This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: RFA: m32r/x support
- To: Catherine Moore <clm at redhat dot com>
- Subject: Re: RFA: m32r/x support
- From: Nick Clifton <nickc at cambridge dot redhat dot com>
- Date: 07 Nov 2001 16:14:30 +0000
- Cc: gcc-patches at gcc dot gnu dot org
- References: <200111071537.HAA21750@cygnus.com>
Hi Catherine,
Why define a couple of new pieces of RTL...
> + /* Express a group of insns that are executed in parallel.
> + 1st operand: Vector of insns.
> + 2nd operand: Available for scheduler use. */
> + DEF_RTL_EXPR(GROUP_PARALLEL, "group_parallel", "E0", 'g')
> +
> + /* Express a group of insns that are executed in sequence.
> + 1st operand: Vector of insns.
> + 2nd operand: Available for scheduler use. */
> + DEF_RTL_EXPR(GROUP_SEQUENCE, "group_sequence", "E0", 'g')
And then define patterns that abort if they are used ?
> + (define_insn "*small_parallel"
> + [(group_parallel [(match_insn 0 "small_insn_p")
> + (match_insn 1 "small_insn_p")])]
> + "TARGET_M32RX"
> + "*
> + {
> + abort ();
> + }"
> + [(set_attr "length" "4")
> + (set_attr "type" "multi")])
> +
> + (define_insn "*long_group"
> + [(group_sequence [(match_insn 0 "large_insn_p")])]
> + ""
> + "*
> + {
> + abort ();
> + }"
> + [(set_attr "length" "4")
> + (set_attr "type" "multi")])
> +
And which, as far as I can tell, will never actually be generated in
the first place ?
Cheers
Nick