Re: [Xen-devel] [PATCH 09/10] tools/x86emul: Advertise more CPUID features for testing purposes

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 27.03.17 at 15:01, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 27/03/17 13:09, Jan Beulich wrote:
>>>>> On 27.03.17 at 11:56, <andrew.cooper3@xxxxxxxxxx> wrote:
>>> --- a/tools/tests/x86_emulator/x86_emulate.c
>>> +++ b/tools/tests/x86_emulator/x86_emulate.c
>>> @@ -73,20 +73,37 @@ int emul_test_cpuid(
>>>           : "a" (leaf), "c" (subleaf));
>>>  
>>>      /*
>>> -     * The emulator doesn't itself use MOVBE, so we can always run the
>>> -     * respective tests.
>>> +     * Some instructions and features can be emulated without specific
>>> +     * hardware support.  These features are unconditionally reported here,
>>> +     * for testing and fuzzing-coverage purposes.
>>>       */
>>> -    if ( leaf == 1 )
>>> -        res->c |= 1U << 22;
>>> -
>>> -    /*
>>> -     * The emulator doesn't itself use ADCX/ADOX/RDPID, so we can always 
>>> run
>>> -     * the respective tests.
>>> -     */
>>> -    if ( leaf == 7 && subleaf == 0 )
>>> +    switch ( leaf )
>>>      {
>>> -        res->b |= 1U << 19;
>>> -        res->c |= 1U << 22;
>>> +    case 1:
>>> +        res->c |= 1U << 22; /* MOVBE */
>>> +        break;
>>> +
>>> +    case 7:
>>> +        switch ( subleaf )
>>> +        {
>>> +        case 0:
>>> +            res->b |= 1U << 11; /* rtm */
>> Upper case?
> 
> I was trying to visually separate the instructions from the features.

Oh, I see. Fine with me then, i.e. I'd adjust ...

>>> +            res->b |= 1U << 19; /* ADCX/ADOX */
>>> +            res->b |= 1U << 20; /* STAC/CLAC */
>> SMAP?

... this to "smap?"

>>> +            res->b |= 1U << 24; /* CLWB */
>>> +
>>> +            res->c |= 1U << 22; /* RDPID */
>>> +            break;
>>> +        }
>>> +        break;
>>> +
>>> +    case 0x80000001:
>>> +        res->c |= 1U << 4; /* cr8_legacy */
>> I think this one is AMD-only, just like ...
>>
>>> +        if ( ctxt->vendor == X86_VENDOR_AMD )
>>> +            res->c |= 1U << 7; /* misalignsse */
>> ... this.
> 
> The difference is that cr8_legacy is a straight "is this specific
> encoding valid to use", while misalignsse causes rather larger changes
> in how SSE arguments get handled.

True.

>> And what about LAHF_LM, LZCNT, and CLFLUSH{,OPT}?
> 
> This list was only the instructions I had encountered.  I can add all of
> these.

Yes please (subject to the other sub-thread).

Jan


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