Re: [Xen-devel] [PATCH v2 19/30] x86/cpu: Rework Intel masking/faulting setup

2016-02-17 Thread Andrew Cooper
On 17/02/16 07:57, Jan Beulich wrote: > >> +/* Indicies of the masking MSRs, or 0 if unavailable. */ >> +static unsigned int __read_mostly msr_basic, msr_ext, msr_xsave; > I think this way __read_mostly applies only to msr_basic, which I > don't think is what you want. Also I think you mean "indice

Re: [Xen-devel] [PATCH v2 19/30] x86/cpu: Rework Intel masking/faulting setup

2016-02-16 Thread Jan Beulich
>>> On 05.02.16 at 14:42, wrote: > --- a/xen/arch/x86/cpu/intel.c > +++ b/xen/arch/x86/cpu/intel.c > @@ -18,11 +18,18 @@ > > #define select_idle_routine(x) ((void)0) > > -static unsigned int probe_intel_cpuid_faulting(void) > +static bool_t __init probe_intel_cpuid_faulting(void) > { >

[Xen-devel] [PATCH v2 19/30] x86/cpu: Rework Intel masking/faulting setup

2016-02-05 Thread Andrew Cooper
This patch is best reviewed as its end result rather than as a diff, as it rewrites almost all of the setup. On the BSP, cpuid information is used to evaluate the potential available set of masking MSRs, and they are unconditionally probed, filling in the availability information and hardware defa