Re: KVM: x86: use kvm_set_cr3/cr4 in ioctl_set_sregs

From: Avi Kivity <avi@redhat.com>
To: Marcelo Tosatti <mtosatti@redhat.com>
Cc: kvm <kvm@vger.kernel.org>
Subject: Re: KVM: x86: use kvm_set_cr3/cr4 in ioctl_set_sregs
Date: Thu, 16 Apr 2009 12:23:56 +0300	[thread overview]
Message-ID: <49E6F92C.2000202@redhat.com> (raw)
In-Reply-To: <20090416091042.GA6709@amt.cnet>

Marcelo Tosatti wrote:
> Humpf. And something like this? Or GP# instead of triple fault?
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 148cde2..3e63bac 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -3986,7 +3986,10 @@ int kvm_arch_vcpu_ioctl_set_sregs(struct kvm_vcpu *vcpu,
>  
>  	vcpu->arch.cr2 = sregs->cr2;
>  	mmu_reset_needed |= vcpu->arch.cr3 != sregs->cr3;
> -	vcpu->arch.cr3 = sregs->cr3;
> +	if (gfn_to_memslot(vcpu->kvm, sregs->cr3 >> PAGE_SHIFT))
> +		vcpu->arch.cr3 = sregs->cr3;
> +	else
> +		set_bit(KVM_REQ_TRIPLE_FAULT, &vcpu->requests);
>  
>  	kvm_set_cr8(vcpu, sregs->cr8); 
>   

Well, that plugs the hole.  Triple fault is better than #GP IMO.

We're still missing checks on reserved bits, etc., but that can come later.

-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.