Return-path: Received: from mail-bk0-f46.google.com ([209.85.214.46]:34255 "EHLO mail-bk0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751520AbaAPLlh convert rfc822-to-8bit (ORCPT ); Thu, 16 Jan 2014 06:41:37 -0500 Received: by mail-bk0-f46.google.com with SMTP id u15so1117119bkz.5 for ; Thu, 16 Jan 2014 03:41:33 -0800 (PST) MIME-Version: 1.0 In-Reply-To: References: <1389787494-7361-1-git-send-email-michal.kazior@tieto.com> <1389787494-7361-10-git-send-email-michal.kazior@tieto.com> Date: Thu, 16 Jan 2014 12:41:33 +0100 Message-ID: (sfid-20140116_124140_652455_029473B0) Subject: Re: [RFC 9/9] mac80211: implement multi-interface CSA From: Michal Kazior To: Eliad Peller Cc: "linux-wireless@vger.kernel.org" , Johannes Berg Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On 16 January 2014 12:04, Eliad Peller wrote: > On Wed, Jan 15, 2014 at 2:04 PM, Michal Kazior wrote: >> This implements a fairly simple multi-interface >> CSA. It doesn't support multiple channel >> contexts so it doesn't support multi-channel. >> >> Once a CSA is started other CSA requests are >> denied until the first one is completed. A single >> CSA may affect multiple interfaces. CSA can happen >> only if it all target CSA chandefs are compatible >> and it affects all interfaces are sharing a single >> channel context exclusively. >> >> A new worker is introduced: csa_complete_work >> which is used to account per-interface countdowns >> and issue the actual channel switch after last >> interface completes its CSA. >> >> Signed-off-by: Michal Kazior >> --- > [...] > >> >> +static int ieee80211_ap_beacon_presp_backup(struct ieee80211_sub_if_data *sdata) >> +{ >> + struct beacon_data *beacon; >> + struct probe_resp *probe_resp; >> + >> + beacon = sdata_dereference(sdata->u.ap.beacon, sdata); >> + if (beacon) { >> + sdata->u.ap.prev_beacon = kmemdup(beacon, sizeof(beacon) + >> + beacon->head_len + >> + beacon->tail_len, GFP_KERNEL); >> + if (!sdata->u.ap.prev_beacon) >> + return -ENOMEM; >> + } >> + > you must also update the pointers here... You're right, thanks! Michał