Implement PM_DEVICE_RUNTIME for gpio_nrfx.c
PM is required to request/release the GPD domain if such exists
for a given GPIO instance. Without this, every single API call
will result in a requesting the GPD domain, which makes it slow
and unusable from ISRs, given GPD requires IPC. Now, user can
simply get the GPIO before using it, and put it once done,
calling the APIs from ISR while resumed.
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
The nRF GPIO hardware does not store the initial output value
set resulting from gpio_pin_configure() and thus, when
gpio_get_config() is used, the initial value is not returned.
This commit just reads the output value and sets the INIT
value to match in gpio_get_config().
Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
This patch allows to _safely_ configure GPIO ports that have their pad
on FAST_ACTIVE1 domain.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Updated API version enables multi-instance GPIOTE driver.
Additionally obsolete symbol that was used to specify
API version in the past was removed.
Affected drivers have been adjusted and appropriate changes
in affected files have been made.
Signed-off-by: Jakub Zymelka <jakub.zymelka@nordicsemi.no>
Change introduces a new configuration option that can be used to disable
GPIO interrupt support to reduce memory footprint.
Signed-off-by: Marek Pieta <Marek.Pieta@nordicsemi.no>
gpio_nrfx_port_get_direction() is meant to check which inputs
are enabled, and does so by checking one bit at a time
and setting that bit into the "input" parameter.
But "input" was never zeroed, so any garbage ones it
may have remains.
Zero it.
This fixes a problem where if the "input" parameter was not
zeroed by the caller, the result of the funcion call
is undefined.
Detected by valgrind on:
tests/drivers/gpio/gpio_get_direction
Conditional jump or move depends on uninitialised value(s)
by gpio_get_direction_test_disconnect (main.c:64)
Conditional jump or move depends on uninitialised value(s)
by gpio_get_direction_test_output (main.c:102)
When running this test for the nrf52_bsim.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
To ease building for workstation tests, instead of using
the ARM CMSIS instructions instrinsics directly,
use the NRFX macro that uses the compiler builtins when
necessary.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Instead of getting the hardcoded address from the DT structure
use its symbolic name (Also from DT) which will be resolved
by the nRF HAL definitions.
This allows the GPIO peripherals' addresses to be redefined
for the simulated targets.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Change automated searching for files using "IRQ_CONNECT()" API not
including <zephyr/irq.h>.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This PR fixes up a race condition which could cause GPIO writes to be
dropped when one thread was toggling a gpio on a port, and a different
thread was also modifying a (different) gpio on the same port.
Signed-off-by: Nickolas Lapp <nickolaslapp@gmail.com>
Inside gpio_nrfx_pin_interrupt_configure, the "ch" variable is allocated
inside the "if", but then it gets dereferenced and the pointer saved in
trigger_config, which is used in the outer context.
This works fine right now, but that memory location could possibly be
reused by another automatic variable if the code gets changed.
Moving it in the function context to avoid any potential problem.
Signed-off-by: Fabio Baltieri <fabio.baltieri@gmail.com>
Now that we have 8 bits reserved for vendor specific GPIO flags,
introduce a new set of flags for nRF platforms to configure pins drive
mode. These new flags are equivalent to the previous existing ones, but
use a naming scheme the fits better with vendor hardware capabilities.
The table below shows the equivalence between old and new flag
| Old flags | New flags |
|---------------------------|-----------------------|
| `NRF_GPIO_DS_DFLT_LOW` | `NRF_GPIO_DRIVE_S0` |
| `NRF_GPIO_DS_DFLT_HIGH` | `NRF_GPIO_DRIVE_S1` |
| `NRF_GPIO_DS_ALT_LOW` | `NRF_GPIO_DRIVE_H0` |
| `NRF_GPIO_DS_ALT_HIGH` | `NRF_GPIO_DRIVE_H1` |
| `NRF_GPIO_DS_DFLT` | `NRF_GPIO_DRIVE_S0S1` |
| `NRF_GPIO_DS_ALT` | `NRF_GPIO_DRIVE_H0H1` |
| `NRF_GPIO_DS_DFLT_LOW \|` | `NRF_GPIO_DRIVE_S0H1` |
| `NRF_GPIO_DS_ALT_HIGH` | |
| `NRF_GPIO_DS_ALT_LOW \|` | `NRF_GPIO_DRIVE_H0S1` |
| `NRF_GPIO_DS_DFLT_HIGH` | |
Documentation has been written to explain in more detail the meaning of
the flags and how they can be used.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The naming of the custom DS flags is not clear for nRF platforms. This
patch removes all the NRF_GPIO_DS* flags. New flags will be
re-introduced in a follow-up commit.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The driver returns -EIO when a pin to be disconnected was not earlier
configured as input or output, what is not in line with the GPIO API.
This commit changes the driver to return 0 in such case.
Also -EIO is incorrectly returned when an interrupt trigger cannot be
configured in the nrfx_gpiote driver. This commit corrects this value
to -EINVAL.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Incorrect GPIOTE channel was being freed because the pin number
being used is not the absolute pin but the pin within the port.
Signed-off-by: Wael Barakat <waelsbarakat@gmail.com>
In order to bring consistency in-tree, migrate all drivers to the new
prefix <zephyr/...>. Note that the conversion has been scripted, refer
to #45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Reserve the upper 8 bits of gpio_dt_flags_t for SoC specific flags and
move the non-standard, hardware-specific GPIO devicetree flags (IO
voltage level, drive strength, debounce filter) from the generic
dt-bindings/gpio/gpio.h header to SoC specific dt-bindings headers.
Some of the SoC specific dt-bindings flags take up more bits than
necessary in order to retain backwards compatibility with the deprecated
GPIO flags. The width of these fields can be reduced/optimized once the
deprecated flags are removed.
Remove hardcoded use of GPIO_INT_DEBOUNCE in GPIO client drivers. This
flag can now be set in the devicetree for boards/SoCs with debounce
filter support. The SoC specific debounce flags have had the _INT part
of their name removed since these flag must be passed to
gpio_pin_configure(), not gpio_pin_interrupt_configure().
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
GPIOTE channel was not freed when pin was reconfigured. This lead to
channel pool draining when pin was frequently reconfigured.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
So far driver was using SENSE mechanism for all or none edge
interrupts. This was not convenient since in some modules may
require IN event to be used and other did not. Converting it to
use a mask specified in the device tree. Pins indicated in the
mask will use sensing.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Make nrfx GPIO driver part of the PRE_KERNEL_1 initialization stage. As
a result, the GPIO driver can now be initialized before UART if
required, a device that is also initialized during PRE_KERNEL_1.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Refactors all of the on-chip GPIO drivers to use a shared driver class
initialization priority configuration, CONFIG_GPIO_INIT_PRIORITY, to
allow configuring GPIO drivers separately from other devices. This is
similar to other driver classes like I2C and SPI.
Most drivers previously used CONFIG_KERNEL_INIT_PRIORITY_DEFAULT or
CONFIG_KERNEL_INIT_PRIORITY_DEVICE, therefore the default for this new
option is the lower of the two, which means earlier initialization.
Driver-specific options for off-chip I2C- or SPI-based GPIO drivers are
left intact because they often need to be initialized at a different
priority than on-chip GPIO drivers.
Signed-off-by: Maureen Helm <maureen.helm@intel.com>
Introduce combined GPIO drive strength flags for GPIO controllers only
supporting either default or alternative drive strength regardless if
the pin is driven to a high or a low level.
Fixes: #30329
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
Remove Kconfig options for enabling device instances in favor of
taking that information only from device tree. Prior to that
change there was a mix of devicetree and Kconfig.
Bring back use of CONFIG_GPIO_NRF_INIT_PRIORITY.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Currently level interrupts are implemented using GPIO SENSE, but edge
interrupts using GPIOTE events. Using GPIOTE events results in increased
power consumption according to product specifications and erratas of
some nRF MCUs. In case of nRF52832 it is <20uA in System ON Idle and
~400-450uA when used in conjunction with SPI or TWI.
Add a user configurable option to select between GPIOTE events and GPIO
SENSE mechanism, for implementing edge interrupts. Selecting GPIO SENSE
option will allow to reduce power consumption in scenarios mentioned by
nRF MCUs erratas.
Additionally GPIO SENSE mechanism (as opposed to GPIOTE event) allows to
detect state changes of pins configured as output.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
The GPIO driver uses a proprietary GPIOTE channel allocator.
This commit makes it use the allocation mechanism provided by nrfx.
Signed-off-by: Jedrzej Ciupis <jedrzej.ciupis@nordicsemi.no>
Now that we generate a header that extern's all possible devicetree
based device struct we can remove DEVICE_DT_DECLARE and
DEVICE_DT_INST_DECLARE as they aren't needed anymore.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Use the devicetree node as the source of object name and other
information used when defining the device structure.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
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>
A call to atomic_set_bit_to() was used for clearing bits in an atomic_t
variable that stores information about allocated GPIOTE channels.
This caused an issue to be reported by Coverity, as the function treats
its first parameter as an array and it was provided with a pointer to
a singleton.
This commit replaces that call with atomic_and(), to prevent the issue
from being reported and for consistency with the way bits are set for
allocated GPIOTE channels (what is done with a call to atomic_or()).
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
This commit removes API functions and macros which were deprecated in
2.2 release. GPIO drivers are updated accordingly.
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
So far, register state was used to determine if GPIOTE channel is busy.
This leads to issues if channel is used in more customized way after
allocation. In particular, if it temporarly disabled since disabled
channel is treated as available and can be allocated to another user.
Added additional mask which tracks allocated channels. After allocation
user can reconfigure the channel.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
When notifying the handler that GPIO interrupt has occurred
pass only pins that are associated with this handler.
Fixes: #24634
Signed-off-by: Pawel Dunaj <pawel.dunaj@nordicsemi.no>
Use gpio_pin_t uniformly when passing pin indexes to the driver. Use
gpio_flags_t uniformly when passing flags to the driver. Change name
of pin configuration function in API function table to be consistent
with other API functions.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The only remaining port operations have dedicated API function table
entries. Remove the defines for access op (mode), and remove support
for access op from all implementations.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The last external reference to these was removed when the pin
write/read functions were deprecated. Remove the syscall support, API
function table entries, and implementation from all drivers.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
These have been replaced by the appropriate call to
gpio_pin_interrupt_configure(). While the disable function could be
implemented using the new functionality, the enable function cannot
because the interrupt mode is not available. Consequently we cannot
replace these with equivalent functionality using the legacy API.
Clean up the internal implementation by removing the inaccessible
port-based enable/disable feature, leaving the pin-based capability in
place.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>