GCC 4.1.x and hardware FP, software FP errors


I’m trying to compile NET+OS with gcc 4.1.x.
My current compiler setup works fine with the Olimex
LPC-P2148 ARM7 processor running FreeRTOS on it.
(fyi: the code produced is little-endian)

In compiling the NET+OS code, i’ve run into a snag where
I’m getting these kinds of errors:

libgcc.a(_udivsi3.o) uses hardware FP, whereas blram.elf uses software FP

I’ve changed the various Makefiles to use -msoft-float,
but i still get these errors.

Any hints?



Got it fixed.
Had to rebuild the GCC 4.1.1 toolchain with the ./configure option of --float=soft.

I hacked the other binaries to build the resulting image.

So, I think I have a working environment on Mac OS X 10.4.7.

Now to figure what flash address to write the image to…
I’m using the Amontec USBJtag.
Unfortunately, the Wi-EM board uses a non-standard JTAG pin outs… :frowning: