Return-path: Received: from mail-ig0-f179.google.com ([209.85.213.179]:36973 "EHLO mail-ig0-f179.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754793AbbG1Drl (ORCPT ); Mon, 27 Jul 2015 23:47:41 -0400 Received: by igbpg9 with SMTP id pg9so104876183igb.0 for ; Mon, 27 Jul 2015 20:47:40 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1437554563-11712-1-git-send-email-martin.blumenstingl@googlemail.com> Date: Mon, 27 Jul 2015 20:47:40 -0700 Message-ID: (sfid-20150728_054744_403307_FC38CA35) Subject: Re: [PATCH RESEND] ath9k: Fix NF CCA limits for AR9287 and AR9227 From: Adrian Chadd To: Martin Blumenstingl Cc: ath9k-devel , ath9k-devel , Felix Fietkau , John Linville , "linux-wireless@vger.kernel.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: Those are some high values (eg from noise/spur somewhere) and the driver is deciding to program in -110. So the radio thinks its deaf. Hm, I think that explains a few other NICs behaving badly too.... -a On 27 July 2015 at 16:55, Martin Blumenstingl wrote: > On Tue, Jul 28, 2015 at 1:50 AM, Adrian Chadd wrote: >> is it never getting any readings before your patch? NF Readings all >> look empty... > Yes, looks like it. Unfortunately I haven't saved a complete kernel > log (with ath9k debug level = 0x00004449) but only a few parts. Here > are the interesting bits: > <7>[ 449.746679] ath: phy1: NF calibrated [ctl] [chain 0] is -106 > <7>[ 449.746697] ath: phy1: NF[0] (-106) > MAX (-110), correcting to MAX > <7>[ 449.746714] ath: phy1: NF calibrated [ctl] [chain 1] is -90 > <7>[ 449.746731] ath: phy1: NF[1] (-90) > MAX (-110), correcting to MAX > <7>[ 449.746748] ath: phy1: NF calibrated [ext] [chain 0] is -103 > <7>[ 449.746765] ath: phy1: NF[3] (-103) > MAX (-110), correcting to MAX > <7>[ 449.746781] ath: phy1: NF calibrated [ext] [chain 1] is -86 > <7>[ 449.746798] ath: phy1: NF[4] (-86) > MAX (-110), correcting to MAX > <7>[ 449.746878] ath: phy1: Calibration @37407 finished: long ani, caldone: true > > I'd assume (because I don't know the ath9k code) that ath9k simply > ignores NF readings that are "out of range" (as suggested by those > messages). > > > Regards, > Martin