* This commit includes changes to the provisioner for what is sent up. Originally only SHA256 was being used, this change includes both.
* This last commit cover the items 2-4 in issue #236. The Provisioner sends up and updated list of pcrs that include 256, not just sha1. The validation and policy pages have been updated. A second pull request will be created to address parsing the information into a baseline.
* This commit includes functioning TPM quote code that is sent to the ACA. In addition it has code to also sent the pcrs list results.
Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.Additional changes to correct code for sending the pcr list over to the ACA.
* Changed the requirement for the field into protobuf to optional from required.
The tpm_version tool can sometimes print a null
value as part of the TPM major version if the major
version is less than 4 characters. These nulls are
now removed before printing.
The TPM 2 Provisioner gets packaged with an application called
tpm_version, which reported hard-coded values for the TPM
manufacturer and version. Now it collects those things from
the TPM and reports them.
There were some warnings generated by Doxygen when compiling.
One was from an undocumented parameter for a method in the
Properties class. Many others were due to a setting that prevented
Doxygen from being able to associate function declarations with
their definitions if their parameters used standard library objects.
There was a problem in the rpm-post-install.sh script
that ran as part of the CentOS7 rpm installation where
a link was being created called libcurl.so which pointed
to libcurl.so.4. If the link could not be created because
it already existed, the script would quit before finishing
and never place hirs-provisioner-tpm2 in a directory on
the PATH.
The proper solution was to link hirs-provisioner against
libcurl.so.4 so that it is clear which version of the API
was compiled against. This was not happening because
we were linking against a version of curl build by the CPR
project which was not properly embedding the SONAME in the
shared object file. By linking instead against the shared
object file distributed in the development package of
libcurl, hirs-provisioner-tpm2 now looks for libcurl.so.4
rather than the generic libcurl.so. This will prevent our
executable from breaking if libcurl.so gets updated to point
to a newer version of libcurl that uses a different API.
Closes#78.
A couple of variables had conflicting names and the inner scope was
shadowing the outer. Style checker was complaining. Deleted one inner
definition and renamed another variable.
* [#46] Ensure Travis mounts repository rather than clones it in Docker
* [#46] Containerize HIRS ACA and prep ACA container for Integration Tests
* [#46] Containerize HIRS TPM2Provisioner and prep TPM2Provisioner container for Integration Tests
* [#46] Replace localinstall with install
* [#46] Prevent rebuilding of packages unnecessarily
* [#46] Finish initial docker compose setup for integration tests
* [#46] Allow for detection of complete Integration Environment Setup
* [#46] Fix Travis CI to allow for detecting Integ Test Environ Stand-Up
* [#46] Fix Initial Integration Test Script
* [#46] Troubleshoot Integration Test script
