Re: [PATCH v5 3/4] ima_tpm.sh: Fix calculating boot aggregate

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Mimi,

TL;DR: thank you for comments, I update commit message.
Details below.

> Hi Petr,

> On Mon, 2020-12-14 at 23:19 +0100, Petr Vorel wrote:
> > for TPM 2.0 or kernel >= v5.8-rc1:
> > 6f1a1d103b48 ima: ("Switch to ima_hash_algo for boot aggregate")

> > Test still fails with newer TPM 2.0 on kernel < v5.8-rc1.

> The above commit was backported in stable.  Do you know if the failing
> systems backported the above patch?
SLES kernel got it.

> I've recently asked for commit
> 20c59ce010f8 ("ima: extend boot_aggregate with kernel measurements")
> also be backported.
I see, it got backported to v5.4 (LTS).

FYI shell API doesn't have yet support for hint for commits to backport
(my TODO: #700 [1]). Once it's implemented, I'll add 20c59ce010f8 as this tag.

[1] https://github.com/linux-test-project/ltp/issues/700

> > Test was failing, because it expect SHA1 (we ignore MD5) hash, but for TPM 2.0
> > is now used IMA default hash algorithm (by default default SHA256).
> > This is similar for entries in IMA measurement list so we can reuse
> > already existing code.

> > Reading other algorithms than SHA1 or support TPM 2.0 requires evmctl
> > >= 1.3.1 (1.3 would also work for test1, but will be required for test2).

> > Although recent evmctl is recommended, to support older kernels and TPMs
> > which support only SHA1, get boot aggregate with old our legacy
> > ima_boot_aggregate.c.

> ^ the LTP legacy ima_boot_aggregate.c still works, without the evmctl
> dependency.
Yes, I meant in my description this LTP legacy ima_boot_aggregate.c.
Test does not require evmctl for SHA1 hash.

So maybe:
Although recent evmctl is recommended, for older kernels and TPMs
which support only SHA1 is still used the legacy ima_boot_aggregate.c
(no evmctl update required).

> > Also fixed cases:
> > * testing with no TPM device:
> > * TPM TPM 2.0 devices which does not export event log
> > (/sys/kernel/security/tpm0/binary_bios_measurements).

> ^ firmware which does not export the TPM 2.0 binary event log
+1

> > Also fixed test without TPM device (when IMA TPM-bypass is tested)
> > as some TPM 2.0 devices does not export event log
> > (/sys/kernel/security/tpm0/binary_bios_measurements).

> This looks like a duplicate of above.  Maybe just add another bullet
> *
> detecting IMA TPM-bypass mode
+1

> > This does not require evmctl at all.

> I assume this comment refers to TPM 2.0 calculating the
> "boot_aggregate" based on the existing PCR values, as opposed to TPM
> 1.2 which first walks the TPM event log, calculating the PCRs.
No, that's meant for IMA TPM-bypass (test1). Because we only check zero.
Both TPM 1.2 and TPM 2.0 require evmctl for reading PCR-10 (test2).
So just a note for * detecting IMA TPM-bypass mode.

> > Also try best to detect TPM major version (1, 2 or none - assume
> > TPM-bypass). This fixes testing with TPM 2.0 device which does not
> > export event log (/sys/kernel/security/tpm0/binary_bios_measurements):
> > not wrongly assuming TPM-bypass when kernel didn't export other TPM
> > 2.0 files we check in get_tpm_version() but bios boot aggregate is
> > correct (i.e. not 0x00s). In that case evmctl ima_boot_aggregate can get
> > boot aggregate even without TPM event log.

> > Co-developed-by: Mimi Zohar <zohar@xxxxxxxxxxxxx>
> > Signed-off-by: Petr Vorel <pvorel@xxxxxxx>

> Thanks, Petr!
Thanks a lot for patient review and info.

Kind regards,
Petr

> Mimi



[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Linux Kernel]     [Linux Kernel Hardening]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]

  Powered by Linux