[PATCH 2/6] LSM: Infrastructure management of the key security blob

[John Johansen]

On 7/9/24 15:08, Paul Moore wrote:
> On Jul  8, 2024 Casey Schaufler <casey at schaufler-ca.com> wrote:
>>
>> Move management of the key->security blob out of the
>> individual security modules and into the security
>> infrastructure. Instead of allocating the blobs from within
>> the modules the modules tell the infrastructure how much
>> space is required, and the space is allocated there.
> 
> Perhaps mention that the key_free hook is being removed as it is not
> currently needed after this change?
> 
>> Signed-off-by: Casey Schaufler <casey at schaufler-ca.com>
>> ---
>>   include/linux/lsm_hooks.h         |  1 +
>>   security/security.c               | 41 +++++++++++++++++++++++++++++--
>>   security/selinux/hooks.c          | 23 +++++------------
>>   security/selinux/include/objsec.h |  7 ++++++
>>   security/smack/smack.h            |  7 ++++++
>>   security/smack/smack_lsm.c        | 33 +++++++++++--------------
>>   6 files changed, 75 insertions(+), 37 deletions(-)
> 
> ...
> 
>> diff --git a/security/security.c b/security/security.c
>> index 5e93a72bdca6..aae37481b7be 100644
>> --- a/security/security.c
>> +++ b/security/security.c
>> @@ -227,6 +227,9 @@ static void __init lsm_set_blob_sizes(struct lsm_blob_sizes *needed)
>>   		blob_sizes.lbs_inode = sizeof(struct rcu_head);
>>   	lsm_set_blob_size(&needed->lbs_inode, &blob_sizes.lbs_inode);
>>   	lsm_set_blob_size(&needed->lbs_ipc, &blob_sizes.lbs_ipc);
>> +#ifdef CONFIG_KEYS
>> +	lsm_set_blob_size(&needed->lbs_key, &blob_sizes.lbs_key);
>> +#endif
> 
> Since the lsm_blob_sizes struct is going to have the lsb_key field
> regardless of CONFIG_KEYS (which is good, I'm not arguing that), we
> should be okay to call lsm_set_blob_size() on the lsb_key field, right?

yes

> 
>>   	lsm_set_blob_size(&needed->lbs_msg_msg, &blob_sizes.lbs_msg_msg);
>>   	lsm_set_blob_size(&needed->lbs_sock, &blob_sizes.lbs_sock);
>>   	lsm_set_blob_size(&needed->lbs_superblock, &blob_sizes.lbs_superblock);
>> @@ -402,6 +405,9 @@ static void __init ordered_lsm_init(void)
>>   	init_debug("file blob size       = %d\n", blob_sizes.lbs_file);
>>   	init_debug("inode blob size      = %d\n", blob_sizes.lbs_inode);
>>   	init_debug("ipc blob size        = %d\n", blob_sizes.lbs_ipc);
>> +#ifdef CONFIG_KEYS
>> +	init_debug("key blob size        = %d\n", blob_sizes.lbs_key);
>> +#endif /* CONFIG_KEYS */
> 
> This one makes sense.
> 
>> @@ -5301,7 +5337,8 @@ int security_key_alloc(struct key *key, const struct cred *cred,
>>    */
>>   void security_key_free(struct key *key)
>>   {
>> -	call_void_hook(key_free, key);
>> +	kfree(key->security);
>> +	key->security = NULL;
>>   }
> 
> A note to future devs, we can add the key_free hook back if needed, but
> currently nobody is using it beyond basic memory management.
> 
>>   /**
>> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
>> index 19346e1817ff..b3de2e941ef7 100644
>> --- a/security/selinux/hooks.c
>> +++ b/security/selinux/hooks.c
>> @@ -6981,6 +6968,9 @@ struct lsm_blob_sizes selinux_blob_sizes __ro_after_init = {
>>   	.lbs_file = sizeof(struct file_security_struct),
>>   	.lbs_inode = sizeof(struct inode_security_struct),
>>   	.lbs_ipc = sizeof(struct ipc_security_struct),
>> +#ifdef CONFIG_KEYS
>> +	.lbs_key = sizeof(struct key_security_struct),
>> +#endif /* CONFIG_KEYS */
> 
> We can probably get rid of the Kconfig conditional.  I understand the
> desire to keep this to only what is needed, but since this only really
> has an impact on boot, and the impact is some basic math, I'd rather
> not run the risk of rot due to conditional compilation.
> 

sure, the counter argument is why isn't this conditional when other parts
for keys is. That inconsistency introduces it own issues that can cause
problems down the road. I really don't have an opinion on which way
is better, just playing devils advocate.


>>   	.lbs_msg_msg = sizeof(struct msg_security_struct),
>>   	.lbs_sock = sizeof(struct sk_security_struct),
>>   	.lbs_superblock = sizeof(struct superblock_security_struct),
> 
> ...
> 
>> diff --git a/security/smack/smack_lsm.c b/security/smack/smack_lsm.c
>> index a931b44bc959..17bcc9cbf584 100644
>> --- a/security/smack/smack_lsm.c
>> +++ b/security/smack/smack_lsm.c
>> @@ -5010,6 +5005,9 @@ struct lsm_blob_sizes smack_blob_sizes __ro_after_init = {
>>   	.lbs_file = sizeof(struct smack_known *),
>>   	.lbs_inode = sizeof(struct inode_smack),
>>   	.lbs_ipc = sizeof(struct smack_known *),
>> +#ifdef CONFIG_KEYS
>> +	.lbs_key = sizeof(struct smack_known *),
>> +#endif /* CONFIG_KEYS */
> 
> See above, but ultimately this is Smack code so it's your call.

I think we should be consistent, either it goes everywhere or it stays
everywhere. It is going to be confusing for other people coming in and
looking at the code as to why one is conditional and one isn't.

> 
>>   	.lbs_msg_msg = sizeof(struct smack_known *),
>>   	.lbs_sock = sizeof(struct socket_smack),
>>   	.lbs_superblock = sizeof(struct superblock_smack),
> 
> --
> paul-moore.com