Re: [PATCH 1/4] x86: replace __ASM_{CL,ST}AC

[Jan Beulich <jbeulich@xxxxxxxx>]

On 28.07.2020 15:59, Andrew Cooper wrote:
> On 27/07/2020 20:47, Jan Beulich wrote:
> > On 27.07.2020 16:55, Roger Pau Monné wrote:
> > > On Wed, Jul 15, 2020 at 12:48:14PM +0200, Jan Beulich wrote:
> > > > --- /dev/null
> > > > +++ b/xen/include/asm-x86/asm-defns.h
> > > Maybe this could be asm-insn.h or a different name? I find it
> > > confusing to have asm-defns.h and an asm_defs.h.
> > While indeed I anticipated a reply to this effect, I don't consider
> > asm-insn.h or asm-macros.h suitable: We don't want to limit this
> > header to a more narrow purpose than "all sorts of definition", I
> > don't think. Hence I chose that name despite its similarity to the
> > C header's one.
> Roger is correct.  Having asm-defns.h and asm_defs.h is too confusing,
> and there is already too much behind the scenes magic here.
>
> What is the anticipated end result, file wise, because that might
> highlight a better way forward.
For one I'm afraid I don't understand "file wise" here. The one meaning
I could guess can't be it: The name of the file.

And then, "the anticipated end result" is at least ambiguous too: You
can surely see what the file contains by the end of this series, so
again this can't be meant. I have no immediate plans beyond this
series, so I can only state what I did say in reply to Roger's remark
already: "all sorts of asm definitions".

I'd also like to emphasize that asm-defns.h really is a companion of
asm_defns.h, supposed to be include only by the latter (as I think
can be seen from the patches). In this role I think its name being
as similar to its "parent" as suggested makes sufficient sense.

Jan