From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932185Ab2GQWaM (ORCPT ); Tue, 17 Jul 2012 18:30:12 -0400 Received: from li9-11.members.linode.com ([67.18.176.11]:39607 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757182Ab2GQWaJ (ORCPT ); Tue, 17 Jul 2012 18:30:09 -0400 Date: Tue, 17 Jul 2012 17:42:19 -0400 From: "Theodore Ts'o" To: Linus Walleij Cc: Sedat Dilek , Tony Luck , Fenghua Yu , Linus Walleij , Samuel Ortiz , Andrew Morton , David Howells , "Srivatsa S. Bhat" , "David S. Miller" , Eric Dumazet , "H. Peter Anvin" , linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-next@vger.kernel.org, Stephen Rothwell Subject: Re: [PATCH -next] random: Remove remaining relicts of rand_initialize_irq() Message-ID: <20120717214219.GF6049@thunk.org> Mail-Followup-To: Theodore Ts'o , Linus Walleij , Sedat Dilek , Tony Luck , Fenghua Yu , Linus Walleij , Samuel Ortiz , Andrew Morton , David Howells , "Srivatsa S. Bhat" , "David S. Miller" , Eric Dumazet , "H. Peter Anvin" , linux-ia64@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-next@vger.kernel.org, Stephen Rothwell References: <1342524793-2522-1-git-send-email-sedat.dilek@gmail.com> <20120717171251.GD6049@thunk.org> 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) X-SA-Exim-Connect-IP: X-SA-Exim-Mail-From: tytso@thunk.org X-SA-Exim-Scanned: No (on imap.thunk.org); SAEximRunCond expanded to false Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Jul 17, 2012 at 08:57:19PM +0200, Linus Walleij wrote: > So the IRQ will be auto-sampled now or something? > > Sorry for not quite following, which commit do I look at to > understand this now... Yes, there was a huge discussion on LKML about two weeks ago. The key commit that will be going in during the next merge window is: http://git.kernel.org/?p=linux/kernel/git/tytso/random.git;a=commitdiff;h=775f4b297b780601e61787b766f306ed3e1d23eb There are a whole lot of other changes, of course, but that's the key change which IRQF_SAMPLE_RANDOM a no-op. People have tested this under extreme conditions (high speed networking tests with no interrupt mitigations, etc.) and the patch as tweaked had no measurable overhead, so we can simply sample every single commit with a low-overhead fast path which uses per-cpu buffers, and then mix into the /dev/random entropy pool less frequently (which is still fast, but requires a few more memory accesses to shared memory that may incur delays due to cache line bounces). - Ted