Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v2 3/4] ima: Use tpm_chip_find() and access TPM functions
 using it
To:     Jason Gunthorpe <jgg@ziepe.ca>
Cc:     Mimi Zohar <zohar@linux.vnet.ibm.com>,
        linux-integrity@vger.kernel.org, jarkko.sakkinen@linux.intel.com,
        linux-kernel@vger.kernel.org
References: <20180620204236.1572523-1-stefanb@linux.vnet.ibm.com>
 <20180620204236.1572523-4-stefanb@linux.vnet.ibm.com>
 <1529614425.23843.20.camel@linux.vnet.ibm.com>
 <f4af7347-a9dd-1683-5c2c-faeb45750e2e@linux.vnet.ibm.com>
 <20180622032536.GB19151@ziepe.ca>
From:   Stefan Berger <stefanb@linux.vnet.ibm.com>
Date:   Fri, 22 Jun 2018 07:40:37 -0400
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.8.0
MIME-Version: 1.0
In-Reply-To: <20180622032536.GB19151@ziepe.ca>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-MW
Message-Id: <495fdcf7-b5b9-0341-796b-66fdf537811b@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 06/21/2018 11:25 PM, Jason Gunthorpe wrote:
> On Thu, Jun 21, 2018 at 04:59:55PM -0400, Stefan Berger wrote:
>> On 06/21/2018 04:53 PM, Mimi Zohar wrote:
>>> On Wed, 2018-06-20 at 16:42 -0400, Stefan Berger wrote:
>>>> Rather than accessing the TPM functions using a NULL pointer, which
>>>> causes a lookup for a suitable chip every time, get a hold of a tpm_chip
>>>> and access the TPM functions using this chip. We call the tpm_chip
>>>> ima_tpm_chip and protect it, once initialization is done, using a
>>>> rw_semaphore called ima_tpm_chip_lock.
>>>>
>>>> Use ima_shutdown to release the tpm_chip.
>>>>
>>>> Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
>>>>   security/integrity/ima/ima.h        |  3 +++
>>>>   security/integrity/ima/ima_crypto.c | 12 ++++++++++--
>>>>   security/integrity/ima/ima_init.c   | 19 ++++++++++++-------
>>>>   security/integrity/ima/ima_queue.c  |  7 +++++--
>>>>   4 files changed, 30 insertions(+), 11 deletions(-)
>>>>
>>>> diff --git a/security/integrity/ima/ima.h b/security/integrity/ima/ima.h
>>>> index 354bb5716ce3..53a88d578ca5 100644
>>>> +++ b/security/integrity/ima/ima.h
>>>> @@ -24,6 +24,7 @@
>>>>   #include <linux/hash.h>
>>>>   #include <linux/tpm.h>
>>>>   #include <linux/audit.h>
>>>> +#include <linux/rwsem.h>
>>>>   #include <crypto/hash_info.h>
>>>>
>>>>   #include "../integrity.h"
>>>> @@ -56,6 +57,8 @@ extern int ima_policy_flag;
>>>>   extern int ima_used_chip;
>>>>   extern int ima_hash_algo;
>>>>   extern int ima_appraise;
>>>> +extern struct rw_semaphore ima_tpm_chip_lock;
>>>> +extern struct tpm_chip *ima_tpm_chip;
>>> ima_add_templatE_entry() synchronizes appending a measurement to the
>>> measurement list and extending the TPM by taking a lock.  Do we really
>>> need to introduce another lock?
>> This lock protects the ima_tpm_chip from going from != NULL to NULL in the
>> ima_shutdown function. Basically, a global pointer accessed by concurrent
>> threads should be protected if its value can change. However, in this case
>> ima_shutdown would be called so late that there shouldn't be concurrency
>> anymore. Though, I found it better to protect it. Maybe someone else has an
>> opinion?
> Why have a shutdown block? There is no harm in holding a kref if the
> machine is shutting down.

Looking around at other drivers' usage of the reboot notifier, I find 
other drivers as well that use spinlocks or mutexes during the shutdown. 
Besides that, we do have the shutdown block already when device_shutdown 
calls tpm_class_shutdown() and we get the ops_sem.

     Stefan
>
> Jason
>