[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <fa96a33641070b1580f21de86fedd5f8da5eff21.camel@linux.ibm.com>
Date: Tue, 28 Jul 2020 14:48:25 -0400
From: Mimi Zohar <zohar@...ux.ibm.com>
To: Scott Branden <scott.branden@...adcom.com>,
Kees Cook <keescook@...omium.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Luis Chamberlain <mcgrof@...nel.org>, Jessica Yu <jeyu@...nel.org>,
SeongJae Park <sjpark@...zon.de>,
KP Singh <kpsingh@...omium.org>, linux-efi@...r.kernel.org,
linux-security-module@...r.kernel.org,
linux-integrity@...r.kernel.org, selinux@...r.kernel.org,
linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 00/19] Introduce partial kernel_read_file() support
On Mon, 2020-07-27 at 12:18 -0700, Scott Branden wrote:
> Hi Mimi/Kees,
>
> On 2020-07-27 4:16 a.m., Mimi Zohar wrote:
> > On Fri, 2020-07-24 at 14:36 -0700, Kees Cook wrote:
> >> v3:
> >> - add reviews/acks
> >> - add "IMA: Add support for file reads without contents" patch
> >> - trim CC list, in case that's why vger ignored v2
> >> v2: [missing from lkml archives! (CC list too long?) repeating changes
> here]
> >> - fix issues in firmware test suite
> >> - add firmware partial read patches
> >> - various bug fixes/cleanups
> >> v1:
> https://lore.kernel.org/lkml/20200717174309.1164575-1-keescook@chromium.org/
> >>
> >> Hi,
> >>
> >> Here's my tree for adding partial read support in kernel_read_file(),
> >> which fixes a number of issues along the way. It's got Scott's firmware
> >> and IMA patches ported and everything tests cleanly for me (even with
> >> CONFIG_IMA_APPRAISE=y).
> > Thanks, Kees. Other than my comments on the new
> > security_kernel_post_load_data() hook, the patch set is really nice.
> >
> > In addition to compiling with CONFIG_IMA_APPRAISE enabled, have you
> > booted the kernel with the ima_policy=tcb? The tcb policy will add
> > measurements to the IMA measurement list and extend the TPM with the
> > file or buffer data digest. Are you seeing the firmware measurements,
> > in particular the partial read measurement?
> I booted the kernel with ima_policy=tcb.
>
> Unfortunately after enabling the following, fw_run_tests.sh does not run.
>
> mkdir /sys/kernel/security
> mount -t securityfs securityfs /sys/kernel/security
> echo "measure func=FIRMWARE_CHECK" > /sys/kernel/security/ima/policy
> echo "appraise func=FIRMWARE_CHECK appraise_type=imasig" >
> /sys/kernel/security/ima/policy
> ./fw_run_tests.sh
>
> [ 1296.258052] test_firmware: loading 'test-firmware.bin'
> [ 1296.263903] misc test_firmware: loading /lib/firmware/test-firmware.bin
> failed with error -13
> [ 1296.263905] audit: type=1800 audit(1595905754.266:9): pid=5696 uid=0
> auid=4294967295 ses=4294967295 subj=kernel op=appraise_data cause=IMA-
> signature-required comm="fw_namespace" name="/lib/firmware/test-firmware.bin"
> dev="tmpfs" ino=4592 res=0
> [ 1296.297085] misc test_firmware: Direct firmware load for test-firmware.bin
> failed with error -13
> [ 1296.305947] test_firmware: load of 'test-firmware.bin' failed: -13
The "appraise" rule verifies the IMA signature. Unless you signed the firmware
(evmctl) and load the public key on the IMA keyring, that's to be expected. I
assume you are seeing firmware measurements in the IMA measuremenet log.
Mimi
Powered by blists - more mailing lists