Set the build type to Release (-O2) when running CPU-intensive tests (ssl-opt,
or unit tests with debug features). A build type of Check (-Os) would be best
when the main objective of the build is to check for build errors or warnings
and there aren't many tests to run; in this commit there are no such test
cases to change. Only use cmake with no build type (which results in not
passing a -O option, and thus missing some GCC warnings) when exercising cmake
features.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Use Release mode (-O2) for component_test_full_cmake_clang which runs SSL
tests.
To have some coverage with Check mode (which enables more compiler warnings
but compiles with -Os), change a few other builds that only run unit tests
at most to Check mode.
Don't add any new builds, to keep the total build volume down. We don't need
extensive coverage of all combinations, just a reasonable set.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
SSL testing benefits from faster executables, so use -O2 rather than -O1.
Some builds use -O1, but that's intended for jobs that only run unit tests,
where the build takes longer than the tests.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Use case pattern matching instead of multiline split, given there is
only the well formatted PIDs to match on this should be safe.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
On machines with more modern kernels (>5.4 from testing so far) the
useage of -b seems to conflict with the usage of -p. Whilst the usage of
-b seems like a good idea to avoid blocks as we are tight looping on it,
the usage of -p seems to require the usage of stat() (specifically in
/proc) which -b forbids. All you get is a load of warnings
(suppressable by -w) but never a positive result, which means that all
servers are reported as "Failed to start". We are not keen on losing
-b, so instead parse the output of lsof (using -F to format it) to
check the if PIDs that it outputs match that we are looking for.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
digits is also a local variable in host_test.function, leading to compilers
complaining about that shadowing the global variable in
test_suite_base64.function.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This is part of the definition of the encoding, not a choice of test
parameter, so keep it with the test code.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Add unit tests for mask_of_range(), enc_char() and dec_value().
When constant-flow testing is enabled, verify that these functions are
constant-flow.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
I had originally thought to support directories with
mbedtls_x509_crt_parse_path but it would have complicated the code more than
I cared for. Remove a remnant of the original project in the documentation.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
To test c <= high, instead of testing the sign of (high + 1) - c, negate the
sign of high - c (as we're doing for c - low). This is a little easier to
read and shaves 2 instructions off the arm thumb build with
arm-none-eabi-gcc 7.3.1.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
n was used for two different purposes. Give it a different name the second
time. This does not seem to change the generated code when compiling with
optimization for size or performance.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Instead of doing constant-flow table lookup, which requires 64 memory loads
for each lookup into a 64-entry table, do a range-based calculation, which
requires more CPU instructions per range but there are only 5 ranges.
I expect a significant performance gain (although smaller than for decoding
since the encoding table is half the size), but I haven't measured. Code
size is slightly smaller.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Document what each local variable does when it isn't obvious from the name.
Don't reuse a variable for different purposes.
This commit has very little impact on the generated code (same code size on
a sample Thumb build), although it does fix a theoretical bug that 2^32
spaces inside a line would be ignored instead of treated as an error.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Instead of doing constant-flow table lookup, which requires 128 memory loads
for each lookup into a 128-entry table, do a range-based calculation, which
requires more CPU instructions per range but there are only 5 ranges.
Experimentally, this is ~12x faster on my PC (based on
programs/x509/load_roots). The code is slightly smaller, too.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Base64 decoding uses equality comparison tests for characters that don't
leak information about the content of the data other than its length, such
as whitespace. Do this with '=' as well, since it only reveals information
about the length. This way the table lookup can focus on character validity
and decoding value.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
When TEST_EQUAL fails, show the two numerical values in the test log (only
with host_test). The values are printed in hexa and signed decimal.
The arguments of TEST_EQUAL must now be integers, not pointers or floats.
The current implementation requires them to fit in unsigned long long
Signed values no larger than long long will work too. The implementation
uses unsigned long long rather than uintmax_t to reduce portability
concerns. The snprintf function must support "%llx" and "%lld".
For this purpose, add room for two lines of text to the mbedtls_test_info
structure. This adds 154 bytes of global data.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Improve the code structure in case we want to add other similar conditions
later. Document better what we're doing, and document why we're doing it.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Palliative for https://github.com/ARMmbed/mbedtls/issues/3377. If a test
case fails due to an unexpected resend, allow retrying, like in the case of
a client timeout.
Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
This was causing some tests using the openssl s_client to not connect -
I suspect this was due to localhost (at least on my machine) resolving
to ::1 rather than 127.0.0.1. Note that the error seen would have been
that the session file specified with -sess_out did not get created.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
Missing wildcards meant that some servers were not identified as DTLS,
which lead to port checking on TCP rather than UDP, and thus mistakenly
cancelling tests as the server had not come up.
Signed-off-by: Paul Elliott <paul.elliott@arm.com>
