GHSA-27VC-2785-HFJH
Vulnerability from github – Published: 2026-10-23 15:31 – Updated: 2026-10-23 15:31In the Linux kernel, the following vulnerability has been resolved:
net: always dissect GSO packets in __virtio_net_hdr_to_skb()
Commit 9e8db5913264 ("net: avoid false positives in untrusted gso validation") added a '&& skb->network_header' check before flow-dissecting GSO packets without VIRTIO_NET_HDR_F_NEEDS_CSUM in __virtio_net_hdr_to_skb(), because some callers (such as tun_get_user(), tun_xdp_one(), virtnet_receive_done(), and raw_verify_header()) called virtio_net_hdr_*_to_skb() before initializing skb->network_header and skb->dev.
Because __alloc_skb() and __build_skb_around() zero-initialize skb->network_header to 0 (unlike mac_header and transport_header which are initialized to ~0U), those four callers always had skb->network_header == 0 and bypassed flow dissection in __virtio_net_hdr_to_skb(). More generally, skb->network_header is an offset from skb->head (where 0 is also a valid offset whenever skb_headroom(skb) is 0), not a boolean flag.
Whenever the 'if (gso_type && skb->network_header)' branch was skipped, the fallback 'else if (gso_type)' only pulled nh_min_len + thlen (40 bytes for TCPv4) without dissecting the packet, without validating ip_proto or n_proto, and without setting skb->transport_header.
If the packet has a malformed network header, it is not rejected and a subsequent skb_probe_transport_header() also fails, leaving skb->transport_header at ~0U (0xffff). Similarly, if an IPv4 packet carries IP options (ihl > 5) or an IPv6 packet carries extension headers, pulling only nh_min_len + thlen can leave the TCP header outside skb->head. In both cases, tcp_hdrlen(skb) in skb_gso_transport_seglen() reads out-of-bounds:
BUG: KASAN: slab-out-of-bounds in skb_gso_transport_seglen Read of size 2 by task poc/133 skb_gso_transport_seglen (net/core/gso.c:155) skb_gso_validate_mac_len (net/core/gso.c:270) tbf_enqueue (net/sched/sch_tbf.c:260) dev_qdisc_enqueue (net/core/dev.c:4227) __dev_queue_xmit (net/core/dev.c:4884)
In addition, checking virtio_net_hdr_match_proto() only inside 'if (!skb->protocol)' before flow dissection both skipped validation when skb->protocol was pre-set by the caller and rejected VLAN-tagged frames whose outer L2 protocol is ETH_P_8021Q or ETH_P_8021AD.
Fix this by: 1. Initializing skb->dev and skb->network_header (plus skb->protocol for IFF_TUN) before virtio_net_hdr_*_to_skb() in tun_get_user(), tun_xdp_one(), virtnet_receive_done(), and raw_verify_header(). In tun_get_user(), drop the redundant skb_reset_mac_header(skb) in the IFF_TUN case since __virtio_net_hdr_to_skb() unconditionally resets mac_header. 2. Removing '&& skb->network_header' and the unvalidated 'else if (gso_type)' fallback in __virtio_net_hdr_to_skb() so all GSO packets without VIRTIO_NET_HDR_F_NEEDS_CSUM are flow-dissected, have their transport header pulled into linear data, and have skb->transport_header set. 3. Moving the virtio_net_hdr_match_proto() check to after skb_flow_dissect_flow_keys_basic(), validating keys.basic.n_proto against hdr_gso_type.
{
"affected": [],
"aliases": [
"CVE-2026-98954"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-10-23T13:17:16Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nnet: always dissect GSO packets in __virtio_net_hdr_to_skb()\n\nCommit 9e8db5913264 (\"net: avoid false positives in untrusted gso\nvalidation\") added a \u0027\u0026\u0026 skb-\u003enetwork_header\u0027 check before flow-dissecting\nGSO packets without VIRTIO_NET_HDR_F_NEEDS_CSUM in\n__virtio_net_hdr_to_skb(), because some callers (such as tun_get_user(),\ntun_xdp_one(), virtnet_receive_done(), and raw_verify_header()) called\nvirtio_net_hdr_*_to_skb() before initializing skb-\u003enetwork_header and\nskb-\u003edev.\n\nBecause __alloc_skb() and __build_skb_around() zero-initialize\nskb-\u003enetwork_header to 0 (unlike mac_header and transport_header which\nare initialized to ~0U), those four callers always had\nskb-\u003enetwork_header == 0 and bypassed flow dissection in\n__virtio_net_hdr_to_skb(). More generally, skb-\u003enetwork_header is an\noffset from skb-\u003ehead (where 0 is also a valid offset whenever\nskb_headroom(skb) is 0), not a boolean flag.\n\nWhenever the \u0027if (gso_type \u0026\u0026 skb-\u003enetwork_header)\u0027 branch was skipped,\nthe fallback \u0027else if (gso_type)\u0027 only pulled nh_min_len + thlen (40 bytes\nfor TCPv4) without dissecting the packet, without validating ip_proto or\nn_proto, and without setting skb-\u003etransport_header.\n\nIf the packet has a malformed network header, it is not rejected and a\nsubsequent skb_probe_transport_header() also fails, leaving\nskb-\u003etransport_header at ~0U (0xffff). Similarly, if an IPv4 packet\ncarries IP options (ihl \u003e 5) or an IPv6 packet carries extension headers,\npulling only nh_min_len + thlen can leave the TCP header outside\nskb-\u003ehead. In both cases, tcp_hdrlen(skb) in skb_gso_transport_seglen()\nreads out-of-bounds:\n\n BUG: KASAN: slab-out-of-bounds in skb_gso_transport_seglen\n Read of size 2 by task poc/133\n skb_gso_transport_seglen (net/core/gso.c:155)\n skb_gso_validate_mac_len (net/core/gso.c:270)\n tbf_enqueue (net/sched/sch_tbf.c:260)\n dev_qdisc_enqueue (net/core/dev.c:4227)\n __dev_queue_xmit (net/core/dev.c:4884)\n\nIn addition, checking virtio_net_hdr_match_proto() only inside\n\u0027if (!skb-\u003eprotocol)\u0027 before flow dissection both skipped validation when\nskb-\u003eprotocol was pre-set by the caller and rejected VLAN-tagged frames\nwhose outer L2 protocol is ETH_P_8021Q or ETH_P_8021AD.\n\nFix this by:\n1. Initializing skb-\u003edev and skb-\u003enetwork_header (plus skb-\u003eprotocol for\n IFF_TUN) before virtio_net_hdr_*_to_skb() in tun_get_user(),\n tun_xdp_one(), virtnet_receive_done(), and raw_verify_header(). In\n tun_get_user(), drop the redundant skb_reset_mac_header(skb) in the\n IFF_TUN case since __virtio_net_hdr_to_skb() unconditionally resets\n mac_header.\n2. Removing \u0027\u0026\u0026 skb-\u003enetwork_header\u0027 and the unvalidated\n \u0027else if (gso_type)\u0027 fallback in __virtio_net_hdr_to_skb() so all GSO\n packets without VIRTIO_NET_HDR_F_NEEDS_CSUM are flow-dissected, have\n their transport header pulled into linear data, and have\n skb-\u003etransport_header set.\n3. Moving the virtio_net_hdr_match_proto() check to after\n skb_flow_dissect_flow_keys_basic(), validating keys.basic.n_proto\n against hdr_gso_type.",
"id": "GHSA-27vc-2785-hfjh",
"modified": "2026-10-23T15:31:32Z",
"published": "2026-10-23T15:31:32Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-98954"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/44378d02c5aaae24b60a75fbbb539b4aa4db503d"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/5f4dfdf0077730e9d3ef5b37ceb17a5a56d6f34c"
}
],
"schema_version": "1.4.0",
"severity": []
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.