Message-ID: <865dfe9e-ee3d-1067-625e-2d93bf96da4b@arista.com>
Date:   Fri, 10 Feb 2023 10:07:02 -0800
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.6.1
Subject: Re: [PATCH] tpm: add vendor flag to command code validation
Content-Language: en-CA
To:     Jarkko Sakkinen <jarkko@kernel.org>
Cc:     linux-kernel@vger.kernel.org, linux-integrity@vger.kernel.org,
        jgg@ziepe.ca, peterhuewe@gmx.de
References: <20230208195836.30175-1-julien@arista.com>
 <Y+WUqcNTc8t0KIyD@kernel.org>
From:   Julien Gomes <julien@arista.com>
In-Reply-To: <Y+WUqcNTc8t0KIyD@kernel.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2023-02-09 4:49 p.m., Jarkko Sakkinen wrote:
> On Wed, Feb 08, 2023 at 11:58:36AM -0800, Julien Gomes wrote:
>> Some TPM 2.0 devices have support for additional commands which are not
>> part of the TPM 2.0 specifications.
>> These commands are identified with bit 29 of the 32 bits command codes.
>> Contrarily to other fields of the TPMA_CC spec structure used to list
>> available commands, the Vendor flag also has to be present in the
>> command code itself (TPM_CC) when called.
>>
>> Add this flag to tpm_find_cc() mask to prevent blocking vendor command
>> codes that can actually be supported by the underlying TPM device.
>>
>> Signed-off-by: Julien Gomes <julien@arista.com>
>> ---
>>   drivers/char/tpm/tpm2-cmd.c | 4 +++-
>>   include/linux/tpm.h         | 1 +
>>   2 files changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/char/tpm/tpm2-cmd.c b/drivers/char/tpm/tpm2-cmd.c
>> index 65d03867e114..93545be190a5 100644
>> --- a/drivers/char/tpm/tpm2-cmd.c
>> +++ b/drivers/char/tpm/tpm2-cmd.c
>> @@ -777,10 +777,12 @@ int tpm2_auto_startup(struct tpm_chip *chip)
>>   
>>   int tpm2_find_cc(struct tpm_chip *chip, u32 cc)
>>   {
>> +	u32 cc_mask;
>>   	int i;
>>   
>> +	cc_mask = 1 << TPM2_CC_ATTR_VENDOR | GENMASK(15, 0);
>>   	for (i = 0; i < chip->nr_commands; i++)
>> -		if (cc == (chip->cc_attrs_tbl[i] & GENMASK(15, 0)))
>> +		if (cc == (chip->cc_attrs_tbl[i] & cc_mask))
>>   			return i;
>>   
>>   	return -1;
>> diff --git a/include/linux/tpm.h b/include/linux/tpm.h
>> index dfeb25a0362d..4dc97b9f65fb 100644
>> --- a/include/linux/tpm.h
>> +++ b/include/linux/tpm.h
>> @@ -265,6 +265,7 @@ enum tpm2_startup_types {
>>   enum tpm2_cc_attrs {
>>   	TPM2_CC_ATTR_CHANDLES	= 25,
>>   	TPM2_CC_ATTR_RHANDLE	= 28,
>> +	TPM2_CC_ATTR_VENDOR	= 29,
>>   };
>>   
>>   #define TPM_VID_INTEL    0x8086
>> -- 
>> 2.39.1
>>
> 
> Just checking: did you run testing/selftests/tpm2?
> 
> BR, Jarkko

I didn't know about these, good call.
Just ran the three test suites on a vm with a swtpm, as I don't have a
physical box with TPM 2.0 able to run latest kernels handy, all passed.

-- 
Julien Gomes