The driver enables the clock of a gpio-port if any of the
pins use the port. This is done by calling pm_device_runtime_get
when a pin is used and pm_device_runtime_put when the pin is not
used anymore.
These calls needs to be balanced. But if a single pin was configured
as GPIO_DISCONNECTED multiple times, every time pm_device_runtime_put
was called.
This caused the clock of the port to be stopped and therefore also
other pins on the same port stopped working.
This commit fixes this by keeping track of which pin on a port
has requested the clock and only call pm_device_runtime_get or
pm_device_runtime_put when the clock-request for the specific pin
changes.
Fixes#77698
Signed-off-by: Jeroen Broersen <jbroersen@interact.nl>
The local function ll_gpio_get_pin_pull() is not always referenced. Add
the __maybe_unused attribute to avoid compiler warnings (which may be
promoted to errors by downstream users).
Signed-off-by: Keith Short <keithshort@google.com>
This commit rebrands the STM32 EXTI API to a more hardware-agnostic
"GPIO interrupt controller" API, in anticipation of the introduction of
new series lacking the EXTI peripheral. The GPIO and EXTI drivers are
updated to match the rebranded API.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
This commit makes the contents of the stm32_exti_line_t data type opaque to
the EXTI GPIO interrupt controller API users. The GPIO driver is updated
to comply with this API change.
N.B.: while some assertions are removed as part of this commit, they were
broken since forever anyways, so nothing of value is lost.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
Move the functions that interact with EXTI configuration registers to
select or get the GPIO port that triggers events on a given EXTI line
to the EXTI driver.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
Modify all functions in the GPIO driver that interact with the EXTI
so that names, signatures and comments match what they do better.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
Use the `gpio_pin_t` type for all variables that hold a pin number (0..15).
Change all `int` to `uint32_t` instead, as signedness is unwanted.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
This commit changes the EXTI driver API to use unsigned types
for all parameters previously typed as `int`, as the signedness
is unneeded and unwanted.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
Removed the unnecessary initialization of the `ret` variable in
`gpio_stm32_clock_request` where its value is guaranteed to be
overwritten by subsequent operations, then simply returned `ret`.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
Introduced `GPIO_DEVICE_INIT_STM32_IF_OKAY` which utilizes `COND_CODE_1`
to reduce the chain of #if DT_NODE_HAS_STATUS(...) #endif
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
Introduce the stm32h7RS serie to the gpio driver,
based on the stm32h7
The SBS controller is used to configure the EXTI line among
the different GPIO port.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Introduce a custom STM32_GPIO_WKUP GPIO flag.
Use the newly introduced stm32_pwr_wkup_pin_cfg_gpio() public
function to configure GPIO pins, that have the STM32_GPIO_WKUP
flag in DT, as sources for STM32 PWR wake-up pins, on the condition
that there is a wake-up pin that corresponds to each of them.
These GPIO pins can then be used to power on the system after Poweroff
like a reset pin.
Signed-off-by: Abderrahmane Jarmouni <abderrahmane.jarmouni-ext@st.com>
Fix a bug where after a standby, it was impossible to reenable a GPIO
clock.
A counter is incremented each time pm_device_runtime_get is called, and
decremented each time pm_device_runtime_put is called. The
clock is only enabled if this counter equals 1.
When configuring a GPIO (as input or output), the timer is incremented, and
when disconnecting it, it is both incremented and decremented. Thus the
next time we try to configuring it, the clock is not enabled (since the
counter will now be equal to 2).
This causes a problem when using low power standby mode: after wakeup all
clocks are disabled and the GPIO clock can not be reenabled.
This commit fixes this bug by not incrementing the counter when disconnect
is asked (or in other words incrementing it only when configuring either
an input or an output).
Signed-off-by: Guillaume Gautier <guillaume.gautier-ext@st.com>
Set suspended as initial power state, only when the
CONFIG_PM_DEVICE_RUNTIME config is enabled.
The initial state was incorrect, when CONFIG_PM_DEVICE=y and
CONFIG_PM_DEVICE_RUNTIME=n. In that case, the power state was SUSPENDED,
but the device was actually enabled.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
remove all #ifdef CONFIG_SOC_SERIES_STM32xx
before to add a st,stm32g0-exti compatible
added to the matching targets:
C0/G0/U5/L5/MP1:
Signed-off-by: Marc Desvaux <marc.desvaux-ext@st.com>
Fix the enabling of the independant IO supply.
Function LL_PWR_EnableVddIO2 is called LL_PWR_EnableVDDIO2 on U5.
Signed-off-by: Guillaume Gautier <guillaume.gautier-ext@st.com>
In STM32U5 as well it is required to enable VDD before use.
Difference is that U5 enables this under PWR_SVMCR_IO2SV flag.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Many device pointers are initialized at compile and never changed. This
means that the device pointer can be constified (immutable).
Automated using:
```
perl -i -pe 's/const struct device \*(?!const)(.*)= DEVICE/const struct
device *const $1= DEVICE/g' **/*.c
```
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The stm32F1x serie does not provide the pin_get_config function,
Even if the CONFIG_GPIO_GET_CONFIG is set, the -ENOSYS error code
is returned by the z_impl_gpio_pin_get_config().
Signed-off-by: Francois Ramu <francois.ramu@st.com>
Define a pin structure to carry the pin configuration elements
such as push/pull, opendrain, input/output, high/low.
This structure will give the gpio_flags_t to the gpio_stm32_get_config.
Only the stm32F1x serie does not provide the pin config get function.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
In the stm32 gpio driver, the gpio_stm32_get_config function
also returns the pin state HIGH or LOW) to match gpio_pin_get_config
This is valid with CONFIG_GPIO_GET_CONFIG.
Signed-off-by: Francois Ramu <francois.ramu@st.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>
Based on introduction of plain GPIO configurations in STM32 pinctrl
bindings, update STM32 pinctrl/gpio drivers to make this functionality
available.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Before updating stm32 pinctrl/gpio drivers to support plain GPIO
feature, rework pin configuration functions headers to provide
more clarity on the arguments and the information they convey:
- pin configuration
- pin function
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Move the `pm_device_runtime_init_*` functions from <pm/device_runtime.h>
to <pm/device.h>. The initial device state should be settable
independently of whether `CONFIG_PM_DEVICE_RUNTIME` is enabled.
This also resolves a compilation error when attempting to use these
functions without also including <pm/device.h>.
Function documentation is also updated to be more general than only
referencing runtime PM, as this also applies to system PM and manually
run actions.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
In order to align with macros used to obtain a device reference (e.g.
DEVICE_DT_GET), align the PM macros to use "GET" instead of "REF". This
change should have low impact since no official release has gone out yet
with the "REF" macros.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
SYS_INIT can be used for the same purpose, what is being run is simply
an initialization hook.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Call `gpio_stm32_clock_request` unconditionally with "true" if no
runtime PM is enabled, "false" otherwise. In case the GPIO was enabled
(e.g. by a bootloader), the device will be left in proper state when
runtime PM is enabled.
Also added pm_device_runtime_init_suspended to indicate that device is
in a suspended state (will prevent pm_device_runtime_enable to suspend
the device again).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The function is only used by the driver itself. This likely comes from
pre-runtime PM API usage.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Make gpio32_stm32_configure use runtime PM API (so that it can be used
externally without further effort). The raw version of the function (no
PM put/get) has been introduced for internal use.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The suspend operation for this device is a clock gating operation (i.e.
fast), so use synchronous put to skip unnecessary overhead.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The PM runtime API (get/put) provides an inline implementation when not
built-in. Such implementation always returns 1, so it is safe to not
guard PM runtime calls and check for < 0.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Change the function pm_device_runtime_enable() to return 0 on
success or an error code in case of error.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Port some drivers to the recently introduced macros to showcase its
usage and be able to do some initial testing (nRF52840).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
