- wrap cq->cqidx_inc based on cq size.
- optimize t4_arm_cq logic.
Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>