Return-path: Received: from smtp4-g19.free.fr ([212.27.42.30]:41987 "EHLO smtp4-g19.free.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750766AbYJNTUQ (ORCPT ); Tue, 14 Oct 2008 15:20:16 -0400 Message-ID: <48F4F0E9.8070706@free.fr> (sfid-20081014_212021_656995_0D69A404) Date: Tue, 14 Oct 2008 21:20:09 +0200 From: Benoit PAPILLAULT MIME-Version: 1.0 To: Johannes Berg CC: Jouni Malinen , John Linville , linux-wireless@vger.kernel.org Subject: Re: [PATCH] mac80211: Fix scan RX processing oops References: <20081011002955.GA15802@jm.kir.nu> <1223685486.29811.17.camel@johannes.berg> <20081011004828.GB15802@jm.kir.nu> <1223686439.29811.27.camel@johannes.berg> In-Reply-To: <1223686439.29811.27.camel@johannes.berg> Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: Johannes Berg a =E9crit : > On Sat, 2008-10-11 at 03:48 +0300, Jouni Malinen wrote: >> On Sat, Oct 11, 2008 at 02:38:06AM +0200, Johannes Berg wrote: >>>> - ieee80211_rx_bss_put(sdata->local, bss); >>>> + if (bss) >>>> + ieee80211_rx_bss_put(sdata->local, bss); >>> I keep falling into that trap, maybe the put function should just h= andle >>> NULL instead... >> I though about that for half a second or so ;-) and ended up doing t= his >> instead after checking that other ieee80211_rx_bss_put() calls were = only >> passing in non-NULL values. Anyway, I would be fine with _put() bein= g >> able to handle NULL, too. >=20 > I looked at it before and for some reason decided against it too. Let= 's > stick to this patch, I'll re-evaluate making it handle NULL. >=20 >> PS. >> >> I don't know what exactly was triggering this oops (or well, what wa= s >> triggering ieee80211_bss_info_update() to return NULL to be more exa= ct), >> but it was happening very consistently in our office (but not anywhe= re >> else I've been this week). >=20 > Strange. You probably have a mesh network with bogus mesh config or m= esh > ID IEs, I can't see the allocation fail consistently in your office ;= ) >=20 >> It was kind of funny to see that oops at the >> very moment when I was convincing people in a meeting that we can ch= ange >> mac80211 and should do so if it is the best location for something a= nd >> makes it easier to implement something in a driver.. ;-) >=20 > Ouch, sorry :) >=20 > johannes Acked-by: Benoit Papillault BTW, This is indeed Mesh Config IE received with incorrect mesh_id_len (8 bytes instead of 19 bytes as expected). I really think the current patch is the correct one. Regards, Benoit -- To unsubscribe from this list: send the line "unsubscribe linux-wireles= s" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html