[uClinux-dev] address error

David McCullough davidm at snapgear.com
Thu Nov 1 18:42:42 EST 2001


Jivin John D. Charlton lays it down ...
> David,
> 
> Thanks for your help.  I am able to build and run the test app which uses 
> only integer operations.  The combination that works is to let the 
> m68k-elf-gcc use its own choice of crt0.o and to specify the libc.a that 
> comes with the uClinux-distribution.  Also, remove the -msoft-float as you 
> suggested.

It doesn't matter which crt0.o you use,  the both work fine.  Just use the
one that the build system in the 20010622 distro chooses,  which should be
the version installed with the compiler.

> The original intent of the testFP.c was to test floating point.  The floating 
> point is not yet working, but I suspect I need to use a later version uClibc 
> or rebuild it with -DFLOATS. There are currently build problems with the 
> version of uClibc in the CVS for the uCdimm platform.  I am following some of 
> the leads posted in the archives for uClinux-dev and uClibc for floats.
...
> make -k 
> m68k-elf-gcc -m68000 -Os -g -fomit-frame-pointer -DCONFIG_LINEO   -Dlinux 
> -D__linux__ -Dunix -D__uClinux__ -DEMBED -fno-builtin -msep-data -I. -I 
> /opt/uClinux/linux-2.4.x/include -I /opt/uClinux/lib/libc/include -o testFP 
> testFP.c -Wl,-elf2flt -L/opt/uClinux/lib/libc 
> -L/usr/local/lib/gcc-lib/m68k-elf/2.95.3/m68000/msep-data -lc -lgcc
> 
> 
> # ./testFP
> c = 9, d = 9
> LIBC:PRINTF float not implementedx = # 


You need to link against libm to get floating point versions of these
functions.  Do that and you will have working FP, be sure to put $(LIBM)
into the link line berfore $(LDLIBS) or libc,

Cheers,
Davidm


-- 
David McCullough:    Ph: +61 7 3435 2815  http://www.SnapGear.com
davidm at snapgear.com  Fx: +61 7 3891 3630  825 Stanley St., W'gabba QLD 4102, Oz
This message resent by the uclinux-dev at uclinux.org list server http://www.uClinux.org/



More information about the uClinux-dev mailing list