Return-path: Received: from c60.cesmail.net ([216.154.195.49]:40264 "EHLO c60.cesmail.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1759880AbYJNAAm (ORCPT ); Mon, 13 Oct 2008 20:00:42 -0400 Subject: Re: [RFC/RFT] p54: Fix for big-endian architecture From: Pavel Roskin To: Johannes Berg Cc: Christian Lamparter , Larry Finger , wireless In-Reply-To: <1223941001.28019.3.camel@dv> References: <48EE20FB.4060700@lwfinger.net> <1223936680.10113.7.camel@johannes.berg> <1223937392.27118.2.camel@dv> <200810140055.34065.chunkeey@web.de> <1223938526.27118.6.camel@dv> <1223939151.27769.3.camel@dv> <1223939271.10113.10.camel@johannes.berg> <1223941001.28019.3.camel@dv> Content-Type: text/plain Date: Mon, 13 Oct 2008 20:00:39 -0400 Message-Id: <1223942439.28019.9.camel@dv> (sfid-20081014_020109_556603_EBF590B3) Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, 2008-10-13 at 19:36 -0400, Pavel Roskin wrote: > p54_config() calls p54_set_freq(), which can fail with -EINVAL. I'm > looking why it can happen. The calibration table only has frequencies in the 2.4GHz band, but the scan goes through the 5GHz band as well. That's how it fails: freq = 5825 i = 0, priv->iq_autocal[i].freq = 2412 i = 1, priv->iq_autocal[i].freq = 2417 i = 2, priv->iq_autocal[i].freq = 2422 i = 3, priv->iq_autocal[i].freq = 2427 i = 4, priv->iq_autocal[i].freq = 2432 i = 5, priv->iq_autocal[i].freq = 2437 i = 6, priv->iq_autocal[i].freq = 2442 i = 7, priv->iq_autocal[i].freq = 2447 i = 8, priv->iq_autocal[i].freq = 2452 i = 9, priv->iq_autocal[i].freq = 2457 i = 10, priv->iq_autocal[i].freq = 2462 i = 11, priv->iq_autocal[i].freq = 2467 i = 12, priv->iq_autocal[i].freq = 2472 i = 13, priv->iq_autocal[i].freq = 2484 phy9: frequency change failed ret = -22 Either we should be using a different table for the 5GHz band, or the driver should not announce 5GHz channels if the calibration table lacks them. -- Regards, Pavel Roskin