_ER. The Debian security tracker lists fixed wpa packages 1.0-3+deb7u3 for wheezy and 2.3-1+deb8u3 for jessie (DSA-3397-1). We found no device vendor advisory that names this CVE; our hostapd and wpa_supplicant firmware guide shows how to read the build options from a binary.
How do you fix or mitigate it?
Ship hostapd and wpa_supplicant 2.5 or later, or merge the w1.fi patch "WPS: Fix HTTP chunked transfer encoding parser". Without an update, remove upnp_iface from the hostapd configuration, do not start WPS ER, or build without CONFIG_WPS_UPNP and CONFIG_WPS_ER. The same WPS UPnP service carries CVE-2020-12695 (CallStranger), and the memory corruption fix guide covers signed length handling. Other hostapd pages: CVE-2019-11555 and CVE-2021-27803.
CVE-2015-4141 is not in the CISA KEV catalog as of 2026-11-18 (catalog version 2026.11.17). FIRST EPSS gives it 0.02994, the 87.6th percentile, in the 2026-11-17 dataset (CISA KEV, FIRST EPSS).
Evidence
- NVD: WPS UPnP in hostapd (WPS AP) and wpa_supplicant (WPS ER) 0.7.0 through 2.4, DoS via negative chunk length, out-of-bounds read or heap overflow; Primary CVSS v2 4.3 AV:N/AC:M/Au:N/C:N/I:N/A:P; CWE-119; published 2015-06-15. — NVD
- w1.fi 2015-2 (2015-05-04): signed chunk length, negative at 0x8000000 or more, memcpy stack read and heap write overflow, process terminates in practice; hostapd 0.7.0-2.4 with CONFIG_WPS_UPNP=y and upnp_iface; wpa_supplicant 0.7.0-2.4 with CONFIG_WPS_ER=y and WPS_ER_START; fix in v2.5; reported by Kostya Kortchinsky. — w1.fi
- oss-security 2015-05-09: CVE request for the hostapd/wpa_supplicant WPS UPnP chunked transfer encoding issue, quoting the w1.fi advisory. — oss-security
- OpenWrt hostapd-full.config: #CONFIG_WPS_UPNP=y (commented out). — OpenWrt
- OpenWrt wpa_supplicant-full.config: #CONFIG_WPS_ER=y (commented out). — OpenWrt
- Debian: wpa fixed 1.0-3+deb7u3 (wheezy) and 2.3-1+deb8u3 (jessie), DSA-3397-1. — Debian security tracker
- CVE-2015-4141 is not in the CISA KEV catalog, version 2026.11.17 (1,782 entries), checked 2026-11-18. — CISA KEV
- EPSS 0.02994, percentile 0.87570, dataset 2026-11-17. — FIRST EPSS

Related advisories
Sources
Frequently asked questions
Which hostapd and wpa_supplicant builds are affected by CVE-2015-4141?
Versions 0.7.0 to 2.4, and only when the WPS UPnP code is built and enabled: hostapd with CONFIG_WPS_UPNP=y and a upnp_iface setting, or wpa_supplicant with CONFIG_WPS_ER=y after a WPS_ER_START command (per w1.fi advisory 2015-2).
Can CVE-2015-4141 be exploited over Wi-Fi from outside the network?
w1.fi says WPS UPnP runs over a trusted IP network, so the attacker needs IP access to the device, normally from the LAN side. The practical effect is a crash of the hostapd or wpa_supplicant process (per w1.fi advisory 2015-2).
How do I mitigate CVE-2015-4141 without updating?
Remove the upnp_iface parameter from the hostapd configuration, do not start WPS ER in wpa_supplicant, or rebuild without CONFIG_WPS_UPNP and CONFIG_WPS_ER (per w1.fi advisory 2015-2).
Is CVE-2015-4141 in the CISA KEV catalog?
No, not as of 2026-11-18 (catalog version 2026.11.17). FIRST EPSS gives it 0.02994, the 87.6th percentile, in the 2026-11-17 dataset (per CISA KEV and FIRST EPSS).
Questions about this page?