The sample program aescrypt2 shows bad practice: hand-rolled CBC
implementation, CBC+HMAC for AEAD, hand-rolled iterated SHA-2 for key
stretching, no algorithm agility. The new sample program pbcrypt does
the same thing, but better. So remove aescrypt2.
Fix#1906
Signed-off-by: Gilles Peskine <Gilles.Peskine@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.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Signed-off-by: Gilles Peskine <gilles.peskine@arm.com>
The baremetal configuration is, among other things, our default
reference point for code size measurements. So disable debugging
features that would not be enabled in production where code size is
limited. In particular, this shrinks the core SSL modules by about
half. Keep debugging features that are solely in their own
modules (MBEDTLS_ERROR_C, MBEDTLS_VERSION_FEATURES) since it's easy to
filter them out.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Make sure line number reported is correct for the overly long line, and
change the message to be more readable.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
As I descovered, a changelog entry with a line length greater than 80
characters would still pass CI. This is a quick change to the script to
make it detect these descrepancies and fail.
Signed-off-by: Paul Elliott <paul.elliott@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>
An SSL client can be configured to insist on a minimum size for the
Diffie-Hellman (DHM) parameters sent by the server. Add several test
cases where the server sends parameters with exactly the minimum
size (must be accepted) or parameters that are one bit too short (must
be rejected). Make sure that there are test cases both where the
boundary is byte-aligned and where it isn't.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Write a simple unit test for mbedtls_ecp_muladd().
Add just one pair of test cases. One of them causes the argument to
fix_negative to have an argument with an all-bits-zero least
significant limb which briefly triggered a branch in Mbed TLS 2.26+.
See https://github.com/ARMmbed/mbedtls/issues/4296 and
https://github.com/ARMmbed/mbedtls/pull/4297.
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>
tests/suites/helpers.function and tests/suites/*_test.function contain
"#line" directives. This causes the TAGS file to contain references
pointing to the file path named in the "#line" directives, which is
relative to the "tests" directory rather than to the toplevel. Fix
this by telling etags to ignore "#line" directives, which is ok since
we aren't actually running it on any generated code.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In a case exprssion, `|` separates patterns so it needs to be quoted.
Also `\` was not actually part of the set since it was quoting another
character.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
If `$FILTER` (`-f`) and `$EXCLUDE` (`-e`) are simple selections that
can be expressed as shell patterns, use a case statement instead of
calling grep to determine whether a test case should be executed.
Using a case statement significantly reduces the time it takes to
determine that a test case is excluded (but the improvement is small
compared to running the test).
This noticeably speeds up running a single test or a small number of
tests. Before:
```
tests/ssl-opt.sh -f Default 1.75s user 0.54s system 79% cpu 2.885 total
```
After:
```
tests/ssl-opt.sh -f Default 0.37s user 0.14s system 29% cpu 1.715 total
```
There is no perceptible difference when running a large number of tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Avoid using the external command grep for simple string-based checks.
Prefer a case statement. This improves performance.
The performance improvement is moderate but noticeable when skipping
most tests. When a test is run, the cost of the associated grep calls
is negligible. In this commit, I focused on the uses of grep that can
be easily replaced and that are executed a large number of times.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.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>
Marked dirty memory ends up in the result buffer after encoding (due to
the input having been marked dirty), and then the final comparison
to make sure that we got what we expected was triggering the constant
flow checker.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Fix a stack buffer overflow with mbedtls_net_poll() and
mbedtls_net_recv_timeout() when given a file descriptor that is beyond
FD_SETSIZE. The bug was due to not checking that the file descriptor
is within the range of an fd_set object.
Fix#4169
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Although the library documentation does not guarantee that calling
mbedtls_entropy_free() twice works, it's a plausible assumption and it's
natural to write code that frees an object twice. While this is uncommon for
an entropy context, which is usually a global variable, it came up in our
own unit tests (random_twice tests in test_suite_random in the
development branch).
Announce this in the same changelog entry as for RSA because it's the same
bug in the two modules.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
These tests validate that an entropy object can be reused and that
calling mbedtls_entropy_free() twice is ok.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_rsa_gen_key() was not freeing the RSA object, and specifically
not freeing the mutex, in some error cases.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When MBEDTLS_THREADING_C is enabled, RSA code protects the use of the
key with a mutex. mbedtls_rsa_free() frees this mutex by calling
mbedtls_mutex_free(). This does not match the usage of
mbedtls_mutex_free(), which in general can only be done once.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
These tests are trivial except when compiling with MBEDTLS_THREADING_C
and a mutex implementation that are picky about matching each
mbedtls_mutex_init() with exactly one mbedtls_mutex_free().
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
random(), and only this function, is thread-safe.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
mbedtls_hmac_drbg_free() left a mutex in the initialized state. This
caused a resource leak on platforms where mbedtls_mutex_init()
allocates resources.
To fix this, mbedtls_hmac_drbg_free() no longer reinitializes the
mutex. To preserve the property that mbedtls_hmac_drbg_free() leaves
the object in an initialized state, which is generally true throughout
the library except regarding mutex objects on some platforms, no
longer initialize the mutex in mbedtls_hmac_drbg_init(). Since the
mutex is only used after seeding, and seeding is only permitted once,
call mbedtls_mutex_init() as part of the seeding process.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
random(), and only this function, is thread-safe.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
mbedtls_ctr_drbg_free() left a mutex in the initialized state. This
caused a resource leak on platforms where mbedtls_mutex_init()
allocates resources.
To fix this, mbedtls_ctr_drbg_free() no longer reinitializes the
mutex. To preserve the property that mbedtls_ctr_drbg_free() leaves
the object in an initialized state, which is generally true throughout
the library except regarding mutex objects on some platforms, no
longer initialize the mutex in mbedtls_ctr_drbg_init(). Since the
mutex is only used after seeding, and seeding is only permitted once,
call mbedtls_mutex_init() as part of the seeding process.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Document the usage inside the library, and relate it with how it's
additionally used in the test code.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Subtract the number of calls to mbedtls_mutex_free() from the number
of calls to mbedtls_mutex_init(). A mutex leak will manifest as a
positive result at the end of the test case.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
If the mutex usage verification framework is enabled and it detects a
mutex usage error, report this error and mark the test as failed.
This detects most usage errors, but not all cases of using
uninitialized memory (which is impossible in full generality) and not
leaks due to missing free (which will be handled in a subsequent commit).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Check that the source address and the frame counter have the expected
length. Otherwise, if the test data was invalid, the test code could
build nonsensical inputs, potentially overflowing the iv buffer.
The primary benefit of this change is that it also silences a warning
from compiling with `gcc-10 -O3` (observed with GCC 10.2.0 on
Linux/amd64). GCC unrolled the loops and complained about a buffer
overflow with warnings like:
```
suites/test_suite_ccm.function: In function 'test_mbedtls_ccm_star_auth_decrypt':
suites/test_suite_ccm.function:271:15: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
271 | iv[i] = source_address->x[i];
| ~~~~~~^~~~~~~~~~~~~~~~~~~~~~
suites/test_suite_ccm.function:254:19: note: at offset [13, 14] to object 'iv' with size 13 declared here
254 | unsigned char iv[13];
```
Just using memcpy instead of loops bypasses this warnings. The added
checks are a bonus.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When using pthread mutexes (MBEDTLS_THREADING_C and
MBEDTLS_THREADING_PTHREAD enabled), and when test hooks are
enabled (MBEDTLS_TEST_HOOKS), set up wrappers around the
mbedtls_mutex_xxx abstraction. In this commit, the wrapper functions
don't do anything yet.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Don't reference the architecture document. This is an LTS branch and
new invasive tests are only going to be introduced as (perhaps partial
or adapted) backports of invasive tests from development anyway.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When this option is enabled, the product includes additional
interfaces that enable additional tests. This option should not be
enabled in production, but is included in the "full" build to enable
the extra tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
We care about the exit code of our server, for example if it's
reporting a memory leak after having otherwise executed correctly.
We don't care about the exit code of the servers we're using for
interoperability testing (openssl s_server, gnutls-serv). We assume
that they're working correctly anyway, and they return 1 (gnutls-serv)
or die by the signal handle the signal (openssl) when killed by a
signal.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This used to be the case a long time ago but was accidentally broken.
Fix <github:nogrep> #4103 for ssl-opt.sh.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
if MBEDTLS_MPI_MAX_SIZE is odd then RSA_PRV_DER_MAX_BYTES will be two less than expected, since the macros are lacking parentheses.
Signed-off-by: Daniel Otte <d.otte@wut.de>
Fix a buffer overflow in mbedtls_mpi_sub_abs() when calculating
|A| - |B| where |B| is larger than |A| and has more limbs (so the
function should return MBEDTLS_ERR_MPI_NEGATIVE_VALUE).
Fix#4042
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Add test cases for mbedtls_mpi_sub_abs() where the second operand has
more limbs than the first operand (which, if the extra limbs are not
all zero, implies that the function returns
MBEDTLS_ERR_MPI_NEGATIVE_VALUE).
This exposes a buffer overflow (reported in #4042).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
These implementations don't necessarily consume entropy the same way the
mbed TLS internal software implementation does, and the 'reference
handshake' test vectors can thus not be applied to an ALT implementation.
Signed-off-by: Steven Cooreman <steven.cooreman@silabs.com>
In component_test_no_hmac_drbg, the fact that HMAC_DRBG is disabled
doesn't affect the SSL code, but the fact that deterministic ECDSA is
disabled does. So run some ECDSA-related SSL tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
mbedtls_rsa_private() could return the sum of two RSA error codes
instead of a valid error code in some rare circumstances:
* If rsa_prepare_blinding() returned MBEDTLS_ERR_RSA_RNG_FAILED
(indicating a misbehaving or misconfigured RNG).
* If the comparison with the public value failed (typically indicating
a glitch attack).
Make sure not to add two high-level error codes.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In order to remove large buffers from the stack, the der data is written
into the same buffer that the pem is eventually written into, however
although the pem data is zero terminated, there is now data left in the
buffer after the zero termination, which can cause
mbedtls_x509_crt_parse to fail to parse the same buffer if passed back
in. Patches also applied to mbedtls_pk_write_pubkey_pem, and
mbedtls_pk_write_key_pem, which use similar methods of writing der data
to the same buffer, and tests modified to hopefully catch any future
regression on this.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
mbedtls_ctr_drbg_set_reseed_interval() and
mbedtls_hmac_drbg_set_reseed_interval() can now be called before
their seed functions and the reseed_interval value will persist.
Previously it would be overwritten with the default value.
*_drbg_reseed_interval is now set in init() and free().
mbedtls_ctr_drbg_free() and mbedtls_hmac_drbg_free() now
reset the drbg context to the state immediately after init().
Tests:
- Added test to check that DRBG reseeds when reseed_counter
reaches reseed_interval, if reseed_interval set before seed
and reseed_interval is less than MBEDTLS_*_DRBG_RESEED_INTERVAL.
Signed-off-by: gacquroff <gavina352@gmail.com>
Fixes an issue where configs that had `MBEDTLS_MPI_MAX_BITS` greater than 256
but smaller than the test that was running (792 bits) the test would fail
incorrectly.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Extra whitespace and a missing newline at end of file was causing an error with
`check_files.py`.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Move dependancy on `MBEDTLS_MPI_MAX_BITS` to apply to the specific test cases
which will break when `MBEDTLS_MPI_MAX_BITS` is too small. This re-enables
previous tests that were turned off accidentally.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Previously `mbedtls_mpi_exp_mod` was tested with values that were over
`MBEDTLS_MPI_MAX_SIZE` in size. This is useful to do as some paths are only
taken when the exponent is large enough however, on builds where
`MBEDTLS_MPI_MAX_SIZE` is under the size of these test values.
This fix turns off these tests when `MBEDTLS_MPI_MAX_SIZE` is too small to
safely test (notably this is the case in config-thread.h).
Signed-off-by: Chris Jones <christopher.jones@arm.com>
In two test cases, the exponentiation computation was not being fully tested
as when A_bytes (the base) == N_bytes (the modulus) -> A = N. When this is the
case A is reduced to 0 and therefore the result of the computation will always
be 0.
This fixes that issue and therefore increases the test coverage to ensure
different computations are actually being run.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Reword test cases to be easier to read and understand.
Adds comments to better explain what the test is doing.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Add two further boundary tests for cases where both the exponent and modulus to
`mbedtls_mpi_exp_mod()` are `MBEDTLS_MPI_MAX_SIZE`, or longer, bytes long.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Adds test cases to ensure that `mbedtls_mpi_exp_mod` will return an error with
an exponent or modulus that is greater than `MBEDTLS_MPI_MAX_SIZE` in size.
Adds test cases to ensure that Diffie-Hellman will fail to make a key pair
(using `mbedtls_dhm_make_public`) when the prime modulus is greater than
`MBEDTLS_MPI_MAX_SIZE` in size.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Add a test case to ensure `mbedtls_mpi_exp_mod` fails when using a key size
larger than MBEDTLS_MPI_MAX_SIZE.
Add a test case to ensure that Diffie-Hellman operations fail when using a key
size larger than MBEDTLS_MPI_MAX_SIZE.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Check that the exponent and modulus is below `MBEDTLS_MPI_MAX_BITS` before
performing a time expensive operation (modular exponentiation). This prevents
a potential DoS from Diffie-Hellman computations with extremely
large key sizes.
Signed-off-by: Chris Jones <christopher.jones@arm.com>
Add missing tag check for algorithm parameters when comparing the
signature in the description part of the cert against the actual
signature whilst loading a certificate. This was found by a
certificate (created by fuzzing) that openssl would not verify, but
mbedtls would.
Regression test added (one of the client certs modified accordingly)
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
"declaration-after-statement" was generated because that code was
backported from the development branch, which currently uses C99.
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
`finish_sha384_t` was made more generic by using `unsigned char*`
instead of `unsigned char[48]` as the second parameter.
This change tries to make the function casting more robust against
future improvements of gcc analysis.
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
This commit fixes the same warning fixed by baeedbf9, but without
wasting RAM. By casting `mbedtls_sha512_finish_ret()`, `padbuf`
could be kept 48 bytes long without triggering any warnings.
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
GCC 11 generated a warning because `padbuf` was too small to be
used as an argument for `mbedtls_sha512_finish_ret`. The `output`
parameter of `mbedtls_sha512_finish_ret` has the type
`unsigned char[64]`, but `padbuf` was only 48 bytes long.
Even though `ssl_calc_finished_tls_sha384` uses only 48 bytes for
the hash output, the size of `padbuf` was increased to 64 bytes.
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
GCC 11 generated the warnings because the parameter `ret_buf`
was declared as `const char[10]`, but some of the arguments
provided in `run_test_snprintf` are shorter literals, like "".
Now the type of `ret_buf` is `const char *`.
Both implementations of `test_snprintf` were fixed.
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
In GCC 11, parameters declared as arrays in function prototypes
cannot be declared as pointers in the function definition. The
same is true for the other way around.
The definition of `mbedtls_aes_cmac_prf_128` was changed to match
its public prototype in `cmac.h`. The type `output` was
`unsigned char *`, now is `unsigned char [16]`.
In `ssl_tls.c`, all the `ssl_calc_verify_*` variants now use pointers
for the output `hash` parameter. The array parameters were removed
because those functions must be compatible with the function pointer
`calc_verify` (defined in `ssl_internal.h`).
Signed-off-by: Rodrigo Dias Correa <rodrigo@correas.us>
Positive tests: test that the RNG has the expected size, given that we
know how many leading zeros it has because we know how the function
consumes bytes and when the test RNG produces null bytes.
Negative tests: test that if the RNG is willing to emit less than the
number of wanted bytes, the function fails.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Complement to 0a8352b4: peer_pmslen is not initialized when decryption
fails, so '|= peer_pmslen' may access uninitialized memory, as indicated
by Frama-C/Eva.
Co-authored-by: Gilles Peskine <gilles.peskine@arm.com>
Signed-off-by: André Maroneze <maroneze@users.noreply.github.com>
Add a build with MBEDTLS_ERROR_STRERROR_DUMMY but not MBEDTLS_ERROR_C.
Previously, both options were enabled by default, but
MBEDTLS_ERROR_STRERROR_DUMMY only matters when MBEDTLS_ERROR_C is
enabled, so its effect was not tested.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Simplify the guards on MBEDTLS_ERROR_C and MBEDTLS_ERROR_STRERROR_DUMMY.
No longer include superfluous headers and definition: string.h and
platform.h are only needed for MBEDTLS_ERROR_C; time_t is not needed
at all.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This is fix for the MBEDTLS_AES_SETKEY_DEC_ALT macro switch is including the aes xts methods
and building with a custom mbedtls_aes_setkey_dec function will disable the aes xts methods.
The fix is separating the aes xts methods and the MBEDTLS_AES_SETKEY_DEC_ALT can only
switch the presence of the mbedtls_aes_setkey_dec function.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
CMake versions less than 3.0 do not support the `target_sources`
command. In order to be able to support v2.8.12.2 of cmake, add the
extra targets directly to the target command.
This is a backport from the development branch, except that the uses in
this branch are simpler, and modifying the SOURCES property directly is
not needed.
Fixes#3801
Signed-off-by: David Brown <david.brown@linaro.org>
Fix `printf "$foo"` which treats the value of `foo` as a printf format
rather than a string.
I used the following command to find potentially problematic lines:
```
git ls-files '*.sh' | xargs egrep 'printf +("?[^"]*|[^ ]*)\$'
```
The remaining ones are false positives for this regexp.
The errors only had minor consequences: the output of `ssl-opt.sh`
contained lines like
```
Renegotiation: gnutls server strict, client-initiated .................. ./tests/ssl-opt.sh: 741: printf: %S: invalid directive
PASS
```
and in case of failure the GnuTLS command containing a substring like
`--priority=NORMAL:%SAFE_RENEGOTIATION` was not included in the log
file. With the current tests, there was no risk of a test failure
going undetected.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Showing a regexp to say that by default all tests are executed is not
particularly helpful.
If we ever add a default exclusion list or a default filter, we can
edit the documentation again.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This way we can have a single call to mbedtls_platform_zeroize, which
saves a few bytes of code size.
Additionally, on my PC, I notice a significant speed improvement
(x86_64 build with MBEDTLS_AESNI_C disabled, gcc 5.4.0 -O3). I don't
have an explanation for that (I expected no measurable difference).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Remove the zeroization of a pointer variable in the AES block
functions. The code was valid but spurious and misleading since it
looked like a mistaken attempt to zeroize the pointed-to buffer.
Reported by Antonio de la Piedra, CEA Leti, France.
Note that we do not zeroize the buffer here because these are the
round keys, and they need to stay until all the blocks are processed.
They will be zeroized in mbedtls_aes_free().
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Set the CMake-observed variable `CTEST_OUTPUT_ON_FAILURE`, so that when
a "make test" run by CMake fails, verbose test output about the detail
of failure is available.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
From now on, external contributions are no longer acknowledged in the
changelog file. They of course remain acknowledged in the Git history.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Starting with commit 49e94e3, the do/while loop in
`rsa_prepare_blinding()` was changed to a `do...while(0)`, which
prevents retry from being effective and leaves dead code.
Restore the while condition to retry, and lift the calls to finish the
computation out of the while loop by by observing that they are
performed only when `mbedtls_mpi_inv_mod()` returns zero.
Signed-off-by: Peter Kolbus <peter.kolbus@garmin.com>
The toplevel directory is actually just ../..: the makefile commands
are executed in the subdirectory. $(PWD) earlier was wrong because it
comes from the shell, not from make. Looking up $(MAKEFILE_LIST) is
wrong because it indicates where the makefile is (make -f), not which
directory to work in (make -C).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This reduces the impact of the code size increase due to the addition
of calls to mbedtls_platform_zeroize.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
Zeroising of local buffers and variables which are used for calculations in
mbedtls_internal_md*_process() and mbedtls_internal_ripemd160_process()
functions to erase sensitive data from memory.
Checked all function for possible missing zeroisation in MD.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
Zeroising of local buffers and variables which are used for calculations in
mbedtls_pkcs5_pbkdf2_hmac() and mbedtls_internal_sha*_process() functions
to erase sensitive data from memory.
Checked all function for possible missing zeroisation in PKCS and SHA.
Signed-off-by: gabor-mezei-arm <gabor.mezei@arm.com>
Description referred to mbedtls_ssl_sent_t callback,
but the callback is named mbedtls_ssl_send_t.
Signed-off-by: Christopher Moynihan <christophm@gmail.com>
Probably the `W[2 << MBEDTLS_MPI_WINDOW_SIZE]` notation is based on a transcription of 2**MBEDTLS_MPI_WINDOW_SIZE.
Signed-off-by: Daniel Otte <d.otte@wut.de>
The test function mbedtls_mpi_lt_mpi_ct did not initialize ret in test
code. If there was a bug in library code whereby the library function
mbedtls_mpi_lt_mpi_ct() did not set ret when it should, we might have
missed it if ret happened to contain the expected value. So initialize
ret to a value that we never expect.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
If test_fail is called multiple times in the same test case, report
the location of the first failure, not the last one.
With this change, you no longer need to take care in tests that use
auxiliary functions not to fail in the main function if the auxiliary
function has failed.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
RFC5280 does not state that the `revocationDate` should be checked.
In addition, when no time source is available (i.e., when MBEDTLS_HAVE_TIME_DATE is not defined), `mbedtls_x509_time_is_past` always returns 0. This results in the CRL not being checked at all.
https://tools.ietf.org/html/rfc5280
Signed-off-by: Raoul Strackx <raoul.strackx@fortanix.com>
Currently the new component in all.sh fails because
mbedtls_ssl_cf_memcpy_offset() is not actually constant flow - this is on
purpose to be able to verify that the new test works.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
The tests are supposed to be failing now (in all.sh component
test_memsan_constant_flow), but they don't as apparently MemSan doesn't
complain when the src argument of memcpy() is uninitialized, see
https://github.com/google/sanitizers/issues/1296
The next commit will add an option to test constant flow with valgrind, which
will hopefully correctly flag the current non-constant-flow implementation.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This paves the way for a constant-flow implementation of HMAC checking, by
making sure that the comparison happens at a constant address. The missing
step is obviously to copy the HMAC from the secret offset to this temporary
buffer with constant flow, which will be done in the next few commits.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
* mbedtls-2.16: (32 commits)
A different approach of signed-to-unsigned comparison
Fix bug in redirection of unit test outputs
Don't forget to free G, P, Q, ctr_drbg, and entropy
Backport e2k support to mbedtls-2.7
compat.sh: stop using allow_sha1
compat.sh: quit using SHA-1 certificates
compat.sh: enable CBC-SHA-2 suites for GnuTLS
Fix license header in pre-commit hook
Update copyright notices to use Linux Foundation guidance
Fix building on NetBSD 9.0
Remove obsolete buildbot reference in compat.sh
Fix misuse of printf in shell script
Fix added proxy command when IPv6 is used
Simplify test syntax
Fix logic error in setting client port
ssl-opt.sh: include test name in log files
ssl-opt.sh: remove old buildbot-specific condition
ssl-opt.sh: add proxy to all DTLS tests
Log change as bugfix
Add changelog entry
...
I might be wrong, but lcc's optimizer is curious about this,
and I am too: shouldn't we free allocated stuff correctly
before exiting `dh_genprime` in this certain point of code?
Signed-off-by: makise-homura <akemi_homura@kurisa.ch>
They are used to generate cert_md*.crt.
Regenerate cert_md5.crt which had previously been generated for a
different key.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
It wasn't working when invoking programs/x509/cert_write or
programs/x509/cert_req due to relying on the current directory rather
than the location of the makefile.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Discover hash algorithms automatically rather than hard-coding a list,
as was previously done in cert_write.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
For some reason, RIPEMD160, SHA224 and SHA384 were not supported.
This fixes the build recipes for tests/data_files/cert_sha224.crt and
tests/data_files/cert_sha384.crt .
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Replace server2.crt with server2-sha256.crt which, as the name implies, is
just the SHA-256 version of the same certificate.
Replace server1.crt with cert_sha256.crt which, as the name doesn't imply, is
associated with the same key and just have a slightly different Subject Name,
which doesn't matter in this instance.
The other certificates used in this script (server5.crt and server6.crt) are
already signed with SHA-256.
This change is motivated by the fact that recent versions of GnuTLS (or older
versions with the Debian patches) reject SHA-1 in certificates by default, as
they should. There are options to still accept it (%VERIFY_ALLOW_BROKEN and
%VERIFY_ALLOW_SIGN_WITH_SHA1) but:
- they're not available in all versions that reject SHA-1-signed certs;
- moving to SHA-2 just seems cleaner anyway.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Recent GnuTLS packages on Ubuntu 16.04 have them disabled.
From /usr/share/doc/libgnutls30/changelog.Debian.gz:
gnutls28 (3.4.10-4ubuntu1.5) xenial-security; urgency=medium
* SECURITY UPDATE: Lucky-13 issues
[...]
- debian/patches/CVE-2018-1084x-4.patch: hmac-sha384 and sha256
ciphersuites were removed from defaults in lib/gnutls_priority.c,
tests/priorities.c.
Since we do want to test the ciphersuites, explicitly re-enable them in the
server's priority string. (This is a no-op with versions of GnuTLS where those
are already enabled by default.)
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
As a result, the copyright of contributors other than Arm is now
acknowledged, and the years of publishing are no longer tracked in the
source files.
Also remove the now-redundant lines declaring that the files are part of
MbedTLS.
This commit was generated using the following script:
# ========================
#!/bin/sh
# Find files
find '(' -path './.git' -o -path './3rdparty' ')' -prune -o -type f -print | xargs sed -bi '
# Replace copyright attribution line
s/Copyright.*Arm.*/Copyright The Mbed TLS Contributors/I
# Remove redundant declaration and the preceding line
$!N
/This file is part of Mbed TLS/Id
P
D
'
# ========================
Signed-off-by: Bence Szépkúti <bence.szepkuti@arm.com>
For explicit proxy commands (included with `-p "$P_PXY <args>` in the test
case), it's the test's writer responsibility to handle IPv6; only fix the
proxy command when we're auto-adding it.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This is a convenience for when we get log files from failed CI runs, or attach
them to bug reports, etc.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
A lot of DTLS test are timing-sensitive, especially those that contain
assertions about retransmission. Sometimes some DTLS test fails intermittently
on the CI with no clear apparent reason; we need more information in the log
to understand the cause of those failures.
Adding a proxy means we'll get timing information from the proxy logs.
An alternative would be to add timing information to the debug output of
ssl_server2 and ssl_client2. But that's more complex because getting
sub-second timing info is outside the scope of the C standard, and our current
timing module only provides a APi for sub-second intervals, not absolute time.
Using the proxy is easier as it's a single point that sees all messages, so
elapsed time is fine here, and it's already implemented in the proxy output.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
In the entries (mbedtls_x509_crl_entry values) on the list constructed
by mbedtls_x509_crl_parse_der(), set entry->raw.tag to
(SEQUENCE | CONSTRUCTED) rather than to the tag of the first ASN.1
element of the entry (which happens to be the tag of the serial
number, so INTEGER or INTEGER | CONTEXT_SPECIFIC). This is doesn't
really matter in practice (and in particular the value is never used
in Mbed TLS itself), and isn't documented, but at least it's
consistent with how mbedtls_x509_buf is normally used.
The primary importance of this change is that the old code tried to
access the tag of the first element of the entry even when the entry
happened to be empty. If the entry was empty and not followed by
anything else in the CRL, this could cause a read 1 byte after the end
of the buffer containing the CRL.
The test case "X509 CRL ASN1 (TBSCertList, single empty entry at end)"
hit the problematic buffer overflow, which is detected with ASan.
Credit to OSS-Fuzz for detecting the problem.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Add a few more negative test cases for mbedtls_x509_crl_parse.
The test data is manually adapted from the existing positive test case
"X509 CRL ASN1 (TBSCertList, sig present)" which decomposes as
305c
3047 tbsCertList TBSCertList
020100 version INTEGER OPTIONAL
300d signatureAlgorithm AlgorithmIdentifier
06092a864886f70d01010e
0500
300f issuer Name
310d300b0603550403130441424344
170c303930313031303030303030 thisUpdate Time
3014 revokedCertificates
3012 entry 1
8202abcd userCertificate CertificateSerialNumber
170c303831323331323335393539 revocationDate Time
300d signatureAlgorithm AlgorithmIdentifier
06092a864886f70d01010e
0500
03020001 signatureValue BIT STRING
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The usage of "!memcmp()" is at least not recommended
and better to use the macro dedicated for buffer
comparisons.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Do not hexify binary data to compare them, do compare
them directly. That simplifies the check code and save
memory.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Remove `hex` in name of variables of type data_t to reserve it
for variables of type char* that are the hexadecimal
representation of a data buffer.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Command to find the files in which lines have gone
larger than 79 characters due to the renaming:
grep '.\{80\}' \
`git diff-tree --no-commit-id --name-only -r HEAD` \
| grep hexcmp
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
This is an LTS branch, C99 isn't allowed yet, it breaks versions of MSVC that
we still support for this branch.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Currently this breaks all.sh component test_memsan_constant_flow, just as
expected, as the current implementation is not constant flow.
This will be fixed in the next commit.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
This option allows to test the constant-flow nature of selected code, using
MemSan and the fundamental observation behind ctgrind that the set of
operations allowed on undefined memory by dynamic analysers is the same as the
set of operations allowed on secret data to avoid leaking it to a local
attacker via side channels, namely, any operation except branching and
dereferencing.
(This isn't the full story, as on some CPUs some instructions have variable
execution depending on the inputs, most notably division and on some cores
multiplication. However, testing that no branch or memory access depends on
secret data is already a good start.)
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Just move code from ssl_decrypt_buf() to the new cf_hmac() function and then
call cf_hmac() from there.
This makes the new cf_hmac() function used and validates that its interface
works for using it in ssl_decrypt_buf().
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
The dummy implementation is not constant-flow at all for now, it's just
here as a starting point and a support for developing the tests and putting
the infrastructure in place.
Depending on the implementation strategy, there might be various corner cases
depending on where the lengths fall relative to block boundaries. So it seems
safer to just test all possible lengths in a given range than to use only a
few randomly-chosen values.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
The condition is a complex and repeated a few times. There were already some
inconsistencies in the repetitions as some of them forgot about DES.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
inv_mod() already returns a specific error code if the value is not
invertible, so no need to check in advance that it is. Also, this is a
preparation for blinding the call to inv_mod(), which is made easier by
avoiding the redundancy (otherwise the call to gcd() would need to be blinded
too).
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
In the next commit, we'll need to draw a second random value, in order to
blind modular inversion. Having a function for that will avoid repetition.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
fix_negative allocates memory for its result. The calling site didn't
check the return value, so an out-of-memory error could lead to an
incorrect calculation. Fix this.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Fix a memory leak in mbedtls_mpi_sub_abs when the output parameter is
aliased to the second operand (X = A - X) and the result is negative.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Run all the addition and subtraction tests with the result aliased to
the first operand and with the result aliased to the second operand.
Before, only some of the aliasing possibilities were tested, for only
some of the functions, with only some inputs.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Address remaining PR comments for #2118
- Add ChangeLog.d/x509write_csr_heap_alloc.txt.
- Fix parameter alignment per Gille's recommendation.
- Update comments to more explicitly describe the manipulation of buf.
- Replace use of `MBEDTLS_MPI_MAX_SIZE` as `sig` buffer size for
call to `x509write_csr_der_internal()` with more intuitive
`MBEDTLS_PK_SIGNATURE_MAX_SIZE`.
- Update `mbedtls_x509write_csr_der()` to return
`MBEDTLS_ERR_X509_ALLOC_FAILED` on mbedtls_calloc error.
Signed-off-by: Simon Leet <simon.leet@microsoft.com>
Using a stack-buffer with a size > 2K could easily produce a stack
overflow for an embedded device which has a limited stack size.
This commit dynamically allocates the large CSR buffer.
This commit avoids using a temporary buffer for storing the OIDs.
A single buffer is used:
a) OIDs are written backwards starting with the end of the buffer;
b) OIDs are memmove'd to the beginning of the buffer;
c) signature over this OIDs is computed and written backwards from the
end of the buffer;
d) the two memory regions are compacted.
Signed-off-by: Doru Gucea <doru-cristian.gucea@nxp.com>
This commit rewrites mbedtls_x509write_crt_pem() to not use
a statically size stack buffer to temporarily store the DER
encoded form of the certificate to be written.
This is not necessary because the DER-to-PEM conversion
accepts overlapping input and output buffers.
The CRT writing routine mbedtls_x509write_crt_der() prepares the TBS
(to-be-signed) part of the CRT in a temporary stack-allocated buffer,
copying it to the actual output buffer at the end of the routine.
This comes at the cost of a very large stack buffer. Moreover, its size
must be hardcoded to an upper bound for the lengths of all CRTs to be
written through the routine. So far, this upper bound was set to 2Kb, which
isn't sufficient some larger certificates, as was reported e.g. in #2631.
This commit fixes this by changing mbedtls_x509write_crt_der() to write
the certificate in-place in the output buffer, thereby avoiding the use
of a statically sized stack buffer for the TBS.
Fixes#2631.
In test functions calling mbedtls_test_unhexify(), change the
type of the associated parameters from `char*` to `data_t`.
That way the `unhexify` operation is done by the test
framework and not by the unit test code.
Use for the new parameters of type data_t the name of the
local variable that used to store the `unhexify` version of
the `char*` parameter.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters
of mbedtls_ccm_star_encrypt_and_tag/auth_decrypt from
`char *` to `data_t` to get rid of the calls to
mbedtls_test_unhexify():
- Change the name of parameters and local variables to
clarify which ones are related to the outputs of the
library functions under test and which ones are
related to the expected values of those outputs.
- Use two different buffers to store the plain and cipher
text as expected by the library functions.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters
of aes_encrypt_ofb() from `char *` to `data_t` to get rid
of the calls to mbedtls_test_unhexify():
- Change the name of parameters and local variables to
clarify which ones are related to the outputs of the
library functions under test and which ones are
related to the expected values of those outputs.
- Add assertion on fragment_size parameter
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters
of mbedtls_nist_kw_wrap/unwrap() from `char *` to `data_t`
to get rid of the calls to mbedtls_test_unhexify():
- Change the name of parameters and local variables to
clarify which ones are related to the outputs of the
library functions under test and which ones are
related to the expected values of those outputs.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters of
test_chacha20() from `char *` to `data_t` to get rid of the
calls to mbedtls_test_unhexify():
- Reduce the size of output[] buffer to 375 as its content
is "ASCII expended" into a buffer of 751 bytes.
- Align naming of variables to store and check the
output of mbedtls_chacha20_crypt(). No *dst* variables
anynore, only *output* variables.
- Use two different buffers to store the expected output
of mbedtls_chacha20_crypt() (expected_output_str[]) and
the ASCII string representation of the output of
mbedtls_chacha20_crypt() (output_string[]). Both were
stored in dst_str[] before.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters
of test_hkdf() from `char *` to `data_t` to get rid of the
calls to mbedtls_test_unhexify():
- Align naming of variables related to the expected okm
- Rename `okm_hex[]` to `okm_string[]`
- Added TEST_ASSERT( expected_okm_len <= sizeof( okm ) ) to check
that the okm[] buffer is large enough for the okm output.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
In preparation of changing the type of some parameters
of some test functions from `char *` to `data_t` to get
rid of the calls to mbedtls_test_unhexify():
- Align the name of source data length local variable
with the name of the local variable containing the
source data, respectively src_str and src_str_len.
- Change the type of length, index local variables
from int to size_t.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Command to find the files in which lines have gone
larger than 79 characters due to the renaming:
grep '.\{80\}' \
`git diff-tree --no-commit-id --name-only -r HEAD` \
| grep hexify
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
The path to source files were relative which triggered
warnings when generating the build system.
Move to absolute paths based on CMAKE_CURRENT_SOURCE_DIR.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Add the missing executable in the list of executables
to install.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Reorder declaration of executables in alphabetic order.
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
2020-06-25 23:23:23 +02:00
363 changed files with 4443 additions and 3571 deletions
