From owner-freebsd-wireless@FreeBSD.ORG Fri Apr 22 15:38:30 2011 Return-Path: Delivered-To: freebsd-wireless@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 9C2AA106564A for ; Fri, 22 Apr 2011 15:38:30 +0000 (UTC) (envelope-from petar@smokva.net) Received: from morrison.andev.ch (morrison.andev.ch [78.47.142.202]) by mx1.freebsd.org (Postfix) with ESMTP id 2F31B8FC16 for ; Fri, 22 Apr 2011 15:38:29 +0000 (UTC) Received: from pintail.smokva.net (84-73-24-182.dclient.hispeed.ch [84.73.24.182]) by morrison.andev.ch (Postfix) with ESMTP id ABF0C5DA91 for ; Fri, 22 Apr 2011 17:38:28 +0200 (CEST) Date: Fri, 22 Apr 2011 17:38:27 +0200 From: Petar Bogdanovic To: freebsd-wireless@freebsd.org Message-ID: <20110422153827.GA26063@pintail.smokva.net> Mail-Followup-To: freebsd-wireless@freebsd.org References: <20110421222524.GA5730@pintail.smokva.net> <20110422090705.GA1183@pintail.smokva.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21+23 (f7160c94ff70) (2010-12-30) Subject: Re: ath: tx-rate stops adapting after a while X-BeenThere: freebsd-wireless@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "Discussions of 802.11 stack, tools device driver development." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 22 Apr 2011 15:38:30 -0000 On Fri, Apr 22, 2011 at 05:25:15PM +0800, Adrian Chadd wrote: > > Have you tried this netbsd station against some other AP? It wasn't really some other AP but the software has changed on both sides during the past couple of years.. first it was netbsd-4 with freebsd-6-3, then netbsd-4 with freebsd-7 and finally netbsd-5-1 with freebsd-8-1. I remember that freebsd-7 had issues with stations that support some sort of power management but that went away with freebsd-8.. other than that, nothing comes to my mind. Although it's probably important to mention that the bump from netbsd-4 to netbsd-5-1 replaced the old binary-hal with the new/open ath-hal. > What hardware is in the netbsd box? It's a HP dc7100 with a AR5212 PCI device. > * somehow the card is incorrectly transmitting OFDM rates (11g) that > some things can pick up and some can't, or > * the netbsd device can't handle OFDM that you're transmitting and > only handles CCK rates correctly. The part that seems a bit weird to me is when [NetBSD] -> [FreeBSD] uses OFDM but [NetBSD] <- [FreeBSD] stubbornly keeps CCK. Is there any way to make this particular layer more verbose, so that one can see why the driver makes this decision and sticks with it forever? Right now, I get 3.4MB/s from the FreeBSD AP to the NetBSD station, which, as far as I remember, is pretty much the upper limit for OFDM54: $ ifconfig wlan0 list sta ADDR AID CHAN RATE RSSI IDLE TXSEQ RXSEQ CAPS FLAG 00:0b:.....NETBSD 2 10 54M 26.0 0 57455 44432 EPS AE RSN 58:b0:.......OSX1 1 10 48M 38.0 30 44418 36896 EPS AQEP RSN WME $ athstats 1 input output altrate short long xretry crcerr crypt phyerr rssi rate 44862732 28007738 5193 0 730634 2253 360540 0 2948 47 54M 1663 2481 0 0 71 0 0 0 0 49 54M 1666 2481 0 0 43 0 1 0 0 48 54M 1662 2511 0 0 58 0 3 0 0 49 54M 1691 2526 0 0 42 0 7 0 0 49 54M 1658 2487 0 0 67 0 0 0 0 47 54M 1664 2488 0 0 76 0 0 0 0 47 54M 1668 2508 0 0 54 0 4 0 0 48 54M 1708 2556 0 0 48 0 0 0 0 47 54M 1686 2508 0 0 56 0 0 0 0 47 54M 1653 2462 0 0 88 0 1 0 0 47 54M 1635 2478 0 0 44 0 1 0 0 46 48M 1592 2360 0 0 56 0 0 0 0 47 48M 1621 2404 0 0 32 0 0 0 0 46 54M 1662 2517 0 0 65 0 0 0 0 46 54M 1650 2466 0 0 60 0 3 0 0 48 54M 1653 2480 0 0 90 0 2 0 0 46 54M 1685 2520 0 0 47 0 0 0 0 48 54M 1684 2533 0 0 43 0 0 0 0 46 54M 1682 2513 0 0 66 0 1 0 0 47 54M 1618 2400 0 0 98 0 0 0 0 48 48M If it only could stay that way.. :) Petar Bogdanovic