Taken from http://sources.redhat.com/ml/crossgcc/2004-02/msg00104.html

 Author: addsub@eyou.com

 Target: ARM

 Fixes http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.c-torture/execute/920501-8.c

 and makes printf("%f", 1.0) work.

 Lennert Buytenhek wrote in http://sources.redhat.com/ml/crossgcc/2004-09/msg00115.html :

  It ... fixes the 'printf("%f\n", 0.5); prints 0.000000' and general 'floating point

  is broken' on my big-endian hardfloat FPA ARM platform. ...

  It's definitely needed for hardfloat.  So I'd think it's needed for

  big-endian systems in any case, and for VFP on little-endian systems

  too.  Someone would have to verify that though.

 Lennert Buytenhek wrote in http://sources.redhat.com/ml/crossgcc/2004-09/msg00123.html

  I just had a look at glibc-20040830, and [this patch] is still needed and useful

  for this version.  glibc-20040830 out-of-the-box still contains the

  following wrong assumptions:

  - sysdeps/arm/bits/endian.h: float word order is big endian (which it is

    not on vfp systems)

  - sysdeps/arm/gmp-mparam.h: IEEE doubles are mixed endian (which they

    are not on big endian systems, neither on vfp systems)

  - sysdeps/arm/ieee754.h: IEEE doubles are in little endian byte order

    (which they are not on big endian systems)

  [This patch] seems the right solution for all of these issues.

 Dimitry Andric wrote in http://sources.redhat.com/ml/crossgcc/2004-09/msg00132.html :

  It's even needed for glibc CVS, AFAICS.

  The patch hunk which modifies glibc.new/sysdeps/arm/bits/endian.h

  (currently at version 1.4) is only needed for proper VFP operation.

  But the hunk which modifies sysdeps/arm/gmp-mparam.h, and the hunk

  that deletes sysdeps/arm/ieee754.h (yes, this IS correct), are needed

  for proper operation of *any* FP model on big endian ARM.

 See also discussion in followups to

 http://sources.redhat.com/ml/crossgcc/2004-05/msg00245.html)

 Message-ID: <276985760.37584@eyou.com>

 Received: from unknown (HELO eyou.com) (172.16.2.2)

  by 0.0.0.0 with SMTP; Tue, 17 Feb 2004 10:42:40 +0800

 Received: (qmail 8238 invoked by uid 65534); 17 Feb 2004 10:42:38 +0800

 Date: 17 Feb 2004 10:42:38 +0800

 Message-ID: <20040217104238.8237.qmail@eyou.com>

 From: "add" <addsub@eyou.com>

 To: dank@kegel.com

 Reply-To: "add" <addsub@eyou.com>

 Subject: Re:   problem while building arm vfp softfloat gcc `

 Hi, Dan, This is a patch I applied to my glibc-2.3.2, then my softfloat

 toolchain can printf("%f\n",1.0). So you may have a try of this

 diff -uNrp glibc.old/sysdeps/arm/bits/endian.h glibc.new/sysdeps/arm/bits/endian.h

 --- glibc.old/sysdeps/arm/bits/endian.h	1999-04-12 11:59:13.000000000 -0400

 +++ glibc.new/sysdeps/arm/bits/endian.h	2004-02-12 09:15:13.000000000 -0500

 @@ -9,4 +9,9 @@

  #else

  #define __BYTE_ORDER __LITTLE_ENDIAN

  #endif

 +

 +#ifdef __VFP_FP__

 +#define __FLOAT_WORD_ORDER __BYTE_ORDER

 +#else

  #define __FLOAT_WORD_ORDER __BIG_ENDIAN

 +#endif

 diff -uNrp glibc.old/sysdeps/arm/gmp-mparam.h glibc.new/sysdeps/arm/gmp-mparam.h

 --- glibc.old/sysdeps/arm/gmp-mparam.h	2001-07-07 15:21:19.000000000 -0400

 +++ glibc.new/sysdeps/arm/gmp-mparam.h	2004-02-12 09:15:13.000000000 -0500

 @@ -26,5 +26,13 @@ MA 02111-1307, USA. */

  #define BITS_PER_SHORTINT 16

  #define BITS_PER_CHAR 8

 -#define IEEE_DOUBLE_BIG_ENDIAN 0

 -#define IEEE_DOUBLE_MIXED_ENDIAN 1

 +#if defined(__ARMEB__)

 +# define IEEE_DOUBLE_MIXED_ENDIAN 0

 +# define IEEE_DOUBLE_BIG_ENDIAN 1

 +#elif defined(__VFP_FP__)

 +# define IEEE_DOUBLE_MIXED_ENDIAN 0

 +# define IEEE_DOUBLE_BIG_ENDIAN 0

 +#else

 +# define IEEE_DOUBLE_BIG_ENDIAN 0

 +# define IEEE_DOUBLE_MIXED_ENDIAN 1

 +#endif

 diff -uNrp glibc.old/sysdeps/arm/ieee754.h glibc.new/sysdeps/arm/ieee754.h

 --- glibc.old/sysdeps/arm/ieee754.h	2001-07-07 15:21:19.000000000 -0400

 +++ glibc.new/sysdeps/arm/ieee754.h	1969-12-31 19:00:00.000000000 -0500

 @@ -1,115 +0,0 @@

 -/* Copyright (C) 1992, 1995, 1996, 1998 Free Software Foundation, Inc.

 -   This file is part of the GNU C Library.

 -

 -   The GNU C Library is free software; you can redistribute it and/or

 -   modify it under the terms of the GNU Lesser General Public

 -   License as published by the Free Software Foundation; either

 -   version 2.1 of the License, or (at your option) any later version.

 -

 -   The GNU C Library is distributed in the hope that it will be useful,

 -   but WITHOUT ANY WARRANTY; without even the implied warranty of

 -   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

 -   Lesser General Public License for more details.

 -

 -   You should have received a copy of the GNU Lesser General Public

 -   License along with the GNU C Library; if not, write to the Free

 -   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA

 -   02111-1307 USA.  */

 -

 -#ifndef _IEEE754_H

 -

 -#define _IEEE754_H 1

 -#include <features.h>

 -

 -#include <endian.h>

 -

 -__BEGIN_DECLS

 -

 -union ieee754_float

 -  {

 -    float f;

 -

 -    /* This is the IEEE 754 single-precision format.  */

 -    struct

 -      {

 -	unsigned int mantissa:23;

 -	unsigned int exponent:8;

 -	unsigned int negative:1;

 -      } ieee;

 -

 -    /* This format makes it easier to see if a NaN is a signalling NaN.  */

 -    struct

 -      {

 -	unsigned int mantissa:22;

 -	unsigned int quiet_nan:1;

 -	unsigned int exponent:8;

 -	unsigned int negative:1;

 -      } ieee_nan;

 -  };

 -

 -#define IEEE754_FLOAT_BIAS	0x7f /* Added to exponent.  */

 -

 -

 -union ieee754_double

 -  {

 -    double d;

 -

 -    /* This is the IEEE 754 double-precision format.  */

 -    struct

 -      {

 -	unsigned int mantissa0:20;

 -	unsigned int exponent:11;

 -	unsigned int negative:1;

 -	unsigned int mantissa1:32;

 -      } ieee;

 -

 -    /* This format makes it easier to see if a NaN is a signalling NaN.  */

 -    struct

 -      {

 -	unsigned int mantissa0:19;

 -	unsigned int quiet_nan:1;

 -	unsigned int exponent:11;

 -	unsigned int negative:1;

 -	unsigned int mantissa1:32;

 -      } ieee_nan;

 -  };

 -

 -#define IEEE754_DOUBLE_BIAS	0x3ff /* Added to exponent.  */

 -

 -

 -/* The following two structures are correct for `new' floating point systems but

 -   wrong for the old FPPC.  The only solution seems to be to avoid their use on

 -   old hardware.  */

 -

 -union ieee854_long_double

 -  {

 -    long double d;

 -

 -    /* This is the IEEE 854 double-extended-precision format.  */

 -    struct

 -      {

 -	unsigned int exponent:15;

 -	unsigned int empty:16;

 -	unsigned int negative:1;

 -	unsigned int mantissa1:32;

 -	unsigned int mantissa0:32;

 -      } ieee;

 -

 -    /* This is for NaNs in the IEEE 854 double-extended-precision format.  */

 -    struct

 -      {

 -	unsigned int exponent:15;

 -	unsigned int empty:16;

 -	unsigned int negative:1;

 -	unsigned int mantissa1:32;

 -	unsigned int mantissa0:30;

 -	unsigned int quiet_nan:1;

 -	unsigned int one:1;

 -      } ieee_nan;

 -  };

 -

 -#define IEEE854_LONG_DOUBLE_BIAS 0x3fff

 -

 -__END_DECLS

 -

 -#endif /* ieee754.h */

 Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>