soc: renesas: ra: configure option settings memory #71017

iandmorris · 2024-04-02T22:35:02Z

An area of flash memory on the RA4M1 MCU is used to store information used to configure the device following a reset. This patch instructs the linker to reserve this memory area and provides kconfig options that are used to populate it (at build time) with the desired device configuration.

NOTES:

The only board that currently uses the RA4M1 soc port is the Ardunio UNO R4 Minima. This board uses a bootloader to load the application. As this bootloader is located in the lower part of the RA4M1 flash memory it handles the reservation (and configuration) of the option setting memory area. In this case (and others where the application does not reside in the area that contains the option setting memory) this functionality can be disabled via the OPTION_SETTING_MEMORY kconfig option, however leaving it enabled does not cause any issues other than wasting 256 bytes of flash memory.

This option MUST be enabled if the application resides in the area that contains the option setting memory (which is usually the case if a bootloader is not used) - otherwise unintended behavior can occur due to misconfiguration of the device.

thedjnK · 2024-04-02T22:44:27Z

soc/renesas/ra/ra4m1/Kconfig.soc

Move to Kconfig

soburi · 2024-04-03T06:16:08Z

soc/renesas/ra/ra4m1/Kconfig.soc

Please refer to the devicetree definitions.

zephyr/dts/arm/renesas/ra/ra-cm4-common.dtsi

Lines 39 to 44 in f643f3a

hoco: hoco {

compatible = "fixed-clock";

clock-frequency = <24000000>;

status = "okay";

#clock-cells = <0>;

};

Please use this definition for determining HOCO frequency.

thedjnK · 2024-04-04T09:23:00Z

soc/renesas/ra/ra4m1/Kconfig

default n not needed

thedjnK · 2024-04-04T09:23:11Z

soc/renesas/ra/ra4m1/Kconfig

these should be dts entries

thedjnK · 2024-04-04T09:23:25Z

soc/renesas/ra/ra4m1/Kconfig

settings here should have SOC_ prefix

soc/renesas/ra/ra4m1/soc.c

soburi · 2024-04-05T01:30:43Z

soc/renesas/ra/ra4m1/soc.c

I couldn't estimate what the design intention from the "Set all fields to 1” description.
Since WDTSTPCTL is 1, the watchdog will not start, so it would be better to state that.

soburi · 2024-04-05T03:53:25Z

soc/renesas/ra/ra4m1/Kconfig

The voltage detecting and enabling HOCO on reset setting seems good to be managed by hwinfo driver.
If you are having difficulty creating a suitable devicetree at the moment,
I think it's okay to set a fixed value at this point.

soburi · 2024-04-06T12:43:10Z

@iandmorris

I have created PR #71174 based on these patches.
I notice that The option setting is a common function in the RA series.
So, I think defining in a standard file in the RA series is good.

soburi · 2024-04-21T05:06:06Z

soc/renesas/ra/ra4m1/Kconfig

We can use the devicetree definition to determine hoco frequency by the following code.

DT_PROP(DT_PATH(clocks, hoco), clock_frequency)

hi @soburi, i'm new to zephyr so i could be wrong about this but is seems like I can't use:

DT_PROP(DT_PATH(clocks, hoco), clock_frequency)

inside a Kconfig file to access a devicetree property, however i can use this within a 'C' file to access a devicetree property.

when i check other Kconfig files that refer to the devicetree they all seem to use:

dt_node_xxx_yyy

see soc/nxp/mcx/mcxnx4x/Kconfig for an example

let me know what you think, I probably missed something obvious :)

I think this part ...

https://github.com/zephyrproject-rtos/zephyr/pull/71017/files#diff-7cd037a8cbd8c3e02942d0cbda9cf90aa08fbab8ab55dcf57b46e12f660662beR8-R18

can rewrite as follows.
We can refer the hoco clock frequency directly in code, so we no need to define hoco freq in Kconfig.

#define HOCO_FREQ DT_PROP(DT_PATH(clocks, hoco), clock_frequency) #if HOCO_FREQ == 24000000 #define OFS1_HOCO_FREQ 0 #elif HOCO_FREQ == 32000000 #define OFS1_HOCO_FREQ 2 #elif HOCO_FREQ == 48000000 #define OFS1_HOCO_FREQ 4 #elif HOCO_FREQ == 64000000 #define OFS1_HOCO_FREQ 5 #else #error "Unsupported HOCO frequency" #endif

soburi · 2024-04-21T05:16:18Z

soc/renesas/ra/ra4m1/Kconfig

Whether the hoco is valid or invalid, I don't think there is any problem with judging it by the following code.

DT_NODE_HAS_STATUS(DT_PATH(clocks, hoco), okay)

If you are considering a case where the not enabled the hoco at reset but is enabled during startup processing, add a property to dts/bindings/clock/renesas,ra-clock-generation-circuit.yaml. I think it's also good. (But this might be overthinking.)

This is same as SOC_OPTION_SETTING_HOCO_FREQ.

https://github.com/zephyrproject-rtos/zephyr/pull/71017/files#diff-7cd037a8cbd8c3e02942d0cbda9cf90aa08fbab8ab55dcf57b46e12f660662beR122

This line can rewrite

.HOCOEN = !DT_NODE_HAS_STATUS(DT_PATH(clocks, hoco), okay),

And we can remove this Kconfig option.

https://github.com/zephyrproject-rtos/zephyr/pull/71017/files#diff-7cd037a8cbd8c3e02942d0cbda9cf90aa08fbab8ab55dcf57b46e12f660662beR10-R20

It seems to be a good. We can remove the config.

config SOC_OPTION_SETTING_HOCO_DISABLE

soburi · 2024-04-21T05:34:31Z

soc/renesas/ra/ra4m1/Kconfig

I think it's better to create dts/bindings/hwinfo/renesas,ra-lvd.yaml and define it in the devicetree node.

I don't think this setting needs to be done in this PR.
Since the hwinfo driver is used to obtain the startup factor after a reset, I think it would be a good idea to create a devicetree definition related to this and use that.
(If you are new to Zephyr development, creating a hwinfo driver brings good experience.)

@soburi, Agreed, I would like to implement this setting via the device tree but do it in another PR. Shall I leave the VDSEL as it is for this PR or maybe remove the option?

Don't leave options that don't work, as they confuse users.

soburi · 2024-04-26T23:29:16Z

soc/renesas/ra/ra4m1/soc.c

The OFS is used widely in the RA series.
It should place a common directory.

@soburi, while it's true that the OFS is used in a lot of RA microcontrollers it seems the contents of the registers, and their memory location is different in pretty much every series. A quick check of the device datasheets shows that RA6 series has different OSF1 register contents than RA4 series, as does the RA0 series. Therefore, I think it is better to leave this where it is...

If so, let's implement the individual implementation in soc.c.

An area of flash memory on the RA4M1 MCU is used to store information used to configure the device following a reset. This patch instructs the linker to reserve this memory area and provides kconfig options that are used to populate it (at build time) with the desired device configuration. Signed-off-by: Ian Morris <[email protected]>

iandmorris · 2024-05-02T16:10:00Z

@soburi, i think i have addressed all the comments. let me know if i missed something or if you have anything further.

soburi

I would like to make this OK for now.

Renesas has officially started supporting it, so there may be things to consider in the future.

#72150

iandmorris · 2024-05-03T02:44:25Z

Hi @thedjnK, I think I have addressed your comments. When you get a minute could you take a look and approve (or let me know if further changes are required)?
Thanks
Ian

thedjnK · 2024-05-05T09:44:38Z

soc/renesas/ra/ra4m1/Kconfig

+config SOC_OPTION_SETTING_MEMORY
+	bool "Option Setting Memory"
+	default y


Shouldn't this have a description? Never used these parts but "option setting memory" to me has no meaning, I have no idea what it is or what disabling it would do. Non-blocking comment and can be updated in a future PR.

zephyrbot added the platforms: Renesas RA label Apr 2, 2024

zephyrbot requested a review from soburi April 2, 2024 22:35

zephyrbot assigned soburi Apr 2, 2024

thedjnK requested changes Apr 2, 2024

View reviewed changes

soc/renesas/ra/ra4m1/Kconfig.soc Outdated

Comment on lines 22 to 44

Copy link

Contributor

thedjnK Apr 2, 2024

Choose a reason for hiding this comment

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

Move to Kconfig

soburi reviewed Apr 3, 2024

View reviewed changes

iandmorris force-pushed the ra_opt_set_mem branch from 2193478 to 950fa1f Compare April 4, 2024 02:44

thedjnK reviewed Apr 4, 2024

View reviewed changes

iandmorris force-pushed the ra_opt_set_mem branch 2 times, most recently from a20b1c7 to 33bb06c Compare April 5, 2024 01:23

soburi reviewed Apr 5, 2024

View reviewed changes

iandmorris force-pushed the ra_opt_set_mem branch from 33bb06c to 680fb60 Compare April 5, 2024 02:39

soburi reviewed Apr 5, 2024

View reviewed changes

soburi mentioned this pull request Apr 6, 2024

boards: arm: Add initial support for Renesas EK-RA2A1 #71174

Merged

iandmorris force-pushed the ra_opt_set_mem branch from 680fb60 to 87f9c94 Compare April 10, 2024 03:00

iandmorris force-pushed the ra_opt_set_mem branch 2 times, most recently from 7426bdf to f8e5701 Compare April 20, 2024 01:39

soburi reviewed Apr 21, 2024

View reviewed changes

soburi reviewed Apr 26, 2024

View reviewed changes

iandmorris force-pushed the ra_opt_set_mem branch 2 times, most recently from eef6ecf to c5e15f7 Compare May 1, 2024 15:35

iandmorris force-pushed the ra_opt_set_mem branch from c5e15f7 to 2468702 Compare May 1, 2024 22:04

soburi approved these changes May 3, 2024

View reviewed changes

iandmorris requested a review from thedjnK May 5, 2024 01:59

thedjnK approved these changes May 5, 2024

View reviewed changes

fabiobaltieri merged commit 0fd5365 into zephyrproject-rtos:main May 6, 2024

iandmorris deleted the ra_opt_set_mem branch May 6, 2024 16:34

	hoco: hoco {
	compatible = "fixed-clock";
	clock-frequency = <24000000>;
	status = "okay";
	#clock-cells = <0>;
	};

soc: renesas: ra: configure option settings memory #71017

soc: renesas: ra: configure option settings memory #71017

Uh oh!

Conversation

iandmorris commented Apr 2, 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!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

soburi commented Apr 6, 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!

soburi Apr 21, 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!

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!

soburi May 1, 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!

iandmorris commented May 2, 2024

Uh oh!

soburi left a comment

Choose a reason for hiding this comment

Uh oh!

iandmorris commented May 3, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

soburi Apr 21, 2024 •

edited

Loading

soburi May 1, 2024 •

edited

Loading