From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Holger Hoffstaette" Subject: 3.7: massive slowdown of deletion performance with discard Date: Sun, 16 Dec 2012 16:32:11 +0100 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit To: linux-fsdevel@vger.kernel.org Return-path: Received: from plane.gmane.org ([80.91.229.3]:44377 "EHLO plane.gmane.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750981Ab2LPQOy (ORCPT ); Sun, 16 Dec 2012 11:14:54 -0500 Received: from list by plane.gmane.org with local (Exim 4.69) (envelope-from ) id 1TkGro-0006xv-6y for linux-fsdevel@vger.kernel.org; Sun, 16 Dec 2012 17:15:04 +0100 Received: from p50869c9e.dip.t-dialin.net ([80.134.156.158]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 16 Dec 2012 17:15:04 +0100 Received: from holger.hoffstaette by p50869c9e.dip.t-dialin.net with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 16 Dec 2012 17:15:04 +0100 Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Hi, I just tested 3.7-rc1 and found that deleting a kernel tree on SSDs now takes approximately way too long to wait for (as in minutes), wherereas previously this took <1s. As soon as I remount with nodiscard the performance becomes tolerable again. It alss happens on two different SSDs (old Intel G2 & Corsair Force GT), both of which have otherwise excellent performance and show no noticeable impact of discard with kernels 3.6.x. Filesystem is ext4, but after reading various kernel changelog summaries it seems to be caused by multiple changes around discard behaviour in general. Is anyone else noticing this? I find it hard to believe that such a massive slowdown would go unnoticed.. thanks Holger