CVE-2026-4247
When no challenge ACK should be sent the function returns and leaks the mbuf.
Does this matter?
High impact if exploited, but EPSS currently rates exploitation as unlikely (1.12%). Schedule it in the normal patch cycle and watch for a rise in EPSS or a public exploit.
Description
When a challenge ACK is to be sent tcp_respond() constructs and sends the challenge ACK and consumes the mbuf that is passed in. When no challenge ACK should be sent the function returns and leaks the mbuf. If an attacker is either on path with an established TCP connection, or can themselves establish a TCP connection, to an affected FreeBSD machine, they can easily craft and send packets which meet the challenge ACK criteria and cause the FreeBSD host to leak an mbuf for each crafted packet in excess of the configured rate limit settings i.e. with default settings, crafted packets in excess of the first 5 sent within a 1s period will leak an mbuf. Technically, off-path attackers can also exploit this problem by guessing the IP addresses, TCP port numbers and in some cases the sequence numbers of established connections and spoofing packets towards a FreeBSD machine, but this is harder to do effectively.
- CVSS 3.1
- 7.5 HIGHCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
- EPSS
- 1.12% probability · 64th percentile
- CISA KEV
- Not listed
- Weakness
- CWE-401
- Affected
- freebsd/freebsd
- Source
- secteam@freebsd.org
References
Source: NVD record, EPSS from FIRST.org, KEV from CISA. Refreshed daily.