Return-path: Received: from smtp06.msg.oleane.net ([62.161.4.6]:57860 "EHLO smtp06.msg.oleane.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752517Ab3F0IJO (ORCPT ); Thu, 27 Jun 2013 04:09:14 -0400 From: "Jean-Pierre Tosoni" To: "'Felix Fietkau'" Cc: References: <1372267723-25072-1-git-send-email-jp.tosoni@acksys.fr> <51CB33D9.1070104@openwrt.org> In-Reply-To: <51CB33D9.1070104@openwrt.org> Subject: RE: [RFC] mac80211: Use libnl-configurable values for retry counts Date: Thu, 27 Jun 2013 10:09:09 +0200 Message-ID: <003001ce730d$9a32e980$ce98bc80$@acksys.fr> (sfid-20130627_100923_688815_3F415100) MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: linux-wireless-owner@vger.kernel.org List-ID: First, thanks for your prompt reply and for pointing out my mistakes. > > Use wiphy->retry_short for management frames. > > Use wiphy->retry_long for other frames. > That seems a bit arbitrary. Indeed. Should I use the following? retry_long for data frames subject to RTS retry_short for all other frames. That looks closer to the descriptions in 802.11-2012. > > + mp->max_retry = hw->conf.long_frame_max_tx_count; > minstrel_alloc is called at rate control init time. You're not making > max_retry configurable here, it'll pick the default value of > hw->conf.long_frame_max_tx_count and never update it again. Is it OK if I replace max_retry with the config values in minstrel_rate_init and minstrel_update_rates instead? Then we can get rid of max_retry completely? (Sorry, I got mixed in the never-ending sequence of callbacks :) > > diff --git a/net/mac80211/rc80211_minstrel.c > b/net/mac80211/rc80211_minstrel.c > > index ac7ef54..502d0c9 100644 > > --- a/net/mac80211/rc80211_minstrel.c > > +++ b/net/mac80211/rc80211_minstrel.c > > @@ -592,11 +592,7 @@ minstrel_alloc( > > /* maximum time that the hw is allowed to stay in one MRR segment */ > > mp->segment_size = 6000; > > > > - if (hw->max_rate_tries > 0) > > - mp->max_retry = hw->max_rate_tries; > > - else > > - /* safe default, does not necessarily have to match hw properties */ > > - mp->max_retry = 7; > > + mp->max_retry = hw->conf.long_frame_max_tx_count;> > - Felix Jean-Pierre