Re: [Xen-devel] [PATCH v8 43/50] x86emul: support AVX512_VNNI insns

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 19.06.19 at 17:01, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 15/03/2019 11:04, Jan Beulich wrote:
>> --- a/xen/arch/x86/x86_emulate/x86_emulate.c
>> +++ b/xen/arch/x86/x86_emulate/x86_emulate.c
>> @@ -479,7 +479,7 @@ static const struct ext0f38_table {
>>      [0x4d] = { .simd_size = simd_scalar_vexw, .d8s = d8s_dq },
>>      [0x4e] = { .simd_size = simd_packed_fp, .two_op = 1, .d8s = d8s_vl },
>>      [0x4f] = { .simd_size = simd_scalar_vexw, .d8s = d8s_dq },
>> -    [0x52 ... 0x53] = { .simd_size = simd_128, .d8s = 4 },
>> +    [0x50 ... 0x53] = { .simd_size = simd_packed_int, .d8s = d8s_vl },
> 
> Hang on - is the previous patch correct?  Shouldn't it have
> simd_packed_int/d8s_vl from the getgo?

Yes, it is. The situation here is the opposite to what we have with
4FMAPS: The special F2-prefixed encoding gets introduced first
(previous patch), and gets converted to a decode special case
here (for consistency such that the "normal" 66-prefixed operand
characteristics appear in the table.

>> @@ -3211,6 +3212,8 @@ x86_decode(
>>  
>>                  switch ( b )
>>                  {
>> +                /* vp4dpwssd{,s} need special casing */
> 
> Special cases should be discussed.

I'll clone what I've added to the 4FMAPS patch.

Jan



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