base: Workload num blocks reduce code size

This patch makes MARS_WORKLOAD_PER_BLOCK 16 instead of 15 to reduce code size as
much as possible by doing operations on power of 2 rather than not. The kernel
and workload queue must assume index 0 of workload_queue_block bits array is
reserved for the mutex lock. This change reduces kernel code size by 584 bytes.

Signed-off-by: Yuji Mano <yuji.mano@am.sony.com>
Acked-by: Kazunori Asayama <asayama@sm.sony.co.jp>