This commit fixes a bug where an already received Rx frame could not be
processed by the IEEE 802.15.4 driver.
In the current implementation, buffer is marked as free and released to
the buffer pool after `nrf_802154_buffer_free_raw` finishes executing.
However, delays caused by thread scheduling might result in a new frame
being already received and provided to the driver before
`nrf_802154_buffer_free_raw` returns. Such a situation ends in an
assertion now.
This commit changes that behavior by marking the buffer as free before
calling `nrf_802154_buffer_free_raw`.
Signed-off-by: Jedrzej Ciupis <jedrzej.ciupis@nordicsemi.no>
This commit increases verbosity of serialization error handler for
nRF5340 application core. The handler prints the error code now.
Signed-off-by: Jedrzej Ciupis <jedrzej.ciupis@nordicsemi.no>
Add binding for zephyr sdmmc disk device, which uses the SD
subsystem to manage an SD memory card.
Fixes#46410Fixes#46266
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
Needed for running tests on intel_adsp_cavs boards, the cavstool.py
are using it to get the host ip address automatically.
Signed-off-by: Enjia Mai <enjia.mai@intel.com>
In the current implementation of segmented messages transmission, the
transport layer will put as many segments as it can into the advertiser
buffer. If a long segmented message is transmitted, there is a chance
that an ack will be received before the transport layer finishes sending
the first batch of segments. This will trigger retransmissions of those
segments that are marked as missing in the ack, that are also the
segments that are scheduled (or about to be scheduled), but not sent
out by the advertiser yet. Thus, the receiver may receive already
received segments.
This commit changes 2 things:
- it prevents rescheduling transmission of missing segments upon
reception of ack. Thus, ack doesn't cause segments duplication in the
advertiser buffer;
- it doesn't put all segments to the advertiser buffer. Thus, ack that
acknowledges all segments can immediately stop rescheduling segments as
they are already received;
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
The control flow around the irq-lockout used for fetching from
`tx_complete` has been simplified so that there is exactly one
line with irq_unlock that is unconditionally called.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
The cast from `sys_snode_t` to `struct bt_conn_tx` assumes `node` is the
first field in `struct bt_conn_tx`. It's better to future-proof this by
using `CONTAINER_OF` to locate `node`.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
As we work to phase out devicetree 'label properties, convert
driver to just use dev->name instead of DT_INST_LABEL.
Signed-off-by: Kumar Gala <galak@kernel.org>
Add handling for DISK_IOCTL_CTRL_SYNC to SD subsystem. Note that
SD caching is not enabled by the SD stack, so the only required
operation to sync the disk is to wait for any active data programming
to complete.
Fixes#46689
Signed-off-by: Daniel DeGrasse <daniel.degrasse@nxp.com>
Instead of using or'ed list of Kconfig options listing the compatible
series with what can be called "stm32h7 variant", use the matching
compatible information.
This will prevent to update the driver next time that a compatible series
is ported into zephyr.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
When required, add "st,stm32h7-spi" compatible on stm32 compatible
series (today: stm32h7, stm32mp1 and stm32mp1).
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Add a stm32h7 spi compatible.
This compatible intends to match all SPI hardware blocks that
can be compatible with the one available in stm32h7 devices,
for instance, but not limited to stm32u5 and stm32mp1.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Even though possible to use external pull-up and open drain buffer,
prefer internal pull-up to reduce power consumption.
Signed-off-by: Jose Alberto Meza <jose.a.meza.arellano@intel.com>
Fixed handling of k_sem_take() return codes.
Also some minor cleanup in broadcast_audio samples.
Signed-off-by: Erik Robstad <erik.robstad@nordicsemi.no>
If the POSIX API is selected via the POSIX_API option, don't also select
the conflicting NET_SOCKETS_POSIX_NAMES and use the POSIX headers instead.
Signed-off-by: Berend Ozceri <berend@recogni.com>
Use consistent name for SPI HW block property so applications
using a device tree overlay work transparently.
Signed-off-by: Jose Alberto Meza <jose.a.meza.arellano@intel.com>
Use consistent name for SPI HW block properties across Microchip
HW, so applications using a device tree overlay work
transparently.
Signed-off-by: Jose Alberto Meza <jose.a.meza.arellano@intel.com>
This update updates the TraceRecorder module from version v4.6.0
(RC1) to v4.6.3. This update fixes some minor bugs where
deleted threads werent removed from the thread table. This
update is also in preparation for updates which changes
some trace hook arguments in Zephyr.
The update to v4.6.3 also improves syscall tracing by
including the id and name as parameters when logging
these events.
Signed-off-by: Torbjörn Leksell <torbjorn.leksell@percepio.com>
If struct tx_structure is too large than the buffer that
expected_dut_to_host references, leading the compiler to generate a warning
about the possibility of buffer overflow.
Leave out the 'struct ec_response_too_big' so that 'struct tx_structure'
fits within the 256 byte buffer.
This was caught when using GCC 11.3.0 building the test for native_posix.
Signed-off-by: Keith Packard <keithp@keithp.com>
bt_dev.sent_cmd is unreferenced and assigned NULL within hci_cmd_done().
Calling net_buf_unref() again causes a null pointer dereference
Signed-off-by: Grant Ramsay <grant.ramsay@hotmail.com>
The codec (or rather codec configuration) is now taken from the
previously received BASE. This also means that the BASE (with
the codec configurations) is now also stored statically.
When the application attempts to synk to the broadcaster
the stack will lookup the codec configuration based on the
bis index, as a BASE may have multiple subgroups
with multiple codec configurations.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
If the preprocessor constant ESP_MAX_DNS is defined as 0,
Coverity reports a condition that can never be true, and some
dead code as well. This also violates MISRA rules on dead code.
This is caused by this segment in drivers/wifi/esp_at/esp.h:
#if defined(CONFIG_WIFI_ESP_AT_DNS_USE)
#define ESP_MAX_DNS MIN(3, CONFIG_DNS_RESOLVER_MAX_SERVERS)
#else
#define ESP_MAX_DNS 0
#endif
Fix this by never setting ESP_MAX_DNS to 0, as is the currently
case. Define ESP_MAX_DNS only if it is configured through
CONFIG_WIFI_ESP_AT_DNS_USE and CONFIG_DNS_RESOLVER_MAX_SERVERS:
#if defined(CONFIG_WIFI_ESP_AT_DNS_USE)
#define ESP_MAX_DNS MIN(3, CONFIG_DNS_RESOLVER_MAX_SERVERS)
#endif
Since CONFIG_DNS_RESOLVER_MAX_SERVERS is 1 or greater, ESP_MAX_DNS
will always be greater than 0.
If, on the other hand, ESP_MAX_DNS is not defined, relevant
functions will be reduced to empty stubs, since in that case
they do not make sense.
There could be a cleaner solution to this, but this one is the
least intrusive (involves less code changes).
Coverity-CID: 219490
Coverity-CID: 219513
Coverity-CID: 219520
Coverity-CID: 219524
Signed-off-by: Aleksandar Markovic <aleksandar.markovic.sa@gmail.com>
uart_async test and spi_loopback test require UART and SPI
with an ASYNC option.
The SAM0 UART/SPI driver requires configuring DMA support.
Arduino MKRZero is not DMA configured like other SAM0 boards.
We exclude it from the tests.
ea36f9b and f4c5bdf already exclude other SAM0 boards from these tests.
Arduino MKRZero also followed these PR policies.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@fujitsu.com>
The Arduino MKR Zero board is SAMD21 development board.
It has microSD card slot and secure element.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@fujitsu.com>
Add MKR header connector that is implemented by Arduino MKR series.
This allows hardware with compatible headers to define the related GPIOs.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@fujitsu.com>
The shell module will now use BT_AUDIO_CONTEXT_TYPE_ANY
instead of BT_AUDIO_CONTEXT_TYPE_UNSPECIFIED as that provides
a more flexible solution.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
If CONFIG_BT_PAC_SNK or CONFIG_BT_PAC_SRC was not enabled, the
function would not build due to missing Kconfig
(CONFIG_BT_PACS_SNK_CONTEXT or CONFIG_BT_PACS_SRC_CONTEXT).
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The HAS depends on the unicast server, and should be
disabled in the `no_unicast_server` as well as depend
on the CONFIG_BT_AUDIO_UNICAST_SERVER config.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Some functions are only used for the unicast server
but was guarded by CONFIG_BT_AUDIO_UNICAST instead of
CONFIG_BT_AUDIO_UNICAST_SERVER.
Some parts of the unicast-server-only code also had
client specific code that would never be called.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Add support for using PACS and/or capabilities for
the broadcast sink role. PACS and capabilities were
originally only supported for the unicast server
role, so the PACS callbacks were moved their own
struct, as if a device supports both the unicast
server role and the broadcast sink role, it will
only have a single PACS instance.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
