Set Log Level to Info when the HTTP socket reports ENETDOWN,
instead of logging an error each time the network is down.
Signed-off-by: Joel Schaller <joel.schaller16@gmail.com>
When clock accuracy is 1KHz or lower, udp_upload rate has
much lower accuracy. eg. 20M rate has 60M actual throughput.
It's because any packet_duration less than 1000us will be
treated as 1 tick. In this case rate cannot control throughput.
Add clock compensate when time clock is lower than 1KHz to fix it.
In every 10 ticks or longer delta time, calculate packets sent
actual number and expected number, and then compensate ticks.
And add compensate ticks to delay ticks count.
If we cannot compensate a minus ticks as delay cannot be negative,
add it to next compensate period.
If compensate sum ticks is lower than -1000, it means current rate
is higher than low level capability. So no need to compensate
in remaining traffic.
Signed-off-by: Fengming Ye <frank.ye@nxp.com>
Kconfig makes sure the NET_PKT_TIMESTAMP_STACK_SIZE with default value
1024 when NET_PKT_TIMESTAMP_THREAD is enabled.
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
This commit adds a new Kconfig option NET_LINK_ADDR_CUSTOM_LENGTH
that allows to set custom link layer address length if your link layer
technology is not supported directly. If this option is set to a value
greater than 0, that value is used as link layer address length.
Signed-off-by: Jani Hirsimäki <jani.hirsimaki@nordicsemi.no>
The NET_MGMT_EVENT_INFO_DEFAULT_DATA_SIZE is used to set the default
size of the data field in the net_mgmt_event_info structure. This
change allows the user to configure the size of the data field
according to their needs.
Signed-off-by: Jani Hirsimäki <jani.hirsimaki@nordicsemi.no>
Rebinding of UDP sockets works just fine and there's really no reason to
disallow it. Keep the restriction for stream sockets only (as our TCP
implementation doesn't support rebinding).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Add support to handle domain match and suffix match parameters
for proper server certification validation.
Signed-off-by: Triveni Danda <triveni.danda@nordicsemi.no>
When OpenThread iface is brought up `ot ifconfig up` there are several
modules that will attempt to open a platform socket and perform bind
and bind to netif operation.
Since now, `sockfd_upd` structure was initialized after the backbone
interface announced connectivity, but this implies that OpenThread
interface will always be brought up only after this event, which is not
true, or imposed.
Signed-off-by: Cristian Bulacu <cristian.bulacu@nxp.com>
When remote peer have closed the PPP link normally, the
PPP stack on Zephyr side switches back to ESTABLISH phase
to be ready for next handshake.
When calling net_if_down() on the interface, it should not
try to initiate LCP link termination, but instead go directly
to DEAD phase.
See https://datatracker.ietf.org/doc/html/rfc1661#section-3.2
Signed-off-by: Seppo Takalo <seppo.takalo@nordicsemi.no>
No need to use Posix eventfd API in core network code as zvfs_eventfd
is compatible with it and we can now avoid using Posix headers
unnecessarily.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
As midi2 is provided by networking subsystem it should not
depend on any features provided by Posix. Convert Posix poll
API calls to zsock poll ones. There is no functionality changes,
only naming changes.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Corrected the RFC2131 section reference in DHCPv4 comments and help text.
The random delay before sending the initial DHCPDISCOVER message is
defined in Section 4.4.1 of RFC2131, not in 4.1.1 as previously stated.
Signed-off-by: Ofir Shemesh <ofirshemesh777@gmail.com>
In rare occasions iface and TCP mutexes could cause a deadlock. As
notifying the interface readiness takes place just before the iface
mutex is released, it should be not harm to release it just before the
TCP is notified about interface going down to avoid the deadlock.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The current approach was buggy, for example the TCP context could be
unrefed twice in case of forced close. Or in case of a race, when the
application closed the socket first, the TCP context wouldn't be
dereferenced at all.
Calling the tcp_conn_close() unconditionally in case of forced close
solves all those issues.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
A listening TCP context has no active connection therefore it has no a
second ref from the TCP stack. Thereby, when shutting down connections
when interface goes down, net_tcp_put() should not be called on a
listening socket to release the ref on the TCP stack behalf.
Instead, report the error via the registered accept_cb callback no
notify the application about the error, which should then close the
socket and release the associated TCP context.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Consecutive resource instance name should be encoded regardless of
whether timestamp is present or not.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
As per OCPP 1.6, "State of charge of charging vehicle in percentage" is
"Soc", not "SoCState".
Fixes: zephyrproject-rtos/zephyr#98870
Signed-off-by: Benjamin Cabé <benjamin@zephyrproject.org>
Fix incorrect format specifiers in lwm2m_read_cached_data() where
size_t variables were printed using %u and %d instead of %zu. This
caused CI failures on ARM64 platforms where size_t is long unsigned
int, triggering -Werror=format warnings.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Check the address family of the packet before passing it to a ICMP
handler, to avoid scenarios where ICMPv4 packet is paseed to a ICMPv6
handler and vice versa.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
If verdict is NET_OK net_pkt may already be unreferenced and reutilized,
so that setting the `l2_processed` flag would be set on a different
packet, corrupting its state. Avoid this situation by only setting the
l2_processed flag for NET_CONTINUE.
Signed-off-by: Tim Pambor <tim.pambor@codewrights.de>
The net socket service implementation permanently opens a file
descriptor, which should be taken into account by the build system.
Signed-off-by: Jordan Yates <jordan@embeint.com>
As the net_pkt might have already been sent by net_if_try_send_data()
function, the pkt might already contain garbage data. So do not try
to access if after that send call but remember the used iface and family
and use them to update the statistics.
The issue was seen with qemu_x86_64 and qemu_cortex_a53 when CONFIG_SMP
was enabled.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
If CONFIG_NET_STATISTICS_PER_INTERFACE is not set, then the
net_shell_print_statistics_all() function was not found and
the compilation was failing.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
The LwM2M SENML JSON/CBOR decoder expects a name entry for every
time-serialized record. While this was already the case for resources,
resource instances were missing their corresponding name entries. This
change ensures consistency by adding name entries for resource instances
as well.
Signed-off-by: Simon Walz <simon.walz@autosen.com>
Wrong printf format was using for size_t variable which caused
compiler warning and a failure.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
This commit fixes an issue introduced in #97531. Multicast routes added
by application are now set to have a prefix length of 16 bits instead of
128. This will allow multicast routing to all IPv6 addresses from those
groups.
For the OpenThread multicast listener callback, route length will be set
to 128 bits, as a MLD event is expected when an address is subscribed.
Route lookup has been removed since one will always be found as the
application registers multicast routes with same scope, but with a
smaller prefix length.
Fixed issue regarding packet forwarding and BBR state. Previously, code
was covering only the case wehn BBR was secondary and packet was
received from backbone interface. A secondary BBR should not forward a
multicast packet from OT interface to backbone interface and vice versa.
Improved unicast forwarding function to return from the beginning if the
packet has a multicast destination.
Signed-off-by: Cristian Bulacu <cristian.bulacu@nxp.com>
If the network interface goes down, close all TCP connections
that are bound to that interface. The reasoning here is that we
need a way to remove IP address from the interface and it might
not be possible if there is a TCP socket that is bound to that address.
If the interface comes back on, we might get the same IP address in
which case the socket closure was not really needed but it is not
possible to know in advance.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
If the network interface goes down, then we do not have any time
to tear down TCP connection gracefully because there is no more
a connection to send data to. In this case the TCP connection is
forcefully closed without going into FIN_WAIT_1 state.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
Using the Wi-Fi credentials system, we need to stop connecting with stored
credentials once a connection has been successful.
Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
Zephyr's long term goal for crypto support is to use only PSA API. This
commit replaces usages of MD with proper PSA API for HMAC.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
Accept the `ASYNC_CTRL_CHAR_MAP` option in configuration messages from
the peer. The map is reset to the default value each time the interface
comes up.
Signed-off-by: Jordan Yates <jordan@embeint.com>
Expose the currently configured value for the PPP peers Asynchronous
Control Character Map through a public function.
Signed-off-by: Jordan Yates <jordan@embeint.com>
The fix in 1264a923f3 was incomplete, because
it doesn't initialize the variable. To quote from opengroup [1]:
address_len
Either a null pointer, if address is a null pointer, or a pointer to a
socklen_t object which on input specifies the length of the supplied
sockaddr structure, and on output specifies the length of the stored
address.
This caused the returned address to be incomplete, because it got truncated
depending on what addrlen_local got initialized with implicitly. This broke
talking to discovered MQTT-SN gateways.
I intend to implement integration tests for the MQTT-SN UDP transport to
prevent such issues in future, but that will be done in a separate PR.
[1] https://pubs.opengroup.org/onlinepubs/9699919799/functions/recvfrom.html
Signed-off-by: Michael Zimmermann <michael.zimmermann@sevenlab.de>
Adjust the way the SKIP option worked. Before this patch, a constant
priority offset was considered "high priority" this had the effect, that
the threads assigned to work on the priority were effectively usesless.
To fix it, consider this immediate handling as a pseudo-queue and compute
the tc-thead-mapping based on the effective count (+1 if skipping is
enabled). This makes it so that all threads are usefull and the
high-priority skip-path is considered as a pseudo tc-thread.
Signed-off-by: Cla Mattia Galliard <clamattia@gmail.com>
This commit refactored IPV6 and IPV4 address events. Events have been
defined independently for each installed callback.
Previously, IPV6 address event was not triggered due to a mix of IPV4
and IPV6 flags.
Signed-off-by: Cristian Bulacu <cristian.bulacu@nxp.com>
Explicitly cast down the string width to int, as otherwise it is
size_t which may have a bigger size.
Avoids the following build error:
subsys/net/lib/shell/dns.c:496:67: error: field precision specifier
‘.*’ expects argument of type ‘int’, but argument 4 has type ‘size_t’
{aka ‘long unsigned int’}
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
