Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S938790AbcKWOZj (ORCPT ); Wed, 23 Nov 2016 09:25:39 -0500 Received: from foss.arm.com ([217.140.101.70]:53838 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756585AbcKWOZd (ORCPT ); Wed, 23 Nov 2016 09:25:33 -0500 Date: Wed, 23 Nov 2016 14:24:40 +0000 From: Mark Rutland To: David Howells Cc: lukas@wunner.de, linux-efi@vger.kernel.org, linux-security-module@vger.kernel.org, keyrings@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 4/6] efi: Get the secure boot status [ver #2] Message-ID: <20161123142440.GJ24624@leverpostej> References: <20161123134219.GH24624@leverpostej> <20161123104757.GE24624@leverpostej> <147986054870.13790.8640536414645705863.stgit@warthog.procyon.org.uk> <147986057768.13790.3027173260868896792.stgit@warthog.procyon.org.uk> <18007.1479900357@warthog.procyon.org.uk> <31974.1479910408@warthog.procyon.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <31974.1479910408@warthog.procyon.org.uk> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1167 Lines: 35 On Wed, Nov 23, 2016 at 02:13:28PM +0000, David Howells wrote: > Mark Rutland wrote: > > > > > if (secure_boot < 0) > > > > pr_efi_err(sys_table, > > > > "could not determine UEFI Secure Boot status.\n"); > > > > > > In which case, should this be moved into efi_get_secureboot() and it return a > > > bool? > > > > That would make sense to me, provided we're only likely to call that > > once (and only log once). > > > > I guess it would also make sense to change the latter case to soemthing > > like: > > > > Could not determine UEFI Secure Boot status. Assuming enabled. > > > > ... so as to make it clear what the effect is. > > Actually, the two arches have a different interpretation on how to deal with > an error. Matthew Garrett's original x86 patch assumes that if we get an > error when trying to read SecureBoot and SetupMode that we're *not* in secure > mode, but ARM assumes the opposite. Ok. IIUC, that x86 patch was never upstream, so is there any need to follow that example? Was there a rationale for that, or can we simply follow the upstream ARM example? Perhaps it's best to ask Matthew? Thanks, Mark.