The ssl-opt.sh test cases using session resumption tend to fail occasionally
on the CI due to a race condition in how ssl_server2 and ssl_client2 handle
the reconnection cycle.
The server does the following in order:
- S1 send application data
- S2 send a close_notify alert
- S3 close the client socket
- S4 wait for a "new connection" (actually a new datagram)
- S5 start a handshake
The client does the following in order:
- C1 wait for and read application data from the server
- C2 send a close_notify alert
- C3 close the server socket
- C4 reset session data and re-open a server socket
- C5 start a handshake
If the client has been able to send the close_notify (C2) and if has been
delivered to the server before if closes the client socket (S3), when the
server reaches S4, the datagram that we start the new connection will be the
ClientHello and everything will be fine.
However if S3 wins the race and happens before the close_notify is delivered,
in S4 the close_notify is what will be seen as the first datagram in a new
connection, and then in S5 this will rightfully be rejected as not being a
valid ClientHello and the server will close the connection (and go wait for
another one). The client will then fail to read from the socket and exit
non-zero and the ssl-opt.sh harness will correctly report this as a failure.
In order to avoid this race condition in test using ssl_client2 and
ssl_server2, this commits introduces a new command-line option
skip_close_notify to ssl_client2 and uses it in all ssl-opt.sh tests that use
session resumption with DTLS and ssl_server2.
This works because ssl_server2 knows how many messages it expects in each
direction and in what order, and closes the connection after that rather than
relying on close_notify (which is also why there was a race in the first
place).
Tests that use another server (in practice there are two of them, using
OpenSSL as a server) wouldn't work with skip_close_notify, as the server won't
close the connection until the client sends a close_notify, but for the same
reason they don't need it (there is no race between receiving close_notify and
closing as the former is the cause of the later).
An alternative approach would be to make ssl_server2 keep the connection open
until it receives a close_notify. Unfortunately it creates problems for tests
where we simulate a lossy network, as the close_notify could be lost (and the
client can't retransmit it). We could modify udp_proxy with an option to never
drop alert messages, but when TLS 1.3 comes that would no longer work as the
type of messages will be encrypted.
Signed-off-by: Manuel Pégourié-Gonnard <manuel.pegourie-gonnard@arm.com>
Goals:
* Build with common compilers with common options, so that we don't
miss a (potentially useful) warning only triggered with certain
build options.
* A previous commit removed -O0 test jobs, leaving only the one with
-m32. We have inline assembly that is disabled with -O0, falling
back to generic C code. This commit restores a test that runs the
generic C code on a 64-bit platform.
Let the caller decide what certificates and keys are loaded (EC/RSA)
instead of loading both for the server, and an unspecified one
for the client. Use only DER encoding.
For each of the crypto-only presets, run the build and check that the
resulting libmbedx509 and libmbedtls are empty.
Don't bother testing, because for each crypto-only preset, another
component builds that plus the x509 and tls parts and tests
everything.
We already have a specific component in all.sh for testing SSLv3, we don't
need to also test it in components that aren't specifically about it.
Previously config.py full enabled SSLv3, but it no longer does since it is
deprecated.
The splitting of this test into two versions depending on whether SHA-1 was
allowed by the server was a mistake in
5d2511c4d4 - the test has nothing to do with
SHA-1 in the first place, as the server doesn't request a certificate from
the client so it doesn't matter if the server accepts SHA-1 or not.
While the whole script makes (often implicit) assumptions about the version of
GnuTLS used, generally speaking it should work out of the box with the version
packaged on our reference testing platform, which is Ubuntu 16.04 so far.
With the update from Jan 8 2020 (3.4.10-4ubuntu1.6), the patches for rejecting
SHA-1 in certificate signatures were backported, so we should avoid presenting
SHA-1 signed certificates to a GnuTLS peer in ssl-opt.sh.
Because two buffers were aliased too early in the code, it was possible that
after an allocation failure, free() would be called twice for the same pointer.
Previously mocked non-blocking read/write was returning 0 when buffer was empty/full. That was causing ERR_SSL_CONN_EOF error in tests which was using these mocked callbacks. Beside that non-blocking read/write was returning ERR_SSL_WANT_READ/_WRITE depending on block pattern set by test design. Such behavior forced to redesign of these functions so that they could be used in other tests
This error occurs when free space in the buffer is in the middle (the buffer has come full circle) and function mbedtls_test_buffer_put is called. Then the arguments for memcpy are calculated incorrectly and program ends with segmentation fault
If there was a fatal error (bizarre behavior from the standard
library, or missing test data file), execute_tests did not close the
outcome file. Fix this.
In a unit test we want to avoid accessing the network. To test the
handshake in the unit test suite we need to implement a connection
between the server and the client. This socket implementation uses
two ring buffers to mock the transport layer.
In a unit test we want to avoid accessing the network. To test the
handshake in the unit test suite we need to implement a connection
between the server and the client. This ring buffer implementation will
serve as the said connection.
The new macro ASSERT_ALLOC allocates memory with mbedtls_calloc and
fails the test if the allocation fails. It outputs a null pointer if
the requested size is 0. It is meant to replace existing calls to
mbedtls_calloc.
MBEDTLS_PK_SIGNATURE_MAX_SIZE is tested in Mbed Crypto. Its effect on
Mbed TLS is also tested via the X.509 tests. The case of
MBEDTLS_MPI_MAX_SIZE < MBEDTLS_ECDSA_MAX_LEN, for which this component
was added as a regression test, is covered by config-suite-b.h which
is tested via test-ref-configs.pl.
When running 'make test' with GNU make, if a test suite program
displays "PASSED", this was automatically counted as a pass. This
would in particular count as passing:
* A test suite with the substring "PASSED" in a test description.
* A test suite where all the test cases succeeded, but the final
cleanup failed, in particular if a sanitizer reported a memory leak.
Use the test executable's return status instead to determine whether
the test suite passed. It's always 0 on PASSED unless the executable's
cleanup code fails, and it's never 0 on any failure.
FixARMmbed/mbed-crypto#303
Some sanitizers default to displaying an error message and recovering.
This could result in a test being recorded as passing despite a
complaint from the sanitizer. Turn off sanitizer recovery to avoid
this risk.
