On Mon, Apr 15, 2024 at 3:08 PM Chao Gao <chao.gao@xxxxxxxxx> wrote:
>
> >+++ b/arch/x86/include/asm/vmx.h
> >@@ -514,6 +514,7 @@ enum vmcs_field {
> > #define VMX_EPT_IPAT_BIT (1ull << 6)
> > #define VMX_EPT_ACCESS_BIT (1ull << 8)
> > #define VMX_EPT_DIRTY_BIT (1ull << 9)
> >+#define VMX_EPT_SUPPRESS_VE_BIT (1ull << 63)
> > #define VMX_EPT_RWX_MASK (VMX_EPT_READABLE_MASK | \
> > VMX_EPT_WRITABLE_MASK | \
> > VMX_EPT_EXECUTABLE_MASK)
> >diff --git a/arch/x86/kvm/mmu/spte.c b/arch/x86/kvm/mmu/spte.c
> >index 6c7ab3aa6aa7..d97c4725c0b7 100644
> >--- a/arch/x86/kvm/mmu/spte.c
> >+++ b/arch/x86/kvm/mmu/spte.c
> >@@ -413,7 +413,9 @@ void kvm_mmu_set_ept_masks(bool has_ad_bits, bool has_exec_only)
> > shadow_dirty_mask = has_ad_bits ? VMX_EPT_DIRTY_BIT : 0ull;
> > shadow_nx_mask = 0ull;
> > shadow_x_mask = VMX_EPT_EXECUTABLE_MASK;
> >- shadow_present_mask = has_exec_only ? 0ull : VMX_EPT_READABLE_MASK;
> >+ /* VMX_EPT_SUPPRESS_VE_BIT is needed for W or X violation. */
> >+ shadow_present_mask =
> >+ (has_exec_only ? 0ull : VMX_EPT_READABLE_MASK) | VMX_EPT_SUPPRESS_VE_BIT;
>
> This change makes !shadow_present_mask checks in FNAME(sync_spte) and
> make_spte() pointless as shadow_present_mask will never be zero.
It makes them wrong, not pointless. :)
The checks verify that there are "some" bits that are different
between non-present and present PTEs. They need to remove
SHADOW_NONPRESENT_MASK from shadow_present_mask.
Paolo
