Replaces calls to psa_unlock_key_slot with calls to psa_unregister_read.
All instances follow a pattern of a call to psa_get_and_lock_key_slot_X,
followed by some code which reads from a slot, followed by a call to psa_unregister_read.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
This does not yet implement destruction while a key is in use for psa_destroy_key;
that will be implemented in a separate pr.
(I am not sure if I am allowed to change the documentation in the include files.)
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Rename to psa_reserve_free_key_slot, as this function reserves a slot which is
free (not always empty) for filling.
Implement necessary state transitions and state checks.
Rename unlocked_persistent_key_slot to unused_persistent_key_slot.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
This inline function is used in every case we want to change the state of a slot,
except for where we do not care about what the state of the slot was before.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Replaces psa_lock_key_slot and psa_unlock_key_slot.
Future commits will remove the calls to locking/unlocking functions,
and add calls to registering/unregistering functions.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Remove psa_is_key_slot_occupied, any function which calls this can just check
the state variable instead.
Replace psa_is_key_slot_locked with psa_key_slot_has_readers.
References to the now removed functions are changed in future commits.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
Remove the `status` field and replace with the `state` field.
Remove the `lock_count` field and replace with the `registered_readers` field.
Add documentation which describes how and why these fields are to be used.
Signed-off-by: Ryan Everett <ryan.everett@arm.com>
secp224k1 is the one with 225-bit private keys.
The consequences of this mistake were:
* We emitted positive test cases for hypothetical SECP_R1_225 and
SECP_K1_224 curves, which were never executed.
* We emitted useless not-supported test cases for SECP_R1_225 and SECP_K1_224.
* We were missing positive test cases for SECP_R1_224 in automatically
generated tests.
* We were missing not-supported test cases for SECP_R1_224 and SECP_K1_225.
Thus this didn't cause test failures, but it caused missing test coverage
and some never-executed test cases.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
All of ECP requires the bignum module and there is no plan to change that,
so guarding a few bits of code is just noise.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
`$(MBEDTLS_TEST_OBJS)` included TLS-specific test support modules in
`tests/Makefile` but not in `programs/Makefile`. This difference is not
actually necessary. What is necessary is that all programs that use
functions from TLS-specific test support modules are linked with those
modules in addition to `-lmbedtls`, and programs that are not linked with
`-lmbedtls` are not linked with TLS-specific test support modules. Since we
always pass `-lmbedtls` when linking programs in `programs/Makefile`, we can
link with the TLS-specific test support modules as well. This keeps things
simpler.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
fixup "Create common.make with LOCAL_CFLAGS and friends"
The code wasn't what I had intended, although it was functionally
equivalent. Make it more readable and more robust.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
CCM/GCM can be either fully accelerated or rely on just the key type
being accelerated. This means that ultimately it is just the key
type which determines if the legacy block cipher modes need to
be auto-enabled or not.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
Legacy CCM and GCM can work even when AES_C is not defined thanks
to the block_cipher module, so we can relax guards in
cipher_wrap.
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
