Hardware statistics retrieval hurts in tight invocation loops.

Avoid extraneous write and enforce strict ordering of writes targeted to
the tally counters dump area address registers.
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Tested-by: NOliver Freyermuth <o.freyermuth@googlemail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>