Return-path: Received: from purkki.adurom.net ([80.68.90.206]:37334 "EHLO purkki.valot.fi" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751059Ab0DMJHM (ORCPT ); Tue, 13 Apr 2010 05:07:12 -0400 To: Juuso Oikarinen Cc: ext Johannes Berg , "linux-wireless\@vger.kernel.org" , "Ylalehto Janne \(Nokia-D\/Tampere\)" Subject: Re: [RFC PATCH] nl80211: Add support for dynamic ps timeout configuration References: <1271059902-27788-1-git-send-email-juuso.oikarinen@nokia.com> <1271061120.3877.6.camel@jlt3.sipsolutions.net> <1271062589.10120.1169.camel@wimaxnb.nmp.nokia.com> <1271063175.3877.13.camel@jlt3.sipsolutions.net> <1271065535.10120.1191.camel@wimaxnb.nmp.nokia.com> <871vekpexo.fsf@purkki.valot.fi> <1271143746.10120.1232.camel@wimaxnb.nmp.nokia.com> <87wrwboirr.fsf@purkki.valot.fi> <1271148573.10120.1257.camel@wimaxnb.nmp.nokia.com> From: Kalle Valo Date: Tue, 13 Apr 2010 12:07:10 +0300 In-Reply-To: <1271148573.10120.1257.camel@wimaxnb.nmp.nokia.com> (Juuso Oikarinen's message of "Tue\, 13 Apr 2010 11\:49\:33 +0300") Message-ID: <87sk6zoh4x.fsf@purkki.valot.fi> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-wireless-owner@vger.kernel.org List-ID: Juuso Oikarinen writes: >> > The problem here is that the latency at least cannot be used in any >> > simple and/or general way I can think of to control the dynamic PS >> > timeout. >> >> Can't you do something like this: >> >> pm_qos >= 1000 -> timeout 0 ms (aka. full power save) >> pm_qos <= 100 -> timeout 100 ms >> pm_qos <= 50 -> timeout 300 ms >> >> That is, just some arbitrary numbers but which affect dynamic ps >> timeout. I haven't thought about the numbers at all, but they actually >> don't matter because it's easy to change them inside mac80211. > > Theoretically I could, but I'm pretty sure whatever values I would > choose would be unacceptable by others, as they would be tuned for a > specific use. I don't see a problem with that. The current use of pm_qos is very limited, I think it's all positive if we start using it more. > Also, although AFAIK barely anyone uses the DTIM interval which is > determined based on th pm_qos, adding arbitrary rules like this to the > side risks breaking something for someone. AFAIK the pm_qos values are not set in stone in any. Applications just request something and kernel can do whatever it wants, even ignore it. So I don't see any harm if we change how mac80211 uses pm_qos values. And most probably this will change many times in the future. -- Kalle Valo