Received-SPF: pass (google.com: domain of linux-kernel+bounces-186337-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Precedence: bulk
Mime-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=UTF-8
Date: Wed, 22 May 2024 17:13:41 +0300
Message-Id: <D1G8HOCIDWTC.2ERVA0CYHLY0B@kernel.org>
Cc: <keyrings@vger.kernel.org>, "Peter Huewe" <peterhuewe@gmx.de>, "Jason
 Gunthorpe" <jgg@ziepe.ca>, "Mimi Zohar" <zohar@linux.ibm.com>, "David
 Howells" <dhowells@redhat.com>, "Paul Moore" <paul@paul-moore.com>, "James
 Morris" <jmorris@namei.org>, "Serge E. Hallyn" <serge@hallyn.com>,
 <linux-kernel@vger.kernel.org>, <linux-security-module@vger.kernel.org>
Subject: Re: [PATCH 1/3] tpm: Disable TCG_TPM2_HMAC by default
From: "Jarkko Sakkinen" <jarkko@kernel.org>
To: "James Bottomley" <James.Bottomley@HansenPartnership.com>, "Vitor
 Soares" <ivitro@gmail.com>, <linux-integrity@vger.kernel.org>
References: <20240519235122.3380-1-jarkko@kernel.org>
 <20240519235122.3380-2-jarkko@kernel.org>
 <850862655008f84ef0b6ecd99750e8dc395304d1.camel@gmail.com>
 <D1F4V8NMSUNZ.2VCTEKHZZ0LB@kernel.org>
 <17dc838120b56ce342c34611596c7b46dcd9ab5a.camel@HansenPartnership.com>
 <2dd8d49516ec9c7cb8c1182b5b8537b1e82d7067.camel@gmail.com>
 <17a5dcd7aceb356587ef7c8f45b0f6359b2d2a91.camel@HansenPartnership.com>
In-Reply-To: <17a5dcd7aceb356587ef7c8f45b0f6359b2d2a91.camel@HansenPartnership.com>

On Wed May 22, 2024 at 4:35 PM EEST, James Bottomley wrote:
> On Wed, 2024-05-22 at 09:18 +0100, Vitor Soares wrote:
> > On Tue, 2024-05-21 at 08:33 -0400, James Bottomley wrote:
> > > On Tue, 2024-05-21 at 10:10 +0300, Jarkko Sakkinen wrote:
> > > > This benchmark could be done in user space using /dev/tpm0.
> > >=20
> > > Let's actually try that.=C2=A0 If you have the ibmtss installed, the
> > > command to time primary key generation from userspace on your tpm
> > > is
> > >=20
> > > time tsscreateprimary -hi n -ecc nistp256
> > >=20
> > >=20
> > > And just for chuckles and grins, try it in the owner hierarchy as
> > > well (sometimes slow TPMs cache this)
> > >=20
> > > time tsscreateprimary -hi o -ecc nistp256
> > >=20
> > > And if you have tpm2 tools, the above commands should be:
> > >=20
> > > time tpm2_createprimary -C n -G ecc256
> > > time tpm2_createprimary -C o -G ecc256
> > >=20
> > > James
> > >=20
> > >=20
> >=20
> > Testing on an arm64 platform I get the following results.
> >=20
> > hmac disabled:
> > =C2=A0 time modprobe tpm_tis_spi
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 0m2.776s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.006s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.015s
> >=20
> > =C2=A0 time tpm2_createprimary -C n -G ecc256
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 0m0.686s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.044s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.025s
> >=20
> > =C2=A0 time tpm2_createprimary -C o -G ecc256
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 0m0.638s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.048s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.009s
> >=20
> >=20
> > hmac enabled:
> > =C2=A0 time modprobe tpm_tis_spi
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 8m5.840s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.005s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.018s
> >=20
> >=20
> > =C2=A0 time tpm2_createprimary -C n -G ecc256
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 5m27.678s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.059s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.009s
> >=20
> > =C2=A0 (after first command)
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 0m0.395s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.040s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.015s
> >=20
> > =C2=A0 time tpm2_createprimary -C o -G ecc256
> > =C2=A0 real=C2=A0=C2=A0=C2=A0 0m0.418s
> > =C2=A0 user=C2=A0=C2=A0=C2=A0 0m0.049s
> > =C2=A0 sys=C2=A0=C2=A0=C2=A0=C2=A0 0m0.009s
>
> That's interesting: it suggests the create primary is fast (as
> expected) but that the TPM is blocked for some reason.  Is there
> anything else in dmesg if you do
>
> dmesg|grep -i tpm
>
> ?
>
> Unfortunately we don't really do timeouts on our end (we have the TPM
> do it instead), but we could instrument your kernel with command and
> time sent and returned.  That may tell us where the problem lies.

If there was possibility to use bpftrace it is trivial to get histogram
of time used where. I can bake a script but I need to know first if it
is available in the first place before going through that trouble.

BR, Jarkko