From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752629AbcIDHTp (ORCPT ); Sun, 4 Sep 2016 03:19:45 -0400 Received: from mail-it0-f48.google.com ([209.85.214.48]:37846 "EHLO mail-it0-f48.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751203AbcIDHTl (ORCPT ); Sun, 4 Sep 2016 03:19:41 -0400 MIME-Version: 1.0 In-Reply-To: References: <20160903.233256.116565014024856604.davem@davemloft.net> From: Julian Calaby Date: Sun, 4 Sep 2016 17:18:55 +1000 Message-ID: Subject: Re: sparc: bpf_jit: Move four assignments in bpf_jit_compile() To: SF Markus Elfring Cc: David Miller , sparclinux , Adam Buchbinder , Alexei Starovoitov , Daniel Borkmann , Rabin Vincent , "linux-kernel@vger.kernel.org" , kernel-janitors , Julia Lawall , Paolo Bonzini Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Markus, On Sun, Sep 4, 2016 at 5:07 PM, SF Markus Elfring wrote: >> I really wish he'd concentrate on the former rather than the latter. > > I am on the way for such a software development adventure. > There are further improvement opportunities to consider besides the main route, > aren't there? You're missing the point here. Find something useful to change. People are going through the kernel finding functions that should be static for example - this is useful as it provides better documentation for those functions (static == not used outside this file) and fixes errors reported by GCC and static code checkers. Those changes are useful. Another useful change would be ensuring that every struct that has a bool member is kzalloc'd instead of kmalloc'd as this eliminates a possible source of undefined behaviour. (A bool is effectively a u8 with behaviour defined for when that value is 0 or 1, kmalloc'd memory can have any data in it, therefore it's possible and likely that a kmalloc'd struct with a bool member will end up with some value for that member that isn't 0 or 1. kzalloc eliminates this possibility.) Moving four assignments because you think they might improve stuff is just annoying people. >> He's not another Nick Krause. > > Are you going to remember me under an other nickname anyhow? Nick Krause submitted patches that make yours look good. At least yours compile. Thanks, -- Julian Calaby Email: julian.calaby@gmail.com Profile: http://www.google.com/profiles/julian.calaby/