GPU VulnDB

Database/Kernel, userspace & hypervisor

Linux kernel KVM arm64: failed vGIC creation leaves redistributor regions uninitialised for teardown

CVSS 5.5CVE-2026-43351Kernel, userspace & hypervisorcurated

Impact

If vgic_allocate_private_irqs_locked() fails, kvm_vgic_create() exits early with dist->rd_regions never initialised, and kvm_vgic_dist_destroy() then walks that uninitialised state while freeing redistributor regions. The result is a host kernel fault triggered from VM creation, taking down every workload on the hypervisor rather than just the VM that failed to start. That matters on arm64 GPU hosts, where one node carries several accelerators and cannot be drained cheaply. The failure needs the allocation to actually fail, so in practice it requires memory pressure; the published vector is availability-only (C:N/I:N/A:H).

Who can reach it

Local, authenticated: anything that can create VMs on an arm64 KVM host - a cloud control plane, or a tenant with VM-creation rights - combined with host memory pressure that makes the private-IRQ allocation fail.

What to do

Take the stable kernel update that moves the static vGIC initialisation early and makes the half-failed path safe to tear down; this is a host kernel patch, so each arm64 hypervisor must be drained and rebooted onto the fixed kernel. No separate mitigation is published.

References

Related entries

All Kernel, userspace & hypervisor entries

This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.