Hi all, I am using yagarto ARMTools chain, Binutils-2.20 Newlib-1.18.0 GCC-4.4.2 GDB-7.0.1 In my make file options(for CortexM3 processor,LPC17series), -mcpu=cortex-m3 -mthumb mfpu=fpa (or) -mfpu=vfp with this options My program going to hangs when i use simple float operations(covert float to int using type cast,substract float with integer etc.This is for supporting ftoa() api). Same code working form me in following makefile options(ARM7 processor,LPC23xx,21xxseries) in my another project, -mcpu=arm7tdmi -mthumb-interwork.I havn't seen any issue past 6months. Please guide me?? If any one already faced this issue please share your knowledge??? FYI:I am using LPC17 seires with FreeRTOS.
Guest
#1654420
Hi, I want to confirm this problem, but I am using an ARM7 LPC2478 and do not use the thumb set. At google I found a similar report concerning x86: http://old.nabble.com/newlib-error-when-printf-float-on-x86-64-td27499676.html Seems to be a general problem which is new in 1.18. Regards, Gottfried
Prabu 84 wrote: > > In my make file options(for CortexM3 processor,LPC17series), > -mcpu=cortex-m3 -mthumb mfpu=fpa (or) -mfpu=vfp > So why would you use -mfpu=fpa of vfp when a Cortex-M3 has neither an FPA not a VFP? These settings will cause the generation of instructions for a math co-processor that does not exist.