From mboxrd@z Thu Jan 1 00:00:00 1970 From: Till Wollenberg Date: Tue, 25 Aug 2015 17:42:18 +0200 Subject: [ath9k-devel] How to get CSI from Atheros Chipset In-Reply-To: References: <55DB34BC.5010301@uni-rostock.de> Message-ID: <55DC8CDA.3000106@uni-rostock.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org Hi Adrian, * On 24.08.2015 18:40 Adrian Chadd wrote: > On 24 August 2015 at 08:14, Till Wollenberg wrote: >> * On 23.08.2015 at 21:57 Adrian Chadd wrote: >>> [Getting CSI] >>> ok, I've found a few pieces about this. What exactly are people after? >> >> SNR per OFDM sub-carrier for a recevied frame would be *very* interesting to me. > ok, > > So the legacy bits include |I| and |Q| but it's not per-frame, it's > set-once-and-get-next-frame. You then read bins from the NIC itself in > some shared memory inside said NIC. I'll go see if this works on the > earlier chips and try to get this published, at least so there's > something to start with. Sounds interesting. So is this per-frame information if the internal registers are read out right after receiving a frame, or is it always an aggregate over several received frames? > I think AR9380 and later support what you saw in that paper - where > you request that you get CSI data as part of some receive frame. I > /think/ it's part of the TxBF support. I'll have to double-check if > that's okay to extract and publish. I'm not the OP so I'm not quite sure what the paper is about. Nevertheless, I plan to use AR9590 and/or AR9390 cards so I guess that these meet the criterion of "AR9380 and later". Best Till.