From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754745AbcEQJEk (ORCPT ); Tue, 17 May 2016 05:04:40 -0400 Received: from mail-wm0-f43.google.com ([74.125.82.43]:37253 "EHLO mail-wm0-f43.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754515AbcEQJEh (ORCPT ); Tue, 17 May 2016 05:04:37 -0400 Date: Tue, 17 May 2016 10:04:34 +0100 From: Matt Fleming To: Linus Torvalds Cc: Ingo Molnar , Alex Thorlton , Linux Kernel Mailing List , Thomas Gleixner , "H. Peter Anvin" , Andrew Morton , Peter Zijlstra , Borislav Petkov , Josh Poimboeuf Subject: Re: [GIT PULL] EFI fix Message-ID: <20160517090434.GA21993@codeblueprint.co.uk> References: <20160516144648.GA22999@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.24+41 (02bc14ed1569) (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 16 May, at 01:05:45PM, Linus Torvalds wrote: > > So that whole 8-vs-16 offset confusion depends on the frame pointer! > If frame pointers were enabled, it will be 16. If they weren't, it > will be 8. That patch that changes it from 8 to 16 will just move the > bug around. Before, it was correct when frame pointers were disabled > and buggy otherwise. Now, it's correct if frame pointers are enabled, > and buggy otherwise. Urgh, right. We didn't always use frame pointers in efi_call(), they were introduced in commit 779c433b8ea5c ("x86/asm/efi: Create a stack frame in efi_call()") earlier this year to stop objtool complaining. I admit I totally missed the part about pulling arguments off the stack when I reviewed that patch. > I may be missing something, but I think that commit is pure garbage. You're correct. > I think the right fix is to just get rid of that silly conditional > frame pointer thing, and always use frame pointers in this stub > function. And then we don't need that (odd) load to get the old stack > pointer into %rax - we can just use the frame pointer. > > Something like the attached completely untested patch. Looks good to me, but I haven't tested it. Now I'm wondering whether other users of FRAME_BEGIN/FRAME_END make this same mistake. Coccinelle might be able to detect it perhaps.