Re: [PATCH v4 03/25] x86/hyperv: Update 'struct hv_enlightened_vmcs' definition

From: Maxim Levitsky
Date: Thu Jul 14 2022 - 05:57:48 EST


On Thu, 2022-07-14 at 11:13 +0200, Vitaly Kuznetsov wrote:
> Updated Hyper-V Enlightened VMCS specification lists several new
> fields for the following features:
>
> - PerfGlobalCtrl
> - EnclsExitingBitmap
> - Tsc Scaling
> - GuestLbrCtl
> - CET
> - SSP
>
> Update the definition. The updated definition is available only when
> CPUID.0x4000000A.EBX BIT(0) is '1'. Add a define for it as well.
>
> Note: The latest TLFS is available at
> https://docs.microsoft.com/en-us/virtualization/hyper-v-on-windows/tlfs/tlfs
>
> Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> ---
>  arch/x86/include/asm/hyperv-tlfs.h | 26 ++++++++++++++++++++++++--
>  1 file changed, 24 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/include/asm/hyperv-tlfs.h b/arch/x86/include/asm/hyperv-tlfs.h
> index 6f0acc45e67a..ebc27017fa48 100644
> --- a/arch/x86/include/asm/hyperv-tlfs.h
> +++ b/arch/x86/include/asm/hyperv-tlfs.h
> @@ -138,6 +138,17 @@
>  #define HV_X64_NESTED_GUEST_MAPPING_FLUSH              BIT(18)
>  #define HV_X64_NESTED_MSR_BITMAP                       BIT(19)
>  
> +/*
> + * Nested quirks. These are HYPERV_CPUID_NESTED_FEATURES.EBX bits.
> + *
> + * Note: HV_X64_NESTED_EVMCS1_2022_UPDATE is not currently documented in any
> + * published TLFS version. When the bit is set, nested hypervisor can use
> + * 'updated' eVMCSv1 specification (perf_global_ctrl, s_cet, ssp, lbr_ctl,
> + * encls_exiting_bitmap, tsc_multiplier fields which were missing in 2016
> + * specification).
> + */
> +#define HV_X64_NESTED_EVMCS1_2022_UPDATE               BIT(0)
> +
>  /*
>   * This is specific to AMD and specifies that enlightened TLB flush is
>   * supported. If guest opts in to this feature, ASID invalidations only
> @@ -559,9 +570,20 @@ struct hv_enlightened_vmcs {
>         u64 partition_assist_page;
>         u64 padding64_4[4];
>         u64 guest_bndcfgs;
> -       u64 padding64_5[7];
> +       u64 guest_ia32_perf_global_ctrl;
> +       u64 guest_ia32_s_cet;
> +       u64 guest_ssp;
> +       u64 guest_ia32_int_ssp_table_addr;
> +       u64 guest_ia32_lbr_ctl;
> +       u64 padding64_5[2];
>         u64 xss_exit_bitmap;
> -       u64 padding64_6[7];
> +       u64 encls_exiting_bitmap;
> +       u64 host_ia32_perf_global_ctrl;
> +       u64 tsc_multiplier;
> +       u64 host_ia32_s_cet;
> +       u64 host_ssp;
> +       u64 host_ia32_int_ssp_table_addr;
> +       u64 padding64_6;
>  } __packed;
>  
>  #define HV_VMX_ENLIGHTENED_CLEAN_FIELD_NONE                    0

All look good now.

I really don't like the new 'online' TLFS spec - as you said,
they can indeed change it any moment without any traces.

Seems it was done with good intentions, and it much easier to use,
but they should also provide a PDF, or at least some form or archive of
these web pages.

Reviewed-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>

Best regards,
Maxim Levitsky