PoC Released for Linux Vhost Flaw Enabling Root Privilege Escalation and VM Escape

A public proof of concept for Vghost, a high-severity Linux vhost flaw tracked as CVE-2026-74580, shows how the bug can let a local user gain root access or, in certain virtual machine setups, break out of a guest and affect the host. The issue has a CVSS v3.1 score of 8.8 and is fixed in updated Linux kernel releases.

Vhost is a Linux kernel system that helps virtual machines handle network and other device traffic more quickly. It uses shared virtual queues, called virtqueues, to carry device data between a guest and the host. The flaw lies in how vhost caches IOTLB mappings used to find those queues in memory.

The problem occurs when a virtual ring is reconfigured. Linux updates the ring’s address or size but doesn’t clear the cached mapping for the ring’s descriptor, available, or used data.

A later cache hit can therefore skip the expected address check and translate the new address using an old mapping. The resulting pointer may reach kernel copy routines with an invalid target.

Researchers on the Seclists mailing list traced the bug to two earlier kernel changes: one introduced the metadata cache, while the other deferred guest I/O address validation.

PoC Released for Linux Vhost Flaw

For local privilege escalation, an attacker first needs permission to open a vhost device such as /dev/vhost-net or /dev/vhost-vsock. The PoC primes the cache, changes a ring address without clearing the cached entry, then triggers the queue. The researchers say this can read and write kernel memory.

The demonstrated chain uses a kernel memory read to work out the address randomization offset, then writes to the kernel’s modprobe_path setting. Triggering a kernel module request can then run an attacker-controlled helper with root privileges.

The researchers report reproducing the chain on Ubuntu 26.04, RHEL 9.8, Amazon Linux 2023, Azure Linux 3.0, and Oracle Linux 10.2. Device permissions differed: the tested Ubuntu system limited access to the kvm group, while the other listed test systems exposed the device to ordinary users.

That difference makes device permissions an important exposure check, but it does not replace patching. Membership in kvm or another group with access to vhost devices should be treated as privileged while a system remains unpatched.

The second path starts inside a virtual machine. It requires a vhost device using a virtual IOMMU, iommu_platform, and the virtio queue-reset feature VIRTIO_F_RING_RESET. A guest can reset and re-enable a queue, leading the host to reconfigure its ring while the stale metadata remains cached.

When the guest then triggers the queue, the stale mapping can send a used-ring write beyond guest memory and into the QEMU process heap. The researchers say the corruption can crash QEMU or affect its control flow. This is a configuration-dependent VM escape risk, not a claim that every virtual machine can be escaped through the flaw.

The upstream fix clears the vring metadata cache when the ring address or size is changed, so the new configuration must pass IOTLB validation.

The first fixed upstream stable releases listed for the affected series are 5.10.265, 5.15.216, 6.1.183, 6.6.152, 6.12.104, 6.18.45, 7.1.9, and 7.2. Administrators should install a fixed kernel or a vendor backport and reboot into it.

Until then, restrict access to all vhost device nodes, including /dev/vhost-net, /dev/vhost-vsock, /dev/vhost-scsi, and /dev/vhost-vdpa-*. Do not pass these devices to untrusted containers.

For unpatched virtualization hosts, avoid exposing vhost devices with iommu_platform behind a virtual IOMMU to untrusted guests. Permission changes do not revoke file descriptors that are already open, and they do not address the guest-triggered path.

For related coverage, see Cyber Security News’ reports on Linux kernel privilege escalation flaws and KVM guest-to-host escape risks. The technical analysis and PoC are available from Tencent Zhuque Lab and GitHub.

Update defenses against new malware as soon as it appears. Integrate TI Feeds with data from 16K SOCs.

The post PoC Released for Linux Vhost Flaw Enabling Root Privilege Escalation and VM Escape appeared first on Cyber Security News.