[v2,RESEND,0/3] tpm: retrieve digest size of unknown algorithms from TPM

Message

Roberto Sassu Sept. 17, 2018, 9:38 a.m. UTC

Resending to maintainer with correct mailing lists in CC.

The TPM driver currently relies on the crypto subsystem to determine the
digest size of supported TPM algorithms. In the future, TPM vendors might
implement new algorithms in their chips, and those algorithms might not
be supported by the crypto subsystem.

Usually, vendors provide patches for the new hardware, and likely
the crypto subsystem will be updated before the new algorithm is
introduced. However, old kernels might be updated later, after patches
are included in the mainline kernel. This would leave the opportunity
for attackers to misuse PCRs, as PCR banks with an unknown algorithm
are not extended.

This patch set provides a long term solution for this issue. If a TPM
algorithm is not known by the crypto subsystem, the TPM driver retrieves
the digest size from the TPM with a PCR read. All the PCR banks are
extended, even if the algorithm is not yet supported by the crypto
subsystem.

Roberto Sassu (3):
  tpm: rename and export tpm2_digest and tpm2_algorithms
  tpm: modify tpm_pcr_read() definition to pass TPM hash algorithms
  tpm: retrieve digest size of unknown algorithms with PCR read

 drivers/char/tpm/tpm-interface.c    | 24 ++++++++----
 drivers/char/tpm/tpm.h              | 18 ++-------
 drivers/char/tpm/tpm2-cmd.c         | 78 +++++++++++++++++++++++++------------
 include/linux/tpm.h                 | 30 +++++++++++++-
 include/linux/tpm_eventlog.h        |  9 +----
 security/integrity/ima/ima_crypto.c | 10 ++---
 6 files changed, 109 insertions(+), 60 deletions(-)

Comments

Mimi Zohar Sept. 26, 2018, 2:40 p.m. UTC | #1

On Mon, 2018-09-17 at 11:38 +0200, Roberto Sassu wrote:
> Resending to maintainer with correct mailing lists in CC.
> 
> The TPM driver currently relies on the crypto subsystem to determine the
> digest size of supported TPM algorithms. In the future, TPM vendors might
> implement new algorithms in their chips, and those algorithms might not
> be supported by the crypto subsystem.
> 
> Usually, vendors provide patches for the new hardware, and likely
> the crypto subsystem will be updated before the new algorithm is
> introduced. However, old kernels might be updated later, after patches
> are included in the mainline kernel. This would leave the opportunity
> for attackers to misuse PCRs, as PCR banks with an unknown algorithm
> are not extended.
> 
> This patch set provides a long term solution for this issue. If a TPM
> algorithm is not known by the crypto subsystem, the TPM driver retrieves
> the digest size from the TPM with a PCR read. All the PCR banks are
> extended, even if the algorithm is not yet supported by the crypto
> subsystem.

Other than checking the digest size before copying the pcrread buffer,
the patches look good.  Please add my Ack on all 3 patches.

(New address) Acked-by: Mimi Zohar <[e-mail]>

Thanks!

Mimi

Mimi Zohar Sept. 26, 2018, 6:03 p.m. UTC | #2

On Wed, 2018-09-26 at 10:40 -0400, Mimi Zohar wrote:
> On Mon, 2018-09-17 at 11:38 +0200, Roberto Sassu wrote:
> > Resending to maintainer with correct mailing lists in CC.
> > 
> > The TPM driver currently relies on the crypto subsystem to determine the
> > digest size of supported TPM algorithms. In the future, TPM vendors might
> > implement new algorithms in their chips, and those algorithms might not
> > be supported by the crypto subsystem.
> > 
> > Usually, vendors provide patches for the new hardware, and likely
> > the crypto subsystem will be updated before the new algorithm is
> > introduced. However, old kernels might be updated later, after patches
> > are included in the mainline kernel. This would leave the opportunity
> > for attackers to misuse PCRs, as PCR banks with an unknown algorithm
> > are not extended.
> > 
> > This patch set provides a long term solution for this issue. If a TPM
> > algorithm is not known by the crypto subsystem, the TPM driver retrieves
> > the digest size from the TPM with a PCR read. All the PCR banks are
> > extended, even if the algorithm is not yet supported by the crypto
> > subsystem.
> 
> Other than checking the digest size before copying the pcrread buffer,
> the patches look good.  Please add my Ack on all 3 patches.
> 
> (New address) Acked-by: Mimi Zohar <[e-mail]>

I've reviewed, and am currently running with these patches.

Even if the IMA changes were in a separate patch, we wouldn't be able
to break up the patch set anyway.  Jarkko, I'd appreciate your
carrying the entire patch set.

Roberto, a similar change needs to be made for tpm_pcr_extend.  Are
you planning on posting those changes as well?

Mimi

Roberto Sassu Sept. 27, 2018, 6:50 a.m. UTC | #3

On 9/26/2018 8:03 PM, Mimi Zohar wrote:
> On Wed, 2018-09-26 at 10:40 -0400, Mimi Zohar wrote:
>> On Mon, 2018-09-17 at 11:38 +0200, Roberto Sassu wrote:
>>> Resending to maintainer with correct mailing lists in CC.
>>>
>>> The TPM driver currently relies on the crypto subsystem to determine the
>>> digest size of supported TPM algorithms. In the future, TPM vendors might
>>> implement new algorithms in their chips, and those algorithms might not
>>> be supported by the crypto subsystem.
>>>
>>> Usually, vendors provide patches for the new hardware, and likely
>>> the crypto subsystem will be updated before the new algorithm is
>>> introduced. However, old kernels might be updated later, after patches
>>> are included in the mainline kernel. This would leave the opportunity
>>> for attackers to misuse PCRs, as PCR banks with an unknown algorithm
>>> are not extended.
>>>
>>> This patch set provides a long term solution for this issue. If a TPM
>>> algorithm is not known by the crypto subsystem, the TPM driver retrieves
>>> the digest size from the TPM with a PCR read. All the PCR banks are
>>> extended, even if the algorithm is not yet supported by the crypto
>>> subsystem.
>>
>> Other than checking the digest size before copying the pcrread buffer,
>> the patches look good.  Please add my Ack on all 3 patches.
>>
>> (New address) Acked-by: Mimi Zohar <[e-mail]>
> 
> I've reviewed, and am currently running with these patches.
> 
> Even if the IMA changes were in a separate patch, we wouldn't be able
> to break up the patch set anyway.  Jarkko, I'd appreciate your
> carrying the entire patch set.
> 
> Roberto, a similar change needs to be made for tpm_pcr_extend.  Are
> you planning on posting those changes as well?

Yes, I was planning to send the patch after this patch set is accepted.

Roberto

Mimi Zohar Sept. 27, 2018, 1:43 p.m. UTC | #4

On Thu, 2018-09-27 at 08:50 +0200, Roberto Sassu wrote:
> On 9/26/2018 8:03 PM, Mimi Zohar wrote:

> > Roberto, a similar change needs to be made for tpm_pcr_extend.  Are
> > you planning on posting those changes as well?
> 
> Yes, I was planning to send the patch after this patch set is accepted.

Great!

Mimi

Jarkko Sakkinen Sept. 27, 2018, 1:50 p.m. UTC | #5

On Thu, Sep 27, 2018 at 08:50:14AM +0200, Roberto Sassu wrote:
> On 9/26/2018 8:03 PM, Mimi Zohar wrote:
> > On Wed, 2018-09-26 at 10:40 -0400, Mimi Zohar wrote:
> > > On Mon, 2018-09-17 at 11:38 +0200, Roberto Sassu wrote:
> > > > Resending to maintainer with correct mailing lists in CC.
> > > > 
> > > > The TPM driver currently relies on the crypto subsystem to determine the
> > > > digest size of supported TPM algorithms. In the future, TPM vendors might
> > > > implement new algorithms in their chips, and those algorithms might not
> > > > be supported by the crypto subsystem.
> > > > 
> > > > Usually, vendors provide patches for the new hardware, and likely
> > > > the crypto subsystem will be updated before the new algorithm is
> > > > introduced. However, old kernels might be updated later, after patches
> > > > are included in the mainline kernel. This would leave the opportunity
> > > > for attackers to misuse PCRs, as PCR banks with an unknown algorithm
> > > > are not extended.
> > > > 
> > > > This patch set provides a long term solution for this issue. If a TPM
> > > > algorithm is not known by the crypto subsystem, the TPM driver retrieves
> > > > the digest size from the TPM with a PCR read. All the PCR banks are
> > > > extended, even if the algorithm is not yet supported by the crypto
> > > > subsystem.
> > > 
> > > Other than checking the digest size before copying the pcrread buffer,
> > > the patches look good.  Please add my Ack on all 3 patches.
> > > 
> > > (New address) Acked-by: Mimi Zohar <[e-mail]>
> > 
> > I've reviewed, and am currently running with these patches.
> > 
> > Even if the IMA changes were in a separate patch, we wouldn't be able
> > to break up the patch set anyway.  Jarkko, I'd appreciate your
> > carrying the entire patch set.
> > 
> > Roberto, a similar change needs to be made for tpm_pcr_extend.  Are
> > you planning on posting those changes as well?
> 
> Yes, I was planning to send the patch after this patch set is accepted.
> 
> Roberto

Mimi, I can carry it yes.

/Jarkko