Hello everybody Currently, I am trying to create a cross compilation tool chain for an embedded target system, based on an ARM-922T processor. The device has the following features: - No FPU - little endian - armv4 instruction set The cross linux from scratch book provided some very good starting points on this task. Nevertheless, there are a few point which remain unclear so far to me, namely: - I take it that I have to compile gcc to apply software floating point in this case - or should it also work to create a toolchain without any additional flags? - I am unsure of the nature of the available software floating point flags. >From what I've seen so far, --with-float=soft can be used when configuring programs. Is it supposed to be passed to the gcc configure file when creating the actual cross compiler targeting a software floating point machine? Or should it be used when configuring the final program running on the software floating point target? - From the manual, I deduced that -msoft-float is to be passed to the compiler when compiling a program running on the soft-float target. How am I best advised to achieve this? Is there a flag I can specify when I create my gcc compiler so that it always creates soft-float programs? Or shouldn't I at all specify any additional flags when creating my toolchain, but only when compiling the actual target programs with it? Thank you very much in advance for any help on this issue and best regards Pascal -- View this message in context: http://old.nabble.com/Cross-GCC-for-ARM-922T-tp26499746p26499746.html Sent from the gcc - Help mailing list archive at Nabble.com.
