I have a question concerning the use of the soft float library for floating point operations in a user level application.  

Is the soft float library automatically used for any floating point operation when the application is compiled using a buildroot that was compiled with ‘SOFT_FLOAT:=true’ in the top level Makefile without explicitly using the compiler flags –msoft-float and -lfloat? 

If the soft float library is not automatically used, what are the proper compiler flags needed to explicitly compile for and link to the soft float library?