I've been working on the TX underrun issues that I've only just noticed!
* I've just made the TX underrun and TX FIFO adjustment code in the
driver match what the reference driver does.
* That is, the thresholds that are being used for when to start
refilling the TX FIFO now match what the reference driver does for the
AR5416 and later chips, rather than using the same thresholds as the
* I've also fixed the 11n TX delimiter / TX data underrun statistics.
The TX FIFO underrun statistics were already okay.
* I haven't yet updated the earlier HAL code to properly enable the
URN interrupt for pre-11n chips; I'll do that soon.
I'd appreciate it if people would update their devices to the most
recent -HEAD and do some iperf testing, just to make sure things
aren't entirely broken. I've done some initial testing on my local
devices but nothing thorough.
You'll see TX FIFO, data and delimiter underruns show up in athstats.
email@example.com mailing list
To unsubscribe, send any mail to "freebsd-wireless-unsubscr...@freebsd.org"