[PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[Roberto Sassu]

Commit c78719203fc6 ("KEYS: trusted: allow trusted.ko to initialize w/o a
TPM") allows the trusted module to be loaded even a TPM is not found to
avoid module dependency problems.

However, trusted module initialization can still fail if the TPM is
inactive or deactivated. This patch ignores tpm_get_random() errors in
init_digests() and returns -EFAULT in pcrlock() if the TPM didn't return
random data.

Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
---
 security/keys/trusted.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/security/keys/trusted.c b/security/keys/trusted.c
index 9a94672e7adc..34f04ffcf2e5 100644
--- a/security/keys/trusted.c
+++ b/security/keys/trusted.c
@@ -389,6 +389,10 @@ static int pcrlock(const int pcrnum)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
+	/* This happens if the TPM didn't return random data */
+	if (!digests)
+		return -EFAULT;
+
 	return tpm_pcr_extend(chip, pcrnum, digests) ? -EINVAL : 0;
 }
 
@@ -1233,10 +1237,8 @@ static int __init init_digests(void)
 	int i;
 
 	ret = tpm_get_random(chip, digest, TPM_MAX_DIGEST_SIZE);
-	if (ret < 0)
-		return ret;
-	if (ret < TPM_MAX_DIGEST_SIZE)
-		return -EFAULT;
+	if (ret < 0 || ret < TPM_MAX_DIGEST_SIZE)
+		return 0;
 
 	digests = kcalloc(chip->nr_allocated_banks, sizeof(*digests),
 			  GFP_KERNEL);
-- 
2.17.1

Re: [PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[Tyler Hicks]

On 2019-08-02 17:07:33, Roberto Sassu wrote:
> Commit c78719203fc6 ("KEYS: trusted: allow trusted.ko to initialize w/o a
> TPM") allows the trusted module to be loaded even a TPM is not found to
                                                   ^ if

> avoid module dependency problems.
> 
> However, trusted module initialization can still fail if the TPM is
> inactive or deactivated. This patch ignores tpm_get_random() errors in
> init_digests() and returns -EFAULT in pcrlock() if the TPM didn't return
> random data.
> 
> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>

The code changes look correct to me.

  Reviewed-by: Tyler Hicks <tyhicks@canonical.com>

For whoever takes this patch through their tree, I think that adding the
following Fixes tag would be useful (as well as cc'ing stable):

  Fixes: 240730437deb ("KEYS: trusted: explicitly use tpm_chip structure...")

I think it is also worth leaving a short note, in the commit message,
for backporters that commit 782779b60faa ("tpm: Actually fail on TPM
errors during "get random"") should be included with any backports of
this patch.

Thanks!

Tyler

> ---
>  security/keys/trusted.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)
> 
> diff --git a/security/keys/trusted.c b/security/keys/trusted.c
> index 9a94672e7adc..34f04ffcf2e5 100644
> --- a/security/keys/trusted.c
> +++ b/security/keys/trusted.c
> @@ -389,6 +389,10 @@ static int pcrlock(const int pcrnum)
>  	if (!capable(CAP_SYS_ADMIN))
>  		return -EPERM;
>  
> +	/* This happens if the TPM didn't return random data */
> +	if (!digests)
> +		return -EFAULT;
> +
>  	return tpm_pcr_extend(chip, pcrnum, digests) ? -EINVAL : 0;
>  }
>  
> @@ -1233,10 +1237,8 @@ static int __init init_digests(void)
>  	int i;
>  
>  	ret = tpm_get_random(chip, digest, TPM_MAX_DIGEST_SIZE);
> -	if (ret < 0)
> -		return ret;
> -	if (ret < TPM_MAX_DIGEST_SIZE)
> -		return -EFAULT;
> +	if (ret < 0 || ret < TPM_MAX_DIGEST_SIZE)
> +		return 0;
>  
>  	digests = kcalloc(chip->nr_allocated_banks, sizeof(*digests),
>  			  GFP_KERNEL);
> -- 
> 2.17.1
> 

Re: [PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[Roberto Sassu]

On 8/2/2019 5:30 PM, Tyler Hicks wrote:
> On 2019-08-02 17:07:33, Roberto Sassu wrote:
>> Commit c78719203fc6 ("KEYS: trusted: allow trusted.ko to initialize w/o a
>> TPM") allows the trusted module to be loaded even a TPM is not found to
>                                                     ^ if
> 
>> avoid module dependency problems.
>>
>> However, trusted module initialization can still fail if the TPM is
>> inactive or deactivated. This patch ignores tpm_get_random() errors in
>> init_digests() and returns -EFAULT in pcrlock() if the TPM didn't return
>> random data.
>>
>> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
> 
> The code changes look correct to me.
> 
>    Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
> 
> For whoever takes this patch through their tree, I think that adding the
> following Fixes tag would be useful (as well as cc'ing stable):
> 
>    Fixes: 240730437deb ("KEYS: trusted: explicitly use tpm_chip structure...")
> 
> I think it is also worth leaving a short note, in the commit message,
> for backporters that commit 782779b60faa ("tpm: Actually fail on TPM
> errors during "get random"") should be included with any backports of
> this patch.

Right, thanks. I wait for Jarkko's comments and I add both the Fixes tag
and the short note in the next version of the patch.

Roberto

-- 
HUAWEI TECHNOLOGIES Duesseldorf GmbH, HRB 56063
Managing Director: Li Peng, Li Jian, Shi Yanli

Re: [PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[Roberto Sassu]

On 8/2/2019 5:34 PM, Roberto Sassu wrote:
> On 8/2/2019 5:30 PM, Tyler Hicks wrote:
>> On 2019-08-02 17:07:33, Roberto Sassu wrote:
>>> Commit c78719203fc6 ("KEYS: trusted: allow trusted.ko to initialize 
>>> w/o a
>>> TPM") allows the trusted module to be loaded even a TPM is not found to
>>                                                     ^ if
>>
>>> avoid module dependency problems.
>>>
>>> However, trusted module initialization can still fail if the TPM is
>>> inactive or deactivated. This patch ignores tpm_get_random() errors in
>>> init_digests() and returns -EFAULT in pcrlock() if the TPM didn't return
>>> random data.
>>>
>>> Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
>>
>> The code changes look correct to me.
>>
>>    Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
>>
>> For whoever takes this patch through their tree, I think that adding the
>> following Fixes tag would be useful (as well as cc'ing stable):
>>
>>    Fixes: 240730437deb ("KEYS: trusted: explicitly use tpm_chip 
>> structure...")
>>
>> I think it is also worth leaving a short note, in the commit message,
>> for backporters that commit 782779b60faa ("tpm: Actually fail on TPM
>> errors during "get random"") should be included with any backports of
>> this patch.
> 
> Right, thanks. I wait for Jarkko's comments and I add both the Fixes tag
> and the short note in the next version of the patch.

Uhm, I was thinking that maybe it is not necessary to mention commit
782779b60faa. This patch would still return 0 even if that commit is not
backported (TPM_ERR_DISABLED < TPM_MAX_DIGEST_SIZE).

Roberto

-- 
HUAWEI TECHNOLOGIES Duesseldorf GmbH, HRB 56063
Managing Director: Li Peng, Li Jian, Shi Yanli

Re: [PATCH v2] KEYS: trusted: allow module init if TPM is inactive or deactivated

[Tyler Hicks]

On 2019-08-02 18:11:09, Roberto Sassu wrote:
> On 8/2/2019 5:34 PM, Roberto Sassu wrote:
> > On 8/2/2019 5:30 PM, Tyler Hicks wrote:
> > > On 2019-08-02 17:07:33, Roberto Sassu wrote:
> > > > Commit c78719203fc6 ("KEYS: trusted: allow trusted.ko to
> > > > initialize w/o a
> > > > TPM") allows the trusted module to be loaded even a TPM is not found to
> > >                                                     ^ if
> > > 
> > > > avoid module dependency problems.
> > > > 
> > > > However, trusted module initialization can still fail if the TPM is
> > > > inactive or deactivated. This patch ignores tpm_get_random() errors in
> > > > init_digests() and returns -EFAULT in pcrlock() if the TPM didn't return
> > > > random data.
> > > > 
> > > > Signed-off-by: Roberto Sassu <roberto.sassu@huawei.com>
> > > 
> > > The code changes look correct to me.
> > > 
> > >    Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
> > > 
> > > For whoever takes this patch through their tree, I think that adding the
> > > following Fixes tag would be useful (as well as cc'ing stable):
> > > 
> > >    Fixes: 240730437deb ("KEYS: trusted: explicitly use tpm_chip
> > > structure...")
> > > 
> > > I think it is also worth leaving a short note, in the commit message,
> > > for backporters that commit 782779b60faa ("tpm: Actually fail on TPM
> > > errors during "get random"") should be included with any backports of
> > > this patch.
> > 
> > Right, thanks. I wait for Jarkko's comments and I add both the Fixes tag
> > and the short note in the next version of the patch.
> 
> Uhm, I was thinking that maybe it is not necessary to mention commit
> 782779b60faa. This patch would still return 0 even if that commit is not
> backported (TPM_ERR_DISABLED < TPM_MAX_DIGEST_SIZE).

The commit message for 782779b60faa shows 379 being returned when
attempting to get random:

  [   18.092103] tpm tpm0: A TPM error (379) occurred attempting get random

I don't know enough about TPM chips to know how common that is...

Tyler

> 
> Roberto
> 
> -- 
> HUAWEI TECHNOLOGIES Duesseldorf GmbH, HRB 56063
> Managing Director: Li Peng, Li Jian, Shi Yanli