---
id: "CVE-2026-98465"
kind: "vulnerability"
aliases: ["CVE-2026-98465"]
title: "Linux kernel RISC-V KVM: race lets an IMSIC attribute access another vCPU's interrupt file"
severity: "unknown"
cvss: null
epss: null
exploited: false
status: "open"
category: "runtime"
products: []
ecosystems: []
tags: ["hypervisor", "kvm", "linux-kernel", "risc-v", "runtime"]
not_affected_if: ["no-riscv-kvm", "x86-or-arm-hosts"]
published_at: "2026-10-18T16:16:56+00:00"
first_seen_at: "2026-10-23T21:57:15+00:00"
updated_at: "2026-10-25T05:27:09+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-98465", "first_seen_at": "2026-10-23T21:57:15+00:00"}]
---

# Linux kernel RISC-V KVM: race lets an IMSIC attribute access another vCPU's interrupt file

A race in the Linux kernel's RISC-V KVM code lets a device ioctl read or write the interrupt file of a different virtual CPU after a concurrent vCPU migration recycles it. The sources describe the bug and its fix, but do not state an attacker's achievable impact.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- unknown

## Fix

- Update the Linux kernel to a release that includes the upstream commit serializing IMSIC attribute access with the vCPU mutex; the sources do not name a version.

## Timeline

- 2026-10-23T21:57:15+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-98465

## From the sources

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

RISC-V: KVM: Serialize IMSIC attributes with vCPU migration

KVM device ioctls are not serialized against KVM_RUN. As a result,
kvm_riscv_aia_imsic_rw_attr() can snapshot the physical CPU and HGEI of
an IMSIC VS-file before a concurrent vCPU migration releases it.

The HGEI can then be allocated to another vCPU before imsic_vsfile_rw()
uses the stale tuple. A GET or SET attribute may consequently access the
new owner's interrupt file.

Serialize the entire IMSIC attribute operation with the target vCPU
mutex. This prevents the VS-file from being migrated and recycled until
the attribute access completes. Acquire the mutex killably so that the
device ioctl remains interruptible while waiting for KVM_RUN to finish.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_