Mark the I2C instances as `PM_DEVICE_ISR_SAFE`, as the transition
operations are short, it saves RAM resources, and the spin-locking fixes
the non-atomic behaviour of the PM usage counter.
Signed-off-by: Jordan Yates <jordan@embeint.com>
Change the callback context to the dev pointer instead of the data
pointer, as the dev pointer is needed by RTIO.
Signed-off-by: Jordan Yates <jordan@embeint.com>
- Driver always initializes the device in the suspended state
- If CONFIG_PM_DEVICE_RUNTIME=n, device PM callback will be called with
RESUME action, thus setting up pins to default state and enabling the
peripheral
NOTE: when CONFIG_PM_DEVICE=n, the pinctrl sleep state will not be
available (-ENOENT) and so never applied, thus avoiding a pin
suspended->active transition.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Use the RTIO work queue to fake the i2c submit calls for drivers which
haven't yet implemented the API. Applications can change the size of
the work queue pool depending on how much traffic they have on the buses.
Signed-off-by: Yuval Peress <peress@google.com>
Initialise the semaphores in the init function so that a copy of the
data structure does not need to exist in `.data`.
Signed-off-by: Jordan Yates <jordan@embeint.com>
Let the caller decide what should be done in case of the transfer failure.
It will reduce the number of error log prints when the i2c scan is
called for the bus where nothing is connected.
Signed-off-by: Bartosz Bilas <b.bilas@grinn-global.com>
Different nRF52 devices have different maximum TWI DMA transfer size,
and it's easy to hit the limit with i2c displays on nrf52832 (8 bit) and
nrf52810 (10 bit). Currently neither the driver or the hal validate the
limit, leading to random NACK errors when trying to transfer more data.
Add a check on the driver to fail gracefully when going over the limit.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
This commit aligns TWIM shim to utilize memory-region property.
The memory-region is not required property that enables user
to specify placement of dma buffers in memory region.
It is done by assigning to memory-region property,
phandle to node with zephyr,memory-region and mimo-sram compatible.
When memory-region property is not specified for given
instance, buffer is placed in default RAM region with other data.
Signed-off-by: Adam Wojasinski <adam.wojasinski@nordicsemi.no>
There are two different i2c node properites `zephyr,flash-buf-max-size`
and `zephyr,concat-buf-size`. In the end max value of that two is used
to define size of the message buffer.
It's redundant to store both values in device config structure.
Changed config structure to contain only bigger value.
Signed-off-by: Adam Wojasinski <adam.wojasinski@nordicsemi.no>
- use CONFIG_HAS_HW_NRF_* symbols consistently in nRF multi-instance
drivers when creating particular driver instances
- remove unnecessary hidden Kconfig options that indicated the type of
peripheral to be used by a given instance (e.g. SPI, SPIM, or SPIS)
and enabled proper nrfx driver instance; instead, use one option per
peripheral type and include the corresponding shim driver flavor into
compilation basing on that option (not the one that enables the nrfx
driver as it was incorrectly done so far in some cases)
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Since PINCTRL and pinctrl-0 is now required, there's no point in doing
extra validation at driver level. Modify the macro to just check that
sleep state is present when needed, since it was the only remaining
assertion that was not covered. Renamed the macro to make it more clear
what it does: NRF_DT_CHECK_NODE_HAS_PINCTRL_SLEEP
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Add a Kconfig option allowing users to configure the transfer timeout
value, as the default 500 ms may not be sufficient in specific cases.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@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 is a follow-up to commits 9974bb043f
and 00ecc66677.
Add one missing `#ifdef CONFIG_PINCTRL` and remove a no longer needed
one to restore the possibility to use this driver without PINCTRL.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
The twim_config structure is no longer modified at runtime, so it can be
placed in driver's config (const).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
TWIM peripheral needs to be disabled for bus recovery, otherwise SCL/SDA
lines will not be released. This patch makes sure to disable peripheral
if active, and, restore its state afterwards (including pin
configuration).
It is worth to note that a better solution would be to:
1. Define scl/sda pins as `-gpios` in DT
2. Use GPIO API in the driver to perform recovery (as some other drivers
do)
3. Potentially use a "gpio" pinctrl state for this case
Unfortunately HAL is doing everything under the hood, so we have little
options to improve this unless we don't use it for such case. GPIO based
recovery should likely be generalized as many drivers seem to replicate
such _algorithm_.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Enable device runtime PM on the I2C driver. Note that this mechanism
replaces the old nrfx_twim_enable/disable calls with
pm_device_runtime_get/put, which means that driver will not save power
unless CONFIG_PM_DEVICE_RUNTIME=y.
Some quick measurements on thingy_52 running the samples/sensor/hts221
show a decrease of ~2-3uA in average when enabling device runtime PM.
Note that the driver already had implicit PM before, so the change for
users will be ~none, except that they now must enable the PM subsystem
features. While this case is not the best example, the PM subsystem will
turn to be beneficial as a whole when all devices in the board implement
it.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The device definition macro defined per-device init function, resulting
in a cluttered and hard to extend init function. Create a per-instance
IRQ connect function instead, which is called at init time by a common
function.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The driver already implemented some sort of runtime PM by
unconditionally calling nrfx_twim_disable/enable. Perform the same
operations in the PM callback instead of doing a full init/deinit every
time.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The driver re-initialized itself on the twim_transfer call if it wasn't.
This condition was likely added because of some manual/custom PM
schemes. Drop it in preparation for runtime PM.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
nrfx HAL does not support updating frequency at runtime, which means we
need to do a full init/deinit to update it when it is a matter of a
simple register write. Fix this by using the lower level HAL.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
When PINCTRL is enabled, the SCL and SDA pin numbers are not available
in configuration structures used for nrfx drivers initialization.
In this case, these pin numbers need to be obtained from peripheral
registers.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
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>
Extend the macro with checks for DT properties related to pin
assignments that are defined but would be ignored, depending on
whether PINCTRL is enabled or not, what presumably indicates
a resulting configuration different from what the user expects.
Add also a possibility to indicate that the pinctrl-1 property
should not be checked because the caller does not support the
sleep state.
Rename the macro so that its name better reflects its function.
Update accordingly all drivers that use it.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Add support for the new pinctrl API to the I2C drivers that handle
the nRF TWI and TWIM peripherals. Update code of the drivers and
related devicetree bindings.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Change the way this shim driver configures the I2C frequency, so that
it is possible to use also 1 MHz on nRF5340 (the nrfx driver performs
extra initialization steps for this frequency, hence it needs to be
reinitialized when the shim is reconfigured).
Correct the shim to handle selection of 1 MHz (or FAST_PLUS) bitrate
both through dts and I2C API.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Replace all get_dev_data()/get_dev_config() accessor utilities with
dev->data and dev->config.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
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>
Adds the ability for I2C drivers to report synchronous transfer stats
using a I2C specific macro to define the device instance.
The macro creates a container for device_state which allows for per
instance device class common data structure to be used in the device
class api (ex: i2c.h). This is used to maintain per driver instance
stats for all i2c drivers. This is a reusable idea across other device
classes as desired.
Using Kconfig device class stats may be turned on/off individually
this way as well, in this case I2C_STATS.
Signed-off-by: Tom Burdick <thomas.burdick@intel.com>
The device PM subsystem _depends_ on device, not vice-versa. Devices
only hold a reference to struct pm_device now, and initialize this
reference with the value provided in Z_DEVICE_DEFINE. This requirement
can be solved with a forward struct declaration, meaning there is no
need to include device PM headers.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
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>
The PM callback is no longer referenced as "pm_control" but
"pm_action_cb", so reflect this new naming on the callbacks.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Both nRF I2C drivers (i2c_nrfx_twi and i2c_nrfx_twim) perform the bus
recovery procedure in reaction to timeout (500 ms) of any requested
message transfer. Add implementation of the I2C API recovery function
in both these drivers so that it is also possible to execute this
procedure directly.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
TWIM peripherals cannot perform write transactions from buffers
located in flash. The content of such buffers needs to be copied
to RAM before the actual transfer can be requested.
This commits adds a new property (zephyr,flash-buf-max-size) that
informs the driver how much space in RAM needs to be reserved for
such copying and adds proper handling of buffers located in flash.
This fixes an issue that caused that e.g. the DPS310 sensor driver
did not work on nRF SoCs that only have TWIM, not TWI peripherals.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Issue an error logging message when the i2c_nrfx_twim driver lacks
a concatenation buffer big enough to properly handle a call to
i2c_burst_write() function, to give the user a hint what is wrong.
Also use by default a 16-bytes long concatenation buffer for every
instance of the i2c_nrfx_twim driver. Such value should cover most
of the simple uses of the i2c_burst_write() function, like those
in the stmemsc sensor drivers, and when a longer buffer is needed,
the user will be provided with the above message pointing to the
property that should be adjusted.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Instead of passing target states, use actions for device PM control.
Actions represent better the meaning of the callback argument.
Furthermore, they are more future proof as they can be suitable for
other PM actions that have no direct mapping to a state. If we compare
with Linux, we could have a multi-stage suspend/resume. Such scenario
would not have a good mapping when using target states.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
According to the documentation the OFF state has to be used when the
devices is fully turned off, ie, power removed. Most drivers were using
a sort of fall-through for all non-active states, leading to behaviors
not following the specifications.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
- Return -ENOTSUP if the requested state is not supported
- Remove redundant "noop style" functions.
- Use switch everywhere to handle requested state (not necessary in all
drivers, but better take off with consistency in place after current
changes).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The verb tense for the suspended state was not consistent with other
states. The likely reason: state was being used as a command/action.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The difference between low power and suspend states is a thin blur line
that is is not clear and most drivers have used indistinctly. This patch
converges to the usage of the suspend state for low power, since
contrary to the low power state, it is used by both system and runtime
device PM. The low power state is still kept, but its future is unclear
and needs some discussion.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The device PM control function will only be called if the requested
state is different from the current one. A significant amount of drivers
were checking for state changes, now unnecessary. This patch removes all
this redundant logic.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Since the state is no longer modified by the device PM callback, just
use the state value.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
