linux-kernel - RE: Portable Executable (PE) Signature Validation and Measurement for KEXEC system call using IMA

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: <DM6PR21MB13056915DA6AFF16C3753915C35E0@DM6PR21MB1305.namprd21.prod.outlook.com>
Date:   Mon, 25 Mar 2019 21:00:36 +0000
From:   Lakshmi Ramasubramanian <nramas@...rosoft.com>
To:     Mimi Zohar <zohar@...ux.ibm.com>,
        "linux-integrity@...r.kernel.org" <linux-integrity@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: Portable Executable (PE) Signature Validation and Measurement for
 KEXEC system call using IMA

Thanks for your response Mimi.

Please see my responses to your queries inline. 

Thanks,
 -lakshmi

>-----Original Message-----
>From: Mimi Zohar <zohar@...ux.ibm.com> 
>Sent: Monday, March 25, 2019 12:28 PM
>To: Lakshmi Ramasubramanian <nramas@...rosoft.com>; linux-integrity@...r.kernel.org; linux-kernel@...r.kernel.org
>Subject: Re: Portable Executable (PE) Signature Validation and Measurement for KEXEC system call using IMA

>Hi Lakshmi,

>On Fri, 2019-03-22 at 17:39 +0000, Lakshmi Ramasubramanian wrote:
>> Hello,
>> 
>> When loading the new kernel image file for executing KEXEC system call, 
>> we would like to verify that the kernel image file is signed and 
>> the signer certificate is valid.
>>

> I'm not sure what is meant by "and the signer certificate is valid".

> The kexec kernel image signature can be verified by keys either on the
> IMA keyring or the platform keyring.  The current method of verifying
> keys being added to the IMA keyring is by requiring them to be signed
> by a key on the builtin trusted keyring.  This provides a signature
> chain of trust from boot to the kernel, based on a HW root of trust,
> and then transitions to the kernel image's embedded keys.  You
> probably already know as to why/how the platform keys are trusted.

[Lakshmi] By "signer certificate is valid" I meant, the key used to sign the kexec kernel image is trusted. 
                  We will verify the signature using the key in the IMA keyring 
                  (like the way you have described above)

>> 
>> If the kernel image file is in Portable Executable (PE) format we want to 
>> validate the PE Signature and measure the signer X.509 certificate 
>> (Extend as part of IMA Template defaulting to PCR 10, if not otherwise set, 
>>  and the IMA measurement log).

> How/when do you plan to "measure the signer X.509 certificate"?  Is
> this when the certificate is being loaded onto the keyring or at use?
> I'm not sure how much of the certificate is available once loaded
> onto the keyring.

[Lakshmi] We will measure the signer certificate when the file is loaded - in other words, "at use".
The kernel image signer information measured in to the IMA log will then be used as one of the attestation criteria.

>> 
>> We plan to use Integrity Measurement Architecture (IMA) for the above.
>> 
>> Please let us know if anyone is already working on a patch set
>> for such a functionality.
>> 
>> I am aware of the work that Thiago Jung Bauermann @ IBM is doing for 
>> "Appended signatures support for IMA appraisal" 
>> (Web link given below)
>> 
>>     https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Flkml.org%2Flkml%2F2018%2F12%2F12%2F1049&amp;data=02%7C01%7Cnramas%40microsoft.com%7Cc36d42cb91214cc3ca1308d6b157f962%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636891388745548879&amp;sdata=OLtkuptXQYZHuvlVmn3ej%2FpEk501TxzToEcbOref0UU%3D&amp;reserved=0

> Other than Thiago, I'm not aware of anyone else working on this.
>  Thiago is actively working on these patches and will be re-posting
> them shortly.

[Lakshmi] Thanks for the information.

> Mimi

Thanks,
 -lakshmi