Return-path: Received: from mail-ig0-f171.google.com ([209.85.213.171]:38645 "EHLO mail-ig0-f171.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752456AbbJFQHW (ORCPT ); Tue, 6 Oct 2015 12:07:22 -0400 Received: by igxx6 with SMTP id x6so83979923igx.1 for ; Tue, 06 Oct 2015 09:07:22 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <87mvvw13rl.fsf@kamboji.qca.qualcomm.com> References: <1437554563-11712-1-git-send-email-martin.blumenstingl@googlemail.com> <87mvvw13rl.fsf@kamboji.qca.qualcomm.com> Date: Tue, 6 Oct 2015 09:07:22 -0700 Message-ID: (sfid-20151006_180729_862660_031A4A61) Subject: Re: [PATCH RESEND] ath9k: Fix NF CCA limits for AR9287 and AR9227 From: Adrian Chadd To: Kalle Valo Cc: Martin Blumenstingl , 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: from qca-wifi-1.2: ar5416phy.h:#define AR_PHY_CCA_NOM_VAL_KIWI_2GHZ -120 ar5416phy.h:#define AR_PHY_CCA_MIN_GOOD_VAL_KIWI_2GHZ -127 ar5416phy.h:#define AR_PHY_CCA_MAX_GOOD_VAL_KIWI_2GHZ -110 .. and: ar5416phy.h:#define AR_PHY_CCA_NOM_VAL_MERLIN_2GHZ -112 ar5416phy.h:#define AR_PHY_CCA_MIN_GOOD_VAL_MERLIN_2GHZ -127 ar5416phy.h:#define AR_PHY_CCA_MAX_GOOD_VAL_MERLIN_2GHZ -97 .. so although the reference driver has the lower values for AR9287, the AR9280 values work better for him. So, hm. I think I recall finding the same when I did the initial AR9287 support and found the AR9280 ANI/CCA values behaved better. It doesn't look like he has any noise spurs in the spectral scan output in his NIC; it just looks like everything needs to be slightly higher to deal with ambient noise. I'm okay with including it but I'd like other feedback/testing from AR9287/AR9227 users. I used the AR9287 in a laptop for about a year w/ the AR9280 values (running FreeBSD-HEAD) and it performed great. -adrian On 6 October 2015 at 07:16, Kalle Valo wrote: > Martin Blumenstingl writes: > >> The FreeBSD driver [0] uses the same 2G values as for the AR9280 chips. >> Using the same values in ath9k results in much better throughput for me. >> >> Before this patch I had a huge amount of packet loss (sometimes up to >> 40%) and the max transfer speed was somewhere around 5Mbit/s. With this >> patch applied I have zero packet loss and ten times the throughput. >> My device uses a AR9227 which is the PCI variant of the AR9287. >> >> [0] http://bxr.su/FreeBSD/sys/dev/ath/ath_hal/ar9002/ar9287.h >> >> Signed-off-by: Martin Blumenstingl > > So what should I do with this patch? Is this good to apply? > > Discussion here: > > https://patchwork.kernel.org/patch/6841051/ > > -- > Kalle Valo > -- > To unsubscribe from this list: send the line "unsubscribe linux-wireless" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html