Return-path: Received: from mail-pa0-f53.google.com ([209.85.220.53]:42402 "EHLO mail-pa0-f53.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753610AbaAFJYw (ORCPT ); Mon, 6 Jan 2014 04:24:52 -0500 MIME-Version: 1.0 In-Reply-To: References: <1388427307-8691-1-git-send-email-Julia.Lawall@lip6.fr> <1388427307-8691-5-git-send-email-Julia.Lawall@lip6.fr> <1388429761.4410.1.camel@jlt4.sipsolutions.net> <1388438724.4573.2.camel@jlt4.sipsolutions.net> <20131230215701.GA4938@khazad-dum.debian.net> <1388445188.18164.0.camel@jlt4.sipsolutions.net> Date: Mon, 6 Jan 2014 10:24:51 +0100 Message-ID: (sfid-20140106_102546_724195_04B77729) Subject: Re: [PATCH 4/11] use ether_addr_equal_64bits From: Geert Uytterhoeven To: Emmanuel Grumbach Cc: Johannes Berg , Henrique de Moraes Holschuh , Julia Lawall , "kernel-janitors@vger.kernel.org" , Emmanuel Grumbach , Intel Linux Wireless , "John W. Linville" , linux-wireless , "netdev@vger.kernel.org" , "linux-kernel@vger.kernel.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Tue, Dec 31, 2013 at 7:26 AM, Emmanuel Grumbach wrote: > On Tue, Dec 31, 2013 at 1:13 AM, Johannes Berg > wrote: >> >> On Mon, 2013-12-30 at 19:57 -0200, Henrique de Moraes Holschuh wrote: >> > On Mon, 30 Dec 2013, Johannes Berg wrote: >> > > On Mon, 2013-12-30 at 20:58 +0100, Julia Lawall wrote: >> > > > > Is there any way we could catch (sparse, or some other script?) that >> > > > > struct reorganising won't break the condition needed ("within a >> > > > > structure that contains at least two more bytes")? >> > > > >> > > > What kind of reorganizing could happen? Do you mean that the programmer >> > > > might do at some time in the future, or something the compiler might do? >> > > >> > > I'm just thinking of a programmer, e.g. changing a struct like this: >> > > >> > > struct foo { >> > > u8 addr[ETH_ALEN]; >> > > - u16 dummy; >> > > }; >> > > >> > > for example. >> > >> > That is easily resolved by: >> > >> > struct foo { >> > u8 addr[ETH_ALEN]; >> > u16 required_padding; /* do not remove upon pain of death */ >> > }; Adding the u16 also changes the alignment of the whole struct. So it may cost one additional byte _in front of_ the struct. While you're at it, why not just making a new 64-bit aligned type for Ethernet addresses, so it'll also work for !CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS? >> That'd be a stupid waste of struct space. If anything, there should be >> *only* a comment saying that at least two bytes are needed - I'd still >> prefer an automated check. > > Frankly I am not sure I like the patch. This flow is not a fast path I also don't like it. To me this sounds like wasting space for nothing. BTW, would it be that more expensive to always do a 32+16 bit comparison? > at all. While I don't really care for the waste in iwlwifi (because > there isn't), I don't see the real point is make the code more > sensitive to changes to earn basically nothing. Thanks to this discussion, my eye fell on: static inline unsigned compare_ether_addr(const u8 *addr1, const u8 *addr2) { const u16 *a = (const u16 *) addr1; const u16 *b = (const u16 *) addr2; BUILD_BUG_ON(ETH_ALEN != 6); return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) != 0; } What if addr1 or addr2 are odd, and this is running on an architecture that doesn't support unaligned accesses at all?? Have we been lucky forever? Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds