How is floating point overflow handled in aeabi library functions such as __aeabi_fadd?
Does the function return a pre-defined value? Or does it update any registers?
Floating-point overflow typically returns an infinite result.
Additional "fenv" state may be set and tested for using something like "fetestexcept(FE_OVERFLOW)" if the appropriate compiler options and library version are used, see the following for more details:
http://infocenter.arm.com/help/topic/com.arm.doc.dui0472m/chr1359124920125.html
Best regards
Simon.