Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757658AbXHUG0V (ORCPT ); Tue, 21 Aug 2007 02:26:21 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751575AbXHUG0L (ORCPT ); Tue, 21 Aug 2007 02:26:11 -0400 Received: from smtp2.linux-foundation.org ([207.189.120.14]:32864 "EHLO smtp2.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752539AbXHUG0J (ORCPT ); Tue, 21 Aug 2007 02:26:09 -0400 Date: Mon, 20 Aug 2007 23:25:13 -0700 (PDT) From: Linus Torvalds To: Arjan van de Ven cc: David Brownell , Michal Piotrowski , linux-usb-devel@lists.sourceforge.net, Greg KH , LKML , "Stuart_Hayes@Dell.com" , Andrew Morton , Daniel Exner Subject: Re: [linux-usb-devel] [4/4] 2.6.23-rc3: known regressions In-Reply-To: <1187675488.2676.3.camel@laptopd505.fenrus.org> Message-ID: References: <46C098FD.1030601@googlemail.com> <200708202102.58508.david-b@pacbell.net> <200708202148.45575.david-b@pacbell.net> <1187675488.2676.3.camel@laptopd505.fenrus.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1580 Lines: 39 On Mon, 20 Aug 2007, Arjan van de Ven wrote: > > > > So it might be much better if we instead re-introduced that kind of "DMA > > latency requirement", and letting different subsystems react to that as > > they may. > > wait.... we HAVE that infrastructure .. see kernel/latency.c ... Heh. Just shows how wellknown that interface is - it seems like it's only used by the ipw2100 driver and "pcm_native". But yes, that looks like the right thing. > and the C-state code will honor it. CPUFREQ doesn't honor it yet but > that's easy to add.. (this assumes the ACPI BIOS informs us correctly > about the cpu behavior, but that's the best we can do obviously unless > you want a table inside the kernel keyed off vendor/model/stepping) Do we actually have the latency information for these things? Especially since I assume a number of people use the specialized direct-hw-access cpufreq drivers.. I realize that we *have* "transition_latency" at the cpufreq layer, and it is supposed to be in ns, but I wonder how likely it is to bear any relationship to reality, considering that I don't think it's really used for anything.. (yeah, it affects the heuristics, but I don't think it has any _hard_ meaning, so I'd worry that it's not necessarily something that people have tried to make accurate). But I dunno. Linus - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/