From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from he.sipsolutions.net ([78.46.109.217]:38275 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750709Ab1I2OG1 (ORCPT ); Thu, 29 Sep 2011 10:06:27 -0400 Message-Id: <20110929140500.078257800@sipsolutions.net> (sfid-20110929_160631_166416_5ECED6DF) Date: Thu, 29 Sep 2011 16:04:31 +0200 From: Johannes Berg To: John Linville Cc: linux-wireless@vger.kernel.org Subject: [PATCH v2 06/16] mac80211: clear more-data bit on filtered frames References: <20110929140425.390322480@sipsolutions.net> Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: From: Johannes Berg It doesn't seem likely, but maybe possible, that the more-data bit needs to be recomputed due to changes in the queued frames. Clear it for filtered frames to ensure that we never send it incorrectly. It'll be set again as necessary when we retransmit this frame. The more likely case is maybe where the station woke up after the filtered frame in which case more-data should be clear when the frame is transmitted to the station since it is now awake. Signed-off-by: Johannes Berg --- net/mac80211/status.c | 10 ++++++++++ 1 file changed, 10 insertions(+) --- a/net/mac80211/status.c 2011-09-28 19:10:13.000000000 +0200 +++ b/net/mac80211/status.c 2011-09-28 19:10:15.000000000 +0200 @@ -65,6 +65,16 @@ static void ieee80211_handle_filtered_fr sta->tx_filtered_count++; + /* + * Clear more-data bit on filtered frames, it might be set + * but later frames might time out so it might have to be + * clear again ... It's all rather unlikely (this frame + * should time out first, right?) but let's not confuse + * peers unnecessarily. + */ + if (hdr->frame_control & cpu_to_le16(IEEE80211_FCTL_MOREDATA)) + hdr->frame_control &= ~cpu_to_le16(IEEE80211_FCTL_MOREDATA); + if (ieee80211_is_data_qos(hdr->frame_control)) { int tid = *ieee80211_get_qos_ctl(hdr) & IEEE80211_QOS_CTL_TID_MASK;