Return-path: Received: from mail-oi0-f52.google.com ([209.85.218.52]:35516 "EHLO mail-oi0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752663AbbFFSZy (ORCPT ); Sat, 6 Jun 2015 14:25:54 -0400 Received: by oihd6 with SMTP id d6so72220562oih.2 for ; Sat, 06 Jun 2015 11:25:54 -0700 (PDT) Message-ID: <55733B2F.9090807@lwfinger.net> (sfid-20150606_202558_867813_DE778C5C) Date: Sat, 06 Jun 2015 13:25:51 -0500 From: Larry Finger MIME-Version: 1.0 To: Kalle Valo , Severn , Taehee Yoo CC: linux-wireless@vger.kernel.org Subject: Re: rtlwifi NULL pointer dereference References: <5570FB00.3040108@xephris.net> <87fv65nfcn.fsf@kamboji.qca.qualcomm.com> In-Reply-To: <87fv65nfcn.fsf@kamboji.qca.qualcomm.com> Content-Type: text/plain; charset=windows-1252; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: On 06/06/2015 08:20 AM, Kalle Valo wrote: > Severn writes: > >> I've been hitting this NULL pointer deref with an rtl8188ee PCIE card. >> It happens 100% of the time when I bring up a WPA2-PSK AP with hostapd >> 2.4. >> >> If I revert 33511b157bbcebaef853cc1811992b664a2e5862, everything seems >> to work properly. >> >> Below is my hostapd.conf and the crash log. This is mainline 4.0.4 >> with no patches. > > Below is the changelog for the commit. Larry & Taehee, what should we > do? > > commit 33511b157bbcebaef853cc1811992b664a2e5862 > Author: Taehee Yoo > Date: Wed Jan 21 16:58:19 2015 +0900 > > rtlwifi: add support to send beacon frame. > > In AP mode, beacon frame is necessary to keep connection. > this patch adds a sending beacon frame routine in initialization routine. > > Signed-off-by: Taehee Yoo > Acked-by: Larry Finger > Signed-off-by: Kalle Valo Kalle, I'm trying to find the missing setup step when the new interface is added to the driver. Whatever it is is causing mac80211 to issue the warning at net/mac80211/driver-ops.h:12. That problem appears to cascade to the NULL pointer dereference. Any help would be appreciated. Reversion of the above commit may be necessary, but I would like to see if we can get a better "fix" as that commit is necessary for proper beaconing. Larry