Bug 717

Suppose the following situation: 

We have two stations A and B and we have two queues at each stations and packets to be sent in both queues at station A (all the packets are addressed to the station B). The second queue (with lower priority) has started to transmit a packet, but this packet has crashed (due to PER for example). After it DcaTxop will be notifyed about transmission failure and m_slrc in WifiRemoteStation shall be incremented. After this event, the first queue (with higher priority) has accessed the medium and started to transmit its frame. After successfull transmission m_slrc counter shall be set to 0, and previously crashed frame shall be transmitted with wrong slrc counter. Now suppose that station A and B can not communicate -- so, the packets from AC_BE queue shall be retried too many times when packets from AC_VO shall be dropped due to retry counter.

Proposed solution - is to add a map of <AccessCategory, retryCounter> rather than keeping retry counter only in WifiRemoteStationManager (but I do not know how will work rate control algorithms), like described in 9.9.1.6 of 802.11-2007, - short and long retry counters shall be supported for each AC