Closed Buanderie closed 7 years ago
There is no "tx rate lock patch". There are patches:
[1] is needed to fix tx power at max supported level and disable CRDA restrictions
[2] enables API for specify bitrate when injecting packets. But for current ieee80211 stack is not sufficient to specify rate in injected header and you also need to set it via iw dev $WLAN set bitrates ht-mcs-5 1 sgi-5
before tx start
[3] removes most of tx timeslot guard intervals to maximize throughput. We use only one-directional broadcast and tx wouldn't listen for ACK's. https://en.wikipedia.org/wiki/Extended_interframe_space
ez-wifibroadcast-1.4-kernel-4.4-patches.diff implements [1] and [3] mac80211-radiotap-bitrate_mcs_rtscts.linux-4.4.patch implements [2]
Also there are AR9271 patches, but I've never try it because I don't have any atheros hardware. My HW setup is RT3572 (rt2800 driver) in 5GHz band.
Thank you for the information...
Well, concerning [1], I have some additional questions. Isn't there a risk that the card will try to output at power exceeding its nominal value (btw, have you ever seen a case where it caused HW damage ?).
I also don't quite get this part:
// CHANGES: set 58 for every data rate for (i = 0; i < Ar5416RateSize; i++) { // ratesArray[i] -= AR5416_PWR_TABLE_OFFSET_DB * 2; ratesArray[i] = 58; }
What does setting this to 58 do ? Also, why 58 ?
Regarding [3], I can't really use that since I need to be able to switch back and from AP mode. I couldn't get [2] to work, yet but it's not really a priority :)
I've didn't test any ARxxxx patches due to lack of hardware. They were taken from https://github.com/bortek/EZ-WifiBroadcast as is. In case of TX powes limit - card select minumum of its supported power (stored in EEPROM) and CRDA power limit, so you shouldn't got HW damage.
If you have recent mainline kernel (4.x) you don't need [2] (it already here). [1] and [3] are optional and needed only to increase transmit range and channel bandwidth.
Moved to wiki
I understand that in AP mode, the minstrel_ht (or the Atheros one) rate selection algorithm will change the rate/modulation dynamically based on the STA RSSI. But in monitor/injection mode, why should the rate change dynamically ? Shouldn't calling "iw dev wlan0 set bitrates... ht-mcs-5 ..." be enough to set the bitrates and MCS ? Also I have some difficulties finding out what the patch is really doing.