From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752358AbaEDSkW (ORCPT ); Sun, 4 May 2014 14:40:22 -0400 Received: from mail-ee0-f42.google.com ([74.125.83.42]:63574 "EHLO mail-ee0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751085AbaEDSkV (ORCPT ); Sun, 4 May 2014 14:40:21 -0400 Date: Sun, 4 May 2014 20:40:16 +0200 From: Ingo Molnar To: Andy Lutomirski Cc: Linus Torvalds , Linux Kernel Mailing List , the arch/x86 maintainers , "H. Peter Anvin" , Steven Rostedt , Peter Zijlstra Subject: Re: [RFC/HACK] x86: Fast return to kernel Message-ID: <20140504184016.GA16438@gmail.com> References: <210a076ea197ae384705d2c02cfff12a951a62f8.1399057218.git.luto@amacapital.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Andy Lutomirski wrote: > > That said, regular *device* interrupts do often return to kernel > > mode (the idle loop in particular), so if you have any way to > > measure that, that might be interesting, and might show some of > > the same advantages. > > I can try something awful involving measuring latency of > hardware-timed packets on a SolarFlare card, but I'll have > calibration issues. I suppose I could see if 'ping' gets faster. > In general, this will speed up interrupts that wake userspace from > idle by about 100ns on my box, since it's presumably the same size > and the speedup per loop in my silly benchmark. To simulate high rate device IRQ you can generate very high frequency lapic IRQs by using hrtimers, that's generating a ton of per CPU lapic IRQs. Thanks, Ingo