From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-fsdevel-owner@vger.kernel.org>
Received: from verein.lst.de ([213.95.11.211]:55554 "EHLO newverein.lst.de"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1750896AbdGERS7 (ORCPT <rfc822;linux-fsdevel@vger.kernel.org>);
        Wed, 5 Jul 2017 13:18:59 -0400
Date: Wed, 5 Jul 2017 19:18:57 +0200
From: Christoph Hellwig <hch@lst.de>
To: Matthew Garrett <mjg59@srcf.ucam.org>
Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>,
        Christoph Hellwig <hch@lst.de>,
        Al Viro <viro@zeniv.linux.org.uk>,
        James Morris <jmorris@namei.org>,
        linux-fsdevel@vger.kernel.org,
        linux-ima-devel@lists.sourceforge.net,
        linux-security-module@vger.kernel.org,
        linux-ima-user <linux-ima-user@lists.sourceforge.net>
Subject: Re: [PATCH v2 10/10] ima: use existing read file operation method
        to calculate file hash
Message-ID: <20170705171857.GA5246@lst.de>
References: <1498069110-10009-1-git-send-email-zohar@linux.vnet.ibm.com> <1498069110-10009-11-git-send-email-zohar@linux.vnet.ibm.com> <20170628144111.GI2359@lst.de> <1499266209.3059.91.camel@linux.vnet.ibm.com> <20170705170215.GA14148@srcf.ucam.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20170705170215.GA14148@srcf.ucam.org>
Sender: linux-fsdevel-owner@vger.kernel.org
List-ID: <linux-fsdevel.vger.kernel.org>

On Wed, Jul 05, 2017 at 06:02:15PM +0100, Matthew Garrett wrote:
> On Wed, Jul 05, 2017 at 10:50:09AM -0400, Mimi Zohar wrote:
> > [Cc'ing linux-ima-users]
> > 
> > On Wed, 2017-06-28 at 16:41 +0200, Christoph Hellwig wrote:
> > > NAK - we'll need an explicit method for the integrity code.
> > > 
> > > And just curious - what filesystem that you care about actually
> > > implements ->read instead of ->read_iter?  We shouldn't be doing that
> > > for real file systems anymore.
> > 
> > Right, pseudo filesystems are using ->read. The existing builtin
> > measurement policies exclude a number of pseudo filesystems, but not
> > efivarfs. �Unfortunately, we do not know what type of custom policies
> > are currently being used.
> 
> efi variables contain information that may influence userspace behaviour 
> and can also be modified out of band, so I think there's a reasonable 
> argument that they should be measured.

Then efivars should grow a ->integrity_read method.