On Thu, Apr 26, 2018 at 12:10:14PM -0700, Jim Mattson wrote:
> As the comment says above the vmcs12 structure definition, "if there
> are changes in this struct, VMCS12_REVISION must be changed."
>
> Actually, VMCS12_REVISION should never be changed, because such a
> change introduces an incompatibility for live migration across kvm
> versions.
Would it make sense to have this patch also remove that comment then?
>
> The various padding fields were reserved to help accommodate future
> growth without changing the vmcs12 layout. They aren't really
> necessary, because one can always add new fields at the end, but since
> they are there, let's use them as they were intended.
>
> Fixes: 27c42a1bb8674 ("KVM: nVMX: Enable VMFUNC for the L1 hypervisor")
> Signed-off-by: Jim Mattson <jmattson@xxxxxxxxxx>
> ---
> arch/x86/kvm/vmx.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index aa66ccd6ed6c..dd985130c540 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -265,7 +265,6 @@ struct __packed vmcs12 {
> u64 virtual_apic_page_addr;
> u64 apic_access_addr;
> u64 posted_intr_desc_addr;
> - u64 vm_function_control;
> u64 ept_pointer;
> u64 eoi_exit_bitmap0;
> u64 eoi_exit_bitmap1;
> @@ -288,7 +287,8 @@ struct __packed vmcs12 {
> u64 host_ia32_pat;
> u64 host_ia32_efer;
> u64 host_ia32_perf_global_ctrl;
> - u64 padding64[8]; /* room for future expansion */
> + u64 vm_function_control;
> + u64 padding64[7]; /* room for future expansion */
> /*
> * To allow migration of L1 (complete with its L2 guests) between
> * machines of different natural widths (32 or 64 bit), we cannot have
> --
> 2.17.0.484.g0c8726318c-goog
>
