From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xiao Guangrong Subject: [PATCH 0/10] KVM MMU: allow more shadow pages become asynchronous Date: Thu, 22 Apr 2010 14:12:07 +0800 Message-ID: <4BCFE8B7.1040004@cn.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , KVM list , LKML To: Avi Kivity Return-path: Received: from cn.fujitsu.com ([222.73.24.84]:53640 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751306Ab0DVGPO (ORCPT ); Thu, 22 Apr 2010 02:15:14 -0400 Sender: kvm-owner@vger.kernel.org List-ID: In current code, shadow page can become asynchronous only if one shadow page for a gfn, this rule is too strict, in fact, we can let all last mapping page(i.e, it's the pte page) become unsync and sync them at invlpg or flush tlb time. Address this thinking, a gfn may have many shadow pages, for performance reason, when one sp need be synced, just write protect sp->gfn and sync this sp but we keep other shadow pages asynchronous, i.e, sp only can be synced just at their own invlpg and flush TLB time. patch 1 ~ 6 are bugfix/cleanup/optimization for current code patch 7 ~ 10 implement this idea