CVE-2026-68124

Description

In the Linux kernel, the following vulnerability has been resolved: mctp: serial: handle zero-length frames to prevent rx buffer overflow The MCTP serial receive state machine reads a frame length byte in mctp_serial_push_header() case 2 and validates it upper-bound-only: if (c > MCTP_SERIAL_FRAME_MTU) { dev->rxstate = STATE_ERR; } else { dev->rxlen = c; dev->rxpos = 0; dev->rxstate = STATE_DATA; ... } A length of zero passes this check, so rxlen is set to 0 and the state machine advances to STATE_DATA. In mctp_serial_push() STATE_DATA, the incoming byte is stored and rxpos incremented before the terminator is dev->rxbuf[dev->rxpos] = c; dev->rxpos++; dev->rxstate = STATE_DATA; if (dev->rxpos == dev->rxlen) { dev->rxpos = 0; dev->rxstate = STATE_TRAILER; } With rxlen == 0 the "rxpos == rxlen" terminator can never fire (rxpos is already 1 on the first data byte), so subsequent bytes are written past the end of the fixed 74-byte rxbuf, which is the last member of the netdev private area. Every following data byte is an attacker-controlled 1-byte out-of-bounds heap write, and the overflow continues until a frame (0x7e) or escape byte resets the parser -- effectively unbounded. Reaching this requires CAP_NET_ADMIN to attach the N_MCTP line discipline and bring the resulting mctpserialN netdev up, after which the bytes arrive via the tty receive path. Route a zero-length frame straight to STATE_TRAILER instead of STATE_DATA. The trailer/framing bytes are still consumed, and the frame resolves to a zero-length skb that the MCTP core rejects; the parser never enters STATE_DATA with rxlen == 0, so the out-of-bounds write can no longer occur. KASAN, on a frame of 0x7e 0x01 0x00 followed by data bytes (before this change): UBSAN: array-index-out-of-bounds in drivers/net/mctp/mctp-serial.c:370 index 74 is out of range for type 'u8 [74]' BUG: KASAN: slab-out-of-bounds in mctp_serial_tty_receive_buf Write of size 1 at addr ... by task kworker/u16:0 mctp_serial_tty_receive_buf tty_ldisc_receive_buf flush_to_ldisc Allocated by task 152: alloc_netdev_mqs mctp_serial_open v2: route zero-length frames to STATE_TRAILER instead of STATE_ERR so the trailer/framing bytes are still consumed (Jeremy Kerr). Found by 0sec automated security-research tooling (https://0sec.ai).

Gravité et risque d'exploitation

CVE-2026-68124 obtient un score CVSS de 9.6/10, soit une gravité critique. Selon le modèle EPSS, la probabilité qu'elle soit exploitée dans les 30 prochains jours est de 0.34 %, ce qui la place devant 26.3 % des CVE référencées.

Que faire ?

• Vérifiez si Linux (versions <36dc6d6964a3b90411cc7944cd9b8b6f67b9807b, <=6.1.*, 7.2, <793b9b729f1e8de57be8c8daf1a9838be96cabed, <=6.12.*, 6.12.101, 0, <=7.1.*, <5.17, <68819427bc07eca7963a9e8be19e5272cc29186c, a0c2ccd9b5ad0a9e838158404e041b5a8ff762dd, <=6.6.*, 6.18.42, 7.1.6, 6.1.183, <06a6b606129c8a25cd457760f5370f3ff01fe05d, <=6.18.*, <=*, <64b96ae7912244d55257aa330d9569ee0a8f8d99, 6.6.148, 5.17, <f80ba170d7b3a44e3d244a2c8e06031d61bf3b23) est présent sur votre parc.
• Appliquez le correctif ou la version corrigée indiqués dans les références de l'éditeur.
• À défaut de correctif, limitez l'exposition (filtrage réseau, désactivation du composant).
• Automatisez la détection : les agents CVE Heimdall (gratuits, open source) corrèlent l'inventaire de vos serveurs avec la base CVE.

Références

• https://git.kernel.org/stable/c/64b96ae7912244d55257aa330d9569ee0a8f8d99
• https://git.kernel.org/stable/c/36dc6d6964a3b90411cc7944cd9b8b6f67b9807b
• https://git.kernel.org/stable/c/68819427bc07eca7963a9e8be19e5272cc29186c
• https://git.kernel.org/stable/c/f80ba170d7b3a44e3d244a2c8e06031d61bf3b23
• https://git.kernel.org/stable/c/06a6b606129c8a25cd457760f5370f3ff01fe05d
• https://git.kernel.org/stable/c/793b9b729f1e8de57be8c8daf1a9838be96cabed