{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2026-72380",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2026-72380"
  ],
  "published": "2026-08-15T06:04:33Z",
  "summary": "xen/pvcalls: bound backend response req_id before indexing rsp[]",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nxen/pvcalls: bound backend response req_id before indexing rsp[]\n\npvcalls_front_event_handler() takes req_id directly from the\nbackend-supplied ring response and uses it to index the fixed-size\nbedata->rsp[] array for a memcpy() and a store, with no range check. A\nmalicious or buggy backend can set req_id past PVCALLS_NR_RSP_PER_RING\nand drive an out-of-bounds write past the bedata allocation.\n\nreq_id was also declared int while the wire field rsp->req_id is u32, so\na range check on the signed value alone is insufficient: a backend\nreq_id of 0xffffffff becomes -1, passes a >= PVCALLS_NR_RSP_PER_RING\ntest and indexes bedata->rsp[-1]. Declare req_id as u32 so a single\nbound covers both ends.\n\nA backend that sends an out-of-range req_id has violated the wire\nprotocol, so rather than silently dropping the response, log once and\nstop trusting the backend: set bedata->disabled. The event handler then\nignores further responses, and the request paths that wait for a\nresponse return -EIO instead of blocking forever. This mirrors the\nfatal-error handling xen-netback uses (xenvif_fatal_tx_err()).\n\nThe pvcalls frontend currently trusts its backend, so this is not a\nclassic-Xen security issue, but it matters for hardening PV frontends\nagainst malicious backends (confidential and disaggregated deployments).",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H"
    }
  ],
  "affected": [
    {
      "package": {
        "ecosystem": "Raspberry Pi OS:12",
        "name": "linux"
      },
      "versions": [
        "1:6.1.47-1+rpt4"
      ],
      "ecosystem_specific": {
        "binaries": [
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-2712",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v6",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v7",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v8",
            "binary_version": "1:6.1.47-1+rpt4"
          }
        ]
      },
      "database_specific": {
        "per_version": {
          "1:6.1.47-1+rpt4": {
            "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e",
            "upstream_version": "6.1.47",
            "assessment": "commit ancestry (strak rule)",
            "upstream_fixed": null,
            "compiled_filter": false
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-72380"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/d1297a9e2fd6ce08678b370d41bc980ca798f809"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/d33846c8dcc06b83b7acdeac1e8bfbb5c0c26cb2"
    }
  ],
  "database_specific": {
    "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git",
    "upstream_modified": "2026-08-17T05:50:24Z",
    "severity": {
      "source": "kernel CNA",
      "cvss_v3_score": 8.8,
      "cvss_v3_severity": "High",
      "scenario": "AV:A - The bug is reached when a malicious Xen PVCalls backend writes crafted responses into the shared grant-table ring and notifies the guest frontend via an inter-domain event channel; this is an adjacent on-host backend-to-guest path, not remote IP delivery to the guest.\nAC:L - A compromised backend fully controls ring response contents and can set any req_id (including 0xffffffff for rsp[-1]) and re-notify the frontend IRQ at will; no races or attacker-uncontrollable heap layout are required to trigger the out-of-bounds write.\nPR:N - No credentials inside the vulnerable guest are needed; once the pvcalls frontend is xenbus-connected to a backend, that backend alone can inject the malformed response. Backend/domain control is the attack position, not privileges obtainable in the guest via user namespaces.\nUI:N - No victim user or application action is required after xenbus binds the pvcalls frontend to its backend; backend-driven ring writes and event-channel interrupts alone reach pvcalls_front_event_handler() without mounts, socket use, or other interactive steps.\nS:U - The out-of-bounds write corrupts kmalloc heap memory and kernel objects within the guest running the pvcalls frontend. This is backend-to-guest hardening, not a demonstrated guest-to-host or hypervisor escape across the VM security boundary.\nC:H - Backend-chosen req_id indexes bedata->rsp[] out of bounds for a ~20-byte memcpy plus req_id store, corrupting adjacent kernel heap memory (including rsp[-1] via the signed req_id bug) and enabling plausible arbitrary kernel memory disclosure if shaped further.\nI:H - The heap out-of-bounds write can overwrite adjacent kmalloc objects or allocator metadata, providing a credible path to arbitrary guest-kernel memory modification and control-flow hijack rather than being limited to a bounded or non-security integrity effect.\nA:H - Corrupting kernel heap structures via the unbounded rsp[] indexing can immediately panic or oops the guest kernel; even unexploited attempts are likely to crash or disable the frontend, causing total loss of guest availability."
    }
  }
}