This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: ifcvt vs. expand_binop

From: Bernd Schmidt <bschmidt at redhat dot com>
To: Oleg Endo <oleg dot endo at t-online dot de>, GCC Patches <gcc-patches at gcc dot gnu dot org>
Cc: Jeff Law <law at redhat dot com>, Kyrill Tkachov <kyrylo dot tkachov at arm dot com>, Richard Henderson <rth at redhat dot com>
Date: Tue, 22 Sep 2015 17:53:18 +0200
Subject: Re: ifcvt vs. expand_binop
Authentication-results: sourceware.org; auth=none
References: <1442928926 dot 2509 dot 23 dot camel at t-online dot de>

On 09/22/2015 03:35 PM, Oleg Endo wrote:

On SH, the result of comparisons etc. is stored in the T_REG.  It's a 1
bit reg but described as SImode.  To get the T_REG into another reg,
there's this insn:

(define_insn "movt"
   [(set (match_operand:SI 0 "arith_reg_dest" "=r")
	(match_operand:SI 1 "t_reg_operand"))]
   "TARGET_SH1"
   "movt	%0"
   [(set_attr "type" "arith")])

where "t_reg_operand" accepts various forms of the T_REG via
reg,subreg,sign_extend,zero_extend (to get better combine results).

Now I'd like to extend the "t_reg_operand" predicate to accept something
like

(set (reg:SI) (ne:SI (reg:SI 147 t) (const_int 0)))

Is there a reason you're not just adding that as a new pattern ratherthan trying to overload the predicate?



Bernd

Follow-Ups:
- Re: ifcvt vs. expand_binop
  - From: Oleg Endo

References:
- ifcvt vs. expand_binop
  - From: Oleg Endo

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]