jolly wrote:
the second problem is wrong scheduling. if you would
have set drlcmac
debugging, you would see when polling is scheduled. the scheduled FN
must be 13 bursts (3 blocks) in advance of the RTS FN. 8 or 9 bursts
later, the scheduler must schedule "free USF for polling" for the next
block, which is 4 or 5 bursts later. since i don't see the drlcmac
debugging in your log, i cannot verify it.
it seems to work in the beginning, but later the transmission seems to
be shifted at the BTS. when the BTS sends the RTS with a given FN, the
PCU assembles a block for that FN and sends it to the BTS. the BTS must
send that block at the given FN (or drop it, if it is too late). it must
not send it at a different FN.