On 15/02/17 11:12, Jan Beulich wrote:
> @@ -6806,6 +6808,23 @@ x86_emulate(
> sfence = true;
> break;
>
> + case X86EMUL_OPC_66(0x0f38, 0x2a): /* movntdqa m128,xmm */
> + case X86EMUL_OPC_VEX_66(0x0f38, 0x2a): /* vmovntdqa mem,{x,y}mm */
> + generate_exception_if(ea.type != OP_MEM, EXC_UD);
> + /* Ignore the non-temporal hint for now, using movdqa instead. */
> + asm volatile ( "mfence" ::: "memory" );
Similarly for the sfence question in patch 3, why do we need to care
here? The guest will have had to execute mfence prior to movntdqa
trapping for emulation if it cared.
~Andrew
> + b = 0x6f;
> + if ( vex.opcx == vex_none )
> + vcpu_must_have(sse4_1);
> + else
> + {
> + vex.opcx = vex_0f;
> + if ( vex.l )
> + vcpu_must_have(avx2);
> + }
> + state->simd_size = simd_packed_int;
> + goto movdqa;
> +
> case X86EMUL_OPC(0x0f38, 0xf0): /* movbe m,r */
> case X86EMUL_OPC(0x0f38, 0xf1): /* movbe r,m */
> vcpu_must_have(movbe);
>
>
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
