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>