Fix psa_key_derivation_output_key not being able to derive ECC keys
without MBEDTLS_BUILTIN ECC key types enabled.
The PSA crypto drivers can generate these keys without requiring the
builtin key types.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Same note as previous commit regarding guards.
Note that we could auto-enable MD_LIGHT only when SELF_TEST is defined,
and even only when SHA1_C is defined too, but somewhere down the line
we'll want to auto-enable it for the sake of other RSA function (not in
selftest and could use any hash), so there's little point in optimizing
the temporary condition, let's use the simple one upfront.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
But, for now, still guard things with MBEDTLS_MD5_C, as md.c can only
compute MD5 hashes when MBEDTLS_MD5_C is defined. We'll change the
guards once that has changed.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Have clearly separated code to:
* determine whether the assembly-based implementation is available;
* determine whether the intrinsics-based implementation is available;
* select one of the available implementations if any.
Now MBEDTLS_AESNI_HAVE_CODE can be the single interface for aes.c and
aesni.c to determine which AESNI is built.
Change the implementation selection: now, if both implementations are
available, always prefer assembly. Before, the intrinsics were used if
available. This preference is to minimize disruption, and will likely
be revised in a later minor release.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Use a single auxiliary function to determine rk_offset, covering both
setkey_enc and setkey_dec, covering both AESNI and PADLOCK. For AESNI, only
build this when using the intrinsics-based implementation, since the
assembly implementation supports unaligned access.
Simplify "do we need to realign?" to "is the desired offset now equal to
the current offset?".
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The padlock feature is enabled if
```
defined(MBEDTLS_PADLOCK_C) && defined(MBEDTLS_HAVE_X86)
```
with the second macro coming from `padlock.h`. The availability of the
macro `MBEDTLS_PADLOCK_ALIGN16` is coincidentally equivalent to
`MBEDTLS_HAVE_X86` but this is not meaningful.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
On some platforms, including modern Linux, Clang with Msan does not
recognize that explicit_bzero() writes well-defined content to its output
buffer. For us, this causes CMAC operations to fail in Msan builds when
mbedtls_platform_zeroize() is implemented over explicit_bzero(). Fix this.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This will be used in the next commit.
While at it, move driver initialization before RNG init - this will be
handy when the entropy module wants to use drivers for hashes.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
For multi-part operations, we want to make the decision to use PSA or
not only once, during setup(), and remember it afterwards. This supports
the introduction, in the next few commits, of a dynamic component to
that decision: has the PSA driver sub-system been initialized yet?
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
When MBEDTLS_MD_xxx_VIA_PSA is enabled (by mbdetls/md.h), route calls to xxx
over PSA rather than through the built-in implementation.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
When passed an empty OID, mbedtls_oid_get_numeric_string would read one
byte from the zero-sized buffer and return an error code that depends on
its value. This is demonstrated by the test suite changes, which
check that an OID with length zero and an invalid buffer pointer does
not cause Mbed TLS to segfault.
Also check that second and subsequent subidentifiers are terminated, and
add a test case for that. Furthermore, stop relying on integer division
by 40, use the same loop for both the first and subsequent
subidentifiers, and add additional tests.
Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
Clang is nice enough to support bitwise operators on __m128i, but MSVC
isn't.
Also, __cpuid() in MSVC comes from <intrin.h> (which is included via
<emmintrin.h>), not <cpuid.h>.
Signed-off-by: Tom Cosgrove <tom.cosgrove@arm.com>
As of this commit, to use the intrinsics for MBEDTLS_AESNI_C:
* With MSVC, this should be the default.
* With Clang, build with `clang -maes -mpclmul` or equivalent.
* With GCC, build with `gcc -mpclmul -msse2` or equivalent.
In particular, for now, with a GCC-like compiler, when building specifically
for a target that supports both the AES and GCM instructions, the old
implementation using assembly is selected.
This method for platform selection will likely be improved in the future.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The configuration symbol MBEDTLS_AESNI_C requests AESNI support, but it is
ignored if the platform doesn't have AESNI. This allows keeping
MBEDTLS_AESNI_C enabled (as it is in the default build) when building for
platforms other than x86_64, or when MBEDTLS_HAVE_ASM is disabled.
To facilitate maintenance, always use the symbol MBEDTLS_AESNI_HAVE_CODE to
answer the question "can I call mbedtls_aesni_xxx functions?", rather than
repeating the check `defined(MBEDTLS_AESNI_C) && ...`.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Uncrustify indents
```
asm("foo"
HELLO "bar"
"wibble");
```
but we would like
```
asm("foo"
HELLO "bar"
"wibble");
```
Make "bar" an argument of the macro HELLO, which makes the indentation from
uncrustify match the semantics (everything should be aligned to the same
column).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This commit introduce a auxiliary function to zeroize
the cache entry, especially the session structure. The
function is called wherever we need to free the entry.
Signed-off-by: Pengyu Lv <pengyu.lv@arm.com>
Perform debug print of encrypted extensions buffer only after the buffer length was checked successfully
Signed-off-by: Jan Bruckner <jan@janbruckner.de>
The warning is only correct if the assembly code for AESNI is built, not if
MBEDTLS_AESNI_C is activated but MBEDTLS_HAVE_ASM is disabled or the target
architecture isn't x86_64.
This is a partial fix for #7236.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This is only partial support. Only 'client' and 'server' values are accepted for peer and user.
Remove support for role.
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
The function reset_checksum() can be called more than once with the same
handshake context (this happens with DTLS clients, and perhaps in other
cases as well). When that happens, we need to free the old MD contexts
before setting them up again.
Note: the PSA path was already doing the right thing by calling abort,
we just needed to do the same on the MD path.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
With the introduction of #7047, ssl_tls.c uses
mbedtls_md_error_from_psa. This complicates
the dependencies for compiling in psa_to_md_errors,
since now these should be ifdeffed also by
MBEDTLS_USE_PSA_CRYPTO followed by a series of or'ed
MBEDTLS_HAS_ALG_SHA_XXX_VIA_MD_OR_PSA_BASED_ON_USE_PSA.
Since this mechanism will be removed soon, we can simplify it to
just MBEDTLS_USE_PSA_CRYPTO.
Signed-off-by: Andrzej Kurek <andrzej.kurek@arm.com>
GCC 4.6+ complains if a deprecated function calls another.
Working around this universally would require a lot of
preprocessing, this seems to be an easier solution.
Copy mbedtls_pk_error_from_psa code without duplicates
instead of calling the function.
Signed-off-by: Andrzej Kurek <andrzej.kurek@arm.com>
Move all error translation utilities to psa_util.c.
Introduce macros and functions to avoid having
a local copy of the error translating function in
each place.
Identify overlapping errors and introduce a
generic function.
Provide a single macro for all error translations
(unless one file needs a couple of different ones).
Signed-off-by: Andrzej Kurek <andrzej.kurek@arm.com>
Changes for interruptible {sign|verify} hash were not merged at the time of the
previous clang 15 /retval fixes, thus this fixes code added at that time.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
The first round of carry reduction can not generate a carry thus the
secound round is not needed. The comments illustrating when the
carry is 1. The reduction is simmetric so the case when the carry is
-1 is similar.
The illustration is trying to calculate the input value starting with
setting the carry to 1 before the second round of the carry reduction.
It calculates backwords and tries to determine the value range of
each word. It ends up with a contradiction that A10 must have the
value of 0 and UINT32_MAX.
Signed-off-by: Gabor Mezei <gabor.mezei@arm.com>
Move the global variable to the PSA layer, and just set that when calling PSA
level functions.
Move the internal ecp set to before each ecp call.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
The name sha512 might have made sense when it was an
mbedtls_sha512_context, but now it's weird to see things like
mbedtls_md_setup(&sha512, ...MBEDTLS_MD_SHA384...);
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
X.509 already depends on MD_C || USE_PSA_CRYPTO, and this is for the
!USE_PSA_CRYPTO branch, so we're free to use MD.
This change supports our ability to use MBEDTLS_MD_CAN_xxx macros
everywhere in the future, once they have been introduced.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
There's little reason for accessing the hash implementation's internal
state, its output contains most of the same information.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Restore same PSK length enforcement in
conf_psk and set_hs_psk, whether the
negotiated protocol is TLS 1.2 or TLS 1.3.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Change name and document to ensure suitability only for "tags" is explicit. Add
support for output size of zero in PSA_SUCCESS case.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Also in the core part change stage to computation after return from psa_driver_wrapper_pake_setup() regardless of the result. At this point driver context is active even if init has failed.
Additionally handle deallocation of password on failure in mbedtls_psa_pake_setup(). The plan was to handle deallocation on core level by calling abort on failure.
Unfortunately in this case when mbedtls_psa_pake_setup() fails with an unsupported result the built-in implementation is executed (if available) and it will reallocate the password leading to the memory leak.
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
Adapt pake test (passing NULL buffers is not allowed).
Passing the null buffer to psa_pake_output results in a hard fault.
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
In the future key attributes will be available for opaque driver via psa_crypto_driver_pake_get_password_key().
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
Driver is now responsible for creating its own copy of the password in the setup function.
After calling pake setup driver entry point core frees memory for password.
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
- Add `alg` and `computation_stage` to `psa_pake_operation_s`.
Now when logic is moved to core information about `alg` is required.
`computation_stage` is a structure that provides a union of computation stages for pake algorithms.
- Move the jpake operation logic from driver to core. This requires changing driver entry points for `psa_pake_output`/`psa_pake_input` functions and adding a `computation_stage` parameter. I'm not sure if this solution is correct. Now the driver can check the current computation stage and perform some action. For jpake drivers `step` parameter is now not used, but I think it needs to stay as it might be needed for other pake algorithms.
- Removed test that seems to be redundant as we can't be sure that operation is aborted after failure.
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
psa_get_and_lock_key_slot_with_policy() becomes public temporarily as part of:
https://github.com/Mbed-TLS/mbedtls/pull/6608
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
Also while at it, fix debug level for existing DEBUG_RET: errors should
always be level 1.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Found by depends.py MBEDTLS_SHA512_C
In principle, the case where neither SHA-256 nor SHA-384 are available
should never occur, as both TLS 1.2 and TLS 1.3 depend on one of those
being defined. However for now dependencies for TLS 1.2 are not as tight
as they should be; this will be fixed later and is tracked as #6441.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
That's the last family of functions. All calls to mbedtls_sha* and
psa_hash_* in library/ssl_tls.c are now checked for errors.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
On top on some calls not being checked, the PSA path was missing a call
to abort() on errors.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
A few functions were changed from returning void to returning int three
commits ago. Make sure their callers check the return values.
This commits was basically a matter of declaring newly-int-returning
functions MBEDTLS_CHECK_RETURN_CRITICAL and then fixing the resulting
warnings. A few functions had to be made int in the process; they were
applied the same process as well.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This function was manually resetting just the hash that would be used;
it's simpler to just call the function that resets all hashes. This also
avoids calling low-level code from TLS 1.3.
While at it, remove the guards about SHA-256 || SHA-384 that were around
update_checksum, as they are redundant: update_checksum already has
appropriate guards (and TLS 1.3 already depends on one of those tow
hashes being present anyway).
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This part can fail, so it shouldn't be intermixed with the part that
can't fail and is there to ensure all structures are in a clean state,
should any error happen.
Fortunately, the part that should be split out already had a function
doing it: reset_checksum. Also, handshake_params_init had only one
calling site to update.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
There are three family of functions: update_checksum, calc_verify,
calc_finished, that perform hashing operations and were returning void
so far. This is not correct, as hashing functions can return errors (for
example, on hardware failure when accelerated). Change them to return
int.
This commit just changes the types: for now the functions always return
0, and their return value is not checked; this will be fixed in the
next few commits.
There is a related function in TLS 1.3,
mbedtls_ssl_reset_transcript_for_hrr, which also handles hashes, and
already returns int but does not correctly check for errors from hashing
functions so far, it will also be handled in the next few commits.
There's a special case with handshake_params_init: _init functions
should return void, so we'll need to split out the part that can return
errors, see the next commit.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
