Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <860501de903bba660adcdad3e3ccf750dc29b432.camel@linux.ibm.com>
Subject: Re: [PATCH v4] integrity: Always reference the blacklist keyring
 with appraisal
From:   Mimi Zohar <zohar@linux.ibm.com>
To:     Eric Snowberg <eric.snowberg@oracle.com>
Cc:     dmitry.kasatkin@gmail.com, paul@paul-moore.com, jmorris@namei.org,
        serge@hallyn.com, roberto.sassu@huawei.com,
        kanth.ghatraju@oracle.com, konrad.wilk@oracle.com,
        linux-kernel@vger.kernel.org, linux-integrity@vger.kernel.org,
        linux-security-module@vger.kernel.org
Date:   Thu, 27 Jul 2023 13:36:43 -0400
In-Reply-To: <20230726211725.393056-1-eric.snowberg@oracle.com>
References: <20230726211725.393056-1-eric.snowberg@oracle.com>
Content-Type: text/plain; charset="ISO-8859-15"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Wed, 2023-07-26 at 17:17 -0400, Eric Snowberg wrote:
> Commit 273df864cf746 ("ima: Check against blacklisted hashes for files with
> modsig") introduced an appraise_flag option for referencing the blacklist
> keyring.  Any matching binary found on this keyring fails signature
> validation. This flag only works with module appended signatures.
> 
> An important part of a PKI infrastructure is to have the ability to do
> revocation at a later time should a vulnerability be found.  Expand the
> revocation flag usage to all appraisal functions. The flag is now
> enabled by default. Setting the flag with an IMA policy has been
> deprecated. Without a revocation capability like this in place, only
> authenticity can be maintained. With this change, integrity can now be
> achieved with digital signature based IMA appraisal.
> 
> Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
> ---
> v4 changes:
>   Fixed typo
>   Also did some performance testing with this patch. With the associative array
>   implementation used within the keyring code, there doesn't seem to be much of a
>   difference between doing  an appraisal with an empty blacklist keyring and one
>   containing 1500 entries.  At this time it is unknown how many entries a user
>   would place in this keyring, but this seemed like a bigger number than most use
>   cases.  With the 1500 entries, there were only 7 lookups to get through the
>   entire list for a file not contained within it.  For something that was
>   on the list, there was an average of 4 lookups and a single string compare.
>   Based on this testing, IMHO, there could be a lot more entries added to the
>   blacklist keyring without any real performance issues.

Thank you for the blacklist performance testing.   The patch is now
queued in next-integrity-testing.

--
thanks,

Mimib