Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756347Ab0HDSDH (ORCPT ); Wed, 4 Aug 2010 14:03:07 -0400 Received: from rcsinet10.oracle.com ([148.87.113.121]:61285 "EHLO rcsinet10.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752423Ab0HDSDF (ORCPT ); Wed, 4 Aug 2010 14:03:05 -0400 To: Mark Lord Cc: Nigel Cunningham , LKML , pm list Subject: Re: 2.6.35 Regression: Ages spent discarding blocks that weren't used! From: "Martin K. Petersen" Organization: Oracle References: <4C58C528.4000606@tuxonice.net> <4C5960B0.7020003@teksavvy.com> Date: Wed, 04 Aug 2010 14:02:10 -0400 In-Reply-To: <4C5960B0.7020003@teksavvy.com> (Mark Lord's message of "Wed, 04 Aug 2010 08:44:32 -0400") Message-ID: User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.2 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Source-IP: acsmt354.oracle.com [141.146.40.154] X-Auth-Type: Internal IP X-CT-RefId: str=0001.0A090203.4C59AB2F.0055:SCFMA4539814,ss=1,fgs=0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 921 Lines: 23 >>>>> "Mark" == Mark Lord writes: Mark> Looks to me like more and more things are using the block discard Mark> functionality, and as predicted it is slowing things down Mark> enormously. Mark> The problem is that we still only discard tiny bits (a single Mark> range still??) per TRIM command, rather than batching larger Mark> ranges and larger numbers of ranges into single TRIM commands. Mark> That's a very poor implementation, especially when things start Mark> enabling it by default. Eg. the swap code, mke2fs, etc.. I'm working on aggregation. But it's harder than we initially thought... -- Martin K. Petersen Oracle Linux Engineering -- 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/