I am trying to write an ATM driver to Linux. The fact that I am not
allowed to use FP operations in the kernel really make it hard to me
caculating rates and timers. Do you know of any good solution? do you
have any good implementation that replaced the floating point? I am
not using actually the FP register, but the emulation of the compiler.
Why is it a problem??

BTW - I dont know how, but I succeded to compile my module with
variable that is "double", but when I inserted the module the kernel
massage "sloating point used in kernel" appeared. How could it be?