Return-path: Received: from arrakis.dune.hu ([78.24.191.176]:41220 "EHLO arrakis.dune.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751865AbaKLIsN (ORCPT ); Wed, 12 Nov 2014 03:48:13 -0500 Message-ID: <54631EC7.1030009@openwrt.org> (sfid-20141112_094816_971387_0D07C64D) Date: Wed, 12 Nov 2014 09:48:07 +0100 From: Felix Fietkau MIME-Version: 1.0 To: Sujith Manoharan CC: "Pan, Miaoqing" , John Linville , "linux-wireless@vger.kernel.org" , ath9k-devel Subject: Re: [PATCH 4/6] ath9k: Fix high tx power in multi-chain mode References: <1415689644-29693-1-git-send-email-sujith@msujith.org> <1415689644-29693-5-git-send-email-sujith@msujith.org> <54621BFF.7080709@openwrt.org> <21602.40415.179942.572517@gargle.gargle.HOWL> <5326b4c51c744de285b390d8c78a47be@NASANEXM01E.na.qualcomm.com> <21602.54599.303861.990586@gargle.gargle.HOWL> In-Reply-To: <21602.54599.303861.990586@gargle.gargle.HOWL> Content-Type: text/plain; charset=windows-1252 Sender: linux-wireless-owner@vger.kernel.org List-ID: On 2014-11-12 04:34, Sujith Manoharan wrote: > Pan, Miaoqing wrote: >> the issue was found on ap143(QCA9531, 11n, 2x2), reported by Korea team. They >> want to make ap143 to work on 1x2 or 2x2 mode. > > Ok, so apparently there are boards that are 2x2 in the HW, but the > tx chainmask is set as 1 in the eeprom. In such cases, the thermometer > has to be enabled for both the chains. So, I think we need this fix. Makes sense. I don't like all these duplicated SREV checks though. Let's store the chip_chainmask variable from ath9k_hw_fill_cap_info somewhere in ath_hw and use that. - Felix