On Fri, Mar 01, 2019 at 02:25:26PM +0200, Lauri Tirkkonen wrote: > On Wed, Feb 27 2019 15:06:49 +0100, Stefan Sperling wrote: > > Otherwise this diff affects 11n access points only, i.e. athn(4). > > It makes my athn(4) AP run circles (up to 15 Mbps) on a 2 GHz channel > > with overlapping networks. > > > > Any more testers? > > Based on cursory testing, seems to work fine on my athn(4) 11n 5GHz AP. > But I didn't observe any improvements in bandwidth, using an Android > phone station as traffic sink (~9MBit/s maximum both before and after). >
That sounds OK. The actual benefits will depend on the environment. I am seeing around 12Mbit/s on average on 5GHz, and about 9Mbit/s on 2. The main benefit for me is on 2 GHz, where I typically saw less than 500kbit/s and hence never used it. What's important is that we now have a baseline for further improvements (such as Tx aggregration and 40 MHz channels) which isn't ridiculously low. Thanks for testing! Meanwhile, the diff has been committed and should be in the next snapshot. I am still interested in additional test reports.
