[PATCH 1/6] x86: allow opcode templates to be templated

[H.J. Lu]

On Mon, Mar 9, 2020 at 2:19 AM Jan Beulich <jbeulich@suse.com> wrote:
>
> On 06.03.2020 15:42, H.J. Lu wrote:
> > On Fri, Mar 6, 2020 at 12:11 AM Jan Beulich <jbeulich@suse.com> wrote:
> >>
> >> In order to reduce redundancy as well as the chance of things going out
> >> of sync (see a later patch for an example), make the opcode table
> >> generator capable of recognizing and expanding templated templates. Use
> >> the new capability for compacting the general purpose conditional insns.
> >>
> >> opcodes/
> >> 2020-03-XX  Jan Beulich  <jbeulich@suse.com>
> >>
> >>         * i386-gen.c (struct template_arg, struct template_instance,
> >>         struct template_param, struct template, templates,
> >>         parse_template, expand_templates): New.
> >>         (process_i386_opcodes): Various local variables moved to
> >>         expand_templates. Call parse_template and expand_templates.
> >>         * i386-opc.tbl (cc): New. Use it for Jcc, SETcc, and CMOVcc.
> >>         * i386-tbl.h: Re-generate.
> >
> > OK.  Thanks.
>
> Now that this is in, a question I have is how far we want to go with
> using this templating. There are certainly more groups of insns which
> are sufficiently similar, but I guess using templates for just two or
> three insns wouldn't be helpful readability wise. An example of where
> I'm on the edge are the {,V}PCLMUL pseudos. Do you have any thoughts
> on the general question, or is it entirely up to me to decide where
> else to use this new functionality?
>

I will leave it entirely up to you.  Thanks for your great work.

-- 
H.J.