Re: [Xen-devel] [PATCH v6.5 24/26] x86/ctxt: Issue a speculation barrier between vcpu contexts

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 04.01.18 at 01:15, <andrew.cooper3@xxxxxxxxxx> wrote:
> --- a/xen/arch/x86/spec_ctrl.c
> +++ b/xen/arch/x86/spec_ctrl.c
> @@ -32,7 +32,7 @@ enum ind_thunk {
>      THUNK_LFENCE,
>      THUNK_JMP,
>  } opt_thunk __initdata = THUNK_DEFAULT;
> -int opt_ibrs __initdata = -1;
> +int opt_ibrs __initdata = -1, opt_ibpb __initdata = -1;

static, but see below.

> --- a/xen/include/asm-x86/cpufeature.h
> +++ b/xen/include/asm-x86/cpufeature.h
> @@ -109,6 +109,7 @@
>  #define cpu_has_aperfmperf      boot_cpu_has(X86_FEATURE_APERFMPERF)
>  #define cpu_has_xen_smep        boot_cpu_has(X86_FEATURE_XEN_SMEP)
>  #define cpu_has_lfence_dispatch boot_cpu_has(X86_FEATURE_LFENCE_DISPATCH)
> +#define cpu_has_xen_ibpb        boot_cpu_has(X86_FEATURE_XEN_IBPB)

Upon second look I'm actually not convinced using a CPU feature
flag here is warranted: You don't key any patching off of it,
so simply making opt_ibpb global (and non-init) should suffice. I
do realize that this would mean a change to an earlier patch.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel