hwmv2: Native targets covert to HW model v2 #69034

aescolar · 2024-02-15T14:59:39Z

Convert all native targets (POSIX arch based) in tree, and their SOC to the HW model v2.
With these commits all native_sim/posix[_64] and nrf*_bsim tests should pass again.

Notes:

It sits directly on top of hwmv2: Fix unit_testing and another cmake fix #69019 as it needs one of its commits
The last 3 commits are not independent/bisectable (I can squash them in one, but this target branch does not seem to be kept bisectable). So I kept them apart to ease reviewing and future understanding (so it is easier for out of tree users to know how to do the same for their posix arch based boards).
By now the 64 bit native_posix/sim variants, and the nrf5340bsim cpu app/net variants are kept as separate targets instead of variants of the same target. So after this PR commits everything works just like before this branch. So there is no breakage cause by this commits. Otherwise this PR would need to contain a massive amount of changes. Changing them to variants can be done safely after without a massive hard change in 1 commit or non-bisectable ones (i.e. the variants can be added, all tests and docs changed, and the old boards removed in separate commits).

Note: This PR fixes the current CI failures in the nrf5x_bsim ~~and unit_testing targets~~ (unit testing fixed by #69019 already)

boards/native/native_posix/Kconfig.defconfig

nordicjm · 2024-02-16T08:42:58Z

boards/native/native_posix/Kconfig

+	  stdout.
+
+config BOARD_NATIVE_POSIX_64
+	bool


This represents an issue @tejlmand

boards/native/native_sim/Kconfig.defconfig

boards/native/native_sim/Kconfig.native_sim

boards/native/native_sim/Kconfig.native_sim_64

boards/native/native_posix/Kconfig.native_posix

boards/native/native_posix/Kconfig.native_posix_64

boards/native/nrf_bsim/Kconfig.nrf52_bsim

boards/native/nrf_bsim/Kconfig.nrf5340bsim_nrf5340_cpuapp

boards/native/nrf_bsim/Kconfig.nrf5340bsim_nrf5340_cpunet

With hwmv2, the ARCH variable is not yet defined when this module is loaded (kconfig is parsed after this). So we cannot rely on it to detect if we are building for a host target. For this case, let's instead detect it by the BOARD or BOARD_DIR which are some of the very few things defined at this point. We retain the old check to support hwmv1 boards which may be in other folders. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

Just fix the path, it was not correct. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

Add a soc.yml and reorganize the Kconfig options Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

With new board.yml files and reorganizing their Kconfig options. Note: native_posix_64 & native_sim_64 remain as their own targets, instead of being variants of the base ones to avoid breakage in this commit, while not having a massive commit. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

With a new board.yml file and reorganizing their Kconfig options. Note: the nrf5340 variants remain as their own targets, instead of being variants of the base ones to avoid breakage in this commit (while not having a massime commit) Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

To match the current placement. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

tejlmand

first round of comments.

tejlmand · 2024-02-16T09:57:04Z

cmake/modules/FindHostTools.cmake

+if((${BOARD_DIR} MATCHES "boards\/native") OR ("${ARCH}" STREQUAL "posix")
+   OR ("${BOARD}" STREQUAL "unit_testing"))


for host tools we just need a generic compiler for preprocessing.

This mean we should be able to move this check to https://github.com/zephyrproject-rtos/zephyr/blob/main/cmake/modules/FindTargetTools.cmake where the arch is known.

This location is probably leftover from former days boilerplate.cmake file.

We can keep it here for now, but then please add a comment in code or create an enhancement issue / comment in #51831 to request looking at this as followup work.

Item added to #51831 list of TODOs

tejlmand · 2024-02-16T10:01:26Z

boards/native/native_posix/CMakeLists.txt

@@ -25,7 +25,7 @@ zephyr_library_include_directories(
  )

 if(CONFIG_HAS_SDL)
-	add_subdirectory(${ZEPHYR_BASE}/boards/boards_legacy/${ARCH}/common/sdl/ ${CMAKE_CURRENT_BINARY_DIR}/sdl)
+	add_subdirectory(${ZEPHYR_BASE}/boards/native/common/sdl/ ${CMAKE_CURRENT_BINARY_DIR}/sdl)


why is this using a path from zephyr base, and not relative to current folder ?
Such as:

add_subdirectory(${CMAKE_CURRENT_LIST_DIR}/../common/sdl/ ${CMAKE_CURRENT_BINARY_DIR}/sdl)

seems a bit safer.
However, current proposal is also ok, as I don't expect we'll be moving boards around so much more.

tejlmand · 2024-02-16T10:15:08Z

boards/native/native_posix/board.yml

+  vendor: Zephyr
+  socs:
+  - name: native
+- name: native_posix_64


should it be the board or the soc that is considered 64bit ?

Meaning, should we instead have:

board: name: native_posix vendor: Zephyr socs: - name: native - name: native_64

Considering the SoC 64bit then that would also remove the need for:

config BOARD_NATIVE_POSIX_64 bool select BOARD_NATIVE_POSIX select 64BIT

as the 64bit is then selected by the soc and there is no need to select BOARD_NATIVE_POSIX.
I also think this aligns more with what is done for qemu boards.

@nordicjm thoughts ?

should it be the board or the soc that is considered 64bit ?

I intended to do that as a follow up. In this PR I hoped to bring everything back to working order in the hwmv2 format while keeping the PR reasonably sized. The next steps were to define this and the nrf5340 as having different socs , change all tests and BT test infra accordingly, and then remove this "backwards compatible" board definitions.

@nordicjm 's pointed issue would still apply for the nrf*bsim boards, there the nrf52 will remain separate from the nrf53 anyhow.

true, and it's a side-effect of allowing multiple boards being defined in board.yml.

Something which would needs to be addressed.

tejlmand · 2024-02-16T10:16:14Z

boards/native/native_sim/board.yml

+  vendor: Zephyr
+  socs:
+  - name: native
+- name: native_sim_64


I guess similar comment regarding the board or soc being 64bit applies here.

tejlmand · 2024-02-16T10:24:10Z

boards/native/nrf_bsim/board.yml

+  vendor: Zephyr
+  socs:
+  - name: native
+- name: nrf5340bsim_nrf5340_cpuapp


this ought to be a soc with a cpucluster if the intention is to have it appear similar to a nRF530 SoC.

Doing so would also remove some other comments which I would otherwise have regarding the Kconfig.

I did notice:

Note: the nrf5340 variants remain as their own
targets, instead of being variants of the base ones
to avoid breakage in this commit

what kind of breakage are you refering to, and are you planning to make more updates to the bsim boards with regards to variants ?

Same answer as above
(clarification w "variants" I meant "socs" or "cpu-clusters")

tejlmand

concerns has been answered and followup plan is in place.

aescolar requested review from tejlmand, gmarull and nordicjm February 15, 2024 15:00

aescolar marked this pull request as ready for review February 15, 2024 15:02

zephyrbot added area: Xtensa Xtensa Architecture area: Toolchains Toolchains area: Build System area: Testsuite Testsuite labels Feb 15, 2024

zephyrbot requested review from 57300, aaronemassey, andyross, asemjonovs, ceolin, dcpleung, jeremybettis, nashif, stephanosio and yperess February 15, 2024 15:03

zephyrbot assigned tejlmand Feb 15, 2024

aescolar force-pushed the native_hwm branch from d5d36b2 to 3a897b7 Compare February 15, 2024 15:40

zephyrbot added the area: Process label Feb 15, 2024

zephyrbot requested a review from MaureenHelm February 15, 2024 15:48

nashif approved these changes Feb 15, 2024

View reviewed changes

nordicjm reviewed Feb 16, 2024

View reviewed changes

aescolar added 6 commits February 16, 2024 10:41

boards: nrf_bsim: Fix path to pinctrl_soc.h

8e62996

Just fix the path, it was not correct. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

soc: native: Convert to HWMv2

8a05889

Add a soc.yml and reorganize the Kconfig options Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

MAINTAINERS: Change paths to native and nrf*bsim boards

aecb641

To match the current placement. Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>

aescolar force-pushed the native_hwm branch from 1ff466c to aecb641 Compare February 16, 2024 09:42

tejlmand requested changes Feb 16, 2024

View reviewed changes

aescolar mentioned this pull request Feb 16, 2024

Hardware model v2 #51831

Closed

11 tasks

tejlmand approved these changes Feb 16, 2024

View reviewed changes

aescolar requested a review from nordicjm February 16, 2024 11:23

henrikbrixandersen changed the title ~~hwmv2: Native targets covert to HW modem v2~~ hwmv2: Native targets covert to HW model v2 Feb 16, 2024

tejlmand merged commit dc9d6e8 into zephyrproject-rtos:collab-hwm Feb 16, 2024
18 checks passed

aescolar deleted the native_hwm branch February 16, 2024 13:01

This was referenced Feb 16, 2024

hwmv2: Add definitions for the nrf5340bsim boards using the proper scheme #69089

Merged

hwmv2: Add definitions for the 64bit native_sim/posix boards as variants #69095

Merged

carlescufi mentioned this pull request Feb 28, 2024

Hardware Model v2 - Meta issue - Enhancements / feature request #69546

Open

20 tasks

tejlmand mentioned this pull request Apr 11, 2024

Zephyr toolchain infrastructure improvement, generic/target.cmake #71368

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

hwmv2: Native targets covert to HW model v2 #69034

hwmv2: Native targets covert to HW model v2 #69034

aescolar commented Feb 15, 2024 •

edited

Loading

nordicjm Feb 16, 2024

tejlmand left a comment

tejlmand Feb 16, 2024

aescolar Feb 16, 2024

tejlmand Feb 16, 2024

tejlmand Feb 16, 2024

aescolar Feb 16, 2024 •

edited

Loading

tejlmand Feb 16, 2024

tejlmand Feb 16, 2024

tejlmand Feb 16, 2024

aescolar Feb 16, 2024

tejlmand left a comment

		if((${BOARD_DIR} MATCHES "boards\/native") OR ("${ARCH}" STREQUAL "posix")
		OR ("${BOARD}" STREQUAL "unit_testing"))

hwmv2: Native targets covert to HW model v2 #69034

hwmv2: Native targets covert to HW model v2 #69034

Conversation

aescolar commented Feb 15, 2024 • edited Loading

Choose a reason for hiding this comment

tejlmand left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

aescolar Feb 16, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tejlmand left a comment

Choose a reason for hiding this comment

aescolar commented Feb 15, 2024 •

edited

Loading

aescolar Feb 16, 2024 •

edited

Loading