S32KのトピックRe: A question about s32k144 MPU

A question about s32k144 MPU

pangyun — Wed, 25 Jun 2025 07:49:50 GMT

My configuration is as follows：

Code:

(void)MPU_EnableRegion(&mpu_pal_config_1_Instance, 3U, false);

//test = *((uint32_t *)ADDRESS_PROTECT);

Results:

There is no code to read,but it will always go to "void HardFault_Handler(void)".

Change the configuration:

I only changed the address.It will be ok! I am confused.

Please tell me the reason,thank you!

Re: A question about s32k144 MPU

danielmartynek — Wed, 02 Jul 2025 13:00:30 GMT

Hi @pangyun,

Can you please share the project so that I can test it?

What do you see in S32_SCB_CFSR and BFAR registers?

Thank you,

BR, Daniel

Re: A question about s32k144 MPU

pangyun — Thu, 24 Jul 2025 03:49:33 GMT

My project's name is "mpu_pal_memory_protection_s32k144" from the example.I only change the adress in the MPU configuration and some code in main.c, I use Debug_Flash,, here‘s the MPU configuration and the code I changed in main.c, you can test.Thank you!

Re: A question about s32k144 MPU

danielmartynek — Wed, 30 Jul 2025 14:14:35 GMT

Hi @pangyun,

I'm sorry for the delay.

I have been testing it and I think I can reproduce it, however, it fails during a read operation.

If I let it run, it ends up in the HardFault_Handler().

MAP file:

Can you confirm you have the same results?

If not, please send your complete project.

Regards,

Daniel

Re: A question about s32k144 MPU

pangyun — Fri, 01 Aug 2025 08:14:52 GMT

Yes ,it is confirmed that I have the same results, but what's the problem?

Re: A question about s32k144 MPU

danielmartynek — Fri, 01 Aug 2025 08:42:22 GMT

Hi @pangyun,

You disabled the descriptor that grants the core access to the flash region, which is now causing read operations to fail. From my perspective, this behavior appears expected, I don't see any problem.

Regards,

Daniel

Re: A question about s32k144 MPU

pangyun — Fri, 01 Aug 2025 09:01:30 GMT

You disabled the descriptor that grants the core access to the flash region,

and the read operation is blocked, but the problem also exist,why

Re: A question about s32k144 MPU

pangyun — Mon, 04 Aug 2025 02:45:16 GMT

I know your means,but I want to realize a funtion that when I disable some regions to ignore Core read permission, some read cann't be allowed ,for example:

test = *((uint32_t *)ADDRESS_PROTECT),but some codes like :PINS_DRV_ReadPins(SW_GPIO) & (1U << SW) ,it will also go to the hard fault too, I don't want this happen,I need it to run normally.

I am confused now, when I use MPU, some regions cann't be set read permission,it wil lead a serious problem,I want to know what these regions are?

Re: A question about s32k144 MPU

danielmartynek — Mon, 04 Aug 2025 13:12:14 GMT

Hi @pangyun,

Certain SDK functions read constant data stored in flash memory. If the MPU is configured to disable read access to those flash regions, these functions may trigger a HardFault.

To avoid this, you can restructure your linker file to define separate memory sections:

Code Section – contains executable functions.
Constant Data Section – contains read-only constants used by the SDK.

By doing this, you can configure the MPU with more precision:

Allow execution in the code section.
Allow read access in the constant data section.

This separation ensures that functions can execute without faulting, even if read access is disabled in the code region, because the constants they rely on are located in a region where reading is permitted.

Regards,

Daniel

Re: A question about s32k144 MPU

pangyun — Thu, 14 Aug 2025 02:39:00 GMT

I have done as you said:

Allow read access in the constant data section.

Allow execution in the code section.

But the problem still exsits, why?

Re: A question about s32k144 MPU

danielmartynek — Fri, 15 Aug 2025 06:24:41 GMT

Hi @pangyun,

Can you check the location of the constants in the MAP file?

Re: A question about s32k144 MPU

pangyun — Mon, 25 Aug 2025 01:17:39 GMT

Yes，I checked the MAP file and it is confirmed that the constant had been moved .

The address I need to protect involves the entire program area that start from address 0,I don't know whether this function can be achieved by using MPU。

Re: A question about s32k144 MPU

danielmartynek — Mon, 25 Aug 2025 07:02:09 GMT

Hi @pangyun,

Can you share the project so that I can easily reproduce it?

If you don't want to share it here, please create a support ticket.

Re: A question about s32k144 MPU

pangyun — Mon, 25 Aug 2025 08:55:22 GMT

I don't know how to share my project with you.

Re: A question about s32k144 MPU

danielmartynek — Mon, 25 Aug 2025 10:31:41 GMT

Hi @pangyun,

Please compress the entire project into a .zip file and attach it either here or to the support ticket.

Re: A question about s32k144 MPU

pangyun — Tue, 26 Aug 2025 00:46:11 GMT

Here my project, you can test, thank you!

Re: A question about s32k144 MPU

danielmartynek — Thu, 28 Aug 2025 07:28:15 GMT

Thanks @pangyun for the project.

I just tested it.

First, I compiled it into the flash, because you were testing it in SRAM

Then I changed

(void)MPU_EnableRegion(&mpu_pal_config_1_Instance, 2U, false);

(void)MPU_EnableRegion(&mpu_pal_config_1_Instance, 3U, false);

So that the descriptor 3 that allows the READ access is disabled only.

After debugging I see that the fault is at 0x2F5C.

This is because the code of MPU_GetError() uses a LDR instruction to load R0 with the value at 0x2F5C - so this is a read operation. Execution of the code in the Flash is allowed by the MPU but reading is not.

Regards,

Daniel

Re: A question about s32k144 MPU

pangyun — Thu, 28 Aug 2025 07:55:55 GMT

Thank you for your test,but now how can I ensure my project runs normally while preventing flash from being read.

Re: A question about s32k144 MPU

pangyun — Thu, 28 Aug 2025 07:58:02 GMT

How can I determine which addresses can be set as read disabled and which addresses cannot be set as read disabled to ensure the normal operation of my program

Re: A question about s32k144 MPU

pangyun — Thu, 28 Aug 2025 08:00:42 GMT

Our design requirement was to set as much program information as possible to be unreadable, but now it is affecting the normal operation of the program itself. What should I do。