DEVICE_AND_API_INIT and DEVICE_DEFINE are identical except that
DEVICE_DEFINE adds a parameter providing the device pm control
function, while DEVICE_AND_API_INIT does not. This requires duplicate
implementations where if CONFIG_DEVICE_POWER_MANAGEMENT is enabled
then DEVICE_AND_API_INIT delegates to DEVICE_DEFINE with a dummy pm
control function, and if it is not enabled then DEVICE_DEFINE discards
the parameter and delegates to DEVICE_AND_API_INIT.
DEVICE_INIT is like DEVICE_AND_API_INIT but doesn't provide an API.
Clean this up by refactoring so that DEVICE_DEFINE is the core
implementation, providing with and without device power management
right next to each other where they can be compared and maintained.
Redefine DEVICE_INIT and DEVICE_AND_API_INIT delegate to
DEVICE_DEFINE.
Also remove duplicate code by extracting the variations due to
enabling device power management into macros.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The driver API should be assigned at the time the device is defined.
Assigning in the init function is redundant and prevents moving the
pointer to a immutable driver structure.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The device structure definition in the code has drifted from the
displayed version in the documentation. Update the documentation to
match.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Extend the provided template with annotated examples of API based on
whether the functions must be invokable from user mode threads, since
there are currently no in-tree examples of the specific techniques
required.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
This documentation is well-intentioned but not good advice.
The reason why we have a vtable-like API abstraction at the
driver subsystem level is to introduce object orientation.
The subsystem defintions implement an abstract class that
specific driver implementations all implement, providing
a common interface for end users. Multiple drivers may be
written for a subsystem that are interchangeable to the end
user.
However, there is no point in introducing a vtable-like
abstraction for APIs which are specific to a single driver.
We are not trying to present an interface abstraction which
could support multiple implementations. There is just one.
Simply declare a function in a header, and implement it in
the driver's C file.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Name all subsystem reference consistently with an '_api' postfix and
clean up naming and folder structure in some cases.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
The original implementation left this function hidden in init.h which
prevented it from showing up in documentation. Move it to kernel.h,
and document it consistent with the other functions that allow caller
customization based on context.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
1) Add cryptographically secure random functions to provide
FIPS 140-2 compliant random functions.
2) Add name to random function choice selectors to ease
selection in SOC .defconfig files
3) Add bulk fill random functions.
Signed-off-by: David Leach <david.leach@nxp.com>
Add a simple diagram on the device driver model.
Instead of saying "this device driver", spell out that the note
is for the random number generator driver, as the visual
presentation does not have clear distinction.
Also update some formatting.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This PR updates the documentation to cover a technical solution to
providing a driver instance that extends the functionality of a
subsystem API. The solution described was mooted in Zephyr PR #17072
and approved by the Technical Steering Committee during its 2019-08-07
meeting.
Relates-to: #11993
Signed-off-by: Peter A. Bigot <pab@pabigot.com>
When init level hierarchy was reworked the documentation was switched to
describe the new levels, but the count of relevant levels was not
changed.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Move guides and APIs into separate directories and cleanup naming
introducing index files rather than named section files.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
