If we receive any data in FIN_WAIT_1, then ack it even if we
are discarding it.
Fixes#33986
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Jim Paris <jim@jim.sh>
The k_fifo_ prefix is meant for kernel API functions, and
not to our socket helper. So remove the k_ prefix in order
to avoid confusion.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we are waiting all the data i.e., the MSG_WAITALL flag is set,
then if we have not yet received all the data at the end of the
receive loop. We must use the condition variable to get the signal
when the data is ready to be received. Otherwise the receive loop
will not release the socket lock and receive_cb will not be able
to indicate that data is received.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Fix a regression when application is waiting data but does
not notice that because socket layer is not woken up.
This could happen because application was waiting condition
variable but the signal to wake the condvar came before the
wait started. Normally if there is constant flow of incoming
data to the socket, the signal would be given later. But if
the peer is waiting that Zephyr replies, there might be a
timeout at peer.
The solution is to add locking in socket receive callback so
that we only signal the condition variable after we have made
sure that the condition variable is actually waiting the data.
Fixes#34964
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Fix deadlock when db_hash_commit has to wait for the delayed work to
finish. This creates a deadlock if the delayed work for database hash
calculation needs to store the hash since the settings API is locked
when calling the commit callback.
Remove call to k_work_cancel_delayable_sync from db_hash_commit in order
to avoid the deadlock. Instead move comparing of the stored hash to the
delayed work and reschedule the work with no wait.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
If there are no sockets in the system, then do not drop the
packet immediately as there can be other L2 network handlers
like gPTP in the system. This will also allow ICMP messages
to pass to local handler.
Fixes#34865
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Coverity detected that a zero divisor can be passed to
ll_create_connection() without parameter sanitization. Conditionally
check the connection creation parameters according to spec.
Fixes#35343.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
On sending write request we should check write response.
If data prepaired to be written by peer or offset are not equal
to the data and offset we sent, we shall send Execute Write
Request with Request Flag set to 0x00 (Cancel All Prepared Writes).
This was affecting GATT/CL/GAW/BI-32-C and GATT/CL/GAW/BI-37-C
Signed-off-by: Krzysztof Kopyściński <krzysztof.kopyscinski@codecoup.pl>
Fix the missing HCI event data len limit check when encoding
incomplete advertising data report.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Max data length has to be adjusted for subevent code, otherwise we'll
hit an assert when trying to add data to event due to insufficient
free space.
Signed-off-by: Andrzej Kaczmarek <andrzej.kaczmarek@codecoup.pl>
Enable the option to pause the fallback hardware watchdog if the MCU is
halted by a debugger.
This fixes issue #33509 where some boards with Nordic MCUs could not be
flashed anymore after using the task watchdog sample.
Signed-off-by: Martin Jäger <martin@libre.solar>
The resource description on the OMA LwM2M registry states that only the
first instance of a particular error should trigger creation of a new
error code instance.
Signed-off-by: Benjamin Lindqvist <benjamin.lindqvist@endian.se>
The hardware watchdog was always fed with channel ID 0. This is correct
in most cases, but we should still use the actual ID returned from
wdt_install_timeout.
Signed-off-by: Martin Jäger <martin@libre.solar>
As EVENT_OVERHEAD_START_US offset is used in ticks unit in
LLL, ULL scheduling using ticker should also use ticks unit
for EVENT_OVERHEAD_START_US when reducing the first
connection event preparation.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Revert the strict preempt ticker start failure check.
Preempt ticker start can fail when enqueuing prepares into
already filled pipeline which has preempt ticker already
started for the first prepare that was added in the
pipeline.
Regression introduced in commit 5b75bdf589 ("Bluetooth:
controller: nRF5: Check preempt event on timeout").
Fixes#35476.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
From le_ext_adv_param_set we will occasionally attempt
to call bt_addr_copy where the `addr` and &adv->random_addr.a
are the same pointer. Doing a memcpy where source and destination
is the same pointer is undefined behavior and should not be
done.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The VCS client shell was never compiled before, and thus
the implementation had a few undetected errors.
This commit adds the VCS client to the shell CMakelists
as well as fixing the issues.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The common API functions in VCS had dead code in specific
configurations, causing coverity issues. Fixed by this
commit.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Fix side effect in assertion when checking a volatile
variable inside assert check.
Fixes#32904, #32923.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
