From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mario Smarduch Subject: Re: [PATCH v4 0/2] arm/arm64: KVM: Optimize arm64 fp/simd, saves 30-50% on exits for non-VHE Date: Thu, 16 Jul 2015 12:22:12 -0700 Message-ID: <55A80464.7000701@samsung.com> References: <1436577547-5646-1-git-send-email-m.smarduch@samsung.com> <20150716155253.GU7845@cbox> <55A7F68C.6020903@samsung.com> <20150716190529.GA14369@cbox> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: kvmarm@lists.cs.columbia.edu, marc.zyngier@arm.com, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org To: Christoffer Dall Return-path: Received: from mailout1.w2.samsung.com ([211.189.100.11]:40963 "EHLO usmailout1.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752344AbbGPTWW (ORCPT ); Thu, 16 Jul 2015 15:22:22 -0400 Received: from uscpsbgex1.samsung.com (u122.gpu85.samsung.co.kr [203.254.195.122]) by mailout1.w2.samsung.com (Oracle Communications Messaging Server 7.0.5.31.0 64bit (built May 5 2014)) with ESMTP id <0NRL009S5HT9F550@mailout1.w2.samsung.com> for kvm@vger.kernel.org; Thu, 16 Jul 2015 15:22:21 -0400 (EDT) In-reply-to: <20150716190529.GA14369@cbox> Sender: kvm-owner@vger.kernel.org List-ID: On 07/16/2015 12:05 PM, Christoffer Dall wrote: > On Thu, Jul 16, 2015 at 11:23:08AM -0700, Mario Smarduch wrote: >> On 07/16/2015 08:52 AM, Christoffer Dall wrote: >>> On Fri, Jul 10, 2015 at 06:19:05PM -0700, Mario Smarduch wrote: >>>> This is a followp to previous iteration but implemented on top of VHE patches. >>>> Only non-VHE path is addressied by this patch. In second patch 32-bit handler >>>> is updated to keep exit handling consistent with 64-bit code, and nothing >>>> has changed. >>>> >>> Why not simply preserve this the way it was in v3 and have it merged >>> first - after all we have reviewed it and I thought it was more or less >>> ready to be merged - I suspect the VHE patches may have a way to go >>> still ? >> >> Definitely, that's a better path. After looking at VHE patches, >> I would probably leave V3 the way it is (keeping deactivate_xxxx: >> symmetric). Marc has Reviewed V3 and you commented either way was >> fine with you, so V3 should be ok. > > Yes, but there was a comment in the assembly file to fix up IIRC. That's right 1/2 and 0/2 header need little bit of editing. > > Can you do a quick respin with that commentary changed and then Marc > can queue that if he agrees? > > Thanks, > -Christoffer >