From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760238AbYFTQLq (ORCPT ); Fri, 20 Jun 2008 12:11:46 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753532AbYFTQLg (ORCPT ); Fri, 20 Jun 2008 12:11:36 -0400 Received: from smtp01.uc3m.es ([163.117.176.131]:58635 "EHLO smtp01.uc3m.es" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753227AbYFTQLg (ORCPT ); Fri, 20 Jun 2008 12:11:36 -0400 From: "Peter T. Breuer" Message-Id: <200806201611.m5KGBDN02249@inv.it.uc3m.es> Subject: Re: [PATCH 1/3] i/o bandwidth controller documentation In-Reply-To: <200806201602.m5KG2Zx32671@inv.it.uc3m.es> from "Peter T. Breuer"at "Jun 20, 2008 06:02:35 pm" To: "linux kernel" Date: Fri, 20 Jun 2008 18:11:13 +0200 (MET DST) X-Anonymously-To: Reply-To: ptb@inv.it.uc3m.es X-message-flag: Had your Outlook virus, today? http://www.counterpane.com/crypto-gram-0103.html#4 X-WebTV-Stationery: Standard\; BGColor=black\; TextColor=black Reply-By: Sat, 1 Apr 2006 14:21:08 -0700 X-Mailer: ELM [version 2.4ME+ PL66 (25)] X-imss-version: 2.051 X-imss-result: Passed X-imss-scanInfo: M:B L:E SM:2 X-imss-tmaseResult: TT:1 TS:-2.8630 TC:1F TRN:23 TV:5.5.1026(15982.007) X-imss-scores: Clean:100.00000 C:0 M:0 S:0 R:0 X-imss-settings: Baseline:1 C:1 M:1 S:1 R:1 (0.0000 0.0000) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > + Block device I/O bandwidth controller How can this work? You will limit the number of available buffer heads per second? Unfortunaely, the problem is the fs above the block device. If the block device is artificially slowed then the fs will still happily allow a process to fill buffers forever until memory is full, while the block device continues to trickle the buffers away. What one wants is for the fs buffering to be linked to the underlying block device i/o speed. One wants the rate at which fs buffers are filled to be no more than (modulu brief spurts) the rate at which the device operates. That way networked block devices have a chance of having some memory left to send the dirty buffers out to the net with. B/w limiting the device itself doesn't seem to me to do any good. Peter