After dispatching a decrypt operation to OCF and receiving the result, the wg(4) driver failed to check whether the MAC verification step succeeded. The driver thus silently accepted packets with an invalid Poly1305 authentication tag.

A remote attacker who can send UDP packets to a WireGuard endpoint, and who can guess the bounds of the receiver's replay window, can inject forged or modified transport data packets into the tunnel.

A remote attacker who can intercept WireGuard packets bound for a FreeBSD host can modify the ciphertext and authenticated data without detection by the receiver.

WireGuard Easy through 15.3.0, fixed in commit 66b292b, contains a cryptographically weak one-time link token generation vulnerability that allows unauthenticated network attackers to recover WireGuard peer credentials by brute-forcing a keyspace of at most 1000 candidate tokens per client ID, as the token is computed using CRC32 over a random value constrained to 0-999. Attackers can enumerate candidate tokens against the unauthenticated /cnf/:oneTimeLink route, which lacks rate limiting and does not validate token expiration, to obtain a peer's PrivateKey and PresharedKey and impersonate that peer on the VPN network.

In the Linux kernel, the following vulnerability has been resolved:

wireguard: socket: free skb in send6 when ipv6 is disabled

I got a memory leak report:

unreferenced object 0xffff8881191fc040 (size 232): comm "kworker/u17:0", pid 23193, jiffies 4295238848 (age 3464.870s) hex dump (first 32 bytes): 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ backtrace: [<ffffffff814c3ef4>] slabpostallochook+0x84/0x3b0 [<ffffffff814c8977>] kmemcacheallocnode+0x167/0x340 [<ffffffff832974fb>] allocskb+0x1db/0x200 [<ffffffff82612b5d>] wgsocketsendbuffertopeer+0x3d/0xc0 [<ffffffff8260e94a>] wgpacketsendhandshakeinitiation+0xfa/0x110 [<ffffffff8260ec81>] wgpackethandshakesendworker+0x21/0x30 [<ffffffff8119c558>] processonework+0x2e8/0x770 [<ffffffff8119ca2a>] workerthread+0x4a/0x4b0 [<ffffffff811a88e0>] kthread+0x120/0x160 [<ffffffff8100242f>] retfromfork+0x1f/0x30

In function wgsocketsendbufferasreplytoskb() or wgsocketsend buffertopeer(), the semantics of send6() is required to free skb. But when CONFIGIPV6 is disable, kfreeskb() is missing. This patch adds it to fix this bug.

Privilege Escalation to Admin via User Self-Update in wg-portal

Summary

Any authenticated non-admin user can become a full administrator by sending a single PUT request to their own user profile endpoint with "IsAdmin": true in the JSON body. After logging out and back in, the session picks up admin privileges from the database.

Tested against wg-portal v2.1.2 (Docker image wgportal/wg-portal:v2).

Root Cause

When a user updates their own profile, the server parses the full JSON body into the user model, including the IsAdmin boolean field. A function responsible for preserving calculated or protected attributes pins certain fields to their database values (such as base model data, linked peer count, and authentication data), but it does not do this for IsAdmin. As a result, whatever value the client sends for IsAdmin is written directly to the database.

Impact

After the exploit, the attacker has full admin access to the WireGuard VPN management portal. They can:

- Read and modify every user account - Create, modify, and delete WireGuard peers on any interface - View WireGuard interface configurations - Disable or lock other user accounts - Access the full user list and their API tokens

Patches The problem was fixed in the latest release, v2.1.3. The docker images for the tag 'latest' built from the master branch also include the fix.

WireGuard, such as WireGuard 0.5.3 on Windows, does not fully account for the possibility that an adversary might be able to set a victim's system time to a future value, e.g., because unauthenticated NTP is used. This can lead to an outcome in which one static private key becomes permanently useless.

The WireGuard client 0.5.3 on Windows insecurely configures the operating system and firewall such that traffic to a local network that uses non-RFC1918 IP addresses is blocked. This allows an adversary to trick the victim into blocking IP traffic to selected IP addresses and services even while the VPN is enabled. NOTE: the tunnelcrack.mathyvanhoef.com website uses this CVE ID to refer more generally to "LocalNet attack resulting in the blocking of traffic" rather than to only WireGuard.