Received-SPF: pass (google.com: domain of linux-kernel+bounces-81068-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Precedence: bulk
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=UTF-8
Date: Mon, 26 Feb 2024 11:38:39 +0200
Message-Id: <CZEWS8QBYSF9.120NE1F4L75VT@suppilovahvero>
Cc: "Ross Philipson" <ross.philipson@oracle.com>, "Kanth Ghatraju"
 <kanth.ghatraju@oracle.com>, "Peter Huewe" <peterhuewe@gmx.de>
Subject: Re: [PATCH 1/3] tpm: protect against locality counter underflow
From: "Jarkko Sakkinen" <jarkko@kernel.org>
To: "Lino Sanfilippo" <l.sanfilippo@kunbus.com>, "Daniel P. Smith"
 <dpsmith@apertussolutions.com>, "Alexander Steffen"
 <Alexander.Steffen@infineon.com>, "Jason Gunthorpe" <jgg@ziepe.ca>, "Sasha
 Levin" <sashal@kernel.org>, <linux-integrity@vger.kernel.org>,
 <linux-kernel@vger.kernel.org>
References: <20240131170824.6183-1-dpsmith@apertussolutions.com>
 <20240131170824.6183-2-dpsmith@apertussolutions.com>
 <CYU3CFW08DAA.29DJY7SJYPJJZ@suppilovahvero>
 <2ba9a96e-f93b-48e2-9ca0-48318af7f9b1@kunbus.com>
 <ae3fecc4-7b76-4607-8749-045e17941923@infineon.com>
 <91f600ef-867b-4523-89be-1c0ba34f8a4c@kunbus.com>
 <8692fcf6-2e67-45b4-b809-7723f30736a2@apertussolutions.com>
 <75642d53-9daa-4be7-8d3d-d1d86bc60b3e@kunbus.com>
In-Reply-To: <75642d53-9daa-4be7-8d3d-d1d86bc60b3e@kunbus.com>

On Sat Feb 24, 2024 at 4:34 AM EET, Lino Sanfilippo wrote:
>
>
> On 23.02.24 02:56, Daniel P. Smith wrote:
>
> >>
> >> Will the TPM TIS CORE ever (have to) request another locality than 0? =
Maybe the best would
> >> be to hardcode TPM_ACCESS(0) and get rid of all the locality parameter=
s that are
> >> passed from one function to another.
> >> But this is rather code optimization and not really required to fix th=
e reported bug.
> >=20
> > Actually, doing so will break the TPM API. The function
> > tpm_tis_request_locality() is registered as the locality handler,
> > =C2=A0int (*request_locality)(struct tpm_chip *chip, int loc), in the t=
is
> > instance of struct tpm_class_ops{}. This is the API used by the Secure
> > Launch series to open Locality2 for the measurements it must record.
> >=20
>
> I dont understand this. How do you use locality 2 with the current mainli=
ne
> API? Do you adjust the mainline code to use locality 2 instead of 0? This=
 would=20
> at least explain how you ran into the underflow issue which from
> the source code seems to be impossible when using locality 0. But then I =
wonder why
> this has not been made clear in this discussion. And then we are talking
> about fixing a bug that does not even exist in the upstream code.=20

Thanks for bringing this up, now I finally figured out what confuses me
in this series.

Daniel, I also have troubles understanding why locality_count would ever
be greater than zero exactly in the mainline kernel, *without* [1]?

[1] https://lore.kernel.org/linux-integrity/20240214221847.2066632-1-ross.p=
hilipson@oracle.com/

BR, Jarkko