sched: walt: Optimize cycle counter reads

The cycle counter read is a bit of an expensive operation and requires
locking across all CPUs in a frequency domain. Optimize this by
returning the same value if the delta between two reads is zero i.e
two reads are done in the same sched context for the same CPU.

Change-Id: I99da5a704d3652f53c8564ba7532783d3288f227
Signed-off-by: Vikram Mulukutla <markivx@codeaurora.org>
[pkondeti@codeaurora.org: limit the optimization to the
same CPU for the sched context]
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>