* development_2.x:
Reword changelog - Test Resource Leak
Fix fd range for select on Windows
Refactor file descriptor checks into a common function
Update changelog formatting - Missing Free Context
Update changelog formatting Missing Free Context
Update changelog formatting - Missing Free Context
Changelog entry for Free Context in test_suite_aes fix
Free context in at the end of aes_crypt_xts_size()
Fix copypasta in test data
Use UNUSED wherever applicable in derive_input tests
Fix missing state check for tls12_prf output
Key derivation: add test cases where the secret is missing
Add bad-workflow key derivation tests
More explicit names for some bad-workflow key derivation tests
Fix mbedtls_net_poll() and mbedtls_net_recv_timeout() often failing with
MBEDTLS_ERR_NET_POLL_FAILED on Windows: they were testing that the file
descriptor is in range for fd_set, but on Windows socket descriptors are not
limited to a small range. Fixes#4465.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Unrelated to RSA (only used in ECP), but while improving one
mbedtls_safe_cond_xxx function, let's improve the other as well.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
mbedtls_mpi_cf_bool_eq() is a verbatim copy of mbedtls_ssl_cf_bool_eq()
Deduplication will be part of a future task.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Rewrite mbedtls_debug_print_mpi to be simpler and smaller. Leverage
mbedtls_mpi_bitlen() instead of manually looking for the leading
zeros.
Fix#4608: the old code made an invalid memory dereference when
X->n==0 (freshly initialized bignum with the value 0).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Calling mbedtls_mpi_cmp_int reveals the number of leading zero limbs
to an adversary who is capable of very fine-grained timing
measurements. This is very little information, but could be practical
with secp521r1 (1/512 chance of the leading limb being 0) if the
adversary can measure the precise timing of a large number of
signature operations.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The idiom "resize an mpi to a given size" appeared 4 times. Unify it
in a single function. Guarantee that the value is set to 0, which is
required by some of the callers and not a significant expense where
not required.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Since the internal function mpi_fill_random_internal() assumes that X
has the right size, there is no need to call grow().
To further simplify the function, set the sign outside, and zero out
the non-randomized part directly.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In real life, min << N and the probability that mbedtls_mpi_random()
fails to find a suitable value after 30 iterations is less than one in
a billion. But at least for testing purposes, it's useful to not
outright reject "silly" small values of N, and for such values, 30
iterations is not enough to have a good probability of success.
Pick 250 iterations, which is enough for cases like (min=3, N=4), but
not for cases like (min=255, N=256).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This comment is no longer in the specific context of generating a
random point on an elliptic curve.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_mpi_random() uses mbedtls_mpi_cmp_mpi_ct(), which requires its
two arguments to have the same storage size. This was not the case
when the upper bound passed to mbedtls_mpi_random() had leading zero
limbs.
Fix this by forcing the result MPI to the desired size. Since this is
not what mbedtls_mpi_fill_random() does, don't call it from
mbedtls_mpi_random(), but instead call a new auxiliary function.
Add tests to cover this and other conditions with varying sizes for
the two arguments.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Instead of generating blinding values and keys in a not-quite-uniform way
(https://github.com/ARMmbed/mbedtls/issues/4245) with copy-pasted code,
use mbedtls_mpi_random().
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
dhm_make_common includes a piece of code that is identical to
dhm_random_below except for returning a different error code in one
case. Call dhm_random_below instead of repeating the code.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
P-1 is as bad as 1 as a blinding value. Don't accept it.
The chance that P-1 would be randomly generated is infinitesimal, so
this is not a practical issue, but it makes the code cleaner. It was
inconsistent to accept P-1 as a blinding value but not as a private key.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Unify the common parts of mbedtls_dhm_make_params and mbedtls_dhm_make_public.
No intended behavior change, except that the exact error code may
change in some corner cases which are too exotic for the existing unit
tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Instead of generating blinding values in a not-quite-uniform way
(https://github.com/ARMmbed/mbedtls/issues/4245) with copy-pasted
code, use mbedtls_mpi_random().
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Since mbedtls_mpi_random() is not specific to ECC code, move it from
the ECP module to the bignum module.
This increases the code size in builds without short Weierstrass
curves (including builds without ECC at all) that do not optimize out
unused functions.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Rename mbedtls_ecp_gen_privkey_sw to mbedtls_mpi_random since it has
no particular connection to elliptic curves beyond the fact that its
operation is defined by the deterministic ECDSA specification. This is
a generic function that generates a random MPI between 1 inclusive and
N exclusive.
Slightly generalize the function to accept a different lower bound,
which adds a negligible amount of complexity.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_ecp_gen_privkey_mx generates a random number with a certain
top bit set. Depending on the size, it would either generate a number
with that top bit being random, then forcibly set the top bit to
1 (when high_bit is not a multiple of 8); or generate a number with
that top bit being 0, then set the top bit to 1 (when high_bit is a
multiple of 8). Change it to always generate the top bit randomly
first.
This doesn't make any difference in practice: the probability
distribution is the same either way, and no supported or plausible
curve has a size of the form 8n+1 anyway. But it slightly simplifies
reasoning about the behavior of this function.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Don't calculate the bit-size of the initially generated random number.
This is not necessary to reach the desired distribution of private
keys, and creates a (tiny) side channel opportunity.
This changes the way the result is derived from the random number, but
does not affect the resulting distribution.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The library rejected an RNG input of all-bits-zero, which led to the
key 2^{254} (for Curve25519) having a 31/32 chance of being generated
compared to other keys. This had no practical impact because the
probability of non-compliance was 2^{-256}, but needlessly
complicated the code.
The exception was added in 98e28a74e3 to
avoid the case where b - 1 wraps because b is 0. Instead, change the
comparison code to avoid calculating b - 1.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
For Montgomery keys, n_bits is actually the position of the highest
bit and not the number of bits, which would be 1 more (fence vs
posts). Rename the variable accordingly to lessen the confusion.
No semantic change.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Put the Montgomery and short Weierstrass implementations of
mbedtls_ecp_gen_privkey into their own function which can be tested
independently, but will not be part of the public ABI/API.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When mbedtls_nist_kw_wrap was called with output=NULL and out_size=0, it
performed arithmetic on the null pointer before detecting that the output
buffer is too small and returning an error code. This was unlikely to have
consequences on real-world hardware today, but it is undefined behavior and
UBSan with Clang 10 flagged it. So fix it (fix#4025).
Fix a similar-looking pattern in unwrap, though I haven't verified that it's
reachable there.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The persistence level PSA_KEY_PERSISTENCE_READ_ONLY can now only be used
as intended, for keys that cannot be modified through normal use of the API.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_psa_get_stats() was written back before lifetimes were
structured as persistence and location. Fix its classification of
volatile external keys and internal keys with a non-default
persistence.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
On space-constrained platforms, it is a useful configuration to be able
to import/export and perform RSA key pair operations, but to exclude RSA
key generation, potentially saving flash space. It is not possible to
express this with the PSA_WANT_ configuration system at the present
time. However, in previous versions of Mbed TLS (v2.24.0 and earlier) it
was possible to configure a software PSA implementation which was
capable of making RSA signatures but not capable of generating RSA keys.
To do this, one unset MBEDTLS_GENPRIME.
Since the addition of MBEDTLS_PSA_BUILTIN_KEY_TYPE_RSA_KEY_PAIR, this
expressivity was lost. Expressing that you wanted to work with RSA key
pairs forced you to include the ability to generate key pairs as well.
Change psa_crypto_rsa.c to only call mbedtls_rsa_gen_key() if
MBEDTLS_GENPRIME is also set. This restores the configuration behavior
present in Mbed TLS v2.24.0 and earlier versions.
It left as a future exercise to add the ability to PSA to be able to
express a desire for a software or accelerator configuration that
includes RSA key pair operations, like signature, but excludes key pair
generation.
Without this change, linker errors will occur when attempts to call,
which doesn't exist when MBEDTLS_GENPRIME is unset.
psa_crypto_rsa.c.obj: in function `rsa_generate_key':
psa_crypto_rsa.c:320: undefined reference to `mbedtls_rsa_gen_key'
Fixes#4512
Signed-off-by: Jaeden Amero <jaeden.amero@arm.com>
TLS code specific to SHA-384 was gated on MBEDTLS_SHA512_C. But SHA-384 also
requires that MBEDTLS_SHA512_NO_SHA384 is not defined. This lead to dead
code in TLS when MBEDTLS_SHA512_C and MBEDTLS_SHA512_NO_SHA384 were both
defined (i.e. when SHA-512 was enabled but not SHA-384).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
They depended on MBEDTLS_SHA512_C only. A check for !MBEDTLS_SHA512_NO_SHA384
was missing.
Fix#4499.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The sequence of calls starts-update-starts-update-finish is not a
guaranteed valid way to abort an operation and start a new one. Our
software implementation just happens to support it, but alt
implementations may very well not support it.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Session-ID based session resumption requires that the resumed session
is consistent with the client's ClientHello in terms of choice of
ciphersuite and choice of compression.
This check was previously assumed to be performed in the session cache
implementation, which seems wrong: The session cache should be an id-based
lookup only, and protocol specific checks should be left to Mbed TLS.
This commit
- adds an explicit ciphersuite and compression consistency check after
the SSL session cache has been queried
- removes the ciphersuite and compression consistency check from
Mbed TLS' session cache reference implementation.
Don't use ssl_check_xxx() for functions with void return
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
Prepare to isolate the Montgomery and short Weierstrass
implementations of mbedtls_ecp_gen_privkey into their own function.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
A previous fix in d596ca8a1e worked with
beta versions of GCC 11, but not with the final 11.1 release.
This time, just disable the warning locally.
Fix#4130
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Zephyr's native posix port define _POSIX_C_SOURCE with a higher value
during the build, so when mbedTLS defines it with a different value
breaks the build.
As Zephyr is already defining a higher value is guaranteed that mbedTLS
required features will be available. So, just define it in case it was
not defined before.
[taken from Zephyr mbedtls module:
76dcd6eeca]
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Signed-off-by: David Brown <david.brown@linaro.org>
Since they became equivalent after moving the is_sign checking back to
the PSA core, they're now redundant, and the generic mac_setup function
can just be called directly.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Since a valid mac operation context would guarantee that the stored
mac size is >= 4, it wasn't immediately obvious that the zero-length
check is meant for static analyzers and a bit of robustness.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The PSA core checks the key type and algorithm combination before
calling the driver, so the driver doesn't have to do this once more.
The PSA core will also not start an operation with a requested length
which is larger than the full MAC output size, so the output length check
in the driver isn't needed as long as the driver returns an error on
mac_setup if it doesn't support the underlying hash algorithm.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
It makes sense to do the length checking in the core rather than expect
each driver to deal with it themselves. This puts the onus on the core to
dictate which algorithm/key combinations are valid before calling a driver.
Additionally, this commit also updates the psa_mac_sign_finish function
to better deal with output buffer sanitation, as per the review comments
on #4247.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
As psa_mac_sign_finish / psa_mac_verify_finish already checks that the
operation structure is valid (id is non-zero), the driver itself doesn't
have to check for that anymore. If the operation has a driver ID assigned,
it means that driver has returned success from its setup function, so the
algorithm value will be set correctly.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The key passed to the driver has been imported by the PSA Core, meaning
its length has already been verified, and the driver can rely on the
buffer length and key attributes being consistent.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
This means there is no longer a need to have an internal HMAC API, so
it is being removed in this commit as well.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Prefix with 'mbedtls_psa' as per the other types which implement some
sort of algorithm in software.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Now renamed to mbedtls_psa_safer_memcmp, it provides a single location
for buffer comparison.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The purpose of key_set was to guard the operation structure from being
used for update/finish before a key was set. Now that the implementation
fully adheres to the PSA API, that function is covered by the `alg`
variable instead. It's set to the algorithm in use when a key is set, and
is zero when the operation is reset/invalid.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
* Early return since there's nothing to clean up
* Get rid of unnecessary local variable
* Check algorithm validity for MAC in the PSA core instead of in the driver
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Apparently it was at some point assumed that there would be
support for MAC algorithms with IV, but that hasn't been
implemented yet. Until that time, these context structure
members are superfluous and can be removed.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Typedef'ed structures are suffixed _t
Also updated the initialiser macro with content that actually
matches the structure's content.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Since HMAC moved into its own compilation unit, the internal API needed
to be documented and finalized. This means no more reaching deep into
the operation structure from within the PSA Crypto core. This will make
future refactoring work easier, since internal HMAC is now opaque to the
core.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Step 3/x in moving the driver. Separate commits should make for easier
review.
Additional changes on top of code movement:
* Copied the implementation of safer_memcmp from psa_crypto into
psa_cipher_mac since the mac_verify driver implementation
depends on it, and it isn't available through external linkage
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Step 2/x in moving the driver. Separate commits should make for easier
review.
Additional changes on top of code movement:
* Early-return success on input with zero-length to mac_update, to
avoid NULL pointers getting passed into the driver dispatch
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Step 1/x in moving the driver. Separate commits should make for easier
review.
Additional changes on top of just moving code:
* Added a sanity check on the key buffer size for CMAC.
* Transfered responsibility for resetting the core members of the
PSA MAC operation structure back to the core (from the driver
wrapper layer)
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
This is a temporary measure. Other operations in the PSA Core which rely
on this internal HMAC API should be rewritten to use the MAC API instead,
since they can then leverage accelerated HMAC should a platform provide
such acceleration support.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The changed logic is to try a sign-message driver (opaque or transparent);
if there isn't one, fallback to builtin sofware and do the hashing,
then try a sign-hash driver. This will enable to the opaque driver
to fallback to software.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
To avoid code duplication of the old-style SE interface usage
call psa_driver_wrapper_sign/verify_hash function instead of
the direct internal functions.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
Use common funtion for psa_sign_hash and psa_sign_message and one for
psa_verify_hash and psa_verify_message to unify them.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
Move the key buffer size calculation code under
tests to avoid check-names.sh to complain about
"likely macros with typos".
This removes the calculation of key buffer
sizes for the test driver from the wrapper based on
static size data. But the code is still there in test
code to be used when we go back to work on the
generation of the driver wrapper.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
pk_get_pk_alg will either return 0 or a pk error code. This means that
the error code will always be a high level module ID and so we just
return ret.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Add implementation for MBEDTLS_PSA_CRYPTO_BUILTIN_KEYS
Merging as it has been ready for four days now and I prefer not having to go through other rebases especially given the coming change of scope of development (3.0 rather than 2.2x).
Improve comments explaining error code checking, fix incorrect comments
and make a small formatting fix.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
* group setting of attributes before calling get_builtin_key
* return early instead of going to exit when no resources are allocated yet
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Leverage the fact that the get_builtin_key entrypoint returns a key's
attributes, such that a proper size for the builtin key's buffer can
be calculated through the driver's get_key_buffer_size hook.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Since the loading attempt of a builtin key might be followed by trying
to load a persistent key, we can only wipe the allocated key data, not
the associated metadata.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The macro always meant 'location', but was mistakenly named 'lifetime'.
Naming it location instead makes much more sense, and drives home the
conceptual differences between location and lifetime values.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Instead of the full attributes struct, it now only takes/returns what it
actually needs to.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Builtin key support for the test driver is always compiled in, and no
longer guarded by MBEDTLS_PSA_CRYPTO_BUILTIN_KEYS.
Parsing the key slot from the buffer by cast and assign instead of memcmp.
For exporting keys, the test driver no longer reaches into the key
identifier in order to check whether a key is builtin, but rather
assumes so based on the key buffer length. It's the driver's
responsibility to be able to detect the key material it returned as part
of the get_builtin_key operation.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
As part of test_psa_crypto_drivers, define a builtin symmetric
plus an ECC key on the test driver lifetime.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
According to the design in psa-driver-interface.md. Compiles without
issue in test_psa_crypto_drivers.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
This reverts commit c75d9f589b.
This was merged by mistake in development instead of development_3.0.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This reverts commit 48f6d0d6e5.
This was merged by mistake in development instead of development_3.0.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This reverts commit 9a86843470.
This was merged by mistake in development instead of development_3.0.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
An incorrect error code addition was spotted by the new invasive testing
infrastructure whereby pk_get_pk_alg will always return a high level
error or zero and pk_parse_key_pkcs8_unencrypted_der will try to add
another high level error, resulting in a garbage error code.
Apply the same fix from ae3741e8a to fix the bug.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Replace all occurences of error code addition in the library with the new
MBEDTLS_ERROR_ADD macro.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Add a missing guard for the definition and declaration of
mbedtls_test_hook_error_add.
Also make the declaration always visible when MBEDTLS_TEST_HOOKS is
enabled. This fixes an issue when MBEDTLS_ERROR_C is not defined but
MBEDTLS_TEST_HOOKS is.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Fix unused parameter warnings when MBEDTLS_TEST_HOOKS is not enabled.
A few issues were caught by check-names.sh namely:
- mbedtls_error_add was not capitalised.
- mbedtls_test_hook_error_add was being defined multiple times as the
definition was in a header.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
The previous implementation of the error addition interface did not comply
with the invasive testing architecture guidelines. This commit fixes that
by:
- Renaming functions/macros/variables to follow the mbedtls_error_xxx or
mbedtls_test_hook_xxx convention.
- Making mbedtls_test_hook_error_add a global variable that can be set
by the testing code.
- Using a static inline function call, as opposed to macro, to keep
discrepancies between debug and production version to a minimum.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Change the signature of the `hook` parameter of `mbedtls_set_err_add_hook`
to use the actual signature of the function as opposed to `void *`. This
fixes a warning when compiling with clang `-pedantic`.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
`error.c` and error.h are the more logical place to keep this code and it
prevents issues with building `common.c` and conflicts with other projects
that use mbedtls (such as mbedOS).
`error.c` has been automatically generated by first adding the code to
`error.fmt` and then running `./scripts/generate_errors.pl`.
Also add parenthesis to the addition in `MBEDTLS_ERR_ADD`.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Fix building by adding `common.c` to the build scripts (both make and Cmake).
Also reworks the hook function pointer (also renamed to `err_add_hook`) to be
a static local to `common.c` with a setter function to set the pointer to a
checking function.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
`error.c` is a file generated from `error.h` and thus cannot contain the code
that was previously added. This commit fixes that issue by moving the
`MBEDTLS_ERR_ADD` macro and associated function and function pointer into
`common.h` and `common.c`.
Also fix a typo in `tests/include/test/helpers.h` where tabs were accidentally
used instead of spaces.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
All occurences of manual error code addition/combination, in `rsa.c`, have
been replaced with the `MBEDTLS_ERR_ADD` macro.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Adds a macro (`MBEDTLS_ERR_ADD`) to add error codes together and check that the
result will not be corrupted. This additional check is only enabled during
testing when `MBEDTLS_TEST_HOOKS` is defined.
Also includes a reference usage example in `rsa.c` where two high-level error
codes could be incorrectly added together under the right conditions. This now
ensures that when this error occurs during testing it will be correctly
reported.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Corresponds better to the validation done in other modules of PSA Crypto.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Capitalise the MPS trace macros, as per the coding style (and make a slight
change to naming convention to avoid a name collision).
Signed-off-by: Dave Rodgman <dave.rodgman@arm.com>
As we want to do Mbed TLS aead operations as a
driver does, aead operations should not access
the key slot as key slots are not available to
drivers.
Second step in this PR: do not unlock the key slot
as part of operation abort.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
As we want to do Mbed TLS aead operations as a
driver does, aead operations should not access
the key slot as key slots are not available to
drivers.
First step in this PR: move key resolution from
aead operation setup to psa_aead_encrypt/decrypt
APIs.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Fix the subtraction in fix_negative, which was incorrectly not looking
for a carry. This caused the result to be wrong when the least
significant limb of N was 0. Fix#4296.
The bug was introduced by d10e8fae9e
"Optimize fix_negative". Thanks to Philippe Antoine (catenacyber) for
reporting the bug which was found by his EC differential fuzzer.
Credit to OSS-Fuzz.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This header file will contain declarations of functions that are not
part of the public ABI/API, and must not be called from other modules,
but can be called from unit tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Move the handling of the sign out of the base-specific loops. This
both simplifies the code, and corrects an edge case: the code in the
non-hexadecimal case depended on mbedtls_mpi_mul_int() preserving the
sign bit when multiplying a "negative zero" MPI by an integer, which
used to be the case but stopped with PR #2512.
Fix#4295. Thanks to Guido Vranken for analyzing the cause of the bug.
Credit to OSS-Fuzz.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In a TLS client, enforce the Diffie-Hellman minimum parameter size
set with mbedtls_ssl_conf_dhm_min_bitlen() precisely. Before, the
minimum size was rounded down to the nearest multiple of 8.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When creating a persistent key or registering a key
with an invalid key identifier return
PSA_ERROR_INVALID_ARGUMENT instead of
PSA_ERROR_INVALID_HANDLE.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
This check was added earlier to avoid useless calls to `memcpy()`
with length `0` in the _frequent_ case where we're not accumulating.
By now, the whole code path has been moved to a branch which is only
executed if the reader is accumulating, and the only time this check
would be relevant is if we happen to feed an empty fragment to the
reader. In this case, the call to memcpy() could be removed, but
since this case is exceptional and the call to memcpy() is still
correct even for a length 0 copy, we remove the check for simplicity
of the code.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
Previously, the semantics of mbedtls_mps_reader_commit() was to invalidate
all buffers previously fetched via mbedtls_mps_reader_get(), forbidding
any further use by the 'consumer'. This was in fact a necessary constraint
for the current implementation, which did some memory moving in
mbedtls_mps_reader_commit().
This commit simplifies the reader's semantics and implementation in
the following way:
- API: A call to mbedtls_mps_reader_commit() does no longer invalidate
the buffers previously obtained via mbedtls_mps_reader_get().
Instead, they can continue to be used until
mbedtls_mps_reader_reclaim() is called.
Calling mbedtls_mps_reader_commit() now only sets a marker
indicating which parts of the data received through
mbedtls_mps_reader_get() need not be backed up once
mbedtls_mps_reader_reclaim() is called. Allowing the user
to call mbedtls_mbedtls_reader_commit() multiple times
before mbedtls_mps_reader_reclaim() is mere convenience:
We'd get exactly the same functionality if instead of
mbedtls_mps_reader_commit(), there was an additional argument
to mbedtls_mps_reader_reclaim() indicating how much data
to retain. However, the present design is more convenient
for the user and doesn't appear to introduce any unnecessary
complexity (anymore), so we stick with it for now.
- Implementation: mbedtls_mps_reader_commit() is now a 1-liner,
setting the 'commit-marker', but doing nothing else.
Instead, the complexity of mbedtls_mp_reader_reclaim()
slightly increases because it has to deal with creating
backups from both the accumulator and the current
fragment. In the previous implementation, which shifted
the accumulator content with every call to
mbedtls_mps_reader_commit(), only the backup from the
fragment was necessary; with the new implementation
which doesn't shift anything in
mbedtls_mps_reader_commit(), we need to do the
accumulator shift in mbedtls_mps_reader_reclaim().
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit adds an implementation of the MPS trace module
based on `printf()`.
The enabling macro MBEDTLS_MPS_TRACE remains unset by default
because MPS tracing is very verbose and consumes unnecessary
space in the CI.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit adds an internal header `library/mps/error.h` related
to error codes in MPS.
For now, those error codes can be considered internal and thus we
don't have to avoid clashes with other Mbed TLS error codes. This
is OK as long as it's true that MPS isn't public API, and its error
codes are never forwarded to the return values of public API calls.
The error code allocation of MPS will likely need revisiting over time.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
Most buffers that MPS deals with are small and representable
with integer types of width 16-bit or more.
For highly memory constrained systems, it is therefore a potential
for significant memory savings to use 16-bit types for buffer sizes
throughout MPS.
In prepraration for this, this commit introduces typdefs
```
mbedtls_mps_size_t
mbedtls_mps_stored_size_t
```
for buffer sizes in the MPS implementation and the MPS structures,
respectively.
So far, those MUST be defined as `size_t`: While an effort has been made
to write most of MPS code in terms of `mbedtls_mps_[stored_]size_t` in a
way that would allow narrower types, those aren't yet supported. Still,
we retain the typedefs in order to avoid unnecessary rewriting of a large
body of the MPS codebase.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit adds the interface fo the MPS reader component as
`library/mps/reader.h`.
Please see the file itself for extensive documentation.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
MPS' tracing module uses four macros:
1) TRACE( type, fmt, ... )
This acts like `printf( fmt, ... )` but also allows
the specification of a type of trace output (comment,
warning, error, ...)
2) TRACE_INIT
This acts like TRACE() but increases the level of
indentation. It will be used at the beginning of
function calls.
3) RETURN( val )
Equivalent to `return( val )` plus a decrement in the
level of indentation. This should be used at the end of
functions that have been started with TRACE_INIT.
4) TRACE_END
This combines a trace output with a decrement of the
level of indentation. It's necessary prior to leaving
functions which have been started with TRACE_INIT
but which don't have a return value.
This commit defines those macros as no-op dummies in
`library/mps/trace.h` for now.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit adds an MPS-specific compile-time option
`MBEDTLS_MPS_TRACE` to the internal MPS header `mps/common.h`.
So far -- this may need revisiting -- MPS comes with its own
internal tracing module which allows to track the operation of
MPS' various layers for the purpose of understanding of it workings
as well as for debugging.
The reasons for the introduction of a module separate from SSL debug
are the following:
1) The SSL debug module requires an SSL context to function because
debug callbacks are part of the runtime configuration of the SSL
module.
The MPS tracing module, in contrast, is not supposed to be used
in production environments, and there is no need for a runtime
configuration. Instead, a compile-time defined tracing callback
is used.
2) In the interest of modularity, MPS' tracing module shouldn't
require having an SSL context around.
3) Purely visually, MPS' tracing module adds support for indentation
according to call-depth and coloring according to which module is
being used, which makes it very useful for what's going on; however,
those features aren't available in the SSL debug module (and they
shouldn't be).
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit adds the compile-time option MBEDTLS_MPS_ENABLE_ASSERTIONS
which controls the presence of runtime assertions in MPS code.
See the documentation in the header for more information.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
This commit introduces the internal MPS header `mps/common.h`
which will subsequently be populated with MPS-specific compile-time
options and helper macros. For now, it's a stub.
Signed-off-by: Hanno Becker <hanno.becker@arm.com>
Remove cipher_generate_iv driver entry point as there
is no known use case to delegate this to a driver.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
The IV length computed in the cipher PSA implementation is
the default IV length thus use the PSA macro PSA_CIPHER_IV_LENGTH
defined to do that.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Use psa_generate_random() to generate IVs instead of
mbedtls_psa_get_random(). mbedtls_psa_get_random() is
meant to be used as the f_rng argument of Mbed TLS
library functions.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Symmetric key management is not intended to be
delegated to drivers. Thus, key management code
for a given symmetric key type should be included
in the library whether or not the support for
cryptographic operations based on that type of
symmetric key may be delegated to drivers.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Isolate the Mbed TLS cipher driver interfaces.
Do the actual cipher operations in utility
functions that are just called by the interface
functions.
The utility functions are intended to be also called
by the cipher test driver interface functions (to be
introduced subsequently) and allow to test the case
where cipher operations are fully accelerated with no
fallback (component test_psa_crypto_config_basic of
all.sh).
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
As per drivers, pass to the Mbed TLS implementation of
the cipher multi-part operation its operation context
and not the PSA operation context.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Move members that are of no use to the PSA crypto core
to the Mbed TLS implementation specific operation context.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
For cipher multi-part operations, dispatch based on
the driver identifier even in the case of the
Mbed TLS software implementation (viewed as a driver).
Also use the driver identifier to check that an
cipher operation context is active or not.
This aligns the way hash and cipher multi-part
operations are dispatched.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Change the operation context to the PSA one to be
able to call the software implementation from
the driver wrapper later on.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Change the signature of
psa_driver_wrapper_cipher_encrypt/decrypt_setup to
that of a PSA driver cipher_encrypt/decrypt_setup
entry point.
Change the operation context to the PSA one to be
able to call the software implementation from
the driver wrapper later on.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Split out the cipher setup based on cipher.c
in psa_cipher_setup_internal() whose signature
is that of a PSA driver cipher_setup entry
point.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Rework psa_cipher_setup in preparation of
calling the cipher setup based on cipher.c
through the interface of a PSA driver
cipher_setup entry point.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Add psa_crypto_cipher.[ch] files to contain the
Mbed TLS implementation of PSA driver cipher driver
entry points.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
To run succesfully the test
"PSA sign: invalid algorithm for ECC key" of
test_suite_psa_crypto when ECDSA support is not included
in the library, always return INVALID_ARGUMENT
in case of an ECC key not used for ECDSA, whether
ECDSA support is present or not.
Then apply the same logic to RSA sign RSA and RSA/ECC
verify for the sake of consistency.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Add missing PSA_WANT_CCM/GCM/CMAC. This completes
the set of PSA_WANT config options given the
current support of PSA crypto in Mbed TLS.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
If an elliptic curve was enabled in the Mbed TLS classic API (#define
MBEDTLS_ECP_DP_xxx), but not enabled in the PSA configuration (#define
PSA_WANT_ECC_xxx), it would still work if you tried to use it through
PSA.
This is generally benign, but could be a security issue if you want to
disable a curve in PSA for some security reason (such as a known bug
in its implementation, which may not matter in the classic API if Mbed
TLS is running in a secure enclave and is only reachable from
untrusted callers through the PSA API). More urgently, this broke
test_suite_psa_crypto_not_supported.generated.
So if a curve is not enabled in the PSA configuration, ensure that
it's treated as unsupported through the PSA software implementation.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Return PSA_ERROR_INVALID_HANDLE instead of
PSA_ERROR_DOES_NOT_EXIST if invalid key is passed for some key
operations.
Signed-off-by: Maulik Patel <Maulik.Patel@arm.com>
The hash driver entry points (and consequentially the hash driver core)
are now always compiled on when PSA_CRYPTO_DRIVER_TEST is turned on.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
The PSA Core is already calling psa_hash_abort, so the driver doesn't
have to do that explicitly.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Drivers (both built-in and external) need to declare their context
structures in a way such that they are accessible by the
to-be-autogenerated crypto_driver_contexts.h file. That file lives in
include/psa, which means all builtin driver context structure
declarations also need to live in include/psa.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
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>
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>
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>
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>
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>
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>
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>
mac size is previously checked to not be less than 4, so it can't be zero
anymore at this point.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Memsan build was reporting a false positive use of uninitialised memory
in x509_crt.c on a struct filled by an _stat function call. According to
the man pages, the element reported has to be filled in by the call, so
to be safe, and keep memsan happy, zero the struct first.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Fix function mbedtls_ecp_mul_shortcuts() to skip multiplication when m
is 0 and simply assignt 0 to R. Additionally fix ecjpake_zkp_read() to
return MBEDTLS_ERR_ECP_INVALID_KEY when the above condintion is met.
Fix#1792
Signed-off-by: TRodziewicz <rodziewicz@gmail.com>
When ECDSA is not supported by the library, prefer
to return NOT_SUPPORTED than INVALID_ARGUMENT when
asked for an ECDSA signature.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Move the check that ECDSA is supported from the
caller of the function responsible for Mbed TLS
ECDSA signatures to this function, namely
mbedtls_psa_ecdsa_sign_hash().
This makes the caller code more readable and is
more aligned with what is expected from a
sign_hash() PSA driver entry point.
Add a negative test case where a deterministic
ECDSA signature is requested while the library
does not support deterministic ECDSA.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Reworked the validation of MAC algorithm with the used key type by
introducing psa_mac_key_can_do, which guarantees that PSA_MAC_LENGTH can
be called successfully after validation of the algorithm and key type.
This means psa_get_mac_output_length is no longer required.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Multiplication is not constant flow on any CPU we are generally
targetting, so replace this with bit twiddling.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Comparing algorithm with its FULL_LENGTH_MAC version doesn't work in
cases where algorithm is a wildcard. Wildcard input is not specified in
the documentation of the function, but in order to test the function
using the same test as PSA_MAC_LENGTH we're mimicking that behaviour here.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
Avoid code duplication. Also update the guarantees made by the function
doc to match the guarantees given by PSA_MAC_LENGTH.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
This makes it more in-line with how psa_key_policy_permits works. It
also adds consistency: the intersection of MAC with default length and
MAC with exact-length is now computed correctly in case the exact length
equals the default length of the algorithm when used with the given
key type.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
