From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753182AbdGJBrf (ORCPT ); Sun, 9 Jul 2017 21:47:35 -0400 Received: from mx1.redhat.com ([209.132.183.28]:39106 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753076AbdGJBre (ORCPT ); Sun, 9 Jul 2017 21:47:34 -0400 DMARC-Filter: OpenDMARC Filter v1.3.2 mx1.redhat.com A3838267C1 Authentication-Results: ext-mx06.extmail.prod.ext.phx2.redhat.com; dmarc=none (p=none dis=none) header.from=redhat.com Authentication-Results: ext-mx06.extmail.prod.ext.phx2.redhat.com; spf=pass smtp.mailfrom=bhe@redhat.com DKIM-Filter: OpenDKIM Filter v2.11.0 mx1.redhat.com A3838267C1 Date: Mon, 10 Jul 2017 09:47:29 +0800 From: Baoquan He To: Kees Cook Cc: LKML , "x86@kernel.org" , Matt Fleming , Thomas Gleixner , Ingo Molnar , "H. Peter Anvin" , izumi.taku@jp.fujitsu.com, fanc.fnst@cn.fujitsu.com, Thomas Garnier Subject: Re: [PATCH v4 4/4] x86/boot/KASLR: Restrict kernel to be randomized in mirror regions Message-ID: <20170710014729.GC13560@x1> References: <1499603862-11516-1-git-send-email-bhe@redhat.com> <1499603862-11516-5-git-send-email-bhe@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.7.0 (2016-08-17) X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.30]); Mon, 10 Jul 2017 01:47:34 +0000 (UTC) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 07/09/17 at 07:11am, Kees Cook wrote: > On Sun, Jul 9, 2017 at 5:37 AM, Baoquan He wrote: > > Signed-off-by: Baoquan He > > +/* Marks if efi mirror regions have been found and handled. */ > > +static bool efi_mirror_found; > > I think this is only ever checked once? How about having > process_efi_entries return bool to indicate if mirror was found? Also, > that function should be behind #ifdef. Let's do something like this: > > > > + > > +static void process_efi_entries(unsigned long minimum, > > + unsigned long image_size) > > #ifdef CONFIG_EFI > /* Returns true if mirror region found (and further scanning should stop) */ Well, here it might be not like that. The mirror regions could be multiple, we need find and process each of them to add slots. > static bool process_efi_entries(...) > { > ... > } > #else > static inline bool process_efi_entries(...) > { > return false; > } > #endif > > Then: > > > +#ifdef CONFIG_EFI > > + process_efi_entries(minimum, image_size); > > + if (efi_mirror_found) > > + return slots_fetch_random(); > > +#endif > > Can become: > > if (process_efi_entries(minimum, image_size)) > return slots_fetch_random() > > and no #ifndef needed here. > > > + > > process_e820_entries(minimum, image_size); > > return slots_fetch_random(); > > } > > @@ -652,7 +708,7 @@ void choose_random_location(unsigned long input, > > */ > > min_addr = min(*output, 512UL << 20); > > > > - /* Walk e820 and find a random address. */ > > + /* Walk available memory entries to find a random address. */ > > random_addr = find_random_phys_addr(min_addr, output_size); > > if (!random_addr) { > > warn("Physical KASLR disabled: no suitable memory region!"); > > -- > > 2.5.5 > > > > Otherwise, if the EFI logic is good, this looks sensible. > > Thanks for splitting up the patches! > > -Kees > > -- > Kees Cook > Pixel Security