Commit graph

15051 commits

Author SHA1 Message Date
Gilles Peskine
f8223abb16 Cover all key types
Generate test cases for all key types. These test cases cover the key
representation (checked with export) and the encoding of the key type and
the bit-size.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
897dff9527 Code to generate storage format test cases
Start generating storage format test cases. This commit introduces two test
data files: test_suite_psa_crypto_storage_format.v0.data for reading keys in
storage format version 0 (the current version at this time), and
test_suite_psa_crypto_storage_format.current.data for saving keys in the
current format (version 0 at this time).

This commit kicks off the test case generation with test cases to exercise
the encoding of usage flags. Subsequent commits will cover other aspects of
keys.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
2352396808 Obtain the values of expressions by running C code
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
e00944807c New python module to encode a PSA key for storage
Construct an object given the attributes and material for a PSA crypto key
and get the Mbed TLS storage representation.

The code to generate the storage representation was written based on the
specification in docs/architecture/mbed-crypto-storage-specification.md,
without looking at the code.

The data in the unit tests is from the AES-128 format_storage_data_check
test case in test_suite_psa_crypto_persistent_key.data, tweaked manually.

This commit creates a basic framework for using symbolic values for
attributes, but does not yet implement obtaining the corresponding numerical
values from an external source.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
00d37d0026 Test code for storage format stability
Save tests are for forward compatibility: import a key in the current format
and check that it has the expected storage format so that future versions
will still be able to read it.

Read tests are for backward compatibility: read a key in the format of a
past version (injected into storage) and check that this version can use it.
Exercise the key unless it is meant to test metadata storage only.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
33c601cb73 Hook up PSAMacroCollector to PSAMacroEnumerator
Make it possible to enumerate the key types, algorithms, etc.
collected by PSAMacroCollector.

This commit ensures that all fields of PSAMacroEnumerator are filled
by code inspection. Testing of the result may reveal more work to be
done in later commits.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
22fcf1b5f5 Move PSAMacroEnumerator to macro_collector
It's useful for more than test_psa_constant_names.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
10ab267afb Add some type annotations
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
6f7ba5f9c4 Split out PSAMacroEnumerator from the test data collection code
Split out the code that enumerates constructors of a PSA crypto type
from the code used to populate the list of constructors for the
specific purpose of testing psa_constant_names.

This commit adds some documentation but otherwise strives to minimize
code changes.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
3d7783909a Use an iterator when constructing test cases
It's more pythonic, and more importantly more readable.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
0298bdae8d generate_psa_tests.py: allow generating each file independently
Generating all files all the time makes debugging one specific target
harder. So support generating a selection of targets only.

As a bonus, it is now more apparent what files this script generates,
and check-generated-files.sh takes advantage of it.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
Gilles Peskine
b94ea51ad0 Break up the god class TestGenerator
Use separate classes for information gathering, for each kind of test
generation (currently just one: not-supported), and for writing output
files.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-03-10 23:22:35 +01:00
David Brown
9b178deedb Fix typo in PSA ECC curve config option
Fix SEC to SECP as the curve name.  This fixes failing tests that
verified the config option was working.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:50 -07:00
David Brown
cc2f869055 Fix generated psa crypto tests
Now that many of these are implemented, update the generated list to
test them.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:50 -07:00
David Brown
3353f81694 Update psa/crypto_config.h to new names
An earlier commit fixes the names of the PSA_WANT_ECC_ macros.  Update
the crypto_config.h file to match these new names.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:50 -07:00
David Brown
dcdde59c6f tests: psa: Change Elliptic curve defines to PSA names
Now that PSA crypto config supports the new PSA_WANT_ECC_xxx defines,
change the psa-specific test suites to use these new names.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:45 -07:00
David Brown
c7b9b2b6d6 Update defines for ECC PSA configs
Use the names as described in
`docs/proposed/psa-conditional-inclusion-c.md which use a transform
like: SECP256R1 -> SECP_R1_256.  The CURVE25519 and CURVE448 become
MONTGOMERY_255 and MONTGOMERY_448.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:13 -07:00
David Brown
20a8c4305a Define PSA_WANT definitions for all ECC curves
Mirror the default non-PSA configuration by enabling all supported ECC
curves.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:13 -07:00
David Brown
8de143e72d Create PSA configs for ECC curves
For each curve defined MBEDTLS_ECP_DP_xxx_ENABLED, we have a
corrsponding PSA config define PSA_WANT_ECC_xxx.  Along with that is a
value MBEDTLS_PSA_ACCEL_ECC_xxx which can be used to allow HW
acceleration of that particular curve.

If the PSA config requests an unaccelerated curve, the corresponding
MBEDTLS_PSA_BUILTIN_ECC_xxx will also be defined.

This commit defines these for all curves currently defined, with the
defines working in either direction, depending on whether
MBEDTLS_PSA_CRYPTO_CONFIG is defined.

Signed-off-by: David Brown <david.brown@linaro.org>
2021-03-10 13:19:13 -07:00
Paul Elliott
b744990717 Fix missed size_t printf
Code was missed due to rework moving duplicated code into a function

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 18:17:12 +00:00
Paul Elliott
9907e2c334 Improve wording of ChangeLog entry
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:14:10 +00:00
Paul Elliott
aa5e132df7 Improve include guards for format attribute
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
61d2209e42 Fix missed invalid specifier in PSA Crypto build
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
48438c758a Fix incorrect long long specifier for win64
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
d48d5c6615 Fix size_t and longlong specifiers for MinGW
MinGW and older windows compilers cannot cope with %zu or %lld (there is
a workaround for MinGW, but it involves linking more code, there is no
workaround for Windows compilers prior to 2013). Attempt to work around
this by defining printf specifiers for size_t per platform for the
compilers that cannot use the C99 specifiers.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
3949065aef Fix incorrect case in changelog entry
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
abb3af7826 Invert gate on printf attribute
Inverted the logic without thinking.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
3891caf1ce Misc review requested fixes
Style fixes and cast certain defines to size_t

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
21c62a2aed Raise level of -Wformat-truncation to 2
This was not possible earlier due to warnings, however now this seems to
be ok.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
3a5a107fa7 Correct fix for potential truncation
This was a false positive caused by the compiler seeing the %08lx
specifiers and judging the output on that, rather than the numbers being
fed in. Given these are going to be maximum 32 bit numbers, then better
to use %08x, which keeps -Wformat-truncation=2 happy as well.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
f8d733e49a Correct include guard for function __attribute
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
9e3256aead Adding printf format warning flags to makefiles
Add printf format warning flags that are supported by GCC 3.0 or later
to the makefiles

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
6f21e11265 Add Changelog entry
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
9f35211774 Fixes for invalid printf format specifiers
Fixes for printf format specifiers, where they have been flagged as
invalid sizes by coverity, and new build flags to enable catching these
errors when building using CMake. Note that this patch uses %zu, which
requires C99 or later.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
dd9e8f6dd0 Fix potential buffer overflow in printf
Printf could potentially produce 2 64 bit numbers here when there is
only space for one, thus causing a buffer overflow. This was caught by
the new warning flags.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
4e589701d8 Declare mbedtls_debug_print_msg as printf-like
We were not getting any warnings on printf format errors, as we do not
explicitly use printf anywhere in the code. Thankfully there is a way
to mark a function as having printf behaviour so that its inputs can be
checked in the same way as printf would be.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
Paul Elliott
ea32d55623 Add printf warning flags to CMake Build
Add extra printf warning flags into the cmake build, only adding those
that are supported by each version of the compiler

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2021-03-10 17:00:32 +00:00
gabor-mezei-arm
9c3b5077e6
Update PSA_CIPHER_ENCRYPT_OUTPUT_SIZE
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
2021-03-10 15:57:44 +01:00
Steven Cooreman
d50db945c4 Add hash acceleration driver testing
Test hash algorithm functions when called through a transparent
driver in all.sh test_psa_crypto_config_basic and
test_psa_crypto_drivers components.

Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:03:36 +01:00
Steven Cooreman
83f300e2f7 Restructure the hash driver content
Apply the right define guards for the right purpose. The 'core' hash
driver is included if any hash algorithm is either to be tested through
the test driver, or if it is requested by a user and not accelerated
(i.e. 'fallback'/'software' driver requested for the algorithm).

Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
25555227e5 Rename hash test driver functions to match auto-naming scheme
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
d029b60770 Move test driver hash function declarations to software driver
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
4f7d0586e1 Setup internal dependency macros for software hash driver
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
0eeb794a2e Initialize status with CORRUPTION_DETECTED and update fallthrough
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
5adf52c72d Correctly void potentially unused arguments
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
b1777312da Make the driver context union a defined type
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
f763810e58 Add test driver for hash operations
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
c8288354a2 move hash update zero-length-input check back into the core
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
dbf8ceda54 Change the way driver context structures are used
Apparently there's a goal to make the PSA Crypto core free from
dynamic memory allocations. Therefore, all driver context structures
need to be known at compile time in order for the core to know their
final size.

This change defines & implements for hashing operations how the context
structures get defined.

Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00
Steven Cooreman
8e9e407fed Clarify documentation of internal hash software driver interface
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
2021-03-10 13:00:25 +01:00