From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1161043AbXCGAmO (ORCPT ); Tue, 6 Mar 2007 19:42:14 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1161048AbXCGAmO (ORCPT ); Tue, 6 Mar 2007 19:42:14 -0500 Received: from smtp-outbound-1.vmware.com ([65.113.40.141]:36650 "EHLO smtp-outbound-1.vmware.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1161043AbXCGAmN (ORCPT ); Tue, 6 Mar 2007 19:42:13 -0500 Message-ID: <45EE0A68.6010406@vmware.com> Date: Tue, 06 Mar 2007 16:42:16 -0800 From: Dan Hecht User-Agent: Thunderbird 1.5.0.2 (X11/20060420) MIME-Version: 1.0 To: tglx@linutronix.de Cc: Zachary Amsden , Ingo Molnar , akpm@linux-foundation.org, ak@suse.de, Virtualization Mailing List , Jeremy Fitzhardinge , Rusty Russell , LKML , john stultz , Dan Hecht Subject: Re: + stupid-hack-to-make-mainline-build.patch added to -mm tree References: <200703060654.l266sVxr014860@shell0.pdx.osdl.net> <45ED16D2.3000202@vmware.com> <20070306084258.GA15745@elte.hu> <20070306084647.GA16280@elte.hu> <45ED2C82.3080008@vmware.com> <1173178774.24738.311.camel@localhost.localdomain> <45EDD82F.90204@vmware.com> <1173225182.24738.507.camel@localhost.localdomain> In-Reply-To: <1173225182.24738.507.camel@localhost.localdomain> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 07 Mar 2007 00:42:11.0890 (UTC) FILETIME=[71A98120:01C76051] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On 03/06/2007 03:53 PM, Thomas Gleixner wrote: >> 2) Virtual interrupts have a relatively high overhead as compared with >> native interrupts. So, in vmitime, we wanted to be able to lower the >> timer interrupt rate at runtime, even if HZ is a compile time constant >> (and set to something high, like 1000hz). While we could hack this in >> by using evt->min_delta_ns, it wouldn't really work since process time >> accounting would be wrong. Instead, we should allow the >> tick_sched_timer in cases (c) and (d) to have runtime configurable >> period, and then scale the time value accordingly before passing to >> account_system_time. This is probably something the Xen folks will want >> also, since I think Xen itself only gets 100hz hard timer, and so it can >> implement at best a oneshot virtual timer with 100hz resolution. Any >> objections to us doing something like this? > > Yes. It's gross hackery. > > 1) We want to have a cleanup of the tick assumptions _all_ over the > place and this is going to be real hard work. > > 2) As I said above. The time accounting for virtualization needs to be > fixed in a generic way. > > I'm not going to accept some weird hackery for virtualization, which is > of exactly ZERO value for the kernel itself. Quite the contrary it will > make the cleanup harder and introduce another hard to remove thing, > which will in the worst case last for ever. > Okay, to confirm I'm on the same page as you, you want to move process time accounting from being periodic sampled based to being trace based? i.e. at the system-call/interrupt boundaries, read clocksource and compute directly the amount of system/user/process time? Do you know if anyone has explored this? I thought there was a discussion about this a while back but it was rejected due to the sample-based approach having much lower overheads on high system call rate workloads.