From: bugzilla-daemon@bugzilla.kernel.org Subject: [Bug 15906] performance regression in "umount" of filesystems using barriers Date: Fri, 21 May 2010 06:08:15 GMT Message-ID: <201005210608.o4L68Frk026105@demeter.kernel.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" To: linux-ext4@vger.kernel.org Return-path: Received: from demeter.kernel.org ([140.211.167.39]:39012 "EHLO demeter.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932270Ab0EUGIQ (ORCPT ); Fri, 21 May 2010 02:08:16 -0400 Received: from demeter.kernel.org (localhost.localdomain [127.0.0.1]) by demeter.kernel.org (8.14.3/8.14.3) with ESMTP id o4L68FcV026106 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Fri, 21 May 2010 06:08:15 GMT In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: https://bugzilla.kernel.org/show_bug.cgi?id=15906 --- Comment #28 from Jens Axboe 2010-05-21 06:08:09 --- Eric is spot on, that latter commit does indeed fix up that issue. The problem is that the caller holds the umount sem when initiating the WB_SYNC_NONE writeback, and thus it passes down ->sb_pinned == 1. But since we clear the work state for WB_SYNC_NONE early, the caller can then drop the umount sem before we are actually done writing out inodes. So the fix is queued up with the original patch, when it goes upstream there should be no warnings. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching the assignee of the bug.