public inbox for cgen@sourceware.org
 help / color / mirror / Atom feed
* insn semantics required?
@ 2011-01-24 14:24 Nathan Froyd
  2011-01-24 14:53 ` Frank Ch. Eigler
  2011-01-24 14:59 ` Jean-Marc Saffroy
  0 siblings, 2 replies; 3+ messages in thread
From: Nathan Froyd @ 2011-01-24 14:24 UTC (permalink / raw)
  To: cgen

I've been looking through the documentation and couldn't determine
conclusively whether defining the semantics of an insn is required.  Is
it?  Auto-magically defining a simulator seems less useful in this day
and age, but an auto-generated assembler and disassembler seems quite
useful.

I suppose as a stopgap one could simply provide an uninteresting
semantics for each insn.

-Nathan

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: insn semantics required?
  2011-01-24 14:24 insn semantics required? Nathan Froyd
@ 2011-01-24 14:53 ` Frank Ch. Eigler
  2011-01-24 14:59 ` Jean-Marc Saffroy
  1 sibling, 0 replies; 3+ messages in thread
From: Frank Ch. Eigler @ 2011-01-24 14:53 UTC (permalink / raw)
  To: Nathan Froyd; +Cc: cgen

Hi -

On Mon, Jan 24, 2011 at 06:23:54AM -0800, Nathan Froyd wrote:
> I've been looking through the documentation and couldn't determine
> conclusively whether defining the semantics of an insn is required.
> [...]
> I suppose as a stopgap one could simply provide an uninteresting
> semantics for each insn.

Yes, you could pass (nop) or perhaps an empty list in the semantics
slot of the (dni ...).

- FChE

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: insn semantics required?
  2011-01-24 14:24 insn semantics required? Nathan Froyd
  2011-01-24 14:53 ` Frank Ch. Eigler
@ 2011-01-24 14:59 ` Jean-Marc Saffroy
  1 sibling, 0 replies; 3+ messages in thread
From: Jean-Marc Saffroy @ 2011-01-24 14:59 UTC (permalink / raw)
  To: Nathan Froyd; +Cc: cgen

Hi Nathan,

On 01/24/2011 03:23 PM, Nathan Froyd wrote:
> I've been looking through the documentation and couldn't determine
> conclusively whether defining the semantics of an insn is required.  Is
> it?

The syntax requires that you define something, although I think "()"
will do if you don't built a simulator. Otherwise, as you noted, a
simple "(nop)" will definitely be sufficient.

>  Auto-magically defining a simulator seems less useful in this day
> and age, but an auto-generated assembler and disassembler seems quite
> useful.

It really depends if you have a platform available for testing. A
simulator (auto-generated or not) is very convenient to run the gcc test
suite for regression testing.


Cheers,
JM

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2011-01-24 14:59 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-01-24 14:24 insn semantics required? Nathan Froyd
2011-01-24 14:53 ` Frank Ch. Eigler
2011-01-24 14:59 ` Jean-Marc Saffroy

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).