Move it to a common file named `zephyr,memory-common.yaml`, which
replaces `zephyr,memory-attr.yaml` and takes its contents as well.
This is so that another binding, e.g., `vnd,memory-region`, can support
being combined with the `zephyr,memory-region` binding like so:
node@deadbeef {
compatible = "vnd,memory-region", "zephyr,memory-region";
zephyr,memory-region = "NAME";
...
};
To allow this, edtlib would require `vnd,memory-region` to include the
property definitions from this new common file.
The same can't be done by including `zephyr,memory-region.yaml` directly
because that file marks the property in question as always required, and
it shouldn't be required whenever the `vnd,memory-region` compatible is
used on its own.
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
Add a new binding that should help with samples or MBOX tests by getting
information from DT.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Added new test for `DT_IRQ_LEVEL` and `DT_INST_IRQ_LEVEL`.
Introduced a new `vnd.cpu-intc` compatible so that we have a
root level interrupt controller that acts as level 1
aggregator, and modified `test_intc` to be level 2 aggregator,
updated test of `DT_IRQN(TEST_I2C_BUS)` accordingly.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
Extend api test suite to cover the new devicetree macros. This
includes extending the devicetree overlay with two new bindings:
- GPIO device which is also an interrupt controller
- interrupt holder using interrupts-extended to point to both
existing interrupt controller test_intc, and the newly added
GPIO device
Signed-off-by: Bjarki Arge Andreasen <bjarki@arge-andreasen.me>
These were first deprecated in v3.2. We have kept support for them
for as long as we can, but now their presence is causing CI failures
in some configurations. Specifically, using the deprecated 'label'
property is causing a warning which is promoted to an error in some
twister runs. This is blocking other forward progress in the
devicetree API.
I tried to rework the tests to avoid this, but it was too much effort
for the time I had to work on the task. Removing the APIs is therefore
unfortunately the best way forward to unblocking other work.
Re-work the test suite a bit to maintain coverage where we are using
the label property to test other macros.
Add a migration guide section to help with the transition.
Signed-off-by: Martí Bolívar <mbolivar@amperecomputing.com>
These worked because edtlib allows 'pinctrl-.*' properties without
specifying them on the bindings. However, this has never been an
anounced pinctrl feature, the reference documents explicitly mention
that usage of pinctrl-device.yaml is mandatory.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
The 'zephyr,memory-region-mpu' property was addede gqas a
convenient way to create and configure MPU regions using information
coming from DT. It has been used a lot since it was introduced so I
guess we can consider it a Zephyr success story ™ .
Unfortunately it has been proved to be a bit limited and with some
important limitations:
1. It was introduced as a property of the compatible
zephyr,memory-region that is used to create linker regions and
sections from DT data. This means that we can actually create MPU
regions only for DT-defined regions and sections.
2. The naming is unfortunate because it is implying that it is used only
for MPU.
3. It is misplaced being in include/zephyr/linker/devicetree_regions.h
and still it has nothing to do with the linker at all.
4. It is exporting a function called LINKER_DT_REGION_MPU that again has
nothing to do with the linker.
Point (1) is also particularly limiting because it is preventing us to
characterize memory regions that are not generated using the
'zephyr,memory-region' compatible, like generic mmio-sram regions.
While we fix all the issues, we also want to extend a bit the range of
usefulness of this property. We are renaming it 'zephyr,memory-attr' and
it is now carrying information about the type of memory the property is
attached to (cacheable, non-cacheable, IO, eXecutable, etc...). The user
can use this property and the DT API coming with it to act on the memory
node it is accompanied by.
We are still providing the DT_MEMORY_ATTR_APPLY() macro that can be used
to create the MPU regions as before, but we are adding also a
DT_MEMORY_ATTR_FOREACH_NODE() macro that can be used to cycle through
the memory nodes and act on those.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Usage of 64-bit address constants from devicetree without a
UINT64_C wrapping macro results in the following warning and the
cut-off of the address value:
"warning: integer constant is so large that it is unsigned"
This change extends devicetree API adding few wrappers over the
address constant getters which add ULL size suffix to an
address integer literal when the appearance of 64-bit address
values is possible
Signed-off-by: Alexander Razinkov <alexander.razinkov@syntacore.com>
Add test for the experimental feature GPIO_ENABLE_DISABLE_INTERRUPT,
which covers APIs gpio_pin_interrupt_enable() and
gpio_pin_interrupt_disable().
Signed-off-by: Sung-Chi Li <lschyi@google.com>
This commit adds access to the string values without a quotes.
Signed-off-by: Radosław Koppel <r.koppel@k-el.com>
Co-authored-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Since we can include other binding files at any level (child,
grand-child, etc.) it makes no sense to maintain two copies of pinctrl
props definitions (pincfg-node/pincfg-node-group). Instead,
pincfg-node.yaml defines props at root level, and it is included where
needed, either child-binding or grandchild-binding.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
DTS property attributes are (by default) not required.
Explicitly specifying `required: false` is redundant.
Perhaps a warning to that effect would be useful.
Signed-off-by: Chris Friedt <cfriedt@meta.com>
This adds a few bits to the devicetree API tests for multi-bus
nodes where a bus can support multiple protocols. This uses
I3C as basis as I3C controller can support both I2C and I3C on
the same bus, while I2C controller cannot support both. So
this needs to make sure the correct bus macros are generated
if appropriate (and not generated if not needed).
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Add emulator functionality to the serial_test driver, so that it can be
used to simulate a device on the other end of the uart.
If you don't set the buffer-size property in the dts node, there should
be effectively no change from the previous behavior.
Signed-off-by: Jeremy Bettis <jbettis@google.com>
This commit adds string token versions of the values also
in items inside string-array.
Signed-off-by: Radosław Koppel <r.koppel@k-el.com>
Co-authored-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Co-authored-by: Kumar Gala <galak@kernel.org>
Don't import the label prop from base.yaml as it is marked deprecated.
This lets the various tests that still use DT_LABEL() to work properly.
Signed-off-by: Kumar Gala <galak@kernel.org>
Use a test compatible so that test can create a DT-based device and
provide a valid choice when building the test.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
All in tree device drivers use some form of DEVICE_DT_GET
so we no longer need to require label properties.
Signed-off-by: Kumar Gala <galak@kernel.org>
added bindings for vnd,reset used for devicetree test reset
added adc temp sensor as supporting reset
Signed-off-by: Andrei-Edward Popa <andrei.popa105@yahoo.com>
This contains accessor macros for getting the maximum bitrate supported
by a CAN controller/transceiver combination.
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
This macro expands to DT_STRING_TOKEN if property exists, otherwise
falls back to default value.
Helpful when a non-required enum binding doesn't mention a default
value, but a default value makes sense to be set in the code.
Including DT_INST_STRING_TOKEN_OR and test code.
Signed-off-by: Bernardo Perez Priego <bernardo.perez.priego@intel.com>
The vnd,gpio-device compatible is now used by the test GPIO driver
(CONFIG_GPIO_TEST) introduced in #41387. This means that we can't define
new devices with this compatible when CONFIG_GPIO=y.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
The pinsN group name can be a confusing in some circumstances, so change
it to groupN. Some platforms (e.g. nrf or gd32) are already using
groupN. Documentation and API tests have been updated.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Add tests for the instance based enum macros DT_INST_ENUM_IDX and
DT_INST_ENUM_IDX_OR.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
One limitation of the current IPM API is that it is assuming that the
hardware is only exporting one single channel through which the data can
be sent or signalling can happen.
If the hardware supports multiple channels, the IPM device must be
instantiated (possibly in the DT) several times, one for each channel to
be able to send data through multiple channels using the same hw
peripheral. Also in the current IPM API only one callback can be
registered, that means that only one driver is controlling all the
signalling happening on all the channels.
This patch is introducing a new MBOX API that is supporting
multi-channel signalling and data exachange leveraging and extending the
previous (and outdated) IPM API.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Add a set of tests to check the API behavior. The API tests can only run
on a platform that does not have an actual pinctrl driver, e.g.
native_posix. The test itself implements a pinctrl mock driver and
provides the required "pinctrl_soc.h" header with required types/macros.
The implementation is used in the tests to verify the behavior of the
API or Devicetree macros.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Currently all the *-names and *-cells properties are derived from the
name of the base <name>s property. This is a limitation because:
- It forces the base property name to be plural ending in -s
- It doesn't allow the english exception of plural words ending in -es
With this patch we add one additional property 'specifier-space' that
can be used to explicitly specify the base property name.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Suggested-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Busy simulator is using counter device and entropy device to
generate random cpu load. Counter device cofiguration can be
used to set cpu load interrupt priority and optional pin that
can be set during the load.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Make sure that child bindings with their own compatibles are treated
as first-class bindings. Do this by making sure that nodes whose
bindings are defined via 'child-binding:' are picked up as bus nodes,
instead of a parent bus node of the same type.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
These need to be gpio controllers and have the necessary specifier
properties for use in gpio specifiers.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Test macro DT_PROP_BY_PHANDLE_IDX_OR. There are two tests, one when
the property exists and other when the property is not set.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
We aren't using vnd,serial for anything currently. Remove requiring
interrurpts property as that requirement isn't needed.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
In case a "required: False" enum binding doesn't mention a
default value, but a default value makes sense to be set in the code,
DT_ENUM_IDX_OR could be used to provide the default idx to be used.
New macro comes with appropriate tests.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Generalize io-channels and io-channel-names devicetree node properties
and move their definitions to base.yaml. Keep binding specific
description where relevant.
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
Add test cases that verify various bits and pieces of the legacy
devicetree macros match the new APIs.
Writing these test cases without giving rise to deprecated macro
warnings which might break people's CI if they build with -Werror
requires turning off the __WARN() generation in
devicetree_legacy_unfixed.h. The entire file is deprecated at this
point and must be explicitly enabled with an opt-in Kconfig option, so
there isn't any harm in doing this.
Nevertheless, take a minimally invasive approach to avoiding __WARN()
generation in gen_legacy_defines.py, to avoid the possibility of
breakage. This code is basically frozen anyway, so hacks like this
won't cause maintainability problems since it isn't being actively
maintained.
Use the new tests as fodder for a migration guide from the old API in
the documentation.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
At some point, "child-binding:" apparently only worked up to 2 levels
deep. That's not the case anymore, but add a regression test to make
sure that doesn't break. 3 levels deep ought to be enough for anyone.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
The macro iterates through the list of child nodes in a DT_DRV_COMPAT
instance and invokes provided macro for each node.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
