Re: [Xen-devel] [PATCH v4 02/20] x86emul: support most remaining AVX2 insns

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 28/02/18 12:58, Jan Beulich wrote:

You're missing a SoB on this version of the patch, but assuming that
gets fixed, Acked-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>

> --- a/xen/arch/x86/x86_emulate/x86_emulate.c
> +++ b/xen/arch/x86/x86_emulate/x86_emulate.c
> @@ -7306,6 +7325,11 @@ x86_emulate(
>          op_bytes = 8 << vex.l;
>          goto simd_0f_ymm;
>  
> +    case X86EMUL_OPC_VEX_66(0x0f38, 0x16): /* vpermps ymm/m256,ymm,ymm */
> +    case X86EMUL_OPC_VEX_66(0x0f38, 0x36): /* vpermd ymm/m256,ymm,ymm */

It seems especially odd that these don't cope with vex.l or vex.w.  I
can only guess that it down to a limitation in the vector pipeline which
wasn't resolved until AVX-512 was introduced.

~Andrew

> +        generate_exception_if(!vex.l || vex.w, EXC_UD);
> +        goto simd_0f_avx2;
> +
>      case X86EMUL_OPC_VEX_66(0x0f38, 0x20): /* vpmovsxbw xmm/mem,{x,y}mm */
>      case X86EMUL_OPC_VEX_66(0x0f38, 0x21): /* vpmovsxbd xmm/mem,{x,y}mm */
>      case X86EMUL_OPC_VEX_66(0x0f38, 0x22): /* vpmovsxbq xmm/mem,{x,y}mm */
>


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