Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752816AbbHUUoZ (ORCPT ); Fri, 21 Aug 2015 16:44:25 -0400 Received: from mail-io0-f182.google.com ([209.85.223.182]:33681 "EHLO mail-io0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752199AbbHUUoX (ORCPT ); Fri, 21 Aug 2015 16:44:23 -0400 MIME-Version: 1.0 In-Reply-To: <197171440188481@webcorp01e.yandex-team.ru> References: <1440177121-12741-1-git-send-email-klamm@yandex-team.ru> <197171440188481@webcorp01e.yandex-team.ru> Date: Fri, 21 Aug 2015 13:44:23 -0700 X-Google-Sender-Auth: irNMMttYUmYr9kX7zUTgM7F1RP4 Message-ID: Subject: Re: [PATCH] mm: use only per-device readahead limit From: Linus Torvalds To: Roman Gushchin Cc: linux-mm , Linux Kernel Mailing List , Raghavendra K T , Jan Kara , Wu Fengguang , David Rientjes , Andrew Morton Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1276 Lines: 28 On Fri, Aug 21, 2015 at 1:21 PM, Roman Gushchin wrote: > > It's just a raid driver. For instance, drivers/md/raid5.c:6898 . Ok. That makes me a bit less nervous. I was worried there was some admin program out there that just ups the readahead on peoples devices, which would mean that ra_pages is some random value chosen by crazy user space people. > So, I like an idea to delegate the readahead limit calculation to the underlying i/o level. Yeah, I'm not against it either. It's just that historically we've had some issues with people over-doing readahead (because it often helps some made-up microbenchmark), and then we end up with latency issues when somebody does a multi-gigabyte readahead... Iirc, we had exactly that problem with the readahead() system call at some point (long ago). But if it's just the default ra_pages, then that should be ok. I think the kernel defaults are generally sane, and I hope there isn't some crazy distro that ends up mucking with this. 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/