From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754188AbdGUNY6 (ORCPT ); Fri, 21 Jul 2017 09:24:58 -0400 Received: from mx1.redhat.com ([209.132.183.28]:51261 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753779AbdGUNY5 (ORCPT ); Fri, 21 Jul 2017 09:24:57 -0400 DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com A7F49A23EA Authentication-Results: ext-mx01.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx01.extmail.prod.ext.phx2.redhat.com; spf=pass smtp.mailfrom=jpoimboe@redhat.com DKIM-Filter: OpenDKIM Filter v2.11.0 mx1.redhat.com A7F49A23EA Date: Fri, 21 Jul 2017 08:24:52 -0500 From: Josh Poimboeuf To: Andrey Ryabinin Cc: Matthias Kaehlcke , Chris J Arges , Borislav Petkov , Thomas Gleixner , Ingo Molnar , "H . Peter Anvin" , "x86@kernel.org" , LKML , Douglas Anderson , Michael Davidson , Greg Hackmann , Nick Desaulniers , Stephen Hines , Kees Cook , Arnd Bergmann , Bernhard =?utf-8?Q?Rosenkr=C3=A4nzer?= Subject: Re: [PATCH] Revert "x86/uaccess: Add stack frame output operand in get_user() inline asm" Message-ID: <20170721132452.ihpws67e3e7ym3al@treble> References: <20170713203416.isvijqbwbcgupgj7@treble> <20170713211245.GG95735@google.com> <20170713213406.gx4ixkx6kxa4ppps@treble> <20170713215704.GJ95735@google.com> <20170719174630.kz5g553evcrnirmr@treble> <20170720151813.5wnpsb5wy7bqrpec@treble> <20170720205652.patvjnfqymrv73ho@treble> <7501b8fc-366d-239d-6358-a403a5bc0eab@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <7501b8fc-366d-239d-6358-a403a5bc0eab@gmail.com> User-Agent: Mutt/1.6.0.1 (2016-04-01) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.25]); Fri, 21 Jul 2017 13:24:57 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Jul 21, 2017 at 12:13:31PM +0300, Andrey Ryabinin wrote: > > Still, unfortunately, I don't think that's going to work for GCC. > > Changing the '__sp' register variable to global in the header file > > causes it to make a *bunch* of changes across the kernel, even in > > functions which don't do inline asm. It seems to be disabling some > > optimizations across the board. > > All I see is just bunch of reordering of independent instructions, like this: > > -ffffffff81012760: 5b pop %rbx > -ffffffff81012761: 31 c0 xor %eax,%eax > +ffffffff81012760: 31 c0 xor %eax,%eax > +ffffffff81012762: 5b pop %rbx > > -ffffffff810c29ae: 48 83 c4 28 add $0x28,%rsp > -ffffffff810c29b2: 89 d8 mov %ebx,%eax > +ffffffff810c29ae: 89 d8 mov %ebx,%eax > +ffffffff810c29b0: 48 83 c4 28 add $0x28,%rsp > > I haven't noticed any single bad/harmful change. The size of .text remained the same. I compiled with -ffunction-sections to make the comparisons easier. The reordering is much more extreme than your example. (This is with GCC 7, btw). And it's not just reordering of instructions. It's control flow changes as well. Also, the text size grew a little: text data bss dec hex filename 10630602 8295074 16461824 35387500 21bf86c vmlinux.before 10634013 8295074 16461824 35390911 21c05bf vmlinux.after A small two-line change, which is supposed to be a noop, or at least should only affect a small number of functions, but which instead affects optimization decisions across the entire kernel, is actively harmful IMO. > And btw, arm/arm64 already use global current_stack_pointer just fine. I wonder if they looked for the impact. -- Josh