AM620-Q1: Deep sleep console log.

Gibbs Shih

Expert 5285 points

Part Number: AM620-Q1

Tool/software:

Hi, Dear Expert

Environment,

SDK 10.0

HW, AM62_LP_EVM & AM62A_LP_EVM

We follow this instruction to enter deep sleep mode.

https://software-dl.ti.com/processor-sdk-linux/esd/AM62AX/10_00_00/exports/docs/linux/Foundational_Components/Power_Management/pm_low_power_modes.html?highlight=sleep%20mode

echo mem > /sys/power/state

Our dump message as below,

But I think standard message as below.

echo mem > /sys/power/state
[  230.181404] PM: suspend entry (deep)
[  230.185406] Filesystems sync: 0.000 seconds
[  230.219094] Freezing user space processes
[  230.224495] Freezing user space processes completed (elapsed 0.001 seconds)
[  230.231506] OOM killer disabled.
[  230.234736] Freezing remaining freezable tasks
[  230.240432] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[  230.247825] printk: Suspending console(s) (use no_console_suspend to debug)
[  230.266309] ti-sci 44043000.system-controller: ti_sci_cmd_set_device_constraint: device: 179: state: 1: ret 0
[  230.266456] ti-sci 44043000.system-controller: ti_sci_cmd_set_device_constraint: device: 178: state: 1: ret 0
[  230.273953] omap8250 2800000.serial: PM domain pd:146 will not be powered off
[  230.274450] ti-sci 44043000.system-controller: ti_sci_cmd_set_device_constraint: device: 117: state: 1: ret 0
[  230.284177] remoteproc remoteproc1: stopped remote processor 79000000.r5f
[  230.287440] Disabling non-boot CPUs ...
[  230.289569] psci: CPU1 killed (polled 0 ms)
[  230.292413] psci: CPU2 killed (polled 4 ms)
[  230.295457] psci: CPU3 killed (polled 0 ms)

It seems that we lost some log message, does deep sleep work ok?

Or Do I miss something?

Thank you very much

Gibbs

8 months ago

0 Anshu Madwesh 8 months ago

TI__Mastermind 18380 points

Hi Gibbs,

Does the SoC resume as expected? It it does, then there is no issue if any kernel logs don't display.

Thanks,

Anshu

0 Gibbs Shih 8 months ago in reply to Anshu Madwesh

TI__Expert 5285 points

Hi, Anshu

I think AM62 do not go to sleep mode, because we found power consumption is still too high

so I create this thread.

different log message is one of the clue.

Ref original thread

https://e2e.ti.com/support/processors-group/processors---internal/f/processors---internal-forum/1454207/am620-q1-power-consumption-is-still-to-high-in-deep-sleep-mode/5591056#5591056

Gibbs

0 Anshu Madwesh 8 months ago in reply to Gibbs Shih

TI__Mastermind 18380 points

Hi Gibbs,

We'll address the power consumption part in the other thread.

Is the SoC able to suspend and resume?

Can you try 'rtcwake -s 10 -m mem'? This will suspend the SoC and resume after 10 seconds.

Best Regards,

Anshu

0 Gibbs Shih 8 months ago in reply to Anshu Madwesh

TI__Expert 5285 points

Hi, Anshu

Thanks you replies

Update my status

* SW : 10.01.10.04

* EVB : AM62_LP_EVM

Even log is different, but it seems go to sleep mode.

* Before Sleep, 410mA

* After Sleep, 60mA

(1) I am not sure the test result is correct, but it seems work on EVM.

(2) Customer say their customized board seems not work for deep sleep mode, I will still check with them. but do we have debug check list for deep sleep mode?

Thank you very much.

Gibbs

+1 Anshu Madwesh 8 months ago in reply to Gibbs Shih

TI__Mastermind 18380 points

Hi Gibbs,

Please be aware a tool that measures the current through the power source is measuring the whole board, not just the SoC. This will cause the power to high than expected.

If you can confirm that the suspend/resume is not working from a software perspective, not from power consumption, we can continue to debug this further. The next step is to review the SYSFW trace: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1323581/faq-processor-sdk-am62x-a-p-how-can-we-debug-deep-sleep-on-am62x-am62a-am62p-using-sysfw-trace

Thanks,

Anshu

0 Gibbs Shih 8 months ago in reply to Anshu Madwesh

TI__Expert 5285 points

This is just an update let some one need to know when he(she) search this thread

Red : Enter deep sleep

Yellow : After mcu gpio trigger, exit deep sleep

How to I/O trigger?

Ref.

software-dl.ti.com/.../pm_wakeup_sources.html

3.3.6.2. MCU GPIO

0 Anshu Madwesh 8 months ago in reply to Gibbs Shih

TI__Mastermind 18380 points

Hi Gibbs,

Thanks for the update. From that log, Deep Sleep seems to work correctly as it suspends and then resumes upon a MCU GPIO wake.

Will keep the thread open for now.

Thanks,

Anshu

Processors

Processors forum

AM620-Q1: Deep sleep console log.

3.3.6.2. MCU GPIO