The data length is already calculated in tcp_in() so no need
to do it again in tcp_data_get(). Just pass the length to the
tcp_data_get() function.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This change enables support for DNS service discovery
(RFC 6763) in the mdns_responder service and sample app.
Fixes#29429
Signed-off-by: Christopher Friedt <chrisfriedt@gmail.com>
This change adds support for DNS Service Discovery (DNS-SD)
as described in RFC 6763.
Fixes#29099
Signed-off-by: Christopher Friedt <chrisfriedt@gmail.com>
This change adds net_context_port_in_use(), which is a simple
wrapper around net_context_check_port() and is used to check
if a particular socket is bound to a given IP address.
Fixes#29649
Signed-off-by: Christopher Friedt <chrisfriedt@gmail.com>
Do not send the original pkt in 6lo based networks as in those
the IPv6 header is mangled and we would not be able to do any
resends of the original pkt. So for 6lo networks, clone the
pkt and send it to peer. The original pkt is kept in sent list
in case we need to resend to peer.
Fixes#29771
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Track the number of dropped TCP data segments and number of dropped
TCP packets in network statistics. It is useful to see these
numbers separately.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add a simple backoff mechanism between consecutive registration attempts
in case of registration failures. Finally, notify the application in
case the registration failed several times.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Refactor the boostrap regstration procedure, by splitting the message
creation and sending into a separate function, in similar manner as
it's done with regular registration.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Currently, when socket errors occur during receive, the LwM2M engine
restarts the state machine and registers again to the server. While this
works in simple use case (only RD client socket open), it's not a valid
approach when more sockets are open (FW update socket).
Fix this by introducing socket fault callback, which is registered by
the LwM2M engine users. This way, a proper socket owner is notified on
error and can pertake appropriate action.
For RD socket errors the behaviour remains the same - the state machine
is reset and the client registers again to the server. For FW update
socket, handle the error by reopening the socket and retransmitting the
last request. This allows to resume the download from the point the
error occured, w/o a need to start from scratch.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Call mbedtls_ssl_conf_max_frag_len() on created TLS context
configuration, so that Maximum Fragment Length (MFL) will be sent to
peer using RFC 6066 max_fragment_length extension. MFL value is
automatically chosen based on MBEDTLS_SSL_OUT_CONTENT_LEN and
MBEDTLS_SSL_IN_CONTENT_LEN mbed TLS macros.
This extension is mostly useful for TLS client side to tell TLS server
what is the maximum supported receive record length.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
This commit corrects the maximum allowed amount of children to
match Thread specification.
Signed-off-by: Eduardo Montoya <eduardo.montoya@nordicsemi.no>
If the peer ACKs data when it closes the connection, update
our sequence number accordinly. The connection would eventually
be terminated but this will avoid extra resends by the peer.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
When a connection is being closed, it is possible that the application
will have a lock to net_context and TCP2 connection lock. If we then
receive a final TCP2 ACK and close the connection, the locking order
get switched and TCP2 will first try to get its own lock and then the
net_context lock. This will lead to deadlock as the locking ordering
is now mixed.
The solution is to unref the TCP connection after releasing the
connection lock. The TCP connection unref function will anyway get the
lock so no need to do double locking.
Fixes#29444
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there are enough tls_context objects in the system (configured by
CONFIG_NET_SOCKETS_TLS_MAX_CONTEXTS), but there are not enough file
descriptors (configured by CONFIG_POSIX_MAX_FDS) to create underneath
TCP/UDP socket, then TLS socket creation fails with leaked tls_context.
Call tls_release() in ztls_socket() error path whenever underneath
TCP/UDP socket creation fails.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
According to LwM2M specfication v1.0.2, par. 5.3.2, the LwM2M client
MUST send an “Update” operation to the LwM2M Server whenever the
lifetime parameter of the Server object changes the server). The same
applies for the object instances created/deleted. The changes in objects
seem to already be handled, but the lifetime was not.
Additionally, the "Update" message shall only contain these parameters
which changed since the last update (including objects). As it's
straightforward to determine if the liftime changed but it's not easy
to tell if there were updates in the object instances, add an
additional parameter to the engine_trigger_update() function, indicating
that new object information shall be sent in the "Update" message.
Eventually add a proper error checking in `sm_send_registration` as the
function is reworked anyway.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The memset on firmware_ctx during PULL FW update initialization will
set the socket descriptor to a valid value of 0. This leads to an error
if parsing of the URI provided by the server fails, and the firware_ctx
is closed - the socket with a descriptor 0 will be accidently closed.
Fix this by invalidating the socket FD after the memset on
initialization.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
It shouldn't be optional to handle separate response, as it's a
mandatory requirement according to the RFC7252:
"The protocol leaves the decision whether to
piggyback a response or not (i.e., send a separate response) to
the server. The client MUST be prepared to receive either."
Therefore, remove the flag as separate responses are handled now
properly.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Separate response handling implemented in the engine was faulty. The
separate response was not acknowledged by the client, resulting in
spurious retransmissions from the server side.
Also, the pending CON message was retransmitted by the client even after
it was acknowledged by an empty ACK, but the respnse haven't arrived
yet. Fix this by adding a new `acknowledged` flag to the `lwm2m_message`
structure. Once acknowledged, the flag is set and the confirmable
message is no longer retransmitted. We keep the message on the pending
list in order to timeout properly in case separate response does not
arrive in time.
Finally, prevent the reply callback from being called twice in case
the response is transmitted separately from ACk. The callback should
only be called on the actual reply, not the empty ACK.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
So far this function existed as a static function in LwM2M PULL FOTA
module. Since such functionality will be needed in other places, make it
an internal API function.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
using CONFIG_NET_BUF_POOL_USAGE monitor avail_count,
this variable should be protect.
Protecting it by using atomic variable
Signed-off-by: Ehud Naim <ehudn@marvell.com>
Some LWM2M backends/servers, such as emxq, expect the sequence numbers
to begin on 0.
This change is in line with how other lwm2m clients, such as Anjay and
Wakama, starts the notification sequence.
Signed-off-by: Viktor Sjölind <viktor.sjolind@endian.se>
The commit 93e5181f ("net: context: Add locking for concurrent
access") added net_context locking to only IPv4 sockets.
That is not enough and we need locking also to other supported
socket address families like IPv6, SocketCAN and packet socket.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Prior to this commit, the LwM2M stack would TLV-encode integers
depending on their internal storage size. An integer with value 5 stored
in an int8_t would be encoded with length 1, but an integer stored in an
int32_t would be encoded as "00 00 00 05" with length 4.
This commit checks if the value is castable to a smaller int and encodes
it as such if so. This is cascading, so even a 64 bit integer with value
5 will be encoded with length 1.
Note that this does not seem to be required by the specification, but
this is how Anjay and the other LwM2M stack seem to do it.
Signed-off-by: Benjamin Lindqvist <benjamin.lindqvist@endian.se>
When we are sending ICMPv6 error message, we need to store the
link local addresses of the received packet somewhere in order
to know where to send the new error message.
Easiest is to store the ll addresses is to the error message
itself, just before where the sent packet will start in memory.
We cannot use the original pkt to store the ll addresses
as that packet might get overwritten if we receive lot of packets.
Fixes#29398
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Currently, there is a case for net_config_init function that for
timeout==0 and when iface is already up, the ip setup is not proceed
and the error message "Timeout while waiting network..." is logged.
This commit fixes it.
Signed-off-by: Lukasz Maciejonczyk <lukasz.maciejonczyk@nordicsemi.no>
Adds the socket option TLS_ALPN_LIST for SOL_TLS sockets
Passes the configured alpn list to the mbedtls config
on mbedtls init
Signed-off-by: Emil Hammarstrom <emil.hammarstrom@assaabloy.com>
If we are sending data directly, we already have TCP lock so
there is no need to do any locking. But when data is re-sent,
the work queue handler is doing the sending so we need to lock
the TCP connection.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
When receiving data that needs to be passed the data to application,
queue it for short time so that we do not have TCP connection lock
held. This way if the application wants to send data, there is no
possibility that the connection lock would prevent sending data.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Release the context lock before passing data to the application
socket as that might cause deadlock if the application is run
before the RX thread and it starts to send data and if the RX
thread is never able to run (because of priorities etc).
Fixes#29347
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
These defines are leftover of old platform settings implementation
and are not used any more.
Signed-off-by: Lukasz Maciejonczyk <lukasz.maciejonczyk@nordicsemi.no>
net_icmp4_input() may net_pkt_unref() a packet. The header mustn't be
accessed after this or the system may crash.
Signed-off-by: Oliver Hitz <oliver@net-track.ch>
Existing openthread_api_mutex_lock()/unlock() functions are
crucial to assure thread safety of an application which
needs to use OT API directly, but some applications may also
require a non-blocking version of the former for less critical
OT-related tasks.
Add openthread_api_mutex_try_lock() which never waits and
exits immediately if the mutex is held by another thread.
Signed-off-by: Damian Krolik <damian.krolik@nordicsemi.no>
If there are more than one RX or TX threads, then make the name
of each of them unique so that it is easier to figure them out
in "kernel stacks" command.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The name of the connection manager thread (conn_mgr_thread) was
unnecessarily long in "kernel stacks" command. So make the name
to "conn_mgr" which fits nicely to the output of that command.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If an address was obtained by IPCP, it should always be removed in
ipcp_down(). This commit replaces the predicate with something slightly
more robust.
Signed-off-by: Benjamin Lindqvist <benjamin.lindqvist@endian.se>
Windows 10 sends ECN-Echo and Congestion Window Reduced (CWR) flags
together with SYN flag in the connection establishment but the code
did not ignore these flags and send just SYN back (instead of SYN|ACK).
This caused the connection establishement in application level to
fail as the application was never notified about it.
Fixes#29258
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit moves IPv6 initialization from OT init to OT start to
avoid unwantedly bringing 802.15.4 radio up.
Previously, even when OT manual start was enabled, the radio would
be receiving frames resulting in unnecessary power consumption and
causing issues for instance when the device just wants to use
Bluetooth for provisioning befor moving to Thread.
Signed-off-by: Eduardo Montoya <eduardo.montoya@nordicsemi.no>
Another team reported that current default values for number of allowed
IP addresses per child (4) and and max number of children (10) are too
small for some customers.
Increased the values allowed configuring child count.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
OpenThread mac counters require rx failed notification to work properly.
Made use of previously implemented notification.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
Clip the timeout to 64 seconds, this avoids the timeout value to
increase to high values (e.g. several years).
Signed-off-by: Armand Ciejak <armand@riedonetworks.com>
This avoids code duplication. The same logic was present
in dhcpv4_send_request() and dhcpv4_send_discover().
Signed-off-by: Armand Ciejak <armand@riedonetworks.com>
The TCP2 was calling accept callback before actually finalizing
the connection attempt.
Fixes#29164
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
A normal websocket close sequence:
close(websock) ->
websocket_close_vmeth() ->
websocket_disconnect()
close(ctx->sock) called in the function websocket_disconnect()
and cause websocket_close_vmeth() called again.
Finally stack overflow by loop close call.
It's maybe a side-effect by PR #27485
Signed-off-by: Jackie Ja <qazq.jackie@gmail.com>
Improve token handling by removing special meaning of tokenlen == 0,
which allows to handle server requests w/o a token (so far such
requests would cause the lwm2m engine to autogenerate token in the
response).
In order to autogenerate token during message initialization, use
special symbol `LWM2M_MSG_TOKEN_GENERATE_NEW`. If no token is wished to
be used, simply set the tokenlen to 0.
Additionally, fix an issue with token autogeneration, where invalid
token len was used (0 instead of 8).
Fixes#28299
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
LwM2M engine did not set response code for the Bootstrap-finish message,
hence it replied with the code copied from the request which is not
correct. Fix this by setting correct code for the Bootstrap-finish
reply.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
So far, `LWM2M_RD_CLIENT_EVENT_BOOTSTRAP_TRANSFER_COMPLETE` event was
reported before the final ACK for the Bootstrap Finish was sent from the
client side. This could cause delays in the ACK sending, in case the
application wanted for instance to store the received data in flash.
Fix this, by reporting the
`LWM2M_RD_CLIENT_EVENT_BOOTSTRAP_TRANSFER_COMPLETE` event on the next
state tansition (before the actual registration starts).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
"Public Key or Identity" resource is of opaque data type, therefore it's
not correct to assume it will be a NULL terminated string (the existing
servers, for instance Leshan, does not include NULL terminator). Use the
actual size associated with the resource instead.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
So far, the resource instance structure kept only the information about
the buffer length provided to the resource (in the `data_len` field).
While this approach might be enough for integer resources, where the
actual data size is fixed, it did not work for opaque resources. It is
impossible to determine the actual opaque resource length after it's
been written into.
Fix this, by replacing the current `data_len` field of the
`lwm2m_engine_res_inst` with `max_data_len`, indicating the buffer
size, and making the `data_len` field to hold the actual data size of
the resource.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Currently, after `CONFIG_LWM2M_RD_CLIENT_SUPPORT_BOOTSTRAP` is enabled,
the LwM2M engine will initiate bootstrap procedure on each run. This
approach limits the flexibility of the application, as it's not always
necessary to go over the bootstrap procedure (for instance, the
application may decide to store the security object obtained during the
bootstrap in flash, and restore it on boot).
Fix this by introducing an additional `flags` parameter to the
`lwm2m_rd_client_start()` function, which provides information whether
to run bootstrap in the current session or not.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
So far, the LwM2M state machine started in the `ENGINE_INIT` state,
which made it exectue the registration/bootstrap registration even when
`lwm2m_rd_client_start()` was not called. With a new `ENGINE_IDLE`
state, the state machine can wait for the application to actually start
the client before proceeding. It also makes sense to stay in the
ENGINE_IDLE state after successfull deregistration, until the
application restarts the client.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Without removing the stale address obtained during IPCP, it will still
be present the next time we do IPCP, marked as "in use" by the network
stack even if it is stale. This turned out to be a showstopper for
restarting the PPP stack on devices without static IP.
Signed-off-by: Benjamin Lindqvist <benjamin.lindqvist@endian.se>
This commit fixes PUSH FOTA when opaque content-format is used.
This consists of the following fixes:
* Moved `struct block_context` to a private header, so that it can be a
part of `struct lwm2m_input_context`. This allows content decoders to
make use of the block context data.
* Removed faulty `get_length_left` function from the plain text
decoder, and replace it with coap_packet_get_payload() to obtain the
actual payload size.
* Introduce `struct lwm2m_opaque_context` as a part of block context,
which allows to keep track of opaque data download progress.
* Simplify `lwm2m_write_handler_opaque()` function. It will now only
make calls to `engine_get_opaque` - it's the decoder responsibility
to update the opaque context according to it's content format (for
instance TLV decoder should only update it with the actual opaque
data size, not the whole TLV).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The L4 connected/disconnected events are usually used to detect
when the application is connected to the network. Unfortunately
if the device has also a static address, then the connected event
might be created (for the static address) even if DHCPv4 is not
ready yet and application would not be able to connect (yet) to the
network. In order to allow the application to fine tune the network
connection creation, generate start, bound and stop events for DHCPv4.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The "tc" field which was holding the traffic class thread
priority is not used nor needed so remove it from the struct
to save some space.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Replace all calls to the assert macro that comes from libc by calls to
__ASSERT_NO_MSG(). This is usefull as the former might be different
depending on the libc used and the later can be customized to reduce
flash footprint.
Signed-off-by: Xavier Chapron <xavier.chapron@stimio.fr>
Using zephyr's internals zsock_ calls make mqtt library more compatible,
now it does not depend on NET_SOCKETS_POSIX_NAMES.
Signed-off-by: Jan Pohanka <xhpohanka@gmail.com>
No real need to assert when the send timer is cancelled. Just
check if there is re-transmission going on and do nothing if
there is not.
Fixes#28758
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The local and accepted socket was not bound which caused the
local address to be set as NULL. This then caused issues when
zsock_getsockname() was called by the application.
Fixes#28735
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Since the info parameter is only read from and never written the const
in the function parameter should be present.
Signed-off-by: Christian Taedcke <christian.taedcke@lemonbeat.com>
This commit adds additional Kconfigs that allow for changing
configuration values for the Zephyr port in OpenThread.
Those values are:
- number of the internal OT message buffers
- number of the state change callbacks
- number of the EID-to-RLOC cache entries
- size of the NCP buffer
Signed-off-by: Rafał Kuźnia <rafal.kuznia@nordicsemi.no>
Instead of directly checking the multicast IPv4 address, use
the net_ipv4_is_addr_mcast() utility function.
Fixes#26584
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
MQTT client state is protected using mutex. That mutex however is
temporarily unlocked when calling event callbacks. This means that in
client_disconnect() transport can already be disconnected, but without
marking it as such in client->internal.state.
When mutex is unlocked in event_notify() function, then there are two
possible paths of failure:
1) First possibility is when RX and TX are called from two separate
threads, so that the other thread gets resumed and functions like
verify_tx_state() (e.g. in mqtt_publish()) allow to continue
communication over disconnected medium.
2) Another possibility is that user calls mqtt_abort() or
mqtt_disconnect() in event handler.
In both cases MQTT library tries to send or receive data, possibly
followed by second close() of underlying file descriptor.
Prevent using disconnected transport by clearing MQTT client state right
after calling mqtt_transport_disconnect(), without releasing mutex, even
for a while.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
When compiling with CONFIG_NET_NATIVE=n and CONFIG_NET_OFFLOAD=n
then the following error is printed.
```
In file included from zephyr/subsys/net/ip/ net_if.c:23:0:
zephyr/subsys/net/ip/net_private.h: In function 'net_context_state':
zephyr/subsys/net/ip/net_private.h:58:27:
error: type of 'context' defaults to 'int' [-Werror=implicit-int]
static inline const char *net_context_state(context)
```
This add `struct net_context *` as type for context.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Remove obsolete include of the shell_uart.h file.
It is sufficient to include the shell.h file.
Signed-off-by: Jakub Rzeszutko <jakub.rzeszutko@nordisemi.no>
Since commit b3a1ede830 OpenThread uses the Zepyhr settings submodule
instead of writing to the flash directly. The flash.c file is not
compiled anymore, so let's just remove it. Also remove the
OT_PLAT_FLASH_PAGES_COUNT Kconfig option which was solely used by that
file.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Check that Ethernet header is in the first net_buf fragment.
This is very unlikely to happen as device driver is expected
to only deliver proper Ethernet frames to upper stack.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We run various TCP function from work queue. Make sure the
connection lock is taken before accessing the connection.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
support longer name like urn:dev:ops:{OUI}-{ProductClass}-{SerialNumber}
or urn:imei-msisdn:###############-###############
Signed-off-by: Pascal Brogle <pascal.brogle@husqvarnagroup.com>
Fixed stack being to small for joiner operations.
Enabled auto joining even in case of manual start.
Fixed attachement of SED on norfic radios.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Instead of casting k_delayed_work directly to k_work, use the
k_work field name. This avoids warnings from Coverity and
allows the code to work even if the k_delayed_work fields are
re-ordered in the future.
Coverity-CID: 214346
Fixes#28659
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add dependency to UDP in DHCPv4 Kconfig option as UDP is needed
in DHPCv4 to work properly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Bug description:
When in tcp_conn_unref(), in case one of the delayed works is already
submitted to sysworkq (after delay period), e.g. send_timer, the check
of k_delayed_work_remaining_get() prevents calling
k_delayed_work_cancel().
This leads to corrupting sysworkq when zeroing struct tcp* conn.
Note that the "next" pointer for the work queue is part of the struct
work (in _reserved field). Which is, in this case, a member of struct
tcp.
Scenario leading to the bug:
(1) net_tcp_connect() is called from a work in sysworkq
(2) net_tcp_connect() submits conn->send_timer to sysworkq
(3) while net_tcp_connect() is waiting on connect_sem, delay period
passes (z_timeout) and send_timer enters sysworkq work slist
(4) also, some other code (app) submits more works to queue, now pointed
by conn->send_timer in sysworkq work list
(5) connection fails (no answer to SYN), causing a call to
tcp_conn_unref()
(6) tcp_conn_unref() is calling tcp_send_queue_flush()
(7) checking k_delayed_work_remaining_get(&conn->send_timer) returns 0
due to delay period end, but send_timer is still in sysworkq work
slist (sysworkq thread still hasn't handled the work)
(8) BUG!: no call to k_delayed_work_cancel(&conn->send_timer)
(9) back in tcp_conn_unref(), a call to memset(conn, 0, sizeof(*conn))
zeroes conn->send_timer
(10) conn->send_timer is pointed to in sysworkq work slist, but is
zeroed, clearing pointer to following works submitted in stage (4)
(11) EFFECT! the works in stage (4) are never executed!!
NOTES:
* k_delayed_work_cancel(), handles both states:
(1) delayed work pends on timeout and
(2) work already in queue.
So there is no need to check k_delayed_work_remaining_get()
* This is also relevant for conn->send_data_timer
Solution:
removing checks of k_delayed_work_remaining_get(), always calling
k_delayed_work_cancel() for work in struct tcp, in unref, before memset
Signed-off-by: David Komel <a8961713@gmail.com>
Instead of hardcoding the priority1 and priority2 values used
in BMCA, let the user tweak the values via Kconfig.
Fixes#28151
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Since conn->send_data_total is of time size_t we need to use %zu or
we'll get build errors in sanitycheck on 64-bit platforms
Fixes#28605
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
If the network driver for some reason did not set the data in
the network packet properly, then just drop it as we cannot do
anything with just plain net_pkt.
Fixes#28131
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is no handler for IPv4 broadcast packet, then ignore it
instead of trying to send an ARP message to resolve the senders
address.
Fixes#21016
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We need to have timer that closes the connection for good if
we do not get the FIN and ACK reponse from the peer.
If there is any pending data when application does close(),
send them before sending FIN.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is some error during connection creation, just bail
out in order to avoid null pointer access.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we try to send data but the sending window is full, then
try to kick the resend of the pending data.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is no space in the sending window, then return -EAGAIN
so that the caller may try later.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
We should have a max value for sending window so that application
is not able to use all our net_bufs for queueing packets.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The socket layer expects to receive -ENOBUFS if we do not have
any network buffers so cannot build a network message to be sent.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If there is an error, the net_context.c:context_sendto() will
free the net_pkt, so we must not do it here.
This commit fixes this error message:
<err> net_pkt: *** ERROR *** pkt 0x20421908 is freed already
(context_sendto():1672)
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Usually the out-of-memory situation will clear itself eventually,
so if that happens in TCP, then keep the connection running and
let the user to decide what to do next.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we run out of network buffers and cannot send data, and if
we have a blocking socket, then wait until new buffers are
available before returning. As this might lead to deadlock,
wait only max 10 seconds and return ENOMEM if we cannot get
buffers in a reasonable amount of time.
Fixes#28216
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of hardcoded value of 3, use the value from Kconfig file
so that user can tweak the TCP retry count.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Use SYS_SLIST_FOR_EACH_CONTAINER_SAFE() macro when searching
the connection list so that we notice if new entries are added
or removed in the list.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of forcing the slist node to be first in the tcp struct,
use the pointer to node when accessing the slist. This way we
can change the ordering of fields in tcp struct.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Change so that the caller of lwm2m_init_message is
responsible for generating a message id and remove message id generation
from lwm2m_init_message. Prevents generating a new id when the caller's
intent is to init a message with id 0.
Fixes#28283
Signed-off-by: Pascal Brogle <pascal.brogle@husqvarnagroup.com>
message id 0 and token 0 have special semantics in the lwm2m engine,
they are used to request generation of new id, mark them as such.
Signed-off-by: Pascal Brogle <pascal.brogle@husqvarnagroup.com>
Fix the OpenThread initialization to prevent the IPv6 interface to
be enabled when `CONFIG_OPENTHREAD_RAW` is set.
Signed-off-by: Eduardo Montoya <eduardo.montoya@nordicsemi.no>
Because the previous commit fixed the destination address check,
we need to handle msghdr (used by sendmsg()) in CAN socket.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
No need to specifically check CAN sockets here. This also means
that SocketCAN will need to support sendmsg() call too. This will
come in following commit.
Fixes#28229
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The timeout variable in `dtls_rx()` was initialized improperly.
Coverity ID: 214219
Fixes#28161
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The default string representing Sensor Type resource in Generic IPSO
object would not fit into the predefined buffer. Increase the buffer
size and add extra BUILD_ASSERT to detect this situation.
Coverity ID: 214225
Fixes#28164
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The timeout log error message condition in wrong. When the timout
happens the "count == -1" and the condition is invalid.
This commit fixes it.
Signed-off-by: Lukasz Maciejonczyk <lukasz.maciejonczyk@nordicsemi.no>
When the connection is terminated, make sure that any pending
data is feed to the application.
Fixes#28057
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Fix for a problem in current lwm2m firmware object implementation.
Transfer should not begin when an empty string is received.
Signed-off-by: Marin Jurjevic <marin.jurjevic@hotmail.com>
Some implementation consist of multiple libraries to be linked instead
of one. Added possibility to pass multiple libraries. Additionally
renamed the config name as it was stateing something different than it
does.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
If we receive a TCP segment with FIN | ACK | PSH flags, then
update the ack values properly.
Fixes#27982
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This new option allows to include `sockets_tls.c` into the build when
socket offloading is enabled.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Given that the offloaded poll handling differs from the poll handling of
native sockets (entire poll function call is offloaded), some
adjustements were needed to make TLS socket work with offloaded poll
calls.
To achieve this, in case socket offloading is used, instead of jumping
directly to the offloaded poll call, a TLS wrapper for the offloaded
poll will be called. This wrapper will do additional checks at the
mbedtls level, to verify that the event is only notified to the caller
when the application data is available (i. e. not to report events
during handshake or when partial data is received, not ready to
decrypt).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Separate TLS sockets implementation from net_context layer. Instead of
calling net_context functions directly, create an underlying TCP/UDP
socket, and call socket functions on it instead. This results in a
simpler design, where we don't need to duplicate specific parts of code
from the native socket implementation. Additionally, this allows to use
a different TCP/UDP stack underneath (i. e. when sockets offloading is
used).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
This commit adds helper function to verify protocol, which was done in
two different places in the code. The function returns the underlying
protocol information on success.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
When connection is closed and we send ACK flag, use proper seq
values so that any data that is still in flight will get acked too.
Currently this assumes that window is still open.
Fixes#27876
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Entropy device, as being unique here, does not need to be passed through
mbedtls and can instead set static locally.
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Now that device_api attribute is unmodified at runtime, as well as all
the other attributes, it is possible to switch all device driver
instance to be constant.
A coccinelle rule is used for this:
@r_const_dev_1
disable optional_qualifier
@
@@
-struct device *
+const struct device *
@r_const_dev_2
disable optional_qualifier
@
@@
-struct device * const
+const struct device *
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Following are the changes to variable names that are matching
with tag names (Rule 5.7 violations)
In kernel.h, event_type is matching with a tag name in
lib/os/onoff.c. Added a _ prefix to event_type and
also to the macro argument names.
In userspace.c, *dyn_obj is matching with the tag name
dyn_obj in the file itslef. Changed it to dyn
In device.h, device_mmio.h, init.h and init.c,
changed the *device to dev. Except for one change in
init.h
Signed-off-by: Spoorthy Priya Yerabolu <spoorthy.priya.yerabolu@intel.com>
Add more tests to verify that we discard malformed packets.
In order to simplify the testing, separate message validation to
dns_validate_msg() function in resolve.c. Allow that function to
be called from unit test. This way we can construct invalid DNS
messages in unit test and verify that they are discarded when
needed.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The ANCOUNT has nothing to do with label count so remove the
original while loop and just go through all the labels until
we have read all of them.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This is not possible with valid DNS messages but is possible if
we receive malformed DNS packet.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Allow user to create SOCK_DGRAM type AF_PACKET socket. This
allows user to send raw IP packets without specifying
L2 (like Ethernet) headers.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The steps_removed field must be placed right after root_system_id
so that priority vector comparision can be done in one memcmp()
call. This fixes the best master clock selection algorithm (BMCA).
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
In 'struct coap_resource' path description:
- the plus symbol represents a single-level wild card in the path;
- the hash symbol represents the multi-level wild card in the path.
This change keeps compatibility with RFC 7252 but allows handling
multiple requests in single function.
Signed-off-by: Eug Krashtan <eug.krashtan@gmail.com>
-Wimplicit-fallthrough=2 requires a fallthrough comment or a compiler
to tells gcc that this happens intentionally.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Added additonal checks in net_ipv6_input to ensure that multicasts
are only passed to the upper layer if the originating interface
actually joined the destination multicast group.
Signed-off-by: Philip Serbin <philip.serbin@lemonbeat.com>
Now that TCP2 is the default stack, make the shell compatible
with both stacks.
Note: net_tcp_foreach is not implemented in TCP2, so related
code depends on TCP1.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
PPP Phase Diagram [1] allows only one way phase change. In current
implementation there is an additional RUNNING phase, which is entered
just after NETWORK phase.
Prevent going back from RUNNING to NETWORK phase when Term-Req was
received, as this is meaningless for overall PPP operation and violates
PPP Phase Diagram property of having one way direction change.
This change also improves Adminitrative Close handling (calling
lcp_close()). This request results in moving into TERMINATE phase. Then
LCP is put down (by calling lcp_down()) and then ppp_link_down() is
called, which so far (before this patch) resulted in moving back to
NETWORK and then to DEAD. Right now (after this patch) we move directly
from TERMINATE to DEAD phase, which is exactly how [1] specifies it.
[1] https://tools.ietf.org/html/rfc1661#section-3.2
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
State Transition Table [1] specifies that Administrative Close should
result in CLOSING state. This is not respected in case of LCP, as
STOPPED state was forced in lcp_close().
Don't force going into STOPPED state in lcp_close() and rely on
ppp_fsm_close() to move to CLOSING state instead.
This patch fixes overall Adminitrative Close procedure and allows to
move back into fully operating PPP connection once again after
Adminitrative Open.
[1] https://tools.ietf.org/html/rfc1661#section-4.1
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Print role information next to port information. Also make
sure that port number is set correctly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Save some memory and separate common string snippets from larger
strings. In this case "does not match" sub-string is printed by
several warning prints so separate the sub-string from the bigger
string.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
net: route: Add prefix-based ipv6 multicast forwarding
This adds/reenables the feature of multicast routing/forwarding.
The forwarding decision is based on the added multicast routes
and the new network interface flag:
NET_IF_FORWARD_MULTICASTS.
Signed-off-by: Jan Georgi <jan.georgi@lemonbeat.com>
This patch implements optional authentication phase, which is done
between link establishment and network phases. It is part of LCP option
negotiation to decide whether authentication is needed and which
protocol will be used. For now we add only PAP support and try to
negotiate it when some other protocol (e.g. CHAP or EAP) is proposed
earlier. For simplicity reason we only add one way authentication
support, which means that we try to authenticate to the other peer, but
do not require authentication from it.
This is an important step to make PPP work with cellular network modems,
because most of them require to provide username and password within PPP
authentication phase. Those credentials are used by modem to login to
cellular network. In most cases however it is enough to provide dummy
values, because they are not verified. For this reason and simplicity of
this patch we hardcode PAP Peer-ID and Password now.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
This module had only some header includes, without actual code. Remove
it, as it is better to create one module per authentication protocol
anyway.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Add enums of PAP, CHAP and EAP authentication protocols. Also add their
string representations, so they will nicely show up in logs during
debugging.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Interpret -EINVAL return value from options' parse() callback as "option
value is not supported". After receiving such value nack() callback will
be executed in order to put supported value in the response.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Given the comment in `ipv6_event_handler`, the init function shall wait
for the last added IPv6 address to be confirmed with DAD. This is
contrary with current logic, which will only wait for the first address
to be confirmed.
This commit adds a break statement in the loop, not to overwrite the
recent address to verify with an older one.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
When verifying the parameters check NULL value separately.
This will avoid nasty warning message to be printed.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add function pointer to vtable and use that directly instead of
routing via ioctl() call. This is done as we are trying to get
rid of ioctl() calls in the system.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Do not route close() calls via ioctl() as that is error prone
and quite pointless. Instead create a callback for close() in
fdtable and use it directly.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of casting struct in6_addr, sin6_addr member is now used
directly, like in every other instance of calling net_ipv6_is_prefix().
Signed-off-by: Christian Taedcke <christian.taedcke@lemonbeat.com>
This macro has no use in code and was most probably replaced by
FSM_TIMEOUT (which resolves to CONFIG_NET_L2_PPP_TIMEOUT) with the same
default value.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Those macros are not used anymore after commit 35a2519091 ("net: l2:
ppp: add generic function for handling Conf-Req"), because we no longer
need preallocated table on stack for parsing option information.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
This option is not used after commit 50b2cafc42 ("net: l2: ppp: use
net_pkt API for replying to Configure-Req") has been applied. We don't
need to preallocate table on stack for parsing option information, so
information about maximum number of supported options is useless.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
This commit allows let build zperf sample with overlay-netusb.conf.
USB subsystem must be enabled by the application.
Signed-off-by: Emil Obalski <emil.obalski@nordicsemi.no>
For platforms using non-native stacks, net_tcp_init() should be
compiled out, similar to how it is done in tcp_internal.h.
Fixes#27463
Signed-off-by: Vincent Wan <vwan@ti.com>
Instead of just showing network interface pointer, print also
network interface index. This helps debugging.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
OT_RADIO_CAPS_SLEEP_TO_TX was added as a radio capability
for ieee802154 radio. Waiting on RX state before transmission
is alternative condition to OT_RADIO_CAPS_SLEEP_TO_TX support
as it was a result of OpenThread architecture and is actually
not needed in the Zephyr. Such change lets to start transmission
faster and lower SED device power consumption in active state
about 30%.
Signed-off-by: Kamil Kasperczyk <kamil.kasperczyk@nordicsemi.no>
Enable the new TCP stack as a default one in order to get people
to use it before Zephyr 2.4 is released.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Zephyr platform does not support MAC retransmissions on its own,
so OT_RADIO_CAPS_TRANSMIT_RETRIES capability was removed.
It should not be enabled basing on IEEE802154_HW_CSMA support,
as these are quite seperate features. Current implementation
assumes that platform performs retransmissions on its own,
what is not provided and leads to lack of MAC retransmissions.
Signed-off-by: Kamil Kasperczyk <kamil.kasperczyk@nordicsemi.no>
Allow user to enable CONFIG_NET_PKT_TXTIME and
CONFIG_NET_PKT_TXTIME_STATS at the same time. This will increase
memory consumption but as the CONFIG_NET_PKT_TXTIME_STATS is
only meant to be enabled for debugging purposes this overhead
can be accepted.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
In TCP, we increase the net_pkt refcount in order to resend
it later if we do not receive ACK in time. Because we are not
getting a new net_pkt, the TXTIME statistics would be calculated
incorrectly. So if we re-send, reset the net_pkt creation time.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add extra net_pkt RX timing information to "net stats" command.
You neeed to enable CONFIG_NET_PKT_RXTIME_STATS and
CONFIG_NET_PKT_RXTIME_STATS_DETAIL config options.
The "net stats" command will print something like this:
Avg RX net_pkt (7) time 459 us [0->23->16->257->160=456 us]
The extra stats is the end of line inside brackets.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Add extra net_pkt TX timing information to "net stats" command.
You neeed to enable CONFIG_NET_PKT_TXTIME_STATS and
CONFIG_NET_PKT_TXTIME_STATS_DETAIL config options.
The "net stats" command will print something like this:
Avg TX net_pkt (5867) time 82 us [0->33->20->27=80 us]
The extra stats information is the end of line inside brackets.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Collect information how long net_pkt has travelled in IP stack
in certain points. See network documentation what these points
are and how to get information about the timings. This initial
commit adds support to TX timing collection.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
CONFIG_NET_CONFIG_INIT_TIMEOUT should not only be available if
CONFIG_NET_CONFIG_AUTO_INIT=y. If *_AUTO_INIT=n then application
may wish to initialize net manually and still use configured
timeout.
Signed-off-by: Emil Obalski <emil.obalski@nordicsemi.no>
If the CONFIG_NET_CONFIG_AUTO_INIT is not set, then let application
to do the initialization manually. In that case the app should call
net_config_init_app() when it is ready to setup the network.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Include directories for ${ARCH} is not specified correctly.
Several places in Zephyr, the include directories are specified as:
${ZEPHYR_BASE}/arch/${ARCH}/include
the correct line is:
${ARCH_DIR}/${ARCH}/include
to correctly support out of tree archs.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The original return type of tcp_data_get() was unsigned and the
return value <0 was not checked properly.
Fixes#25723
Coverity-CID: 210559
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The LwM2M server might modify the lifetime value while the device is
registered, hence it's needed to obtain the value directly from the
Server object instance, before each Register Update.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
A proper way to match a Security object instance with a Server object
instance is via Short Server ID resource. Both coupled object instances
should carry the same value of this resource in order to me considered
matched.
This was not implemented in the LwM2M library and it was incorrectly
assumed that the Security object instance index corresponds to the
Server object instance index. While such apporach works is simple
scenario, it might yield incorrect results when bootstrap is used.
Fix this, by verifyng the Short Server ID resource in the Secuirty
instance used, and finding a matching Server instance. The server object
instance is stored for future use in the engine.
Additionally, remove an extra Server object instance that was created
when the bootstrap procedure was used. Since the boostrap Security
object instance does not have the corresponding Server object, it's
enough to have a single Server instance.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Remove the static buffer for TCP options. Make sure that the
options were read properly to temp buffer.
Fixes: #25729
Coverity-CID: 210056
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This set of functions seem to be there just because of historical
reasons, stemming from Kbuild. They are non-obvious and prone to errors,
so remove them in favor of the `_ifdef()` ones with an explicit
`CONFIG_` condition.
Script used:
git grep -l _if_kconfig | xargs sed -E -i
"s/_if_kconfig\(\s*(\w*)/_ifdef(CONFIG_\U\1\E \1/g"
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Currently, OpenThread builds produce warning from the OT logger module,
because we don't use one of the parameters provided by logger macro.
Explicitly mark the parameter as unused to prevent warnings being
thrown.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Drop unused ppp_parse_options_array() function and 'struct
ppp_option_pkt' data type. Both were used in initial PPP implementation,
but are not longer needed after recent PPP refactoring.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Introduce new ppp_config_info_req() function that can be used in order
to handle options received within Conf-Req packet. As an input it takes
array of supported options. If received Conf-Req packet contains unknown
options, then a Conf-Rej packet is automatically generated with all of
those options. If all of received options are supported, then function
continues to parse each provided option individually by calling option
specific callbacks.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Use a generic handler for received options in Conf-Rej packet. This will
mark all those options as rejected, to they will no longer be included
in subsequence Conf-Req packets.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Use a generic handler for received options in Conf-Rej packet. This will
mark all those options as rejected, to they will no longer be included
in subsequence Conf-Req packets.
This allows to communicate with PPP peers, which do not support DNS1 and
DNS2 options.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
So far there was no generalized way how to handle negotiation of "my
options" (those sent with Conf-Req packet to peer). Peer response for
proposed options and their values were not tracked in any way, so
rejected (by peer) options were sent over and over again in subsequent
Conf-Req packets. In case of IPCP it means that all sent options such as
IP_ADDRESS, DNS1, DNS2 were mandatory for being supported by peer in
order to successfully finish negotiation. For example if 'pppd' was the
configured peer (e.g. when using with QEMU and net-tools scripts),
without ms-dns being configured (which is the default), then "my
options" negotiation failed.
Introduce generalized mechanism for negotiating "my options", so it is
easier to maintain current set of supported options and easily add new
options to the implementation in future. FSM instance inititialization
function can provide information about supported "my options".
Information is is passed as a table of 'struct ppp_my_option_info'
containing:
* option code (type),
* callback for adding option to Conf-Req packet,
* callback for received Conf-Ack with that option,
* callback for received Conf-Nak with that option.
Store also runtime information about each negotiated option. First of
all remember which options have been rejected, so they are not sent to
peer once again in the next Conf-Req packet. This will solve issue with
IPCP negotiation when peer doesn't support DNS1 or DNS2 options. Keep
also track about which options have been acked, so such information can
be verified before bringing FSM instance up.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Rework implementation of ppp_parse_options() to be more generic and
flexible to use. Pass callback and user data for parsing each option
separately.
Keep old functionality of ppp_parse_options() accessible via
ppp_parse_options_array() function. This will make sure that old code
using this function doesn't need to be changed now. There are plans
however to remove ppp_parse_options_array() once there are no more users
of it in future.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Use net_pkt API to assemble Configure-Req messages. Compared to net_buf
API it allows us to simplify code.
Remove append_to_buf() helper functions, as their functionality is now
totally replaced by net_pkt API. Additionally net_pkt API handles data
that wraps several net_buf packets, which was not the case with
append_to_buf().
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Use net_pkt instead of net_buf API for replying to Configure-Req. We use
the fact that for now we reply with either Configure-Ack or
Configure-Rej only. In both cases we can allocate net_pkt ahead, because
we know its maximum length (which is equal to length of received
Configure-Req packet).
Make also an improvement in generic FSM code and reply with
Configure-Rej to all Configure-Req for which there is no config_info_req
callback set. Use that to drop LCP specific Conf-Req handling code,
because there is no option properly supported there yet.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
So far ppp_parse_options() has returned enum net_verdict. This type was
never used directly by network stack. Instead, caller was always
checking for NET_OK and returning error code in case of failure.
Change implementation of ppp_parse_options(), so it returns error code
in case of error and 0 when succeeded. This better fits its generic
purpose.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
There is no allocation in ppp_l2.c file, which mean that
BUF_ALLOC_TIMEOUT macros is not used anywhere. Drop it.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Make sure that only those threads that have been granted access
to net_if objects, can call the functions that modify net_if data.
The CONFIG_NET_IF_USERSPACE_ACCESS config option is also removed
as it is no longer needed after this change.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Make net_if a kernel object with type K_OBJ_NET_IF so that we
can restrict access to it.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
There was a race condition when ppp_fsm_open() was called in CLOSED
state. Conf-Req was sent first, then state was changed to
REQUEST_SENT. In the meantime however we have already received Conf-Req
to which we responded with Term-Ack.
Change state before sending Conf-Req, so we handle Conf-Req from peer
properly instead of dropping it.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
OpenThread API is not thread safe, therefore it shall be protected
from being preempted by OT thread, or other thread issuing API calls.
The problem showed up after a recent OpenThread upmerge, where changes
in the Joiner class made this problem visible. W/o extra protection,
`otJoinerStart` call can be preempted by the OT thread, leading to an
unexpected behavior.
Introduce new function to allow to lock any API operations for others.
Anyone willing to call OT APIs, shall lock the mutex first to get
exclusive access to the stack.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
`otLinkRawSetEnable` was removed and Link Raw mode is now enabled with
`otLinkRawSetReceiveDone` which requires to register a callback
function. Since it makes little sense for OT L2 to register a Link Raw
callback in current setup, leave it up to the application to register
the callback.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Bugfix: in tcp_conn_unref(), the conn was zeroed before removing it
from the connection list (tcp_conns).
Zeroing conn, results in zeroing its 'next' member,
which in effect removes all its following connections referred to
in tcp_conns linked list.
The solution is to move the memset() after sys_slist_find_and_remove().
Signed-off-by: David D <a8961713@gmail.com>
The network interface check was invalid when IPv6 Router
Solicitation message was timeout and interface was not found.
This is highly unlikely but needs to be checked properly.
Fixes#27145
Coverity-CID: 211511
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Currently, the application could receive a duplicate CONNACK event, in
case the server rejected the connection at MQTT level (with an error
code provided with CONNACK message). A subsequent connection close (with
`mqtt_abort` for instance) would produce the duplicate event.
Fix this by reporting back to the MQTT engine, that the connection was
refused, so it can close the connection rightaway. Rework the event
notification logic, so that DISCONNECT event instead of a duplicate
CONNACK event is notified in that case.
Also, prevent the MQTT engine from notyfing DISCONNECT event in case of
socket errors during initial connection phase (i. e. before
`mqtt_connect` function finished).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
is_init field is useless, because there is only single code path that
always sets it to true before using it.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Make sure IPv6 address is removed from network interface in IPV6CP
protocol down handler. This makes sure that application can receive
high-level notification about missing network connection.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Make sure IPv4 address is removed from network interface in IPCP
protocol down handler. This makes sure that application can receive
high-level notification about missing network connection.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Added the 'fall through' comment to switch to quiet compiler
and coverity warnings.
Fixes#25724Fixes#25726
Signed-off-by: David Leach <david.leach@nxp.com>
To allow hardware crypto acceleration custom mbedtls library needs to be
prepared. However current implementation forces the default library to
be used.
This patch allows not using the default library and passing custom
mbedtls target to use with openthread.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
The TCP2 stack does operations directly on the packet data which may
or may not be aligned. The unaligned access causes a fault on the
Cortex-M0+ so use the UNALIGNED_* macros instead.
Signed-off-by: Michael Hope <mlhx@google.com>
Putting data to local buffer before transmission was removed
to optimize operation. Local buffering was not needed, as passed
buffer cannot be modified until sending is finished.
Signed-off-by: Kamil Kasperczyk <kamil.kasperczyk@nordicsemi.no>
Currently, functions for poll sock_fds array management are buggy, in
case there is another socket open (for instance the socket for firmware
update download), it could get overwritten, if the LwM2M socket was
closed and re-opened in a meantime (e. g. on registration timeout).
Fix this, by appending new entries to the sock_fds in continuous manner.
In case of removal, the deleted entry is overwritten by the last one,
and the last one is cleared.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
During FW update, the application expects a consecutive data stream.
Therefore retransmitted blocks shall not be forwarded to the
application, but ignored. In case blocks are received out of order,
return an error and do not handle this block.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
When FW update in PUSH mode is used, the firmware is encapsulated in the
TLV as an opaque data, according to the LMWM2M satandard, and then
sliced into blocks and transferred block by block in several
transactions. Therefore, the TLV header is only present in the initial
message.
Current implementation did not handle this case well, reporting errors
on consecutive blocks, therefore making the FW update in PUSH mode
broken.
This commit fixes this issue with following changes:
* The TLV is only assumed to be present in the initial block, while
consecutive blocks will be processed directly into the appropriate
handler,
* 32-bit variables shall be used whenever dealing with the opaque data
length, since the firmware size can easily exceed the 16-bit range,
* Additional information, required for the FW block transfer to work
properly were added to the block context structure,
* The application shall only be notified of the actual data length, and
not the total block size (the total TLV size including header).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
When the dupplicate address detection mechanism is not used, we should
not wait for IPv6 address configuration. Because of the bug we are
waiting for it, when CONFIG_NET_CONFIG_MY_IVP6_ADDR is not set.
This commit fixes it.
Signed-off-by: Lukasz Maciejonczyk <lukasz.maciejonczyk@nordicsemi.no>
Remove the `sm_bootstrap_reg_done` function, which produced duplicated
logs. The bootstrap registration done event is already logged on state
transition. Additionally, in case bootstrap procedure took longer time,
the duplicated log message was printed on each `lwm2m_rd_client_service`
call (500 miliseconds by default).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The security object instance used should be cleared on engine reset,
otherwise we might end up using invalid object instance for the
registration (i. e. if the engine was restarted during the bootstrap
procedure due to socket errors, the registration attempt will use
bootstrap security object instead of finding a proper one).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
During the bootstrap procedure, when Boostrap Finish was received, the
response message was not initialized properly, resulting in a socket
error (NULL pointer porovided) and the response not being sent.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The Z_STRUCT_SECTION_ITERABLE() provides an alignment so remove
the 32 byte alignment for net_if and ppp_protocol_handler structs.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Use system provided Z_STRUCT_SECTION_FOREACH() and
Z_STRUCT_SECTION_ITERABLE() macros instead of manually coding
everything for network sections.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Different switch was used for build to allow including all shim body
in zephyr without the need for building openthread from this repo. This
allows developer to include custom OpenThread sources as part of the
application.
This change is needed as Thread is a subject of certification and
sources provided by zephyr may not necesairly pass certification or be
precertified. User is allowed to use certified OpenThread version
this way.
Signed-off-by: Marek Porwisz <marek.porwisz@nordicsemi.no>
Fix compilation error when the DIAG module is enabled.
All core functionalities are provided.
Signed-off-by: Piotr Szkotak <piotr.szkotak@nordicsemi.no>
When this happens, ip6_addr_cb and ipv6_addr_event_handler() both look
unused, since they are passed to macros which discard them.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
In order to implement a blocking connect, add a semaphore
and block on it in net_tcp_connect().
The semaphore is released when ESTABLISHED state is reached.
In case tcp_conn_unref() is called while waiting on the semaphore,
defer the unreference, tcp_conn_unref() will be called from
net_tcp_connect().
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
In order to improve readability, refactor and simplify
the control flow in net_tcp_connect().
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
In order for events to be correctly interpreted in SYN_SENT,
check for SYN and ACK simultaneosly.
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
net_tcp_put() can be called before ESTABLISHED state
is reached, send FIN only in ESTABLISHED state.
Signed-off-by: Oleg Zhurakivskyy <oleg.zhurakivskyy@intel.com>
Single work object for whole fsm was not being able to handle more than
single packet at a time. Because of that we have overwritten already
scheduled packets, resulting in fsm timeout and net_pkt leak.
Use net_pkt work object instead, so we can safely schedule more than a
single packet.
This commit also drops workaround for qemu_x86 unit testing.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Current implementation of net_ipv6_send_rs() uses
net_ipv6_addr_create_ll_allnodes_mcast() for creating RS packet.
As result we send RS packet to all-nodes destination [FF02::1],
but RFC4291 requires all-routers destination [FF02::2] for RS packets.
Signed-off-by: Eug Krashtan <eug.krashtan@gmail.com>
`addrlen` parameter is updated on each `recvfrom` call, indicating the
actual address length returned. In case both, IPv4 and IPv6 are used on
different sockets (i. e. on regular LWM2M socket and FOTA socket), the
returned address length will differ.
In case `from_addr_len` is not reinitialized on each iteration, the
value stored in the `from_addr_len` variable will eventually indicate
the smaller IPv4 address size, therefore resulting in a failure in a
consecutive call on an IPv6 socket.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Although LwM2M engine uses cooperative threads, the internal `send()`
implementation might trigger context switch when it calls a kernel
function, therefore resulting in `send()` call being entered from both
the LwM2M thread and the retransmit work.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Currently there is a single function that handles both Configure-Ack and
Configure-Rej messages. This is obviously wrong for Configure-Rej,
because implementation applies options received in the message.
Remove Configure-Rej callback, so those frames are simply ignored for
the time no valid handling code exists.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Clear negotiated options in protocol down handler. That way all
addresses are properly requested (by sending 0.0.0.0 in Configure-Req)
in the subsequent option negotiation phases.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Unit tests were failing to build because random header was included by
kernel_includes.h. The problem is that rand32.h includes a generated
file that is either not generated or not included when building unit
tests. Also, it is better to limit the scope of this file to where it is
used.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
The context should only clear messages it owns, not all of them. Since
both context (LwM2M and FOTA) share common message pool, they might
interrupt their operation otherwise (i. e. cancel retransmissions).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The FOTA socket was not closed when download finished or an error
occured.
Additionally, fix the socket fd verification (it was assumed 0 is not a
valid fd which is not correct).
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
