From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: KVM: MMU: rmap_write_protect() hugepage iteration bug Date: Sun, 08 Jun 2008 10:30:45 +0300 Message-ID: <484B8AA5.4040202@qumranet.com> References: <20080608002736.GA25582@dmt.cnet> <20080608015452.GC8321@duo.random> <20080608044853.GA1408@dmt.cnet> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Andrea Arcangeli , kvm-devel To: Marcelo Tosatti Return-path: Received: from il.qumranet.com ([212.179.150.194]:50017 "EHLO il.qumranet.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753673AbYFHHaq (ORCPT ); Sun, 8 Jun 2008 03:30:46 -0400 In-Reply-To: <20080608044853.GA1408@dmt.cnet> Sender: kvm-owner@vger.kernel.org List-ID: Marcelo Tosatti wrote: > How bad you think restarting is? Unless there are workloads with very > large hugepage chains most of the data should be cached in L1. Also the > fact that a 4K page has just been shadowed inside a large region makes > it unlikely for the same chain to be examined again soon. > Non-pae guests will have a chain entry for every mm_struct, since the large ptes reside in the pgd and that is replicated for every mm_struct. So it's easy to have very long chains. -- error compiling committee.c: too many arguments to function