linux-kernel - Re: [PATCH v0 2/2] IMA: Call queue functions to measure keys

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <e89dcb1c-c463-919a-aabb-e285d884a914@linux.microsoft.com>
Date:   Tue, 3 Dec 2019 08:09:20 -0800
From:   Lakshmi Ramasubramanian <nramas@...ux.microsoft.com>
To:     Mimi Zohar <zohar@...ux.ibm.com>, linux-integrity@...r.kernel.org
Cc:     eric.snowberg@...cle.com, dhowells@...hat.com,
        matthewgarrett@...gle.com, sashal@...nel.org,
        jamorris@...ux.microsoft.com, linux-kernel@...r.kernel.org,
        keyrings@...r.kernel.org
Subject: Re: [PATCH v0 2/2] IMA: Call queue functions to measure keys

Thanks for reviewing the changes Mimi. I'll address your comments in the 
next update.

> 
> Overwriting the initial policy is highly recommended, but not everyone
> defines a custom policy.  Should there be a time limit or some other
> criteria before deleting the key measurement queue?
> 
> Mimi

For the above, I feel checking for the presence of custom policy, if 
that is possible, would be a more deterministic approach compared to 
having a time limit.

On my machine, systemd loads the custom IMA policy from 
/etc/ima/ima-policy if that file is present. Is this the recommended way 
to configure custom IMA policy? If yes, can the IMA initialization 
function check for the presence of the above file?

Another way to address this issue is to define a new CONFIG parameter to 
determine whether or not to support deferred processing of keys. If this 
config is chosen, custom IMA policy must be defined.

Least preferred option would be to leave the queued keys as is if custom 
policy is not defined - at the cost of, maybe, a non-trivial amount of 
kernel memory consumed.

If detection of custom policy is not possible, then define a timer to 
drain the key measurement queue.

Please let me know which approach you think is optimal.

thanks,
  -lakshmi