Re: [PATCH] x86: always supply .cpuid() handler to x86_emulate()

["Jan Beulich" <JBeulich@suse.com>]

>>> On 11.11.16 at 15:16, <andrew.cooper3@citrix.com> wrote:
> On 10/11/16 12:30, Jan Beulich wrote:
>> --- a/xen/arch/x86/traps.c
>> +++ b/xen/arch/x86/traps.c
>> @@ -2755,6 +2755,24 @@ static int priv_op_write_msr(unsigned in
>>      return X86EMUL_UNHANDLEABLE;
>>  }
>>  
>> +int pv_emul_cpuid(unsigned int *eax, unsigned int *ebx, unsigned int *ecx,
>> +                  unsigned int *edx, struct x86_emulate_ctxt *ctxt)
>> +{
>> +    struct cpu_user_regs regs = *ctxt->regs;
> 
> You need a CPUID faulting check here, matching the hvm side of things,
> or you will leave a latent bug which gets exposed when switching to
> using full x86_emulate() for PV guests.

That addition is part of the patch actually switching to x86_emulate().
I don't think it belongs here, as for now we don't mean to emulate
CPUID insns.

> Would it be wise to add a fail_if(!ops->cpuid) to x86_emulate() to catch
> other misuses?

I don't think so - there may be cases where it's indeed unneeded
(and I foresee such arising quickly once the ->validate() hooks is
in place, addition of which is now too part of aforementioned patch).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel