soc: nordic: Add LRCCONF management #79067

adamkondraciuk · 2024-09-26T15:28:17Z

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required.

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Upstream PR: zephyrproject-rtos/zephyr#79067 Signed-off-by: Adam Kondraciuk <[email protected]>

anangl · 2024-09-27T07:08:14Z

drivers/clock_control/clock_control_nrf2_common.h

Please remove also the struct clock_lrcconf_sink definition. There is no need to keep it as it can be replaced with just sys_snode_t.

anangl · 2024-09-27T07:08:30Z

drivers/clock_control/clock_control_nrf2_fll16m.c

Suggested change

#include <soc/nordic/common/soc_lrcconf.h>

#include <soc_lrcconf.h>

anangl · 2024-09-27T07:09:10Z

soc/nordic/common/soc_lrcconf.c

These should be static.

anangl · 2024-09-27T07:20:05Z

soc/nordic/nrf54h/CMakeLists.txt

Why is it added only for ARM cores?
I think that instead of the above entry, the following should be added in soc/nordic/common/CMakeLists.txt (so that nRF92 could also use it):

zephyr_library_sources_ifdef(NRF_PLATFORM_HALTIUM soc_lrcconf.c)

It was correct to only have it for arm cores, or rather for local domains, as global domains don't have an NRF_LRCCONF010 instance and it breaks their builds.

anangl · 2024-09-27T07:30:35Z

soc/nordic/nrf54h/power.c

Suggested change

#if !IS_ENABLED(CONFIG_SOC_NRF54H20_CPURAD)

#if !defined(CONFIG_SOC_NRF54H20_CPURAD)

IS_ENABLED should only be used in C expressions. See:

zephyr/include/zephyr/sys/util_macro.h

Lines 117 to 118 in 7e892f8

* Note: Use of IS_ENABLED in a <tt>\#if</tt> statement is discouraged

* as it doesn't provide any benefit vs plain <tt>\#if defined()</tt>

anangl · 2024-09-27T07:37:03Z

soc/nordic/nrf54h/soc.c

This looks like a leftover from some previous approach. node is not used by the code below.

anangl · 2024-09-27T07:38:15Z

soc/nordic/nrf54h/power.h

typo: "Pepare"

anangl · 2024-09-27T07:39:09Z

soc/nordic/nrf54h/power.h

This sentence seems to need rephrasing.

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Upstream PR: zephyrproject-rtos/zephyr#79067 Signed-off-by: Adam Kondraciuk <[email protected]>

karstenkoenig · 2024-09-27T10:31:28Z

soc/nordic/nrf54h/power.c

This only releases the poweron request when PM_STATE_SUSPEND_TO_IDLE was used to call k_cpu_idle() via the pm subsystem, but if the pm subsystem decided against suspending the system it will call k_cpu_idle() here instead and then not call the release.
This can happen if the is no suspend-to-idle state defined in devicetree or it has a minimum residency time that isn't met, e.g. a k_msleep(10) for with this pm configuration https://github.com/nrfconnect/sdk-nrf/blob/84492d6d9264745050a018d01f4f333b285864c1/tests/benchmarks/multicore/idle/boards/nrf54h20dk_nrf54h20_cpuapp_s2ram.overlay#L12C39-L12C39

This is still way better than losing power, but needs to be followed up to have all paths covered eventually

There is a mechanism to supply an exit hook on arm cores

zephyr/arch/arm/Kconfig

Line 56 in abd9008

config ARM_ON_EXIT_CPU_IDLE

there is this mechanism but it is quite ugly and requires special header with a macro. It was added as a workaround for PAN where code need to be executed immediately after WFI (so function cannot be used). I have #71667 which adds function hook symmetric to on_enter hook but PR stuck.

Maybe your PR would have a clean usecase now and it could be argued for again to avoid that assembly mess?

anangl · 2024-09-27T10:20:31Z

soc/nordic/nrf54h/power.h

This function does not seem to be needed anymore. It looks a bit like a hack actually.

anangl · 2024-09-27T10:30:02Z

soc/nordic/nrf54h/Kconfig

Shouldn't this be selected only when actually needed, so if PM || POWEROFF?

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Upstream PR: zephyrproject-rtos/zephyr#79067 Signed-off-by: Adam Kondraciuk <[email protected]>

nordic-krch · 2024-09-27T11:30:19Z

soc/nordic/common/soc_lrcconf.c

any reason why simple reference counter is not used here? I only see that the difference is that you keep list of clients (but its not used anywhere) and allow to have non-symmetric requests (second request from the same client is not counted) but it can potentially cause some issues. Reference counter would be simpler and use less resources.

At least right now there is non-symmetric requests so it couldn't be dropped

I think the code is symmetric now, maybe this would benefit from an onoff manager style thing for reference counting instead of manually tracking the references?

nordic-krch · 2024-09-27T11:37:57Z

soc/nordic/nrf54h/power.c

there is this mechanism but it is quite ugly and requires special header with a macro. It was added as a workaround for PAN where code need to be executed immediately after WFI (so function cannot be used). I have #71667 which adds function hook symmetric to on_enter hook but PR stuck.

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Upstream PR: zephyrproject-rtos/zephyr#79067 Signed-off-by: Adam Kondraciuk <[email protected]>

Add `idle` and `s2ram` power states for nRF54H20 cpuapp and cpurad. Also the substate `idle_cache_disable` added. Upstream PR: zephyrproject-rtos/zephyr#79067 Signed-off-by: Adam Kondraciuk <[email protected]>

karstenkoenig · 2024-10-28T11:01:41Z

This latest change supporting EngB needs to be applied downstream as well

adamkondraciuk · 2024-10-28T13:47:04Z

This latest change supporting EngB needs to be applied downstream as well

nrfconnect/sdk-nrf#18284

anangl · 2024-11-25T10:18:06Z

soc/nordic/nrf54h/soc.h

This looks quite weird. Why not just use extern sys_snode_t soc_node; in power.c, for example? Is this node intended to be used by any other modules/subsystems?

bjarki-andreasen

With the added lrcconf API being reference tracking, one node should not be shared between modules, like soc_pd_sys_snode_get(). Each module creates their own sys_snode_t and calls soc_lrcconf_poweron_request independently, soc_lrcconf_poweron_request tracks both and ensures thread safety :)

bjarki-andreasen · 2024-11-25T13:14:20Z

soc/nordic/common/soc_lrcconf.c

nit: space needed here

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Signed-off-by: Adam Kondraciuk <[email protected]>

Add `idle` and `s2ram` power states for nRF54H20 cpuapp and cpurad. Also the substate `idle_cache_disable` added. Signed-off-by: Adam Kondraciuk <[email protected]>

IRQs must be disabled before starting any procedures to prepare for low-power states. Signed-off-by: Adam Kondraciuk <[email protected]>

anangl · 2024-11-25T14:59:24Z

soc/nordic/nrf54h/soc.c

 #define HSFLL_NODE DT_NODELABEL(cpurad_hsfll)
 #endif

+sys_snode_t soc_node;


A comment explaining that this is not static because it is also accessed from power.c would not hurt.

This node really should not be shared in the first place... Instead, add a power_init() function to power.c which requests the power domain, moving the node to power.c and making it static, and call power_init() from soc.c

or, just move power_domain_init() to power.c?

This is a bit complicated, because power.c is only included in the build for CONFIG_PM OR CONFIG_POWEROFF. That's why I suggested this shortcut with extern.
But I agree, it's probably worth putting a little more effort to do it cleanly.

How about #ifdefed power_init() call?

Not neccesary to block it

adamkondraciuk requested a review from nordic-krch as a code owner September 26, 2024 15:28

zephyrbot added platform: nRF Nordic nRFx area: Clock Control labels Sep 26, 2024

zephyrbot requested review from anangl, jaz1-nordic, kl-cruz, magp-nordic, masz-nordic and nika-nordic September 26, 2024 15:29

zephyrbot assigned nordic-krch Sep 26, 2024

anangl reviewed Sep 27, 2024

View reviewed changes

adamkondraciuk force-pushed the idle_lrcconf_config branch from 46ac139 to f7ea669 Compare September 27, 2024 10:17

karstenkoenig reviewed Sep 27, 2024

View reviewed changes

anangl reviewed Sep 27, 2024

View reviewed changes

nordic-krch reviewed Sep 27, 2024

View reviewed changes

adamkondraciuk force-pushed the idle_lrcconf_config branch from f7ea669 to c13c926 Compare September 27, 2024 14:40

adamkondraciuk force-pushed the idle_lrcconf_config branch from c13c926 to f046e9a Compare September 30, 2024 07:35

adamkondraciuk force-pushed the idle_lrcconf_config branch from aec753f to 19a2be2 Compare October 28, 2024 10:49

adamkondraciuk requested review from 57300, karstenkoenig, kl-cruz and nordic-segl October 28, 2024 10:50

karstenkoenig previously approved these changes Oct 28, 2024

View reviewed changes

adamkondraciuk dismissed karstenkoenig’s stale review via 7e627b3 November 13, 2024 11:17

adamkondraciuk force-pushed the idle_lrcconf_config branch 3 times, most recently from 8a27e3d to a65320d Compare November 13, 2024 19:55

nordic-segl approved these changes Nov 14, 2024

View reviewed changes

nordic-krch previously approved these changes Nov 25, 2024

View reviewed changes

anangl reviewed Nov 25, 2024

View reviewed changes

bjarki-andreasen previously requested changes Nov 25, 2024

View reviewed changes

soc/nordic/common/soc_lrcconf.c Outdated

Comment on lines 26 to 28

Copy link

Contributor

bjarki-andreasen Nov 25, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: space needed here

adamkondraciuk dismissed nordic-krch’s stale review via 2e39284 November 25, 2024 14:46

adamkondraciuk force-pushed the idle_lrcconf_config branch from a65320d to 2e39284 Compare November 25, 2024 14:46

adamkondraciuk requested review from anangl, bjarki-andreasen and nordic-krch November 25, 2024 14:47

adamkondraciuk added 3 commits November 25, 2024 15:49

soc: nordic: Add LRCCONF management

11800d7

Due to the possibility of simultaneous accesess to LRCCONF registers, additional management is required. Signed-off-by: Adam Kondraciuk <[email protected]>

dts: arm: nordic: Add power states for nRF54H20

adccd16

Add `idle` and `s2ram` power states for nRF54H20 cpuapp and cpurad. Also the substate `idle_cache_disable` added. Signed-off-by: Adam Kondraciuk <[email protected]>

soc: nordic: nrf54h: disable IRQ before PM config

2dc89a7

IRQs must be disabled before starting any procedures to prepare for low-power states. Signed-off-by: Adam Kondraciuk <[email protected]>

adamkondraciuk force-pushed the idle_lrcconf_config branch from 2e39284 to 2dc89a7 Compare November 25, 2024 14:53

anangl approved these changes Nov 25, 2024

View reviewed changes

kl-cruz approved these changes Nov 26, 2024

View reviewed changes

nordic-krch approved these changes Nov 26, 2024

View reviewed changes

fabiobaltieri merged commit be1f405 into zephyrproject-rtos:main Nov 26, 2024
26 checks passed

	#include <soc/nordic/common/soc_lrcconf.h>
	#include <soc_lrcconf.h>

	#if !IS_ENABLED(CONFIG_SOC_NRF54H20_CPURAD)
	#if !defined(CONFIG_SOC_NRF54H20_CPURAD)

	* Note: Use of IS_ENABLED in a <tt>\#if</tt> statement is discouraged
	* as it doesn't provide any benefit vs plain <tt>\#if defined()</tt>

soc: nordic: Add LRCCONF management #79067

soc: nordic: Add LRCCONF management #79067

Uh oh!

Conversation

adamkondraciuk commented Sep 26, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

karstenkoenig commented Oct 28, 2024

Uh oh!

adamkondraciuk commented Oct 28, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjarki-andreasen left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjarki-andreasen Nov 25, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anangl Nov 25, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

10 participants

bjarki-andreasen Nov 25, 2024 •

edited

Loading

anangl Nov 25, 2024 •

edited

Loading