This has several advantages:
1. we don't need any hard assumptions about symbol length. The
current hard-coded limit of 32 characters might well be not true.
2. replaces a lot of code for reading those names with a single
call to a 1-line function to calculate string location.
3. eliminates the need to allocate buffers for exported symbol names
by replacing them with a simple pointer assignment.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Instead of first allocating all sections and then copying then,
create a helper function to allocate and copy a single section and
call it for all appropriate sections. We need to call this function
from another location when copying string sections.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Some compilers (e.g. riscv32) does not handle well complex macros
for logging. Generated code is bigger than expected (e.g. riscv32
code is almost twice bigger than cortex-m code). Use of logging can
lead to unexpected code increase.
To handle that an analysis of the zephyr code base was performed and
it shown that 75-80% of logs are simple strings with 0 arguments
(~45%), one 32 bit argument (~26%) or two 32 bit arguments (~6%).
Given that a set of dedicated macro were created which are applied
to those 3 cases which on 32 bit platform create very simple log
messages without padding or alignment needed.
Such dedicated macros save up to 40% of code (riscv32) and also
executes 30% faster (arm cortex and riscv32).
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
Add Z_CBPRINTF_IS_WORD_NUM(x) which determines if argument is a
number that can fit in 32 bit word. It is used by logging macros
optimized for handling strings with numeric arguments.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
Extend frontend API with optional set of functions which can
be used when simplified log message handling is enabled. If this
mode is enabled then there are dedicated macros for processing the
most common messages (string + 0-2 word arguments). Using this API
can speed up the processing of messages that are the most common.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
In preparation for new feature which optimizes handling of
the most common log messages (0-2 32 bit word arguments) add
functions dedicated for that purpose.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
This fixes missing unreference of network buffer allocated
when there is no data to send in the ring buffer.
The code have been refactored, so that we check whether there is
anything to send first, and if there is, the TX buffer is allocated.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
The default is that the high time for open-drain clk is one
PPBAUD, which is typically very short. Some device require
a longer high time during the open-drain address phase so
add a property to allow device tree to override the default.
Signed-off-by: Mike J. Chen <mjchen@google.com>
Remove the MCTRLDONE wait in mcux_i3c_request_auto_ibi().
I've seen this code getting stuck where the MCTRLDONE
bit is never set in the MSTATUS register by the controller
and this function spins forever. Documentaiton of the
MCTRLDONE bit only mentions it being set for EmitStartAddr
and ProcessDAA, but not for AutoIBI requests.
All the calls to this function do completion checks
afterwards, and with a timeout, so I believe the MCTRLDONE
check is not needed (and may not even be correct).
Signed-off-by: Mike J. Chen <mjchen@google.com>
At high i3c rates, the mcux_i3c_do_one_xfer_read()
could get into an infinite loop where the rx_count
kept returning 0 but the complete status bit
was never set. I believe the problem was that
the function was not emptying the FIFO fast enough,
so tighten the loop that processes the FIFO.
Signed-off-by: Mike J. Chen <mjchen@google.com>
mcux_i3c_configure() was saving values to a ctrl_config_hal
struct, but config_get() was not returning the values in
that struct. Remove that struct from the static data of
the driver and instead just have it on the stack. We init
that struct as needed just before calling the SDK API
I3C_MasterInit(). There's no reason to keep it around.
Change mcux_i3c_configure() to save a copy of the configuration
in the static data common.ctrl_config, which is what is
returned by config_get().
Signed-off-by: Mike J. Chen <mjchen@google.com>
Enable using a BlackMagicProbe for flashing and debugging XIAO BLE
boards.
Tested successfully with BMP 1.9.2 running on ST-Link/v2 against a
`xiao_ble` board.
Signed-off-by: Augusto Zanellato <augusto.zanellato@gmail.com>
Fix the problem not apply pinctrl eventhough the config is defined.
In practice, the setting is equals to soc default.
So, there is no apparent change in behavior.
Signed-off-by: TOKITA Hiroshi <tokita.hiroshi@gmail.com>
- Add extension to get git metadata (date, SHA-1, ...) regarding the
latest update made to a page
- Add date of last "actual" update to each manually authored doc page
- Add admonition inviting to report issues
- Add button in breadcrumb to report issue
Fixes#60622.
Signed-off-by: Benjamin Cabé <benjamin@zephyrproject.org>
This adds a vcs_link_get_open_issue_url Jinja filter to craft the URL
allowing to report an issue for a given Sphinx page
Signed-off-by: Benjamin Cabé <benjamin@zephyrproject.org>
Update the vcs_link extension to also support generating "edit" URL
vcs_link_version is also now a Sphinx config making its value available
to extensions.
Signed-off-by: Benjamin Cabé <benjamin@zephyrproject.org>
This adds a if CONFIG_SOC_FAMILY_NXP_ADSP guard in
Kconfig.defconfig for nxp_adsp. Or else all of its default
get applied everywhere. For example, qemu_xtensa fails
kernel.logging.message_capture tests because
CONFIG_TEST_LOGGING_DEFAULTS is disabled in
nxp_adsp/Kconfig.defconfig which should not have applied
to qemu_xtensa at all. So put a guard in there.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Adds a new entry for multi-function device (mfd) drivers to ensure that
incoming PRs to this area get labeled appropriately and have reviews
requested.
Signed-off-by: Maureen Helm <maureen.helm@analog.com>
The optlen parameter passed to getsockopt() should be of (socklen_t *)
type. Using int produced build errors on 64 bit platforms.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Recently added tests require more stack on certain platforms
(nrf52840dk_nrf52840 in this case), hence increase the ZTEST stack size.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Convert this sample application build to sysbuild,
and add support for the nrf5340bsim target.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Enable building this sample on the simulated nrf5340 net
core.
Note that this sample does nothing on its own, as it
requires a companion application core sample to driver it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
PM should not be enabled by default at the soc level. This is an
application decision and not a platform decision whether to enable or
disable PM features.
Enabling PM by default will result in descripencies and test scope and
failures in tests that do not account for PM being enabled.
Fixes#60359
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Convert this sample application build to sysbuild,
and update documentation.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
