{
  "affected": [
    {
      "ecosystem_specific": {
        "urgency": "not yet assigned"
      },
      "package": {
        "ecosystem": "Debian:13",
        "name": "linux"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "6.12.101-1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "ecosystem_specific": {
        "urgency": "not yet assigned"
      },
      "package": {
        "ecosystem": "Debian:14",
        "name": "linux"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "7.1.6-1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "ecosystem_specific": {
        "urgency": "not yet assigned"
      },
      "package": {
        "ecosystem": "Debian:12",
        "name": "linux-6.12"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "6.12.101-1~deb12u1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "details": "In the Linux kernel, the following vulnerability has been resolved:  net: gre: fix lltx regression for GRE tunnels with SEQ/CSUM  Before commit 00d066a4d4ed (\"netdev_features: convert NETIF_F_LLTX to dev-\u003elltx\"), NETIF_F_LLTX was set unconditionally in both __gre_tunnel_init() and ip6gre_tnl_init_features() alongside GRE_FEATURES:      dev-\u003efeatures |= GRE_FEATURES | NETIF_F_LLTX;  When that commit converted NETIF_F_LLTX to the dev-\u003elltx flag, it placed 'dev-\u003elltx = true' after the SEQ/CSUM early returns instead of before them. This causes GRE/GRETAP/ip6gre tunnels with SEQ or CSUM+encap to lose lockless TX, reintroducing _xmit_lock acquisition around their ndo_start_xmit. Since GRE xmit re-enters the stack via ip_tunnel_xmit(), holding _xmit_lock risks ABBA deadlock with the underlay device.    CPU0                        CPU1   ----                        ----   lock(\u0026qdisc_xmit_lock_key#6);                               lock(\u0026qdisc_xmit_lock_key#3);                               lock(\u0026qdisc_xmit_lock_key#6);   lock(\u0026qdisc_xmit_lock_key#3);  Fix by moving dev-\u003elltx = true before the early returns in both functions, restoring the original unconditional behavior.",
  "id": "DEBIAN-CVE-2026-68296",
  "modified": "2026-09-14T16:47:42.779965097Z",
  "published": "2026-08-10T13:20:19.053Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://security-tracker.debian.org/tracker/CVE-2026-68296"
    }
  ],
  "upstream": [
    "CVE-2026-68296"
  ]
}