From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sheng Yang Subject: Re: [PATCH 3/3] KVM: x86: Use FPU API Date: Mon, 17 May 2010 16:22:13 +0800 Message-ID: <201005171622.14483.sheng@linux.intel.com> References: <4BEE561E.5020607@redhat.com> <1274083741-25856-1-git-send-email-sheng@linux.intel.com> <4BF0FBCE.5010002@redhat.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , kvm@vger.kernel.org, "H. Peter Anvin" , Ingo Molnar To: Avi Kivity Return-path: Received: from mga06.intel.com ([134.134.136.21]:12271 "EHLO orsmga101.jf.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752677Ab0EQIY3 (ORCPT ); Mon, 17 May 2010 04:24:29 -0400 In-Reply-To: <4BF0FBCE.5010002@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Monday 17 May 2010 16:18:22 Avi Kivity wrote: > On 05/17/2010 11:09 AM, Sheng Yang wrote: > > Convert KVM to use generic FPU API. > > > > Signed-off-by: Sheng Yang > > --- > > Like this? (Drop patch 1) > > Will be more readable with a patch that converts host_fx_image to > unlazy_fpu(), and a second patch that converts guest_fx_image to the fpu > API. OK. > I think unlazy_fpu() is even a performance win in case userspace doesn't > do a lot of floating point (which is the case with qemu). I wonder why > we didn't think of it before. ... -- regards Yang, Sheng