Building with clang warns:
tests/net/socket/tcp/src/main.c:377:13: error: variable 'c_sock' is used
uninitialized whenever 'if' condition is false
[-Werror,-Wsometimes-uninitialized]
} else if (family == AF_INET6) {
^~~~~~~~~~~~~~~~~~
tests/net/socket/tcp/src/main.c:396:15: note: uninitialized use occurs here
test_connect(c_sock, s_saddr, addrlen);
^~~~~~
tests/net/socket/tcp/src/main.c:377:9: note: remove the 'if' if its
condition is always true
} else if (family == AF_INET6) {
^~~~~~~~~~~~~~~~~~~~~~~~
tests/net/socket/tcp/src/main.c:360:12: note: initialize the variable
'c_sock' to silence this warning
int c_sock;
^
= 0
tests/net/socket/tcp/src/main.c:377:13: error: variable 's_sock' is used
uninitialized whenever 'if' condition is false
[-Werror,-Wsometimes-uninitialized]
} else if (family == AF_INET6) {
^~~~~~~~~~~~~~~~~~
tests/net/socket/tcp/src/main.c:387:12: note: uninitialized use occurs here
test_bind(s_sock, s_saddr, addrlen);
^~~~~~
tests/net/socket/tcp/src/main.c:377:9: note: remove the 'if' if its
condition is always true
} else if (family == AF_INET6) {
^~~~~~~~~~~~~~~~~~~~~~~~
tests/net/socket/tcp/src/main.c:361:12: note: initialize the variable
's_sock' to silence this warning
int s_sock;
^
= 0
Not really needed since we have zassert_unreachable(), but doesn't hurt
to initialize the variables.
Signed-off-by: Tom Hughes <tomhughes@chromium.org>
After recent kernel changes there's some thread scheduling race when
running tests, therefore add k_yield() at the end of the test to make
sure the network stack has a chance to run in between tests.
Make sure CONFIG_NET_TCP_TIME_WAIT_DELAY is set to 0 so that TCP
connections are released immediately.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Make sure that buffer configuration and buffer lengths provided in
struct msghdr are left intact after being processed by recvmsg().
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Make sure that buffer configuration and buffer lengths provided in
struct msghdr are left intact after being processed by recvmsg().
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The conversion from IPv4 string presentation to numeric value
did not check if the individual address value was between 0 and 255
inclusive.
Add also test case for this.
Fixes#84593
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
CONFIG_PSA_WANT_KEY_TYPE_ECC_KEY_PAIR_BASIC is promptless so it cannot
be selected. Moreover it's also automatically enabled by
CONFIG_PSA_WANT_KEY_TYPE_ECC_KEY_PAIR_GENERATE in the same overlay file
so there would be no need to explicitly enable it.
As for the IMPORT, EXPORT, DERIVE they are needed for the TLS connection
to work properly. Previously it was working because at least IMPORT and
EXPORT are internally enabled by Mbed TLS at build time. So here we
are basically doing the same enablements with Kconfigs in clear.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
The tests overflowed system workqueue stack slightly when running on
nRF52840, increase the stack size globally for the tests to avoid
potential issues on other platforms as well.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The test interfaces registered addresses within the same subnet. As now
only the subnet part of the address is verified when choosing source
address or interface, it was not deterministic which interface would be
selected. Therefore, make the two interfaces use two different subnets
so that tests that verified which address/interface was used work
reliably as earlier.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Add minimum flash requirement to the tests. This will effectively
exclude nrf5340dk/nrf5340/cpuapp/ns as it does not have
enough flash for the application.
Fixes#81862
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Add minimum flash requirement to the tests. This will effectively
exclude nrf5340dk/nrf5340/cpuapp/ns as it does not have
enough flash for the application.
Fixes#81608
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
The main problem of MBEDTLS_PSA_CRYPTO_LEGACY_RNG is that it
brings in some legacy modules (entropy + ctr_drbg/hmac_drbg)
which means extra ROM/RAM footprint.
MBEDTLS_PSA_CRYPTO_EXTERNAL_RNG instead simply calls to the
CSPRNG which makes it definitely smaller.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
remove k_work related code and change
the argument of the callback to `struct net_socket_service_event`.
Signed-off-by: Fin Maaß <f.maass@vogl-electronic.com>
CONFIG_ZVFS_POLL_MAX is now used to control the maximum number of poll()
entires. Thereby, CONFIG_NET_SOCKETS_POLL_MAX is redundant and shall
be deprecated.
Modify the defaults for NET_SOCKETS_POLL_MAX and ZVS_POLL_MAX so that
the deprecation actually makes sense instead of symbol removal. In case
the application still sets the old config, it will modify the
ZVS_POLL_MAX default.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Remove the `work_q` parameter from `NET_SOCKET_SERVICE_SYNC_DEFINE` and
`NET_SOCKET_SERVICE_SYNC_DEFINE_STATIC` as this feature was dropped
during review but the removal was not 100% complete.
Signed-off-by: Jordan Yates <jordan@embeint.com>
Make sure that CONFIG_ZVFS_OPEN_MAX in the tests and samples is at
least the same value as CONFIG_NET_SOCKETS_POLL_MAX so that we
cannot have more polled sockets than there are available file
descriptors.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Since fc007eeef5 TLS sockets support
timeout based on CONFIG_NET_SOCKETS_CONNECT_TIMEOUT value. The TLS test
suite reduced this config value to speed up the tests, however it turns
out the 100ms timeout is too low to finalize the handshake on slower
platforms (nrf52840 specifically). Therefore increase the timeout value
to 200ms to prevent test failures on those.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Since the same test might be run in parallel on different platforms,
the one that it's started later will fail due to port re-use for
the OpenSSL server.
Since it's not really required to run the tests on native_sim and
native_sim/native/64, the latter is removed in order to solve this
problem.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
This commit adds a test to showcase how to configure TLS 1.2 and 1.3
socket connections against an OpenSSL server. Only a limited number
of combinations are available for now:
- TLS 1.2
- RSA certificate and key exchange.
- ECDSA certificate and ECDHE key exchange.
- TLS 1.3 only supports ephemeral (ECDHE) key exchange with/without
session tickets.
Since the goal is to test TLS connection and not low level ethernet
functionalities or similar, the only supported platform is "native_sim"
where Linux sockets are used to connect to the OpenSSL server locally.
The idea is that the Zephyr application acts a client and tries to connect
to the OpenSSL server running on the same PC. For sake of simplificity a
bash script is provided to start the OpenSSL server properly.
For completeness a bash script is also provided to re-generate
certificates and keys.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
Move the network buffer header file from zephyr/net/buf.h to
zephyr/net_buf.h as the implementation now lives outside of the networking
subsystem.
Add (deprecated) zephyr/net/buf.h header to maintain compatibility with old
file path.
Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Currently, the test doesn't clear the control data buffer before
calling recvmsg. This leads to recvmsg being unable to add the new
control data, which corresponds to the current received data.
This commit aims to clear the control data buffer to match most use
cases, when the control data buffer is empty before calling recvmsg.
Signed-off-by: Axel Le Bourhis <axel.lebourhis@nxp.com>
The test case for SO_BINDTODEVICE socket option was flaky, the client
socket always sent the datagram to the IP address of the second
interface, so in theory every packet should end up on that interface.
In practice though, due to imperfect loopback packet handling, the test
worked as the packet ended up on the interface it was sent from.
The test should send datagrams to the IP addresses of the interface 1
and 2 alternatively. The server socket binds to ANY address, so w/o
interface binding it should receive all datagrams, so it allows to
verify if SO_BINDTODEVICE filtering works fine.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
As found in PR #75525, we should not modify the polled fd array
in multiple places. Because of this fix, the async version of
the socket service could start to trigger while it is being handled
by the async handler. This basically means that the async version
cannot work as intended so remove its support.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Make sure we at least build test the network socket tracing
support. The tracing tests do not enable networking so do the
socket tracing tests here.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
For each of the fdtable.h functions listed below, convert the
z_ prefixed semi-private functions to use the zvfs_ prefix.
ZVFS stands for Zephyr Virtual File System and
is intended to be a common library used by the C library,
POSIX API, Networking, Filesystem, and other areas.
There are already a few functions in fdtable.h that use the
zvfs_ prefix, so this change is mostly about unifying them in
a way that uses a suitable prefix ("namespace") so that it can
be considered a public API.
- z_alloc_fd
- z_fdtable_call_ioctl
- z_finalize_fd
- z_finalize_typed_fd
- z_free_fd
- z_get_fd_obj
- z_get_fd_obj_and_vtable
- z_get_obj_lock_and_cond
- z_reserve_fd
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
Fill-in the mode field of the fd_entry so that the
implementation can be made aware that the specific file
descriptors created are sockets.
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
The preprocessor gives this error when calling zsock_sendto
in user mode when there is a mismatch with amount of parameters
to the sendto wrapper call.
tests/net/socket/udp/src/main.c: In function ‘comm_sendto_recvfrom’:
tests/net/socket/udp/src/main.c:125:46: error: macro "zsock_sendto" \
requires 6 arguments, but only 5 given
125 | 0, &addr, addrlen);
| ^
In file included from include/zephyr/net/socket.h:1381,
from tests/net/socket/udp/src/main.c:15:
include/generated/zephyr/syscalls/socket.h:266: note: \
macro "zsock_sendto" defined here
266 | #define zsock_sendto(sock, buf, len, flags, dest_addr, addrlen)
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
The POSIX_MAX_FDS option does not correspond to any standard
POSIX option. It was used to define the size of the file
descriptor table, which is by no means exclusively used by
POSIX (also net, fs, ...).
POSIX_MAX_FDS is being deprecated in order to ensure that
Zephyr's POSIX Kconfig variables correspond to those defined in
the specification, as of IEEE 1003.1-2017. Namely,
POSIX_OPEN_MAX. CONFIG_POSIX_MAX_OPEN_FILES is being deprecated
for the same reason.
To mitigate any possible layering violations, that option is
not user selectable. It tracks the newly added
CONFIG_ZVFS_OPEN_MAX option, which is native to Zephyr.
With this deprecation, we introduce the following Kconfig
options that map directly to standard POSIX Option Groups by
simply removing "CONFIG_":
* CONFIG_POSIX_DEVICE_IO
Similarly, with this deprecation, we introduce the following
Kconfig options that map directly to standard POSIX Options by
simply removing "CONFIG":
* CONFIG_POSIX_OPEN_MAX
In order to maintain parity with the current feature set, we
introduce the following Kconfig options.
* CONFIG_POSIX_DEVICE_IO_ALIAS_CLOSE
* CONFIG_POSIX_DEVICE_IO_ALIAS_OPEN
* CONFIG_POSIX_DEVICE_IO_ALIAS_READ
* CONFIG_POSIX_DEVICE_IO_ALIAS_WRITE
Gate open(), close(), read(), and write() via the
CONFIG_POSIX_DEVICE_IO Kconfig option and move
implementations into device_io.c, to be conformant with the
spec.
Lastly, stage function names for upcoming ZVFS work, to be
completed as part of the LTSv3 Roadmap (e.g. zvfs_open(), ..).
Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>
test_packet_sockets_dgram create ETH_P_TSN sockets but
send and recv ETH_P_IP packet.
Fix it by sending ETH_P_TSN packet.
Signed-off-by: Fengming Ye <frank.ye@nxp.com>
In order to be compatible with Linux AF_PACKET socket calls, the
protocol field needs to be in network byte order.
So for example, if user wants to receive all packets, then the
protocol field needs to be set as "htons(ETH_P_ALL)".
See Linux manual page at
https://www.man7.org/linux/man-pages/man7/packet.7.html
for details.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
This commit fixes#73337.
Before #72243 Mbed TLS was not using
USE_PSA and all PSA features were not enabled. After #72243
if BUILD_WITH_TFM is set:
- USE_PSA in Mbed TLS is enabled by default and
- all PSA features are enabled.
This commits reverts both changes for net.socket.register.tls
test case.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
Remove the `_MAC` part because those Kconfig options enable only hash
algorithms, nothing MAC-related, and the `_ENABLED` part to align the
naming to the Mbed TLS defines (plus we don't need such a part).
As a bonus, enabling SHA-256 does not automatically enable SHA-224
anymore.
See the migration guide entries for more details on the practical
changes.
Signed-off-by: Tomi Fontanilles <tomi.fontanilles@nordicsemi.no>
The ztest stack was to low to execute tests on some platforms (spotted
on nrf52840dk/nrf52840), hence increase it.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The native_posix board does not work well with eventfd so disable
it with tests that use socket service API.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
The LOG_MODULE_REGISTER() was not called anywhere so if one tried
to enable logging, there was a linker error about missing logger
functions.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Instead of using net_context API directly, the DNS resolver is
changed to use the socket service API. This allows DNS access
for offloaded sockets and can lower overall memory consumption.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
According to 'man 7 socket' about SO_PRIORITY:
Setting a priority outside the range 0 to 6 requires the CAP_NET_ADMIN
capability.
So use 6 instead of 8, in order to make UDP socket tests pass with native
offloaded sockets on native_sim platform.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
A few tests involving TCP were missing TCP teardown delay. Adding these
improves tests stability when testing in a loop for longer periods (i.
e. with CONFIG_ZTEST_SHUFFLE enabled).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Make sure the IPv6 source address selection works as expected
when using the IPV6_ADDR_PREFERENCES socket option.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
One should not call a function directly from within zassert()
as it can cause the value of errno to be lost as seen by this
failed test output
Assertion failed at tests/net/socket/reuseaddr_reuseport/src/main.c:151:
test_bind_success: (zsock_bind(sock, addr, addrlen) not equal to 0)
bind() failed with error 0
FAIL - test_ipv4_udp_bad_both_not_set in 0.000 seconds
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Fix the failures introduced in #72078 by manually enabling all the
hash algorithms as they used to be.
Signed-off-by: Tomi Fontanilles <tomi.fontanilles@nordicsemi.no>
