From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-106310-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 6636 invoked by alias); 6 Dec 2004 21:59:56 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 6594 invoked from network); 6 Dec 2004 21:59:52 -0000
Received: from unknown (HELO mail.cs.umass.edu) (128.119.243.168)
  by sourceware.org with SMTP; 6 Dec 2004 21:59:52 -0000
Received: from localhost (IDENT:F7ejTznEeu96xLb1lRL7JLQ+EW4bVLN7@loki.cs.umass.edu [128.119.243.168])
	by mail.cs.umass.edu (8.12.11/8.12.5) with ESMTP id iB6Lxl2U032742;
	Mon, 6 Dec 2004 16:59:48 -0500
Date: Mon, 06 Dec 2004 21:59:00 -0000
Message-Id: <20041206.165932.23013187.kazu@cs.umass.edu>
To: e9925248@stud4.tuwien.ac.at
Cc: gcc@gcc.gnu.org
Subject: Re: Unreachable code in reload.c
From: Kazu Hirata <kazu@cs.umass.edu>
In-Reply-To: <20041206213801.GA29867@zuhause>
References: <20041206213801.GA29867@zuhause>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-SW-Source: 2004-12/txt/msg00272.txt.bz2

Hi Matrin,

> reload.c, function decompose, line 2355 (CVS HEAD version):
> 
>      val.start = true_regnum (x);
>       if (val.start < 0)
>         {
>           /* A pseudo with no hard reg.  */
>           val.start = REGNO (x);
>           val.end = val.start + 1;
>         }
>       else
>         /* A hard reg.  */
>         val.end = val.start + hard_regno_nregs[val.start][GET_MODE (x)];
> 
> 
> The then-part of the if statement can never be reached

Neither is a part of the SUBREG case pointed to with an arrow.

    case SUBREG:
      if (!REG_P (SUBREG_REG (x)))
	/* This could be more precise, but it's good enough.  */
	return decompose (SUBREG_REG (x));
      val.reg_flag = 1;
      val.start = true_regnum (x);
      if (val.start < 0)
	return decompose (SUBREG_REG (x));    <---
      else
	/* A hard reg.  */
	val.end = val.start + hard_regno_nregs[val.start][GET_MODE (x)];
      break;

because true_regnum returns a non-negative number for a SUBREG with
REG in it.  Perhaps you can convert these "if"s to gcc_assert like so?

  gcc_assert (val.start >= 0);

Kazu Hirata