CUDA 3.1 can't call std functions in kernel code?

Accelerated Computing CUDA CUDA Programming and Performance
1

I updated to CUDA 3.1.
CUDA 3.1 can’t call std::numeric_limitsstd::size_t::max() in kernel code.

global
void test()
{
size_t x =std::numeric_limitsstd::size_t::max();
}

I get compile error,
error: calling a host function from a device/global function is not allowed

CUDA 3.0 can call std functions.
For this bugs, I can’t reuse many std codes, it is very serious problem.
NVIDIA, please bug fix !