From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751287AbdBTWDA (ORCPT ); Mon, 20 Feb 2017 17:03:00 -0500 Received: from mx4-phx2.redhat.com ([209.132.183.25]:36084 "EHLO mx4-phx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750919AbdBTWC6 (ORCPT ); Mon, 20 Feb 2017 17:02:58 -0500 Date: Mon, 20 Feb 2017 17:02:53 -0500 (EST) From: Paolo Bonzini To: Andy Lutomirski Cc: kvm list , linux-kernel@vger.kernel.org, X86 ML , "H. Peter Anvin" , Borislav Petkov Message-ID: <890292613.23897064.1487628173163.JavaMail.zimbra@redhat.com> In-Reply-To: References: <38d613dc-5b46-1f4f-04d1-53c01932e6d6@redhat.com> Subject: Re: RFC: Getting rid of LTR in VMX MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Originating-IP: [10.4.164.1, 10.5.100.50] X-Mailer: Zimbra 8.0.6_GA_5922 (ZimbraWebClient - FF51 (Linux)/8.0.6_GA_5922) Thread-Topic: Getting rid of LTR in VMX Thread-Index: 0b7gueMvOy/lLU3lUBtk/R5HhP6L7Q== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org > > Yes. But 150-200 clock cycles are nothing compared to the cache misses > > you get from preemption, so I'd ignore that. Saving 300 clock cycles on > > userspace exits from TR+GSBASE would be about 5% on my Haswell. > > That's still 5% :) Yes, 5% on userspace exits is good (though they're rare). OTOH, 300 clock cycles on preemption are nothing to write home about. Paolo