[PATCH 1/1] [CVE-2012-1601] [HARDY] KVM: Ensure all vcpus are consistent with in-kernel irqchip settings

[Tim Gardner]

On 04/25/2012 11:53 AM, Brad Figg wrote:
> From: Avi Kivity <avi at redhat.com>
> 
> CVE-2012-1601
> 
> BugLink: http://bugs.launchpad.net/bugs/971685
> 
> If some vcpus are created before KVM_CREATE_IRQCHIP, then
> irqchip_in_kernel() and vcpu->arch.apic will be inconsistent, leading
> to potential NULL pointer dereferences.
> 
> Fix by:
> - ensuring that no vcpus are installed when KVM_CREATE_IRQCHIP is called
> - ensuring that a vcpu has an apic if it is installed after KVM_CREATE_IRQCHIP
> 
> This is somewhat long winded because vcpu->arch.apic is created without
> kvm->lock held.
> 
> Based on earlier patch by Michael Ellerman.
> 
> Signed-off-by: Michael Ellerman <michael at ellerman.id.au>
> Signed-off-by: Avi Kivity <avi at redhat.com>
> (backported from commit 3e515705a1f46beb1c942bb8043c16f8ac7b1e9e upstream)
> Signed-off-by: Brad Figg <brad.figg at canonical.com>
> ---
>  arch/x86/kvm/x86.c       |    9 +++++++++
>  include/linux/kvm_host.h |    2 ++
>  virt/kvm/kvm_main.c      |    5 +++++
>  3 files changed, 16 insertions(+)
> 
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 2085040..f036054 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -1582,6 +1582,9 @@ long kvm_arch_vm_ioctl(struct file *filp,
>  		break;
>  	}
>  	case KVM_CREATE_IRQCHIP:
> +		r = -EINVAL;
> +		if (atomic_read(&kvm->online_vcpus))
> +			goto out;
>  		r = -ENOMEM;
>  		kvm->arch.vpic = kvm_create_pic(kvm);
>  		if (kvm->arch.vpic) {
> @@ -3350,6 +3353,11 @@ void kvm_arch_check_processor_compat(void *rtn)
>  	kvm_x86_ops->check_processor_compatibility(rtn);
>  }
>  
> +bool kvm_vcpu_compatible(struct kvm_vcpu *vcpu)
> +{
> +	return irqchip_in_kernel(vcpu->kvm) == (vcpu->arch.apic != NULL);
> +}
> +
>  int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
>  {
>  	struct page *page;
> @@ -3435,6 +3443,7 @@ static void kvm_free_vcpus(struct kvm *kvm)
>  		}
>  	}
>  
> +	atomic_set(&kvm->online_vcpus, 0);
>  }
>  
>  void kvm_arch_destroy_vm(struct kvm *kvm)
> diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
> index 958e003..01055ae 100644
> --- a/include/linux/kvm_host.h
> +++ b/include/linux/kvm_host.h
> @@ -121,6 +121,7 @@ struct kvm {
>  	struct kvm_memory_slot memslots[KVM_MEMORY_SLOTS +
>  					KVM_PRIVATE_MEM_SLOTS];
>  	struct kvm_vcpu *vcpus[KVM_MAX_VCPUS];
> +	atomic_t online_vcpus;

This seems spurious. Its not part of the upstream patch, nor is it used
anywhere.

>  	struct list_head vm_list;
>  	struct file *filp;
>  	struct kvm_io_bus mmio_bus;
> @@ -306,5 +307,6 @@ struct kvm_stats_debugfs_item {
>  	struct dentry *dentry;
>  };
>  extern struct kvm_stats_debugfs_item debugfs_entries[];
> +bool kvm_vcpu_compatible(struct kvm_vcpu *vcpu);
>  
>  #endif
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index 240156e..b128dcc 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -798,12 +798,17 @@ static int kvm_vm_ioctl_create_vcpu(struct kvm *kvm, int n)
>  		goto vcpu_destroy;
>  
>  	mutex_lock(&kvm->lock);
> +	if (!kvm_vcpu_compatible(vcpu)) {
> +		r = -EINVAL;
> +		goto vcpu_destroy;
> +	}
>  	if (kvm->vcpus[n]) {
>  		r = -EEXIST;
>  		mutex_unlock(&kvm->lock);
>  		goto vcpu_destroy;
>  	}
>  	kvm->vcpus[n] = vcpu;
> +	atomic_inc(&kvm->online_vcpus);

See my comment above.

>  	mutex_unlock(&kvm->lock);
>  
>  	/* Now it's all set up, let userspace reach it */

I also noticed that openvz and xen probably need the same patch set.

rtg
-- 
Tim Gardner tim.gardner at canonical.com