drivers: nrf: add support for forcing out-of-tree drivers

[thomasstenersen]

An external project extending the Zephyr RTOS and its drivers may have
subsystems that must use its own specific driver(s) when active. One
example is the nRF5x NVMC that must be scheduled in between radio
operations. A subsystem may also be dependent on its own drivers for
security, real-time and/or because of hardware constrains.

In order to not introduce non-Zephyr specific code into the Zephyr tree,
an option is added to disable the in-tree drivers in Zephyr. Because
Kconfig does not support a good way of de-selecting other symbols, a
variable on the form <DRIVER>_OUT_OF_TREE_DRIVER is added as a
dependency for each <DRIVER>. For example, the out-of-tree subsystem
will select FLASH_OUT_OF_TREE_DRIVER to disable the in-tree driver. A
solution for issue #8181 would open up for a more general solution,
however #8181 requires significant effort.

Support for out-of-tree drivers is added to Nordic drivers for
clock_control, entropy and flash.

See #13631 for earlier discussion.

Signed-off-by: Thomas Stenersen [email protected]

[codecov-io]

Codecov Report

Merging #14509 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #14509   +/-   ##
=======================================
  Coverage   51.96%   51.96%           
=======================================
  Files         309      309           
  Lines       45576    45576           
  Branches    10554    10554           
=======================================
  Hits        23683    23683           
  Misses      17083    17083           
  Partials     4810     4810

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8aba2d5...b127536. Read the comment docs.

[carlescufi]

This is a workaround so the following items need to be addressed:

• Open an issue describing the more general problem and link to it in the commit an dPR
• Change the names so that they are NRF-only
• Depend on SOC_COMPATIBLE_NRF

[ulfalizer]

@aescolar filled me in a bit re. the dev. meeting.

I think it's better to keep this as a local one-off hack until it's needed more broadly (hopefully it won't be). Just be careful so that the one-off doesn't get copied to a bunch of places. :)

[thomasstenersen]

The shippable failed, but when skimming through the log it looked like xtensa things failed. I rebased with changes from master and pushed again. @carlescufi is there anything missing?

[nashif]

I propose to drop the out-of-tree naming and call it alternate or something along those lines. This concept should also work with 2 compatible drivers in the same tree.

[carlescufi]

@thomasstenersen @nashif I agree with the proposal, I never liked OUT_OF_TREE. I suggest:

CLOCK_CONTROL_NRF_ALT or CLOCK_CONTROL_NRF_REPLACE or CLOCK_CONTROL_NRF_FORCE_ALT

or we could use the compatible concept from DT:
CLOCK_CONTROL_NRF_COMPATIBLE, or CLOCK_CONTROL_NRF_ALT_COMPAT or CLOCK_CONTROL_NRF_USE_COMPAT

[carlescufi]

Waiting for agreement on naming

[thomasstenersen]

If you are OK with CLOCK_CONTROL_NRF_FORCE_ALT I will use that, @nashif.

[thomasstenersen]

@nashif @carlescufi I changed the name to _FORCE_ALT.

[carlescufi]

Can you update the descriptions so that they reflect "force alternative implementation"?

[carlescufi]

i.e. not refer to out-of-tree but rather keep it focused on what it actually does

[thomasstenersen]

Changed.

[thomasstenersen]

Shippable failed (timed out apparently?), rebased on top of master to try this again.