yann@1: This patch (C) 2007 Yann E. MORIN
yann@1: Licensed under GPL v2.
yann@1: 
yann@1: First hunk of this patch solves compiling uClibc-0.9.28{,.1} :
yann@1: 
yann@1: make[1]: Entering directory `/home/ymorin/x-tools/build/armeb-unknown-linux-uclibc/gcc-4.1.1-uClibc-0.9.28.1/build-libc'
yann@1:   LD libuClibc-0.9.29.so
yann@1: libc/libc_so.a(difftime.os): In function `difftime':
yann@1: difftime.c:(.text+0x8): undefined reference to `__floatsidf'
yann@1: difftime.c:(.text+0x2c): undefined reference to `__subdf3'
yann@1: libc/libc_so.a(_fpmaxtostr.os): In function `_fpmaxtostr':
yann@1: _fpmaxtostr.c:(.text+0xd4): undefined reference to `__nedf2'
yann@1: _fpmaxtostr.c:(.text+0xf8): undefined reference to `__eqdf2'
yann@1: _fpmaxtostr.c:(.text+0x114): undefined reference to `__divdf3'
yann@1: _fpmaxtostr.c:(.text+0x120): undefined reference to `__ltdf2'
yann@1: _fpmaxtostr.c:(.text+0x1c4): undefined reference to `__muldf3'
yann@1: _fpmaxtostr.c:(.text+0x388): undefined reference to `__gedf2'
yann@1: _fpmaxtostr.c:(.text+0x430): undefined reference to `__adddf3'
yann@1: libc/libc_so.a(__psfs_do_numeric.os): In function `__psfs_do_numeric':
yann@1: __psfs_do_numeric.c:(.text+0x520): undefined reference to `__truncdfsf2'
yann@1: libc/libc_so.a(strtof.os): In function `strtof':
yann@1: strtof.c:(.text+0x1c): undefined reference to `__extendsfdf2'
yann@1: /home/ymorin/x-tools/build/armeb-unknown-linux-uclibc/gcc-4.1.1-uClibc-0.9.28.1/cc_core_prefix/lib/gcc/arm-linux-uclibc/4.1.1/libgcc.a(_fixunsdfsi.o):
yann@1: In function `__fixunsdfsi':
yann@1: libgcc2.c:(.text+0x34): undefined reference to `__fixdfsi'
yann@1: make[2]: *** [lib/libc.so] Error 1
yann@1: make[1]: *** [lib/libc.so.0] Error 2
yann@1: make[1]: Leaving directory `/home/ymorin/x-tools/build/armeb-unknown-linux-uclibc/gcc-4.1.1-uClibc-0.9.28.1/build-libc'
yann@1: make: *** [/home/ymorin/x-tools/build/armeb-unknown-linux-uclibc/gcc-4.1.1-uClibc-0.9.28.1/build-libc/lib/libc.a]
yann@1: Error 2
yann@1: 
yann@1: The second hunk fixes building final gcc because soft-float implies using -lfloat (which we don't have).
yann@1: 
yann@1: 
yann@1: diff -dur gcc-4.1.1.no-softfloat-fix/gcc/config.gcc gcc-4.1.1/gcc/config.gcc
yann@1: --- gcc-4.1.1.no-softfloat-fix/gcc/config.gcc	2007-02-01 21:57:17.000000000 +0100
yann@1: +++ gcc-4.1.1/gcc/config.gcc	2007-02-01 22:11:06.000000000 +0100
yann@1: @@ -690,7 +690,7 @@
yann@1:  	    default_use_cxa_atexit=yes
yann@1:  	    ;;
yann@1:  	*)
yann@1: -	    tmake_file="$tmake_file arm/t-linux"
yann@1: +	    tmake_file="$tmake_file arm/t-linux arm/t-arm-elf"
yann@1:  	    ;;
yann@1:  	esac
yann@1:  	tm_file="$tm_file arm/aout.h arm/arm.h"
yann@1: diff -dur gcc-4.1.1.-lfloat/gcc/config/arm/linux-elf.h gcc-4.1.1/gcc/config/arm/linux-elf.h
yann@1: --- gcc-4.1.1.-lfloat/gcc/config/arm/linux-elf.h	2007-02-01 21:57:17.000000000 +0100
yann@1: +++ gcc-4.1.1/gcc/config/arm/linux-elf.h	2007-02-01 23:00:42.000000000 +0100
yann@1: @@ -63,7 +63,7 @@
yann@1:     %{shared:-lc} \
yann@1:     %{!shared:%{profile:-lc_p}%{!profile:-lc}}"
yann@1:  
yann@1: -#define LIBGCC_SPEC "%{msoft-float:-lfloat} %{mfloat-abi=soft*:-lfloat} -lgcc"
yann@1: +#define LIBGCC_SPEC "-lgcc"
yann@1:  
yann@1:  #ifdef USE_UCLIBC
yann@1:  #define LINUX_TARGET_INTERPRETER "/lib/ld-uClibc.so.0"