From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: regression between 0.12.1.2 and 0.12.2 Date: Tue, 26 Jan 2010 17:21:06 +0100 Message-ID: <4B5F1672.1040802@siemens.com> References: <201001210949.29185.toralf.foerster@gmx.de> <4B5A0C70.6020603@web.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: qemu-devel@nongnu.org, kvm , Avi Kivity To: =?ISO-8859-1?Q?Toralf_F=F6rster?= Return-path: Received: from goliath.siemens.de ([192.35.17.28]:22124 "EHLO goliath.siemens.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754660Ab0AZQV3 (ORCPT ); Tue, 26 Jan 2010 11:21:29 -0500 In-Reply-To: <4B5A0C70.6020603@web.de> Sender: kvm-owner@vger.kernel.org List-ID: Jan Kiszka wrote: > Toralf F=F6rster wrote: >> Hi, >> >> under a mostly stable Gentoo I observed this new msg : >> >> tfoerste@n22 ~/virtual/kvm $ qemu -hda gentoo_kdevm.img -hdb portage= _kdeprefix.img -hdd swap.img -smp 2 -m 768 -vga std -soundhw es1370 = = =20 >> BUG: kvm_dirty_pages_log_enable_slot: invalid parameters = =20 >> BUG: kvm_dirty_pages_log_disable_slot: invalid parameters = =20 >> .. >> >> The kvm image can be derived from http://dev.gentooexperimental.org/= ~wired/kvm/ . >> >> My system is a : >> tfoerste@n22 ~/virtual/kvm $ uname -a >> Linux n22 2.6.32.4 #1 SMP Mon Jan 18 20:20:38 CET 2010 i686 Intel(R)= Core(TM)2 Duo CPU P8600 @ 2.40GHz GenuineIntel GNU/Linux >> >> >=20 > That's a pre-0.12.1.2 qemu-kvm issue, upstream is not affected - or i= s > at least not reporting it. It's already in my todo queue, just waitin= g > to be dequeued. I've looked into this a bit, and the bug message that pops up is in fac= t new for your scenario (0.12.1.2->0.12.2, -vga std), it just happens to trigger for me as well in a slightly different setup (CONFIG_FB_CIRRUS)= =2E This is "mostly harmless" (the bug is gracefully handled), indicating that qemu-kvm tries to enable/disable dirty logging for a VGA memory area that was just unregistered. And that is because qemu-kvm tries to keep support for old host kernels that had bugs and required workaround approaches, but that code is bit-rotting a bit. Avi, we should get rid of these messages, either by suppressing them in qemu-kvm for now (stick your head into the sand...) or by finally dropping all those dirty-logging diffs to upstream (in theory, there is a third option: fixing the workaround code, but I don't think it's wort= h the effort). What could be a road map for dropping? What distro kernels are you aware of that may become unusable then? Jan --=20 Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux