« Volver al listado

CVE-2026-74560

Estado: RecibidaSin puntuar—

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

xsk: fix buffer leak in xsk_drop_skb() for AF_XDP multi-buffer Tx

This patch is inspired by the check[1] from sashiko. It says when overflow happens, the address of cq to be published is invalid. Actually the severer thing is the whole process of publishing the address of cq in this particular case is not right: it should truely publish the address and advance the cached_prod in cq as long as it reads descriptors from txq.

The following is the full analysis. xsk_drop_skb() is called in three places, which all discard a partially built multi-buffer skb: 1) xsk_build_skb() -EOVERFLOW error path: packet exceeds MAX_SKB_FRAGS 2) __xsk_generic_xmit() post-loop cleanup: an invalid descriptor in the TX ring prevents the partial packet from completing 3) xsk_release(): socket close while xs->skb holds an incomplete packet

Leer descripción completaMostrar menos

In all three cases, the TX descriptors for the already-processed frags have been consumed from the TX ring (xskq_cons_release), and CQ slots have been reserved. However, xsk_drop_skb() calls xsk_consume_skb() which cancels the CQ reservations via xsk_cq_cancel_locked(). Since the buffer addresses never appear in the completion queue, userspace permanently loses track of these buffers.

Fix this by letting consume_skb() trigger the existing xsk_destruct_skb destructor, which already submits buffer addresses to the CQ via xsk_cq_submit_addr_locked().

Note that cancelling the descriptors back to the TX ring (via xskq_cons_cancel_n) is not a appropriate option because an oversized packet that always exceeds MAX_SKB_FRAGS would be retried indefinitely, which is an obviously deadlock bug in the TX path.

Also move the desc->addr assignment in xsk_build_skb() above the overflow check so that the current descriptor's address is recorded before a potential -EOVERFLOW jump to free_err, consistent with the zerocopy path in xsk_build_skb_zerocopy().

[1]: https://lore.kernel.org/all/20260425041726.85FB3C2BCB2@smtp.kernel.org/

CVSS

NVD no ha asignado puntuación CVSS a esta CVE (habitual desde el cambio de política de abril de 2026).

Probabilidad de explotación (EPSS)

EPSS (Exploit Prediction Scoring System, de FIRST) estima la probabilidad de que una vulnerabilidad sea explotada en 30 días. Complementa a CVSS (impacto) y a CISA KEV (explotación confirmada).

Tecnologías afectadas (1)

⚠ Inferidas por IA a partir de la descripción — NVD aún no ha analizado esta CVE; no son CPE verificados.

Linux — Linux KernelAI

Referencias

JSON original (NVD)

Mostrar
{
  "id": "CVE-2026-74560",
  "cveTags": [],
  "metrics": {},
  "affected": [
    {
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
      "affectedData": [
        {
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "product": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "cf24f5a5feeaae34c1a34d1e04f8ac697290427a",
              "lessThan": "05e283466b86e0b25a5209a3eafe87aa9886064a",
              "versionType": "git"
            },
            {
              "status": "affected",
              "version": "cf24f5a5feeaae34c1a34d1e04f8ac697290427a",
              "lessThan": "a0528ab6af62bf57153caaf8ea28e5fcb0e18fd4",
              "versionType": "git"
            },
            {
              "status": "affected",
              "version": "cf24f5a5feeaae34c1a34d1e04f8ac697290427a",
              "lessThan": "a3c8382ebce4780c6b3ace2c09bc342313ac0186",
              "versionType": "git"
            }
          ],
          "programFiles": [
            "net/xdp/xsk.c"
          ],
          "defaultStatus": "unaffected"
        },
        {
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "product": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "6.6"
            },
            {
              "status": "unaffected",
              "version": "0",
              "lessThan": "6.6",
              "versionType": "semver"
            },
            {
              "status": "unaffected",
              "version": "6.18.44",
              "versionType": "semver",
              "lessThanOrEqual": "6.18.*"
            },
            {
              "status": "unaffected",
              "version": "7.1.8",
              "versionType": "semver",
              "lessThanOrEqual": "7.1.*"
            },
            {
              "status": "unaffected",
              "version": "7.2",
              "versionType": "original_commit_for_fix",
              "lessThanOrEqual": "*"
            }
          ],
          "programFiles": [
            "net/xdp/xsk.c"
          ],
          "defaultStatus": "affected"
        }
      ]
    }
  ],
  "published": "2026-08-15T13:18:01.483",
  "references": [
    {
      "url": "https://git.kernel.org/stable/c/05e283466b86e0b25a5209a3eafe87aa9886064a",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    },
    {
      "url": "https://git.kernel.org/stable/c/a0528ab6af62bf57153caaf8ea28e5fcb0e18fd4",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    },
    {
      "url": "https://git.kernel.org/stable/c/a3c8382ebce4780c6b3ace2c09bc342313ac0186",
      "source": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
    }
  ],
  "vulnStatus": "Received",
  "descriptions": [
    {
      "lang": "en",
      "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nxsk: fix buffer leak in xsk_drop_skb() for AF_XDP multi-buffer Tx\n\nThis patch is inspired by the check[1] from sashiko. It says when\noverflow happens, the address of cq to be published is invalid.\nActually the severer thing is the whole process of publishing the\naddress of cq in this particular case is not right: it should truely\npublish the address and advance the cached_prod in cq as long as it\nreads descriptors from txq.\n\nThe following is the full analysis.\nxsk_drop_skb() is called in three places, which all discard a partially\nbuilt multi-buffer skb:\n1) xsk_build_skb() -EOVERFLOW error path: packet exceeds MAX_SKB_FRAGS\n2) __xsk_generic_xmit() post-loop cleanup: an invalid descriptor in\n   the TX ring prevents the partial packet from completing\n3) xsk_release(): socket close while xs->skb holds an incomplete packet\n\nIn all three cases, the TX descriptors for the already-processed frags\nhave been consumed from the TX ring (xskq_cons_release), and CQ slots\nhave been reserved. However, xsk_drop_skb() calls xsk_consume_skb()\nwhich cancels the CQ reservations via xsk_cq_cancel_locked(). Since\nthe buffer addresses never appear in the completion queue, userspace\npermanently loses track of these buffers.\n\nFix this by letting consume_skb() trigger the existing xsk_destruct_skb\ndestructor, which already submits buffer addresses to the CQ via\nxsk_cq_submit_addr_locked().\n\nNote that cancelling the descriptors back to the TX ring (via\nxskq_cons_cancel_n) is not a appropriate option because an oversized\npacket that always exceeds MAX_SKB_FRAGS would be retried indefinitely,\nwhich is an obviously deadlock bug in the TX path.\n\nAlso move the desc->addr assignment in xsk_build_skb() above the\noverflow check so that the current descriptor's address is recorded\nbefore a potential -EOVERFLOW jump to free_err, consistent with the\nzerocopy path in xsk_build_skb_zerocopy().\n\n[1]: https://lore.kernel.org/all/20260425041726.85FB3C2BCB2@smtp.kernel.org/"
    }
  ],
  "lastModified": "2026-08-17T06:19:54.100",
  "sourceIdentifier": "416baaa9-dc9f-4396-8d5f-8c081fb06d67"
}