On Mon, Sep 06, 2021, Zhenzhong Duan wrote:
> Host value of TSX_CTRL_CPUID_CLEAR field should be unchangable by guest,
> but the mask for this purpose is set to a wrong value. So it doesn't
> take effect.
It would be helpful to provide a bit more info as to just how bad/boneheaded this
bug is. E.g.
When updating the host's mask for its MSR_IA32_TSX_CTRL user return entry,
clear the mask in the found uret MSR instead of vmx->guest_uret_msrs[i].
Modifying guest_uret_msrs directly is completely broken as 'i' does not
point at the MSR_IA32_TSX_CTRL entry. In fact, it's guaranteed to be an
out-of-bounds accesses as is always set to kvm_nr_uret_msrs in a prior
loop. By sheer dumb luck, the fallout is limited to "only" failing to
preserve the host's TSX_CTRL_CPUID_CLEAR. The out-of-bounds access is
benign as it's guaranteed to clear a bit in a guest MSR value, which are
always zero at vCPU creation on both x86-64 and i386.
> Fixes: 8ea8b8d6f869 ("KVM: VMX: Use common x86's uret MSR list as the one true list")
Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Zhenzhong Duan <zhenzhong.duan@xxxxxxxxx>
> ---
> arch/x86/kvm/vmx/vmx.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index 927a552393b9..36588b5feee6 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -6812,7 +6812,7 @@ static int vmx_create_vcpu(struct kvm_vcpu *vcpu)
> */
> tsx_ctrl = vmx_find_uret_msr(vmx, MSR_IA32_TSX_CTRL);
> if (tsx_ctrl)
> - vmx->guest_uret_msrs[i].mask = ~(u64)TSX_CTRL_CPUID_CLEAR;
> + tsx_ctrl->mask = ~(u64)TSX_CTRL_CPUID_CLEAR;
Egad, that's a horrific oversight on my part.
Reviewed-by: Sean Christopherson <seanjc@xxxxxxxxxx>
