Return-path: Received: from mx2.redhat.com ([66.187.237.31]:42322 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751396AbZERSf4 (ORCPT ); Mon, 18 May 2009 14:35:56 -0400 Subject: Re: iwlagn broken in next-20090514 and next-20090515 From: Dan Williams To: Johannes Berg Cc: reinette chatre , Stephen Rothwell , Nico -telmich- Schottelius , "linux-next@vger.kernel.org" , LKML , "John W. Linville" , "linux-wireless@vger.kernel.org" In-Reply-To: <1242670211.29049.10.camel@johannes.local> References: <20090515153733.4384db9f.sfr@canb.auug.org.au> <20090515102621.GA5735@ikn.schottelius.org> <20090518171505.d30d2f64.sfr@canb.auug.org.au> <1242662635.32358.523.camel@rc-desk> <1242663867.27425.1.camel@johannes.local> <1242669029.3845.0.camel@localhost.localdomain> <1242670211.29049.10.camel@johannes.local> Content-Type: text/plain Date: Mon, 18 May 2009 14:35:20 -0400 Message-Id: <1242671720.3845.4.camel@localhost.localdomain> Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, 2009-05-18 at 20:10 +0200, Johannes Berg wrote: > On Mon, 2009-05-18 at 13:50 -0400, Dan Williams wrote: > > > > > > > Also wpa_supplicant is not happy: > > > > > > > > > > > > + wpa_supplicant -B -Dwext -iwlan0 -c /home/user/nico/ethz/wlan/wpa_supplicant.conf > > > > > > ioctl[SIOCSIWENCODEEXT]: No such file or directory > > > > > > ioctl[SIOCSIWENCODEEXT]: No such file or directory > > > > > > ioctl[SIOCSIWENCODEEXT]: No such file or directory > > > > > > ioctl[SIOCSIWENCODEEXT]: No such file or directory > > > > > > > > > > > > > > There were some changes recently to this ioctl, but I am not familiar > > > > with details. Johannes will know. > > > > > > Yeah, I'm not sure -- this happens when clearing keys. I'll look into > > > it, we probably need to accept removing a key that doesn't exist. > > > > Why bother (if that's actually the root cause)? Maybe the supplicant > > should be fixed. Other drivers return ENOENT in this case, so mac80211 > > certainly isn't out-of-common-spec here now. > > I just looked at the code in wpa_supplicant, it doesn't care a bit, just > prints the error. Problem is users attribute any other breakage to the > error it prints since our behaviour changed... I meant "make wpa_supplicant not print anything out on -ENOENT". Dan