Re: 802.11n MIMO support in -current

2016-12-10 Thread Scott Bonds

wow, that's awesome!
I've been rocking a athn lately but I'll swap back to iwm to help test

On 12/10, Stefan Sperling wrote:

The net80211 stack and iwm(4) driver now support MIMO in -current.

In my own testing, things work just fine. But I have gotten used
to breaking other people's wifi without being aware of it.
So please test -current and let me know about any regressions.

Because iwm(4) devices have 2 antennas MCS 15 is the maximum Tx rate
the hardware will support. Support for 40MHz channels and Tx aggregation
are left for future work.

I owe several people attribution:

This work would have been impossible without the help from Theo Buehler (tb@).
He made significant contributions to the implementation of a new rate
scaling algorithm which supports MIMO.

Thanks to the researchers who released the MiRA paper to the internet.
I have studied a dozen papers or so, and eventually decided to implement
this one because it was the only paper which documented enough details
and which didn't impose impractical requirements.

Genua GmbH has funded most of the time I spent working on this project.
Without this kind of support I would not have attempted this project.




802.11n MIMO support in -current

2016-12-10 Thread Stefan Sperling
The net80211 stack and iwm(4) driver now support MIMO in -current.

In my own testing, things work just fine. But I have gotten used
to breaking other people's wifi without being aware of it.
So please test -current and let me know about any regressions.

Because iwm(4) devices have 2 antennas MCS 15 is the maximum Tx rate
the hardware will support. Support for 40MHz channels and Tx aggregation
are left for future work.

I owe several people attribution:

This work would have been impossible without the help from Theo Buehler (tb@).
He made significant contributions to the implementation of a new rate
scaling algorithm which supports MIMO. 

Thanks to the researchers who released the MiRA paper to the internet.
I have studied a dozen papers or so, and eventually decided to implement
this one because it was the only paper which documented enough details
and which didn't impose impractical requirements.

Genua GmbH has funded most of the time I spent working on this project.
Without this kind of support I would not have attempted this project.