Return-path: Received: from mail-yx0-f174.google.com ([209.85.213.174]:63836 "EHLO mail-yx0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752721Ab1HAJn6 convert rfc822-to-8bit (ORCPT ); Mon, 1 Aug 2011 05:43:58 -0400 Received: by yxi11 with SMTP id 11so3023109yxi.19 for ; Mon, 01 Aug 2011 02:43:58 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <4E367432.60601@openwrt.org> References: <1312191173-23963-1-git-send-email-helmut.schaa@googlemail.com> <1312191173-23963-2-git-send-email-helmut.schaa@googlemail.com> <4E367432.60601@openwrt.org> Date: Mon, 1 Aug 2011 11:43:58 +0200 Message-ID: (sfid-20110801_114402_601358_E95F576A) Subject: Re: [PATCHv2 2/2] mac80211: Don't use EAPOL frames for rate sampling From: Helmut Schaa To: Felix Fietkau Cc: John Linville , linux-wireless@vger.kernel.org, Johannes Berg Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Aug 1, 2011 at 11:38 AM, Felix Fietkau wrote: > On 2011-08-01 11:32 AM, Helmut Schaa wrote: >> >> Signed-off-by: Helmut Schaa >> --- >> >> Haven't done much testing yet but this looks safe to me. >> Helmut >> >> ?net/mac80211/rc80211_minstrel_ht.c | ? ?8 +++++++- >> ?1 files changed, 7 insertions(+), 1 deletions(-) >> >> diff --git a/net/mac80211/rc80211_minstrel_ht.c >> b/net/mac80211/rc80211_minstrel_ht.c >> index 66a1eeb..2158838 100644 >> --- a/net/mac80211/rc80211_minstrel_ht.c >> +++ b/net/mac80211/rc80211_minstrel_ht.c >> @@ -608,7 +608,13 @@ minstrel_ht_get_rate(void *priv, struct ieee80211_sta >> *sta, void *priv_sta, >> ? ? ? ? ? ? ? ?return mac80211_minstrel.get_rate(priv, sta,&msp->legacy, >> txrc); >> >> ? ? ? ?info->flags |= mi->tx_flags; >> - ? ? ? sample_idx = minstrel_get_sample_rate(mp, mi); >> + >> + ? ? ? /* Don't use EAPOL frames for sampling on non-mrr hw */ >> + ? ? ? if (mp->hw->max_rates == 1&& >> + ? ? ? ? ? txrc->skb->protocol == cpu_to_be16(ETH_P_PAE)) > > I think checking for mp->hw->max_rates < 3 would be better here, because > then it also covers the primary rate + fallback rate case (as being used by > b43), which also doesn't include the max_prob_rate in the mrr series. The mp->hw->max_rates == 2 case is sample_rate -> max_prob_rate and max_tp_rate -> max_prob_rate at the moment. So, these devices should be fine already. Helmut