Message-ID: <0fa1a12388681b49167b58e349cb7c9e996dcd05.camel@linux.ibm.com>
Date:   Thu, 12 Nov 2020 17:19:29 -0500
From:   Mimi Zohar <zohar@...ux.ibm.com>
To:     Tushar Sugandhi <tusharsu@...ux.microsoft.com>,
        stephen.smalley.work@...il.com, casey@...aufler-ca.com,
        agk@...hat.com, snitzer@...hat.com, gmazyland@...il.com,
        paul@...l-moore.com
Cc:     tyhicks@...ux.microsoft.com, sashal@...nel.org, jmorris@...ei.org,
        nramas@...ux.microsoft.com, linux-integrity@...r.kernel.org,
        selinux@...r.kernel.org, linux-security-module@...r.kernel.org,
        linux-kernel@...r.kernel.org, dm-devel@...hat.com
Subject: Re: [PATCH v5 2/7] IMA: update process_buffer_measurement to
 measure buffer hash

On Thu, 2020-11-12 at 13:47 -0800, Tushar Sugandhi wrote:
> > On Sun, 2020-11-01 at 14:26 -0800, Tushar Sugandhi wrote:
> >> process_buffer_measurement() currently only measures the input buffer.
> >> In case of SeLinux policy measurement, the policy being measured could
> >> be large (several MB). This may result in a large entry in IMA
> >> measurement log.
> > 
> > SELinux is an example of measuring large buffer data.  Please rewrite
> > this patch description (and the other patch descriptions in this patch
> > set) without using the example to describe its purpose [1].
> > 
> > In this case, you might say,
> > 
> > The original IMA buffer data measurement sizes were small (e.g. boot
> > command line), but new buffer data measurement use cases are a lot
> > larger.  Just as IMA measures the file data hash, not the file data,
> > IMA should similarly support measuring the buffer data hash.
> > 
> Sure. Thanks a lot for giving an example wording for us. Will update.
> >>
> >> Introduce a boolean parameter measure_buf_hash to support measuring
> >> hash of a buffer, which would be much smaller, instead of the buffer
> >> itself.
> > 
> >> To use the functionality introduced in this patch, the attestation
> >> client and the server changes need to go hand in hand. The
> >> client/kernel would know what data is being measured as-is
> >> (e.g. KEXEC_CMDLINE), and what data has it’s hash measured (e.g. SeLinux
> >> Policy). And the attestation server should verify data/hash accordingly.
> >>
> >> Just like the data being measured in other cases, the attestation server
> >> will know what are possible values of the large buffers being measured.
> >> e.g. the possible valid SeLinux policy values that are being pushed to
> >> the client. The attestation server will have to maintain the hash of
> >> those buffer values.
> > 
> > Each patch in the patch set builds upon the previous one.   (Think of
> > it as a story, where each chapter builds upon the previous ones.)
> > With rare exceptions, should patches reference subsequent patches. [2]
> > 
> > [1] Refer to Documentation/process/submitting-patches.rst
> > [2] Refer to the section "8) Commenting" in
> > Documentation/process/coding-style.rst
> > 
> I am not sure if you have any concerns about the last two paragraphs.
> The description about the attestation client and server (the last two
> paragraphs) was added for information/clarification purpose only, as per
> your feedback on previous iterations. The subsequent patches don’t have
> any code pertaining to attestation client/server.
> 
> *Question*
> Maybe the last two paragraphs are confusing/redundant. Could you please
> let me know if I should remove the above two paragraphs altogether? 
> (starting with “To use the functionality introduced in this patch ...”)
> 
> If we decide to keep the paragraphs, I will remove the specific examples
> (KEXEC_CMDLINE, SeLinux etc.) as you mentioned elsewhere.

Instead of the above two paragraphs, perhaps explain how measuring the
file data hash differs from measuring the buffer data hash.  Keep the
explanation generic, short and simple.

Mimi

Powered by blists - more mailing lists