Commit graph

1193 commits

Author SHA1 Message Date
Ronald Cron
219f978097
Merge pull request #7059 from ronald-cron-arm/psa-crypto-misc
PSA cryptography miscellaneous
2023-04-04 10:54:03 +02:00
Ronald Cron
789cef87f8 Fix documentation
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:07:57 +02:00
Ronald Cron
135f2ae312 Pacify doxygen.sh
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:07:57 +02:00
Ronald Cron
f6236f032a psa: Introduce psa/build_info.h
The PSA cryptography repository is based to
start with on the PSA cryptography implementation
in Mbed TLS but with a different directority
structure, build system and build configuration.

The build-time configuration information in the
PSA cryptography repository is psa/build_info.h.
This commit introduces this file in Mbed TLS to
be used in place of mbedtls/build_info.h (but
basically just an alias to it) in PSA headers.
This is to ease the update of the PSA cryptography
repository.

Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:07:57 +02:00
Ronald Cron
2f10fceccd psa: include: Move key derivation builtin and driver contexts
Move key derivation builtin and driver contexts
to specific header files as key derivation
contexts depend both on primitive and composite
crypto algorithms contexts.

Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:01:45 +02:00
Ronald Cron
e7cde181ea psa: include: Move some Mbed TLS headers
The structures related to CMAC, GCM, CCM and
CHACHAPOLY operations are defined in
crypto_builtin_composites.h not in
crypto_struct.h. Thus move the cmac.h, gcm.h,
ccm.h and chachapoly.h header inclusions from
crypto_struct.h to crypto_builtin_composites.h.

Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:01:45 +02:00
Ronald Cron
fe8e135816 psa: Remove unnecessary headers
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
2023-03-31 09:01:45 +02:00
Valerio Setti
271c12e256 psa: use only PSA_WANT symbols for PSA_VENDOR_ECC_MAX_CURVE_BITS
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
2023-03-29 10:27:18 +02:00
Valerio Setti
8f1e98a971 psa: set PSA_VENDOR_ECC_MAX_CURVE_BITS based on both SW and accelerated support
Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
2023-03-29 10:27:18 +02:00
Przemek Stekiel
c0e6250ff9 Fix documentation and tests
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-14 11:49:36 +01:00
Przemek Stekiel
fde112830f Code optimizations and documentation fixes
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-13 16:28:27 +01:00
Przemek Stekiel
18cd6c908c Use local macros for j-pake slient/server strings
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-12 16:59:28 +01:00
Przemek Stekiel
d7f6ad7bc8 Minor fixes (comments, cleanup)
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-12 16:59:28 +01:00
Przemek Stekiel
e9254a0e55 Adapt driver dispatch documentation for user/peer getters
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>

Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-10 09:18:03 +01:00
Przemek Stekiel
1e7a927118 Add input getters for jpake user and peer
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-10 09:18:03 +01:00
Przemek Stekiel
26c909d587 Enable support for user/peer for JPAKE
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>
2023-03-10 09:18:02 +01:00
Dave Rodgman
bf4016e5d5
Merge pull request #6567 from mprse/ecjpake-driver-dispatch 2023-03-09 19:23:05 +00:00
Przemek Stekiel
57580f2539 Use proper enum types for pake state/sequence/step
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-07 10:50:09 +01:00
Przemek Stekiel
4aa99403f4 Fix configuration for accelerated jpake
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-07 10:50:09 +01:00
Przemek Stekiel
e3ef3a15cd Further pake code optimizations
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-03-06 17:24:32 +01:00
Dave Rodgman
1f39a62ce6
Merge pull request #7151 from gilles-peskine-arm/psa-headers-alt
Allow alternative names for overridable PSA headers
2023-03-03 12:37:51 +00:00
Gilles Peskine
802ff1b116
Merge pull request #7147 from paul-elliott-arm/interruptible_sign_hash_codestyle_drivers
Remove driver entry points for psa_{get|set}_max_ops()
2023-03-01 10:46:09 +01:00
Paul Elliott
15d7d43904 Pacify Clang 15
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>
2023-02-27 17:25:57 +00:00
Paul Elliott
a16ce9f601 Remove driver entry points for {get|set}_max_ops().
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>
2023-02-24 14:44:18 +00:00
Paul Elliott
a3b625b0a1
Merge pull request #7098 from gilles-peskine-arm/retval-non-empty
Pacify Clang 15 about empty \retval
2023-02-24 09:10:53 +00:00
Gilles Peskine
95c915201e Move the implication of MBEDTLS_PSA_CRYPTO_CLIENT where it belongs
If MBEDTLS_PSA_CRYPTO_C is enabled, we always enable
MBEDTLS_PSA_CRYPTO_CLIENT, since the client-side functions are part of the
full PSA crypto feature set. Historically, we didn't have a good place for
configuration modification, so we did this early in the crypto.h include
tree. Since Mbed TLS 3.0, we have mbedtls/build_info.h for that.

Addresses https://github.com/Mbed-TLS/mbedtls/issues/7144 .

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2023-02-23 17:18:33 +01:00
Gilles Peskine
361b5f992f Make sure the configuration is always included
Before, if psa/crypto_platform.h was overridden and the override didn't
include "mbedtls/build_info.h", it was possible to end up with parts of
the headers not taking the library configuration into account, if no
mbedtls header was included before "psa/crypto.h". Make sure that
the mbedtls configuration is visible from the start, no matter what is
or is not in the platform header.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2023-02-23 17:18:33 +01:00
Gilles Peskine
b1176f2583 Allow alternative names for overridden PSA headers
Integrators of Mbed TLS may override the header files
"psa/crypto_platform.h" and "psa/crypto_struct.h" by overwriting the files
or by placing alternative versions earlier in the include file search path.
These two methods are sometimes inconvenient, so allow a third method which
doesn't require overwriting files or having a precise order for the include
path: integrators can now specify alternative names for the headers.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2023-02-22 22:07:28 +01:00
Przemek Stekiel
6b64862ef7 Documentation fixes and code adaptation
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
251e86ae3f Adapt names to more suitable and fix conditional compilation flags
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
ac067d779e Fix PSA_PAKE_OPERATION_INIT macro
Needs to be adapted for the clang -Wall -Wextra.
Requirea to explicitly initialize all the members of the struct that is the first member in the union.

Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
9dd2440c95 Change pake input: key_lifetime -> key attributes
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>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
1c3cfb4fb0 Introduce PSA_PAKE_OPERATION_STAGE_SETUP to optimize out alg checks
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
dde6a910bb Optimize out psa_pake_computation_stage_t
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
38b4e1761d Remove typedef for enum
Workaround for CI error:
Parsing source code...
Compiling...
=============
All symbols in header: PASS
Naming patterns of public_macros: PASS
Naming patterns of internal_macros: PASS
Naming patterns of enum_consts: FAIL

    > include/psa/crypto_extra.h:1857: 'return' does not match the required
      pattern '^(MBEDTLS|PSA)_[0-9A-Z_]*[0-9A-Z]$'.
      |
 1857 |     return cipher_suite->algorithm;
      |     ^^^^^^

Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
ca8d2b2589 Add get-data functions for inputs + tests
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
0f50f689b7 Remove redundant dummy fields inside unions
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
7b730175b3 Simplify psa_pake_computation_stage_s structure
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:32 +01:00
Przemek Stekiel
b09c487546 Combine core pake computation stage(step,sequence,state) into single driver step
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:31 +01:00
Przemek Stekiel
5798003b99 Add enum that presents computation stage as single value
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:31 +01:00
Przemek Stekiel
e12ed36a6c Move JPAKE state machine logic from driver to core
- 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>
2023-02-22 11:30:31 +01:00
Przemek Stekiel
96ae8b939d Move pake definitions to headers for composite algorithms
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:30:31 +01:00
Przemek Stekiel
51eac53b93 Divide pake operation into two phases collecting inputs and computation.
Functions that only set inputs do not have driver entry points.

Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:25:30 +01:00
Przemek Stekiel
d91bcb7673 Refactor structures for pake driver support
Signed-off-by: Przemek Stekiel <przemyslaw.stekiel@mobica.com>
2023-02-22 11:25:30 +01:00
Paul Elliott
21c3951139 Add reference to mbedtls_ecp_set_max_ops() to docs
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-16 12:15:39 +00:00
Paul Elliott
93d9ca83ea Move num_ops ECP abstraction fully into internal implementation
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-16 12:15:39 +00:00
Paul Elliott
fe9e77ff7a Better formatting of include guard comments
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-15 23:34:29 +00:00
Paul Elliott
84329464d5 Replace allocated hash buffer with array
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-15 23:34:29 +00:00
Paul Elliott
c9774411d4 Ensure that operation is put into error state if error occurs
If an error occurs, calling any function on the same operation should return
PSA_ERROR_BAD_STATE, and we were not honouring that for all errors. Add extra
failure tests to try and ratify this.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-15 23:34:29 +00:00
Paul Elliott
f9c91a7fb5 Store the hash, rather than the pointer
For sign and verify, the pointer passed in to the hash is not guaranteed to
remain valid inbetween calls, thus we need to store the hash in the
operation. Added a test to ensure this is the case.

Signed-off-by: Paul Elliott <paul.elliott@arm.com>
2023-02-15 23:34:29 +00:00