On 13/02/20 22:30, Chia-I Wu wrote:
> Better reflect the structure of the code and metion why we could not
> always honor the guest.
>
> Signed-off-by: Chia-I Wu <olvaffe@xxxxxxxxx>
> Cc: Gurchetan Singh <gurchetansingh@xxxxxxxxxxxx>
> Cc: Gerd Hoffmann <kraxel@xxxxxxxxxx>
> ---
> arch/x86/kvm/vmx/vmx.c | 27 +++++++++++++++++----------
> 1 file changed, 17 insertions(+), 10 deletions(-)
>
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index 3be25ecae145..266ef87042da 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -6854,17 +6854,24 @@ static u64 vmx_get_mt_mask(struct kvm_vcpu *vcpu, gfn_t gfn, bool is_mmio)
> u8 cache;
> u64 ipat = 0;
>
> - /* For VT-d and EPT combination
> - * 1. MMIO: always map as UC
> - * 2. EPT with VT-d:
> - * a. VT-d without snooping control feature: can't guarantee the
> - * result, try to trust guest.
> - * b. VT-d with snooping control feature: snooping control feature of
> - * VT-d engine can guarantee the cache correctness. Just set it
> - * to WB to keep consistent with host. So the same as item 3.
> - * 3. EPT without VT-d: always map as WB and set IPAT=1 to keep
> - * consistent with host MTRR
> + /* We wanted to honor guest CD/MTRR/PAT, but doing so could result in
> + * memory aliases with conflicting memory types and sometimes MCEs.
> + * We have to be careful as to what are honored and when.
> + *
> + * For MMIO, guest CD/MTRR are ignored. The EPT memory type is set to
> + * UC. The effective memory type is UC or WC depending on guest PAT.
> + * This was historically the source of MCEs and we want to be
> + * conservative.
> + *
> + * When there is no need to deal with noncoherent DMA (e.g., no VT-d
> + * or VT-d has snoop control), guest CD/MTRR/PAT are all ignored. The
> + * EPT memory type is set to WB. The effective memory type is forced
> + * WB.
> + *
> + * Otherwise, we trust guest. Guest CD/MTRR/PAT are all honored. The
> + * EPT memory type is used to emulate guest CD/MTRR.
> */
> +
> if (is_mmio) {
> cache = MTRR_TYPE_UNCACHABLE;
> goto exit;
>
This is certainly an improvement, especially the part that points out
how guest PAT still allows MMIO to be handled as WC.
Thanks,
Paolo
