From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from s3.sipsolutions.net ([5.9.151.49]:52284 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S942895AbcJZNMI (ORCPT ); Wed, 26 Oct 2016 09:12:08 -0400 Message-ID: <1477487505.4059.49.camel@sipsolutions.net> (sfid-20161026_151230_149504_21512E0B) Subject: Re: [PATCH][RFC] nl80211/mac80211: Rounded RSSI reporting From: Johannes Berg To: Denis Kenzior , Andrew Zaborowski , linux-wireless@vger.kernel.org Date: Wed, 26 Oct 2016 15:11:45 +0200 In-Reply-To: <580A2438.7030106@gmail.com> (sfid-20161021_163810_476016_D9B31C64) References: <1477010947-6207-1-git-send-email-andrew.zaborowski@intel.com> <1477038639.4068.28.camel@sipsolutions.net> <580A2438.7030106@gmail.com> (sfid-20161021_163810_476016_D9B31C64) Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Fri, 2016-10-21 at 09:20 -0500, Denis Kenzior wrote: > > It's actually not clear to me that this is really how it should be. > > There's a point to be made that taking a more holistic "link > > quality" would be a better choice. That's related, but maybe can be > > a separate discussion. > Can you elaborate on this 'link quality' idea? Well, I didn't really want to - getting 3 system folks into a room will result in 4 different ways of doing it - but you can take into account not just the RSSI, but also the bitrate you can reasonably use on the channel/with the AP, the noise you can perhaps detect (if you can), the amount of packet loss or retransmissions you experience, etc. I think that some systems (Android, maybe Windows) already do something more complex than pure RSSI indicators, but I don't really know for sure. > > Yes, this would be ideal. > >  > > [...] see my other email > This sounds really brittle.  Furthermore, we also need a facility to  > know when signal strength is getting low to trigger roaming > logic.  This would mean sharing CQM facility between roaming & signal > strength notifications.  As you wrote above, things become quite > impractical. This would likely go through the supplicant anyway, so it could manage proper range overlaps etc. for this. It does seem brittle if we just have a single value, but if we add low/high thresholds (with hysteresis) then I think we can do this, and gain more flexibility in the process. But let's discuss more details over in the other email I just sent :) johannes