Re: [PATCH v10 5/9] x86emul: support MOVDIR{I,64B} insns

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 25/05/2020 15:28, Jan Beulich wrote:
> Introduce a new blk() hook, paralleling the rmw() one in a certain way,
> but being intended for larger data sizes, and hence its HVM intermediate
> handling function doesn't fall back to splitting the operation if the
> requested virtual address can't be mapped.
>
> Note that SDM revision 071 doesn't specify exception behavior for
> ModRM.mod == 0b11; assuming #UD here.

Once again - I don't think this wants calling out.  That encoding space
will be used for a new Grp at some point in the future, and be a
different instruction.

>
> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> Reviewed-by: Paul Durrant <paul@xxxxxxx>

Acked-by: Andrew Cooper <andrew.cooper@xxxxxxxxxx>, although with one
recommendation...

> --- a/xen/include/public/arch-x86/cpufeatureset.h
> +++ b/xen/include/public/arch-x86/cpufeatureset.h
> @@ -241,6 +241,8 @@ XEN_CPUFEATURE(AVX512_VPOPCNTDQ, 6*32+14
>  XEN_CPUFEATURE(TSXLDTRK,      6*32+16) /*a  TSX load tracking suspend/resume 
> insns */
>  XEN_CPUFEATURE(RDPID,         6*32+22) /*A  RDPID instruction */
>  XEN_CPUFEATURE(CLDEMOTE,      6*32+25) /*A  CLDEMOTE instruction */
> +XEN_CPUFEATURE(MOVDIRI,       6*32+27) /*A  MOVDIRI instruction */
> +XEN_CPUFEATURE(MOVDIR64B,     6*32+28) /*A  MOVDIR64B instruction */

I'd be tempted to leave these as 'a' for now, seeing as we have the ability.

These instructions aren't actually of any use for domains without PCI
devices, and a "default" will be more migrateable as a consequence.

We're going to need further toolstack changes to make CXL-passthrough
viable, so instruction adjustments can be part of that work.

~Andrew