We currently account a '0' dispatch, and anything above that still falls
below the range set by BLK_MQ_MAX_DISPATCH_ORDER. If we dispatch more,
we don't account it.
Change the last bucket to be inclusive of anything above the range we
track, and have the sysfs file reflect that by including a '+' in the
output: