From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx0a-001b2d01.pphosted.com ([148.163.156.1]:41994 "EHLO mx0a-001b2d01.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753594AbdKJTum (ORCPT ); Fri, 10 Nov 2017 14:50:42 -0500 Received: from pps.filterd (m0098409.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.21/8.16.0.21) with SMTP id vAAJnJ2e038089 for ; Fri, 10 Nov 2017 14:50:41 -0500 Received: from e06smtp15.uk.ibm.com (e06smtp15.uk.ibm.com [195.75.94.111]) by mx0a-001b2d01.pphosted.com with ESMTP id 2e5fxxqtx3-1 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=NOT) for ; Fri, 10 Nov 2017 14:50:40 -0500 Received: from localhost by e06smtp15.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 10 Nov 2017 19:50:38 -0000 Subject: Re: [RFC PATCH] fw_lockdown: new micro LSM module to prevent loading unsigned firmware From: Mimi Zohar To: "Luis R. Rodriguez" , "AKASHI, Takahiro" Cc: David Howells , linux-security-module , linux-fsdevel , linux-kernel , James Bottomley , David Woodhouse , Johannes Berg , Andy Lutomirski , Linus Torvalds , Kees Cook Date: Fri, 10 Nov 2017 14:50:30 -0500 In-Reply-To: <20171110193507.GP22894@wotan.suse.de> References: <1510336703.3404.17.camel@linux.vnet.ibm.com> <20171110193507.GP22894@wotan.suse.de> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <1510343430.3404.25.camel@linux.vnet.ibm.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Fri, 2017-11-10 at 20:35 +0100, Luis R. Rodriguez wrote: > On Fri, Nov 10, 2017 at 12:58:23PM -0500, Mimi Zohar wrote: > > Hi David, > > > > If you are interested in preventing the loading of unsigned firmware, > > the patch below is straight forward.  The patch has ONLY been tested > > with IMA-appraisal enabled, and works as intended - allowing only > > signed firmware to be loaded. > > Very nice! This is the sort of thing that I mean by LSM'ifying fw access > through a system policy. > > We currently handle the LSM aspect for firmware through > kernel_read_file_from_path() and so the kernel_read_file LSM hook, so why a new > hook here? kernel_read_file(), itself, is not an LSM hook, but calls two LSM hooks named security_kernel_read_file(), prior to reading a file, and security_kernel_post_read_file(), post reading a file. In this case, we want to reject even reading the file if it isn't signed, so we're using the security_kernel_read_file() LSM hook. > > Where does this plug in? This is a standalone, micro LSM that can be configured at build.  For now I left it is an optional Kconfig parameter, but at some point, you might want to consider making it required. Mimi