From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH 0/8 v3] KVM support for 1GB pages Date: Mon, 22 Jun 2009 12:55:05 +0300 Message-ID: <4A3F54F9.9060506@redhat.com> References: <1245417389-5527-1-git-send-email-joerg.roedel@amd.com> <4A3CC1FA.2070900@redhat.com> <20090622094013.GD5139@amd.com> <4A3F524C.8030902@redhat.com> <20090622094953.GE5139@amd.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , kvm@vger.kernel.org, linux-kernel@vger.kernel.org To: Joerg Roedel Return-path: Received: from mx2.redhat.com ([66.187.237.31]:43174 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750859AbZFVJyR (ORCPT ); Mon, 22 Jun 2009 05:54:17 -0400 In-Reply-To: <20090622094953.GE5139@amd.com> Sender: kvm-owner@vger.kernel.org List-ID: On 06/22/2009 12:49 PM, Joerg Roedel wrote: >>> After the rework it even fails to boot a guest with gbpages enabled. This week >>> is a bit tight but I will debug this further next week. >>> >>> >> Does the failure happen only when both guest/host use gbpages, or does >> it also happen if gbpages is disabled on either one? >> > > It happens only when the guest uses gbpages. If I disable them at boot the > linux guest boots fine and runs stable. > This seems to point the finger at the write protection logic, either we create a gbpage even though there are shadowed page tables in its area, or (more likely) we don't break up a gbpage when we write protect a page. -- error compiling committee.c: too many arguments to function