Problem in printf

Hello,

I try to use a Double-precision floating-point custom instruction with a NIOS II.So, I get a problem in 'printf' when I rebuild the retoune __eqdf2() to emulate the floating point operation . So, The attachement figure shows the result.

This problem is not happened when I have rebuilded the others floating-point routines: e.g __nedf2(), __adddf3, __subdf3, ...

I think there is a conflict between libraries (standard floating point library and the new rebuided library).

can someone help me??

Thank you for all members.