When writing the new .data file, first write the new content, then replace
the target. This way, there isn't a temporary state in which the file is
partially written. This temporary state can be misleading if the build is
interrupted. It's annoying if you're watching changes to the output and the
changes appear as emptying the file following by the new version appearing.
Now interrupted builds don't leave a file that appears to be up to date but
isn't, and when watching the output, there's a single transition to the new
version.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Add a .gitattributes file that tells GitHub to highlight all .function
files as if they were .c files. This aids in reviewing changes to tests.
Signed-off-by: David Horstmann <david.horstmann@arm.com>
Instead of fully validating beforehand
signature algorithms with regards to the
private key, do minimum validation and then
just try to compute the signature. If it
fails try another reasonable algorithm if any.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
When providing proof of possession of
an RSA private key, allow the usage for RSA
PSS signatures of a hash with a security
level lower that the security level of the
RSA private key.
We did not allow this in the first place to
align with the ECDSA case. But as it is not
mandated by the TLS 1.3 specification (in
contrary to ECDSA), let's allow it.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
When selecting the server certificate based on
the signature algorithms supported by the client,
check the signature algorithms as close as possible
to the way they are checked to compute the
signature for the server to prove it possesses
the private key associated to the certificate.
That way we minimize the odds of selecting a
certificate for which the server will not be
able to compute the signature to prove it
possesses the private key associated to the
certificate.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
The EC J-PAKE module the ability to "fall back" to PSA when MD is not
present a few PRs ago, but the dependency of this key exchange on
SHA-256 wasn't updated at the time.
(Note: the crypto primitive doesn't depend on SHA-256, only its use in
the TLS key exchange does.)
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Stating from the default config means a few things are implicitly
excluded; starting from the full config makes it all fully explicit.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
CMakeLists.txt was calling generate_psa_tests.py and siblings to list the
generated test data files with a --directory option, intended the output to
be this argument textually. This used to work, but no longer does, because
the --directory argument is relative to the current directory when the
Python script is invoked, and the script now shows an absolute path.
CMakeLists.txt now completely ignores the directory part of the listed data
file paths and builds its own. The base_xxx_files variables now contain
actual base names, without a "suites/" prefix. This makes it more robust
with respect to the behavior of the Python script, but it will break if
we put data files in multiple different directories one day.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When pylint runs in parallel, it loses the ability to detect duplicated code
across modules. Duplicated code is usually a bad thing, so give pylint the
opportunity to let us know.
This makes pylint slightly slower, but going from 2 threads to 1 does not
make it anywhere close to twice as slow. On my machine, with Python 3.5,
pylint -j2 takes about 12s while single-threaded pylint takes about 16s
of wall clock time.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
We had 4 identical copies of the check_repo_path function. Replace them by a
single copy in the build_tree module where it naturally belongs.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_vsnprintf replacement works like mbedtls_snprintf replacement, so
copy the requirements for MBEDTLS_PLATFORM_VSNPRINTF_ALT.
(MBEDTLS_PLATFORM_xxx_MACRO shouldn't require MBEDTLS_PLATFORM_C, but that's
a separate preexisting problem which I do not try address at this time.)
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
We were using absolute imports under the assumption that the /scripts
directory is in the path. This worked in normal use because every one of our
Python scripts either were in the /scripts directory, or added the /scripts
directory to the module search path in order to reference mbedtls_dev.
However, this broke things like
```
python3 -m unittest scripts/mbedtls_dev/psa_storage.py
```
Fix this by using relative imports.
Relative imports are only supposed to be used inside a package (Python
doesn't complain, but Pylint does). So make /scripts/mbedtls_dev a proper
package by creating __init__.py.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The option to --directory was intended to be relative to the current
directory when the script is invoked, which is the intuitive behavior. But
this was not implemented correctly, and it was actually interpreted relative
to the mbedtls root (which the script chdir's into). Fix this.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
