On 08/11/2017 08:56, Haozhong Zhang wrote:
> Some reserved pages, such as those from NVDIMM DAX devices, are not
> for MMIO, and can be mapped with cached memory type for better
> performance. However, the above check misconceives those pages as
> MMIO.  Because KVM maps MMIO pages with UC memory type, the
> performance of guest accesses to those pages would be harmed.
> Therefore, we check the host memory type in addition and only treat
> UC/UC- pages as MMIO.
> 
> Signed-off-by: Haozhong Zhang <[email protected]>
> Reported-by: Cuevas Escareno, Ivan D <[email protected]>
> Reported-by: Kumar, Karthik <[email protected]>

WC should be allowed as well, because the combination of EPT_PAT=UC and
gPAT=WC gives WC effective memory type.

Maybe it's better after all if the lookup_memtype call remains in
kvm_is_mmio_pfn, like

        if (!pfn_valid(pfn))
                return true;

        if (is_zero_pfn(pfn) || !PageReserved(pfn_to_page(pfn))
                return false;

        /* ... long comment ... */
        if (!pat_enabled()
                return true;
        else {
                cm = lookup_memtype(PFN_PHYS(pfn));
                return cm == ...
        }

or something like that.

Thanks,

Paolo

> ---
>  arch/x86/kvm/mmu.c | 15 ++++++++++++++-
>  1 file changed, 14 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
> index 0b481cc9c725..7715476bc5c9 100644
> --- a/arch/x86/kvm/mmu.c
> +++ b/arch/x86/kvm/mmu.c
> @@ -2708,7 +2708,20 @@ static bool mmu_need_write_protect(struct kvm_vcpu 
> *vcpu, gfn_t gfn,
>  static bool kvm_is_mmio_pfn(kvm_pfn_t pfn)
>  {
>       if (pfn_valid(pfn))
> -             return !is_zero_pfn(pfn) && PageReserved(pfn_to_page(pfn));
> +             return !is_zero_pfn(pfn) && PageReserved(pfn_to_page(pfn)) &&
> +                     /*
> +                      * Some reserved pages, such as those from
> +                      * NVDIMM DAX devices, are not for MMIO, and
> +                      * can be mapped with cached memory type for
> +                      * better performance. However, the above
> +                      * check misconceives those pages as MMIO.
> +                      * Because KVM maps MMIO pages with UC memory
> +                      * type, the performance of guest accesses to
> +                      * those pages would be harmed. Therefore, we
> +                      * check the host memory type in addition and
> +                      * only treat UC/UC- pages as MMIO.
> +                      */
> +                     (!pat_enabled() || pat_pfn_is_uc_or_uc_minus(pfn));
>  
>       return true;
>  }
> 

Reply via email to