Return-path: Received: from nbd.name ([46.4.11.11]:43578 "EHLO nbd.name" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750911Ab3DOO1Y (ORCPT ); Mon, 15 Apr 2013 10:27:24 -0400 Message-ID: <516C0E40.3020502@openwrt.org> (sfid-20130415_162727_416971_E76F39A5) Date: Mon, 15 Apr 2013 16:27:12 +0200 From: Felix Fietkau MIME-Version: 1.0 To: Johannes Berg CC: linux-wireless@vger.kernel.org Subject: Re: [PATCH 3/5] mac80211: improve the rate control API References: <1365890626-86895-1-git-send-email-nbd@openwrt.org> <1365890626-86895-2-git-send-email-nbd@openwrt.org> <1365890626-86895-3-git-send-email-nbd@openwrt.org> <1366033473.8361.26.camel@jlt4.sipsolutions.net> In-Reply-To: <1366033473.8361.26.camel@jlt4.sipsolutions.net> Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On 2013-04-15 3:44 PM, Johannes Berg wrote: > On Sun, 2013-04-14 at 00:03 +0200, Felix Fietkau wrote: > >> +EXPORT_SYMBOL(rate_control_set_rates); > > I don't think that needs to be exported (at least not now)? Just thought I'd let driver-implemented rate control modules access it as well if they want to. > I suspect that the answer to my question in patch 2 lies here, if the > rate control updates the table then would it not just update the table > when the protection mechanism needs to be changed? There should be > triggers for that, rather than having to check the "use_cts" variables > on every packet, and if they aren't complete that should be fixed? I guess we could do some further work on that, but for now I wanted to keep things simple by having only the rate control module touch the table (and not have it care about CTS protection). - Felix