K 10
svn:author
V 6
adrian
K 8
svn:date
V 27
2011-08-10T16:13:43.066972Z
K 7
svn:log
V 739
Fix the (current) major issue where BAW tracking was failing, causing the TX
side to hang.

When ath_tx_draintxq() was called, flushing hardware-queued TX frames, they'd
be immediately freed. The bf_comp handler wasn't being called. So frames
in an aggregation session would be freed without their seqnum being updated
in the BAW. If enough frames were skipped, the current seqnum would be outside
the BAW and thus TX on that TID would stop.

This fixes the initial issue of hanging TX during an aggregation session.

This also causes the receiver to have to slide the BAW over as there will be
gaps in what's transmitted but without a BAR being sent to force the RX side
to update. That's (for now) preferable to having things just hang.

END