CONFIG_PM Freezes my application

[jurecki0]

Describe the bug

Using a custom board with STM32WB55RGV6, I have an application that uses lorawan, it connects, then it enters a loop of sending a payload every couple of minutes.
Now, I need to make the device go into STOP2 for the couple of minutes it does nothing and just adding CONFIG_PM=y to my conf file breaks everything.
I don't see logs, chirpstack doesn't receive any packets, nothing, not even a request to join.
I tried resuming usart1 to see logs, I tried pm_device_runtime_get(), and pm_device_runtime_put(), I tried,

    // if (device_is_ready(uart))
    // {
    //     pm_device_action_run(uart, PM_DEVICE_ACTION_RESUME);
    // }

And don't know what to do next.
I'm on Zephyr 4.1.0 right now, on zephyr 3.7.0 the workaround code above worked but not in this version.

Regression

• This is a regression.

Steps to reproduce

Just enabling CONFIG_PM in prj.conf for STM32WB55RGV6.

Relevant log output

Impact

Major – Severely degrades functionality; workaround is difficult or unavailable.

Environment

• OS : Windows 11
• Board : Custom board with STM32WB55RGV6
• Zephyr 4.1.0

Additional Context

No response

[github-actions]

Hi @jurecki0! We appreciate you submitting your first issue for our open-source project. 🌟

Even though I'm a bot, I can assure you that the whole community is genuinely grateful for your time and effort. 🤖💙

[jurecki0]

Update: the application freezes if I combine CONFIG_LORA and CONFIG_PM.

[sayoojkkarun]

Can you check if the device is going to any power save mode and not returning back?

[sayoojkkarun]

Different power save modes are defined in device tree, check them and see which one causes the problem.
Are you using CONFIG_PM_DEVICE or CONFIG_PM_DEVICE_RUNTIME as well?

[jurecki0]

I'm gonna check for the power save modes.
As for the config I tried with and without CONFIG_PM_DEVICE_RUNTIME , nothing changed.

What strange thing I observed is that when I add CONFIG_LOG_DEFAULT_LEVEL=4 so the DBG logs should show up, the device connects to the LoRaWAN gateway, the payload has been sent etc.. The moment I switch to another log level, it just breaks.

[sayoojkkarun]

Okay. the debug logs are preventing your system to go into power save mode, hence it is behaving as you have explained. You can try to narrow-down which mode of power save is causing this.

[jurecki0]

I've tried all 3 substate_id's so STOP0, STOP1 and STOP2, same behavior but I can't confirm if it's actually switching to those modes, the power consumption is the same.

I use
pm_state_force(0, &(struct pm_state_info){ .state = PM_STATE_SUSPEND_TO_RAM, .substate_id = 3, });

How can I reliably check if it's entering the actual power state?

[sayoojkkarun]

I've tried all 3 substate_id's so STOP0, STOP1 and STOP2, same behavior but I can't confirm if it's actually switching to those modes, the power consumption is the same.

I use pm_state_force(0, &(struct pm_state_info){ .state = PM_STATE_SUSPEND_TO_RAM, .substate_id = 3, });

How can I reliably check if it's entering the actual power state?

I generally check with prints before entering the power state and the power consumption with power profiler kit.

[jurecki0]

The issue is that I cannot print anything If I enable CONFIG_PM and CONFIG_LORA, nothing gets printed. Even with an empty main.

[erwango]

Some details would be useful:

• STM32WB55 doesn't come with native LORA support, can you provide more information on your set up ?
• Are you able to test other use cases when CONFIG_LOAR=n, CONFIG_PM=y ?