The test framework used to treat them specially (but no longer does). Add
these test cases as non-regression for how the test framework allows "?"
and especially "??" (which I think in the very distant path needed special
handling because the test data was embedded in a .c file, and thus ?? could
be interpreted as the prefix of a trigraph).
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Now that the C code supports the full range of intmax_t, allow any size of
signed integer type in the .data file parser.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Change the type of signed integer arguments from int32_t to intmax_t.
This allows the C code to work with test function arguments with a range
larger than int32_t. A subsequent commit will change the .datax generator
to support larger types.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Use normalization the equality comparisons instead of loose regular
expressions to determine the type of an argument of a test function.
Now declarations are parsed in a stricter way: there can't be ignored junk
at the beginning or at the end. For example, `long long unsigned int x`
was accepted as a test function argument (but not `long long unsigned x`),
although this was misleading since the value was truncated to the range of
int. Now only recognized types are accepted.
The new code is slightly looser in that it accepts `char const*` as well as
`const char*`.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
The test framework stores size_t and int32_t values in the parameter store
by converting them all to int. This is ok in practice, since we assume int
covers int32_t and we don't have test data larger than 2GB. But it's
confusing and error-prone. So make the parameter store a union, which allows
size_t values not to be potentially truncated and makes the code a little
clearer.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
In the .datax parser, since we're calling strtol() anyway, rely on it for
verification. This makes the .datax parser very slightly more
liberal (leading spaces and '+' are now accepted), and changes the
interpretation of numbers with leading zeros to octal.
Before, an argument like :0123: was parsed as decimal, but an argument like
:0123+1: was parsed as a C expression and hence the leading zero marked an
octal representation. Now, a leading zero is always interpreted according to
C syntax, namely indicating octal. There are no nonzero integer constants
with a leading zero in a .data file, so this does not affect existing test
cases.
In the .datax generator, allow negative arguments to be 'int' (before, they
were systematically treated as 'exp' even though they didn't need to be).
In the .datax parser, validate the range of integer constants. They have to
fit in int32_t. In the .datax generator, use 'exp' instead of 'int' for
integer constants that are out of range.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Treat backslash as a universal escape character: "\n" is a newline,
backslash escapes any non-alphanumeric character.
This affects some test cases that had "\," standing for backslash-comma.
With the new uniform treatment of backslashes, this needs to be "\\,".
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
No intended behavior change. This commit is mainly to satisfy pylint, which
complains that gen_from_test_data now has too many variables. But it's a
good thing anyway to make the function a little more readable.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
We're using the non-standard function strcasecmp() just so that the case
of digits beyond 9 can be different in the library and in the test data.
Use matching case in the test data, and use a standard function for the
comparison.
Signed-off-by: Gilles Peskine <Gilles.Peskine@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>
ssl_helpers.c is treated with W3 warning level in MSVC complier.
So that it's reported as error for warning of conversion from
size_t to int. This change fixes all this type of warning seen in
Microsoft Visual Studio 12.0. Besides, some potential problems of
type conversion are also handled.
Signed-off-by: Yanray Wang <yanray.wang@arm.com>
During test of component build_arm_linux_gnueabi_gcc_arm5vte and
build_arm_none_eabi_gcc_m0plus. It fails with
- error: implicit declaration of function
‘test_fail_if_psa_leaking’
It happens because test_fail_if_psa_leaking is defined in
helpers.function. This block of code is not converted into C code
while compiling ssl_helpers.c. The function has been moved to
psa_crypto_helpers.c in order to fix this build error.
Signed-off-by: Yanray Wang <yanray.wang@arm.com>
tests/src/ssl_helpers.c depends on functions defined
in library/*.c. If it's complied as an OBJECT with other c files,
cmake complains undefined reference in link stage under programs/.
Therefore, tests/src/test_helpers/ is created to hold c files with
dependency of library/*.c. Besides, tests/src/test_helper/*.c is
separated into another OBJECT, mbedtls_test_helpers, as sources
to build all test suite executables.
In addition, everest header directory is included in case
MBEDTLS_ECDH_VARIANT_EVEREST_ENABLED is enabled.
Signed-off-by: Yanray Wang <yanray.wang@arm.com>
