Problems with arm-elf with soft-float

[Øyvind Harboe <oyvind dot harboe at zylin dot com>]

I'm trying to configure an arm-elf toolchain for AT91 (using the EB40a
evaluation board).


Hopefully I'm at the last obstacle:


I've configured and compiled the compiler/newlib with the scripts below, but
when I try to link an example application, I get a linker error since libgcc.a is
hard floating point and my code is soft-floating point.


I tried to modify gcc/config/t-arm-elf to compile a permutation of 
libgcc.a with and without hardware floating point, but to no avail. t-arm-elf
is modified to produce the interwork/thumb permutations.


My latest attempt was to configure gcc + newlib w/--nfp


Output from linking problems:


arm-elf-ld -Map thumb/lst/led_swing_remote.map -N -v -T ../../../targets/eb40a/l
dscript_remote thumb/cstartup_remote.o thumb/wait_irq.o  thumb/led_swing.o thumb
/gcc_main.o ../../../drivers/lib_drv/thumb/lib_drv_thumb.a ../../../parts/r40008
/thumb/r40008_thumb.a  /cygdrive/c/arm-elf/arm-elf/lib/thumb/interwork/libc.a /c
ygdrive/c/arm-elf/lib/gcc-lib/arm-elf/3.3/thumb/interwork/libgcc.a -o thumb/led_
swing_remote.elf
GNU ld version 2.13.1
arm-elf-ld: ERROR: /cygdrive/c/arm-elf/lib/gcc-lib/arm-elf/3.3/thumb/interwork/l
ibgcc.a(_udivsi3.o) uses hardware FP, whereas thumb/led_swing_remote.elf uses so
ftware FP
File format not recognized: failed to merge target specific data of file /cygdri
ve/c/arm-elf/lib/gcc-lib/arm-elf/3.3/thumb/interwork/libgcc.a(_udivsi3.o)
arm-elf-ld: ERROR: /cygdrive/c/arm-elf/lib/gcc-lib/arm-elf/3.3/thumb/interwork/l
ibgcc.a(_dvmd_tls.o) uses hardware FP, whereas thumb/led_swing_remote.elf uses s
oftware FP
File format not recognized: failed to merge target specific data of file /cygdri
ve/c/arm-elf/lib/gcc-lib/arm-elf/3.3/thumb/interwork/libgcc.a(_dvmd_tls.o)
make: *** [thumb/led_swing_remote.elf] Error 1





My scripts(I'm using CygWin, but I don't know if thats relevant):



# this script sets up the compiler and library before
# compiling the compilers.
#
#
# Note. the arm-elf configure files must manually be
# modified tos et the MULTILIB_OPTIONS to generate 
# the interwork library permutations

# Make directories where we do the build. 
mkdir build-binutils 
mkdir build-gcc-bare 
mkdir build-gcc 
mkdir build-newlib 
mkdir build-gdb 
mkdir install 
export TARGET=arm-elf 
export PREFIX=`pwd`/install 
export PATH=$PATH:$PREFIX/bin 
export NEWLIBPATH=`pwd`/../gcc33/newlib-1.11.0 
export BINUTILSPATH=`pwd`/../arm/binutils-2.13.1
export GCCSRC=`pwd`/../gcc33
# we need the object file format tools for arm-elf 
cd build-binutils 
$BINUTILSPATH/configure --enable-interwork --target=$TARGET --prefix=$PREFIX 
make all 2>&1 | tee ../binutils-make.log 
make install 2>&1 | tee ../binutils-install.log 
cd .. 
# for some reason we need to build and install the runtime 
# libraries before we compile the real compiler 
cd build-gcc-bare 
$GCCSRC/gcc/configure --enable-multilib --enable-interwork --target=$TARGET --prefix=$PREFIX --with-headers=$NEWLIBPATH/newlib/libc/include --with-gnu-as --with-gnu-ld --disable-shared --enable-languages=c --nfp
make all-gcc 2>&1 | tee ../gcc-bare-make.log
make install-gcc 2>&1 | tee ../gcc-bare-install.log
cd ..
cd build-newlib
$NEWLIBPATH/configure --nfp --enable-multilib --target=$TARGET --prefix=$PREFIX
make all install 2>&1 | tee ../newlib.log
cd ..


# Make directories where we do the build. 
export TARGET=arm-elf 
export PREFIX=`pwd`/install 
export PATH=$PATH:$PREFIX/bin 
export NEWLIBPATH=`pwd`/../gcc33/newlib-1.11.0 
export GCCSRC=`pwd`/../gcc33
# Finally we compile our compiler 
cd build-gcc 
#disable-nls 
#disable-jvmpi 
#disable-libffi 
#disable-interpreter 
#disable-getenv-properties 
$GCCSRC/gcc/configure --disable-libffi --disable-nls --disable-jvmpi --disable-getenv-properties --disable-interpreter --without-x --disable-java-awt --target=$TARGET --prefix=$PREFIX --with-gnu-as --with-gnu-ld --with-newlib=yes --with-headers=$NEWLIBPATH/newlib/libc/include --enable-languages=c,c++ --enable-interwork --enable-multilib --nfp
make all 2>&1 | tee ../gcc-make.log 
make install 2>&1 | tee ../gcc-install.log 
cd .. 
echo Done! 



modified t-arm-elf:


LIB1ASMSRC = arm/lib1funcs.asm
LIB1ASMFUNCS = _udivsi3 _divsi3 _umodsi3 _modsi3 _dvmd_tls _bb_init_func _call_via_rX _interwork_call_via_rX
# We want fine grained libraries, so use the new code to build the
# floating point emulation libraries.
FPBIT = fp-bit.c
DPBIT = dp-bit.c
fp-bit.c: $(srcdir)/config/fp-bit.c
 echo '#define FLOAT' > fp-bit.c
 echo '#ifndef __ARMEB__' >> fp-bit.c
 echo '#define FLOAT_BIT_ORDER_MISMATCH' >> fp-bit.c
 echo '#endif' >> fp-bit.c
 cat $(srcdir)/config/fp-bit.c >> fp-bit.c
dp-bit.c: $(srcdir)/config/fp-bit.c
 echo '#ifndef __ARMEB__' > dp-bit.c
 echo '#define FLOAT_BIT_ORDER_MISMATCH' >> dp-bit.c
 echo '#define FLOAT_WORD_ORDER_MISMATCH' >> dp-bit.c
 echo '#endif' >> dp-bit.c
 cat $(srcdir)/config/fp-bit.c >> dp-bit.c

MULTILIB_OPTIONS     = marm/mthumb
MULTILIB_DIRNAMES    = arm thumb
MULTILIB_EXCEPTIONS  = 
 
# MULTILIB_OPTIONS     += mlittle-endian/mbig-endian
# MULTILIB_DIRNAMES    += le be
# MULTILIB_EXCEPTIONS  = 
# MULTILIB_MATCHES     = mbig-endian=mbe mlittle-endian=mle
# 
#MULTILIB_OPTIONS    += mhard-float/msoft-float
#MULTILIB_DIRNAMES   += fpu soft
#MULTILIB_EXCEPTIONS += *mthumb/*mhard-float*
# 
# MULTILIB_OPTIONS    += mapcs-32/mapcs-26
# MULTILIB_DIRNAMES   += 32bit 26bit
# MULTILIB_EXCEPTIONS += *mthumb/*mapcs-26*
# 
MULTILIB_OPTIONS    += mno-thumb-interwork/mthumb-interwork
MULTILIB_DIRNAMES   += normal interwork
MULTILIB_EXCEPTIONS += *mapcs-26/*mthumb-interwork*

#MUTLILIB_EXTRA_OPTS = msoft-float
# 
# MULTILIB_OPTIONS    += fno-leading-underscore/fleading-underscore
# MULTILIB_DIRNAMES   += elf under
# 
# MULTILIB_OPTIONS    += mcpu=arm7
# MULTILIB_DIRNAMES   += nofmult
# MULTILIB_EXCEPTIONS += *mthumb*/*mcpu=arm7*
# # Note: the multilib_exceptions matches both -mthumb and
# # -mthumb-interwork
# #
# # We have to match all the arm cpu variants which do not have the
# # multiply instruction and treat them as if the user had specified
# # -mcpu=arm7.  Note that in the following the ? is interpreted as
# # an = for the purposes of matching command line options.
# # FIXME: There ought to be a better way to do this.
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm7d
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm7di
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm70
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm700
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm700i
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm710
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm710c
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm7100
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm7500
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm7500fe
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm6
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm60
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm600
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm610
# MULTILIB_MATCHES    += mcpu?arm7=mcpu?arm620
EXTRA_MULTILIB_PARTS = crtbegin.o crtend.o crti.o crtn.o
# If EXTRA_MULTILIB_PARTS is not defined above then define EXTRA_PARTS here
# EXTRA_PARTS = crtbegin.o crtend.o crti.o crtn.o
LIBGCC = stmp-multilib
INSTALL_LIBGCC = install-multilib
# Currently there is a bug somewhere in GCC's alias analysis
# or scheduling code that is breaking _fpmul_parts in fp-bit.c.
# Disabling function inlining is a workaround for this problem.
TARGET_LIBGCC2_CFLAGS = -Dinhibit_libc -fno-inline
# Assemble startup files.
$(T)crti.o: $(srcdir)/config/arm/crti.asm $(GCC_PASSES)
 $(GCC_FOR_TARGET) $(GCC_CFLAGS) $(MULTILIB_CFLAGS) $(INCLUDES) \
 -c -o $(T)crti.o -x assembler-with-cpp $(srcdir)/config/arm/crti.asm
$(T)crtn.o: $(srcdir)/config/arm/crtn.asm $(GCC_PASSES)
 $(GCC_FOR_TARGET) $(GCC_CFLAGS) $(MULTILIB_CFLAGS) $(INCLUDES) \
 -c -o $(T)crtn.o -x assembler-with-cpp $(srcdir)/config/arm/crtn.asm

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com