According to the specificaion, resources are not predefined to use 32 or
64 bit floating point numbers, but should rather accept any of them (as
indicated by the size of the TLV in the message). This lead to issues
for instance with Eclipse Leshan LWM2M server, where Leshan sent 64-bit
value, while Zephyr expected 32-bit, making it impossible to write
the resource.
Therefore, unify the float usage to 32-bit float representation and fix
the TLV parsing functions, to accept values sent as either 32 or 64 bit
float.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
The `poll()` function did not report POLLHUP if the peer ended the DTLS
session, making it impossible to detect such event on the application
side.
On the other hand, TLS erroneusely reported POLLHUP along with each
POLLIN event, as the 0 returned by the `recv()` socket call was
wrongly interpreted (it was expected to get 0 in return as 0 bytes were
requested).
Fix this by introducing a helper function to process the mbedtls context
and verify if new application data is pendingi or session has ended.
Use this new function in the poll handler, instead of a socket `recv()`
call, to remove any ambiguity in the usage, for both TLS and DTLS.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Notify the application when DTLS client session ends by returning
ENOTCONN on such event. Additionally, reset the mbed TLS session
structures, allowing to reinstante the session on the next send() call.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
ECONNABORTED was returned in case tls_mbedtls_reset() function for
resetting session failed, which can be caused by memory shortage. Return
ENOMEM instead.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
MCUboot fork was updated in order to bring bugfixes from its
upstream:
- introduce MCUBOOT_CPU_IDLE() for support low power single thread
(fixes single thread power consumption)
- Allow not working secondary image device and boot form primary
device if image available
- fixes memory alignment of the RAM buffer that is used to temporarily
store data during swap.
- update devicetree py package lib files include path in assembly
- serial recovery: cbor_encoder: fix str encoding macros
- zephyr: Kconfig: fix board references
- do not set defaults for LOG_IMMEDIATE Kconfig
- boot: Do not use `irq_lock()` if using arm cleanup
- Kconfig: fix deadlock on cryptolib selectors
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Fix Periodic Advertising Sync Establishment to accept
synchronization establishment to device listed in the
Periodic Advertisers List when filter policy was used.
Fixes#38520.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The IMXRT685 evaluation board uses a PMIC to reduce the signalling
voltage for a connected SD card to 1.8V, which Zephyr does not currently
support. This results in Zephyr failing to interface with any SD card
that supports 1.8V signalling when run on the RT685 eval board. This
commit disables support for 1.8V signalling for the RT685.
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
Snprintf calls should be avoid when not necessary, instead snprintk
should be used as it offers the same functionnalities at a lower cost
in flash.
Signed-off-by: Xavier Chapron <xavier.chapron@stimio.fr>
there is a memory hole from address 0x10000000-0x12150fff
in the ram on up_squared, we don't have access to read/write
this range, so limit the memory range below 0x10000000.
Signed-off-by: Chen Peng1 <peng1.chen@intel.com>
Fix incorrect calculation of received common extended header
length. Due to this a zero length advertising data is
reported in Extended Advertising and Periodic Advertising
report as one byte AD data.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The sample from samples/drivers/misc/ft800/ causes twister to
timeout as it has no pass conditions defiend nor producess any
output that can be verified. What is more, it requires a shield
which is not reflected in the sample.yaml. This fix sets test harness
to display and adds an entry for the required fixture.
Fixes: #38536
Signed-off-by: Maciej Perkowski <Maciej.Perkowski@nordicsemi.no>
Add #defines for the maximum and minimum allowed values for object
IDs.
Moves the #define for the directory listing object ID to the public
header file
Having these limits available is useful for applications wanting to
ensure they pass in (or handle) valid object ID values.
Signed-off-by: Asbjørn Sæbø <asbjorn.sabo@nordicsemi.no>
Object IDs are 48 bits, but are often carried in uint64_t variables.
This commit defines a mask value that retains the least significant 48
bits.
Signed-off-by: Asbjørn Sæbø <asbjorn.sabo@nordicsemi.no>
This commit adds FMAC HAL and LL drivers to the Kconfig system
Required to enable Filter Math Accelerator (FMAC)
on STM32G4/H7 devices
Signed-off-by: Tom Owen <tom.owen@zepler.net>
Zephyr SDK now features an openocd version compatible
with stm32g0 targets.
Set openocd as the default zephyr runner for this target.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Zephyr SDK now features an openocd version compatible
with stm32g4 targets.
Set openocd as the default zephyr runner for this target.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
For ECFC L2CAP test we user overlay with disabled EATT and
bt_eatt_disconnect_one() is available only with EATT enabled.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
For some unknown reasons, using the new modified HPET timer would
render shell unresponsive under QEMU. So for now, disable QEMU
icount when shell is enabled.
Fixes#37672
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The Bluetooth Core Specification, version 5.3, has introduced multiple
changes to several widely-used terms in order to make them inclusive.
Update the public API to reflect this, excluding hci.h, which will be
done in a subsequent commit.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
commit id ec2b9d42af missed that ivshmem
uses pcie_msi_enable as well, thus fixing it now.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Update the NXP HAL to bring in fix for the FlexCAN driver on the NXP
Kinetis MKE1xF.
Fixes: #38442
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
Implements the Bluetooth appropriate language mapping for the Bluetooth
mesh subsystem.
Changes the following terms:
- Master security credentials -> Flooding security credentials
- Whitelist filter -> Accept filter
- Blacklist filter -> Reject filter
- Removes CDB's NODE_BLACKLISTED, which was not in use.
Signed-off-by: Trond Einar Snekvik <Trond.Einar.Snekvik@nordicsemi.no>
Add a warning if the image is unconfirmed. Add a delay before rebooting
so that the user knows what happened.
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
The current HAWKBIT_PROBE & HAWKBIT_PROBE_DEPLOYMENT_BASE response
handlers aren't able to handle multipacket response from Hawkbit
server. This commit fixes it by using the implementation from
HAWKBIT_DOWNLOAD.
Signed-off-by: Yong Cong Sin <yongcong.sin@gmail.com>
Signed-off-by: Joep Buruma <burumaj50@gmail.com>
Twister was failing for samples/drivers/audio/dmic with timeout
because there were no pass conditions defined. This patch sets
harness to console and adds regex entries to be checked.
Fixes: #38532
Signed-off-by: Maciej Perkowski <Maciej.Perkowski@nordicsemi.no>
Call __reset instead of directly calling __initialize from the common
RISC-V privileged SoC vectors __start. This allows injection of SoC
specific reset code just after setting up the machine trap vector.
RISC-V privilege SoCs without the need for custom reset code can set
CONFIG_INCLUDE_RESET_VECTOR=y to include a __reset stub which simply
calls __initialize.
Fixes: #38396
Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
net_packet_socket_input() was changed to hardcode the return of
NET_CONTINUE and that caused a segmentation fault/crash in
net_core/process_data(), in cases when pkt was unreferred and
NET_OK was returned from net_conn_input()
This happened with socket combo of: AF_PACKET+SOCK_RAW+IPPROTO_RAW.
Signed-off-by: Jani Hirsimäki <jani.hirsimaki@nordicsemi.no>
CONFIG_PM_DEVICE was a de-facto requirement when enabling CONFIG_PM=y
since some device, i.e. UART, used the PM device hooks to block
suspension process while the device was busy finishing transmission.
This has now been fixed using constraints, so CONFIG_PM=y can be enabled
without further requirements.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Extracted new Kconfig file for the Extended Advertising
and for the Bluetooth Isochronous Channel from the
Bluetooth subsystem Kconfig.
Signed-off-by: Kamil Gawor <Kamil.Gawor@nordicsemi.no>
The BAS serivce Kconfig uses now a log template instead of
defining logger configuration itself.
Signed-off-by: Kamil Gawor <Kamil.Gawor@nordicsemi.no>
