From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail2.candelatech.com ([208.74.158.173]) by bombadil.infradead.org with esmtp (Exim 4.80.1 #2 (Red Hat Linux)) id 1WsDw9-0003Av-KG for ath10k@lists.infradead.org; Wed, 04 Jun 2014 16:21:14 +0000 Message-ID: <538F4761.1030906@candelatech.com> Date: Wed, 04 Jun 2014 09:20:49 -0700 From: Ben Greear MIME-Version: 1.0 Subject: Re: RE : RE : RE : ath10k: firmware crash in station mode & problem DFS References: <1399565896-29791-1-git-send-email-greearb@candelatech.com> , , , <5386059B.7040301@candelatech.com> , <53889A89.90802@candelatech.com> , <538CAACA.2030202@candelatech.com> , <538E9D26.4040009@candelatech.com> In-Reply-To: List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "ath10k" Errors-To: ath10k-bounces+kvalo=adurom.com@lists.infradead.org To: Vu Hai NGUYEN Cc: Janusz Dziedzic , Patrick CARNEIRO RODRIGUEZ , "ath10k@lists.infradead.org" , Janusz Dziedzic On 06/04/2014 02:20 AM, Vu Hai NGUYEN wrote: >> I have a script to build every commit now, but except for the first few >> and the last few images, you would need various driver changes to work with them. > >> I can rebase my firmware changes so that hopefully all (or at least most) of the >> images will work on standard kernels, but it may be a few days before I >> can get that done. > >> In the meantime, please try these two: > >> Compiled from original source received from QCA: > >> http://www.candelatech.com/downloads/firmware-2-community-commit-0-207ce4b.bin > >> It appeared to me that that code had a regression, so next commit fixed >> that: > >> http://www.candelatech.com/downloads/firmware-2-community-commit-1-5717c65.bin > >> If either of these work with DFS, then some change I made later broke it, and I can >> build the rest of the images so we can figure out where. > > Both firmware in the link that you gave me work well, my station associate directly with the AP after running wpa_supplicant. > But then I re-put your firmware version non commercial, my STA did not associate directly after running wpa_supp, it keep print the message SCAN EVENT STARTED of wlan0. > I run "iw wlan0 scan" and some times it can see and associate with my AP, some times it can not see the network with channel DFS, I mean it is not stable. Ok, I will work on a series of firmwares for each of my commits so we can find the bad commit and fix it. > And I had other question too, I thought that you've done some test with ath10k in mode 802.11ac by iperf? What is the maximum rate that you obtain? > I'm using 2 PC and connect each one to my AP and STA, but the max rate that I had is around 120Mbps only, in both 802.11n and 802.11ac. > Strange things is that I config 3 antennas in used and leave the rate automatically (Don't set rates using "iw"), my station dump and file "fw_stats" in debugfs show me the rates of MCS 7 or 15 (in 802.11n) and NSS1 or 2 (in 802.11ac). > It never set in MCS 23 or NSS3, though my module wifi support 3 STREAMS, 1 can see it from "iw list": > VHT RX MCS set: > 1 streams: MCS 0-9 > 2 streams: MCS 0-9 > 3 streams: MCS 0-9 > But it also show me: VHT RX highest supported: 0 Mbps and VHT TX highest supported: 0 Mbps, is there something wrong with it? > If I use "iw" to fixed rate, the info from station dump (TX rates is always false because of firmware issue), but RX rate and the info from "fw_stats" is same thing as what I fixed (I can fixe rate at MCS23 and NSS 3 too, I also verified with dmesg | grep mac to see debug message from ath10k). But as I said above, my rate is around 120 Mbps in iperf. We can get 500+Mbps download, and about 400Mbps upload, TCP traffic using our own traffic generator, which is somewhat similar to iperf. Use large tx/rx buffers, large write sizes, and 10+ streams in iperf for more optimal throughput. We often get rates reported at 1.3Gbps, so it seems in our tests we train up well. Unless you are using a modified driver with my firmware, you will not get useful tx-rate stats reported, but I think the rx-rate is accurate with latest upstream driver. Thanks, Ben > > NGUYEN Vu Hai > Acita-Sodielec > Route de Mayres - B.P. 9 > 12100 St GEORGES DE LUZENCON > FRANCE > -- Ben Greear Candela Technologies Inc http://www.candelatech.com _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k