Commit Graph

16554 Commits

Author SHA1 Message Date
Gilles Peskine
9110809c46 Fix mistake in the sample implementation of MBEDTLS_CHECK_RETURN
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-30 19:24:36 +02:00
Gilles Peskine
f315d6a718
Merge pull request #4957 from gilles-peskine-arm/make-tests-v-2.2x
Backport 2.x: Facilitate reproducing unit tests from all.sh
2021-09-30 16:08:56 +02:00
Przemyslaw Stekiel
6c0ec0eb52 Adapt return status to 2.x
Signed-off-by: Przemyslaw Stekiel <przemyslaw.stekiel@mobica.com>
2021-09-30 15:51:05 +02:00
Przemyslaw Stekiel
7ab374f6b5 Add change-log: fix-mbedtls_cipher_crypt-aes-ecb.txt
Signed-off-by: Przemyslaw Stekiel <przemyslaw.stekiel@mobica.com>
2021-09-30 13:12:22 +02:00
Przemyslaw Stekiel
39f4e75120 Address review comments
Signed-off-by: Przemyslaw Stekiel <przemyslaw.stekiel@mobica.com>
2021-09-30 12:29:27 +02:00
Przemyslaw Stekiel
f0fa86eee5 Add PSA support for MBEDTLS_CIPHER_AES_128_ECB
Signed-off-by: Przemyslaw Stekiel <przemyslaw.stekiel@mobica.com>
2021-09-30 12:29:22 +02:00
Przemyslaw Stekiel
6c866689e3 test_suite_cipher: add tests for mbedtls_cipher_setup_psa() with ECB
Signed-off-by: Przemyslaw Stekiel <przemyslaw.stekiel@mobica.com>
2021-09-30 12:27:34 +02:00
Ronald Cron
8489f64f37
Merge pull request #4991 from gilles-peskine-arm/psa-aead-invalid_nonce_length-2.x
2.x only: PSA AEAD invalid nonce length tests
2021-09-29 22:47:03 +02:00
Mateusz Starzyk
15a7420d3c Silence warnings about unused return value
This macro is introduced here for use in deprecated functions. It may also
be useful in user code, so it is in a public header.

Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:57:33 +02:00
Gilles Peskine
8472a10594 New configuration option MBEDTLS_CHECK_RETURN_WARNING
MBEDTLS_CHECK_RETURN_TYPICAL defaults off, but is enabled if
MBEDTLS_CHECK_RETURN_WARNING is enabled at compile time.
(MBEDTLS_CHECK_RETURN_CRITICAL is always enabled.)

The default is off so that a plausible program that builds with one version
of Mbed TLS in the default configuration will still build under the next
version.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:42:51 +02:00
Gilles Peskine
e568ebade1 Use reserved identifier for warn_unused_result
This is normally equivalent, but works even if some other header defines a
macro called warn_unused_result.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:41:01 +02:00
Gilles Peskine
6b5c0f0e44 Better default for MBEDTLS_CHECK_RETURN in config.h
An empty expansion is possible, but as documented its effect is to disable
the feature, so that isn't a good example. Instead, use the GCC
implementation as the default: it's plausible that it could work even on
compilers that don't advertise themselves as sufficiently GCC-like to define
__GNUC__, and if not it gives users a concrete idea of what the macro is
supposed to do.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:40:55 +02:00
Gilles Peskine
ce555e4fad Change DES and AES functions to MBEDTLS_CHECK_RETURN_TYPICAL
For all of these functions, the only possible failures are a hardware
accelerator (not possible unless using an ALT implementation), an internal
error or runtime corruption.

Exception: the self-tests, which serve little purpose if their status isn't
tested.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:40:45 +02:00
Gilles Peskine
ee0a4435f7 Define indirection macros MBEDTLS_CHECK_RETURN_xxx
Define macros MBEDTLS_CHECK_RETURN_CRITICAL, MBEDTLS_CHECK_RETURN_TYPICAL
and MBEDTLS_CHECK_RETURN_OPTIONAL so that we can indicate on a
function-by-function basis whether checking the function's return value is
almost always necessary (CRITICAL), typically necessary in portable
applications but unnecessary in some reasonable cases (TYPICAL), or
typically unnecessary (OPTIONAL).

Update the documentation of MBEDTLS_CHECK_RETURN accordingly. This is split
between the user documentation (Doxygen, in config.h) and the internal
documentation (non-Doxygen, in platform_util.h, of minor importance since
the macro isn't meant to be used directly).

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:40:33 +02:00
Mateusz Starzyk
1ef29fcf47 Add MBEDTLS_CHECK_RETURN description to config.h
Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:40:33 +02:00
Gilles Peskine
377a310da4 Catch failures of AES or DES operations
Declare all AES and DES functions that return int as needing to have
their result checked, and do check the result in our code.

A DES or AES block operation can fail in alternative implementations of
mbedtls_internal_aes_encrypt() (under MBEDTLS_AES_ENCRYPT_ALT),
mbedtls_internal_aes_decrypt() (under MBEDTLS_AES_DECRYPT_ALT),
mbedtls_des_crypt_ecb() (under MBEDTLS_DES_CRYPT_ECB_ALT),
mbedtls_des3_crypt_ecb() (under MBEDTLS_DES3_CRYPT_ECB_ALT).
A failure can happen if the accelerator peripheral is in a bad state.
Several block modes were not catching the error.

This commit does the following code changes, grouped together to avoid
having an intermediate commit where the build fails:

* Add MBEDTLS_CHECK_RETURN to all functions returning int in aes.h and des.h.
* Fix all places where this causes a GCC warning, indicating that our code
  was not properly checking the result of an AES operation:
    * In library code: on failure, goto exit and return ret.
    * In pkey programs: goto exit.
    * In the benchmark program: exit (not ideal since there's no error
      message, but it's what the code currently does for failures).
    * In test code: TEST_ASSERT.
* Changelog entry.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 20:40:31 +02:00
Gilles Peskine
1b8a294d7e Test invalid nonce length for one-shot AEAD decryption
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 17:07:21 +02:00
Gilles Peskine
ffc2a80b36
Merge pull request #4983 from mstarzyk-mobica/backport_hash_message_psa_flags
Backport 2.x: PSA MAC computation with _HASH flag implies _MESSAGE.
2021-09-29 16:07:32 +02:00
Gilles Peskine
f313336c01 Remove barely-used redirect functions
redirect_out was no longer used and redirect_err was only used to
quiet dd. Change the dd invocation to only print diagnostics on
error (on platforms where this is possible).

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
7c220d7d37 Keep going after a shell "[" a.k.a. "test" fails
This is necessary to actually keep going and finish the
component-specific cleanup in component_test_cmake_out_of_source if
ssl-opt.err is non-empty.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
1d55995d32 Remove code that is useless now that components run in a subshell
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
e8056c5e31 Complain if an unsupported component is explicitly requested
In all.sh, when an explicit list of components is specified, error out
if one of the components is not known or not supported. Patterns that
happen to match zero components are still effectively ignored.

Fix #2783

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
39a3b11006 Better not function
In the `not` function, in keep-going mode, arrange to report the
failing command (rather than `"$@"`).

Note that the `!` keyword should not be used, because failures with
`!` are not reported properly.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
4848d7bb45 Fix double reporting when the last command of a function fails
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:59:00 +01:00
Gilles Peskine
f6d29c6a9e Stop dispatching through obsolete functions
Remove the obsolete functions record_status and if_build_succeeded.
They didn't affect error detection, but they made error reporting
worse since $BASH_COMMAND would be the unexpanded "$@".

Keep the function definitions for the sake of pull requests using them
that may still be in flight.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:58:55 +01:00
Gilles Peskine
0a7984f1dd component_test_cmake_out_of_source: simplify and fix error handling
Remove ssl-opt.err even if it's empty.

Call cat unconditionally: it'll have no visible effect if the file is
empty.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:55:21 +01:00
Gilles Peskine
8ab2994eb5 Detect errors on the left-hand side of a pipeline
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:55:21 +01:00
Gilles Peskine
7105a33906 Run each component in a subshell and handle errors more robustly
This commit completely rewrites keep-going mode. Instead of relying
solely on "set -e", which has some subtle limitations (such as being
off anywhere inside a conditional), use an ERR trap to record errors.

Run each component in a subshell. This way a component can set
environment variables, change the current directory, etc., without
affecting other components.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:55:21 +01:00
Gilles Peskine
a5eb22d434 Add --error-test option to test error detection and reporting
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:55:21 +01:00
Gilles Peskine
3de7be8b88 Switch all.sh to bash
This will let us use bash features that are not found in some other sh
implementations, such as DEBUG and ERR traps, "set -o pipefail", etc.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-29 14:55:21 +01:00
Gilles Peskine
27bb62bc81
Merge pull request #4985 from gilles-peskine-arm/check-names-rewrite-2.2x
Backport 2.x: Rewrite check-names.sh in python
2021-09-29 12:43:09 +02:00
Gilles Peskine
1c39975a27 Fix typo
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-28 10:13:45 +02:00
Gilles Peskine
ee20f3698a Remove check-names.sh and now-unused helper scripts
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-27 20:14:12 +02:00
Gilles Peskine
31da67beb7 Switch to the new Python implementation of check-names
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-27 20:13:00 +02:00
Gilles Peskine
d47f636a19 Adapt source file names from Mbed TLS 3.0 to 2.27
* There's no compat-2.x.h in Mbed TLS 2.x, but there's a compat-1.3.h which
  similarly needs to be excluded.
* mbedtls_config.h is called config.h.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-27 20:12:00 +02:00
Yuto Takano
5b4caf21de Fix typos pointed out by check_names
Signed-off-by: Yuto Takano <yuto.takano@arm.com>
2021-09-27 20:08:01 +02:00
Gilles Peskine
7bf5205581 More robust handling of excluded files
Don't try to enumerate excluded files. List included files, and remove names
from the list if they match an excluded-file pattern.

This resolves the problem that the script could get into an infinite loop
due to the use of recursive globbing. Unfortunately, Python's recursive
globs follows symbolic links to directories, which leads to an infinite loop
if a symbolic link points to an ancestor of the directory that contains it.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-27 20:01:24 +02:00
Gilles Peskine
8266b5b0b4 Copy check_names.py and friends from development
Copy the following files:
    tests/scripts/check_names.py
    tests/scripts/list_internal_identifiers.py
    tests/scripts/list-identifiers.sh

Copy the version from b19be6b5f3c9a92b5b17fa27e16901f132f1a310, which is the
result of merging https://github.com/ARMmbed/mbedtls/pull/1638 into the
development branch.

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2021-09-27 20:01:08 +02:00
Mateusz Starzyk
e6e02b6bae Extend mac_key_policy test.
Add checks for psa_mac_compute and psa_mac_verify.

Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
2021-09-27 14:48:38 +02:00
Mateusz Starzyk
25e65db1ce Use separate expected results for MAC sign and verify key policy.
Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
2021-09-27 14:48:38 +02:00
Mateusz Starzyk
18f662b0af Fix mac_key_policy test function
Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
2021-09-27 14:48:38 +02:00
Mateusz Starzyk
dd55b25bd5 Require MESSAGE flag in PSA MAC setup.
Signed-off-by: Mateusz Starzyk <mateusz.starzyk@mobica.com>
2021-09-27 14:48:38 +02:00
Gilles Peskine
186c0216b0
Merge pull request #4978 from davidhorstmann-arm/2.x-fix-aarch64-asm-constraints
Backport 2.x: Fix aarch64 assembly for bignum multiplication
2021-09-27 09:01:12 +02:00
Gilles Peskine
97cd76988d
Merge pull request #4979 from gilles-peskine-arm/doc-use-psa-crypto-2.x
Backport 2.x: Document effects of MBEDTLS_USE_PSA_CRYPTO
2021-09-25 09:25:36 +02:00
Manuel Pégourié-Gonnard
fec7ef8270 Mention areas that are not (well) tested.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00
Manuel Pégourié-Gonnard
ee20baf6e1 Clarify that 1.3 is excluded
Don't mention "TLS 1.2 only" for PSK, as that could give the impression
that the other things about TLS are supported beyond 1.2, which isn't
the case currently.

Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00
Manuel Pégourié-Gonnard
12ab49aaf7 Improve wording and fix some typos.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00
Manuel Pégourié-Gonnard
b4113bac9a Clarify wording of "not covered" section
The section is about things that are not covered, but some lists are
about things that are covered, which was very confusing.

Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00
Manuel Pégourié-Gonnard
1fa923a5bc Fix inaccuracy in key exchange summary
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00
Manuel Pégourié-Gonnard
6cf7d94ab4 Document parts not covered by USE_PSA_CRYPTO
Also, remove the section about design considerations for now. It's
probably more suitable for a developer-oriented document that would also
include considerations about possible paths for the future, which would
better be separated from user documentation (separating the certain that
is now, from the uncertain that might or might not be later).

Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
2021-09-24 20:44:03 +02:00