We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask(), and by defining
it, the old arch_send_call_function_ipi is defined by the core code.

We also take the chance to change send_IPI_mask() and use the new
for_each_cpu() iterator.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>