Part Number: AM62P
This is a follow-up to https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1640288/am62p-am62p-clarification-on-expected-power-consumption-in-suspend-deep-sleep-mode
While my colleague wanted to know what power consumption should be reachable, I have tried to figure out why we cannot reach this.
One of first things I have realized is that we don't go into the deepest suspend mode:
I couldn't measure that PMIC_LPM_EN0 goes LOW and on the kernel log it said:
[ 288.274548] ti-sci 44043000.system-controller: ti_sci: wakeup source:0x80, pin:0xff, mode:0x0
If I have read it correclt, mode:0x0 means that the deepest power mode was not reached.
I took a closer look at wake up sources:
root@am62px-var-som:~# for i in $(find /sys/ -iname 'wakeup' | grep -v kernel); do if [ ! -d "$i" ]; then echo -n "$i: "; cat $i; fi; done
/sys/devices/platform/bus@f0000/2860000.serial/2860000.serial:0/2860000.serial:0.0/tty/ttyS3/power/wakeup: disabled
/sys/devices/platform/bus@f0000/2860000.serial/power/wakeup: disabled
/sys/devices/platform/bus@f0000/20010000.i2c/i2c-2/2-0038/power/wakeup: enabled
/sys/devices/platform/bus@f0000/f900000.usb/f900000.usb:connector/power/wakeup: disabled
/sys/devices/platform/bus@f0000/f900000.usb/f900000.usb:connector/power_supply/usb-charger/power/wakeup: enabled
/sys/devices/platform/bus@f0000/f900000.usb/power/wakeup: enabled
/sys/devices/platform/bus@f0000/20100000.spi/spi_master/spi6/spi6.0/power/wakeup: enabled
/sys/devices/platform/bus@f0000/2820000.serial/power/wakeup: disabled
/sys/devices/platform/bus@f0000/2820000.serial/2820000.serial:0/2820000.serial:0.0/tty/ttyS1/power/wakeup: disabled
/sys/devices/platform/bus@f0000/fa20000.mmc/mmc_host/mmc2/mmc2:0001/mmc2:0001:1/power/wakeup: enabled
/sys/devices/platform/bus@f0000/2800000.serial/2800000.serial:0/2800000.serial:0.0/tty/ttyS0/power/wakeup: disabled
/sys/devices/platform/bus@f0000/2800000.serial/power/wakeup: disabled
/sys/devices/platform/bus@f0000/2850000.serial/power/wakeup: disabled
/sys/devices/platform/bus@f0000/2850000.serial/2850000.serial:0/2850000.serial:0.0/serial0/power/wakeup: disabled
/sys/devices/platform/bus@f0000/f910000.usb/power/wakeup: enabled
/sys/devices/platform/bus@f0000/f910000.usb/31100000.usb/xhci-hcd.0.auto/usb1/power/wakeup: disabled
/sys/devices/platform/bus@f0000/f910000.usb/31100000.usb/xhci-hcd.0.auto/power/wakeup: disabled
After deactivating both USB:
echo disabled > /sys/devices/platform/bus@f0000/f900000.usb/power/wakeup
echo disabled > /sys/devices/platform/bus@f0000/f910000.usb/power/wakeup
pm-suspend would (apparently) go into a deeper suspend mode, as I can measure that PMIC_LPM_EN0 goes LOW for 11us.
However, the complete system goes into a reset.
I have been able to measure some power rails, and while the voltage for the RAM is never lost, VDD_CORE drops from 850mv to 0mV, but only after a 1ms delay after the PMIC_LPM_EN0.
Can you advice on how to continue? Are there specific signals that need / should be measured next?
Is the PMIC signalling the AM62px core that something is not correct or is it the other way round or are some additional power rails that I have not yet measured probably the root cause for this?
In general:
The PMIC has been integrated similar to the SK-AM62P-LP. This should be true for both the HW side, as well as the SW integration.
I have realized that the Linux kernel is not really "aware" of the PMIC - there is no Device Tree for it.
I assume this is normal as only the Device Manager will communicate with the PMIC driver. Is this assumption correct?
Your power measurement in the other thread has been done based on the 11.01. release, I'm currently working on the 11.09 release. Are there any updates in that release which should "force me" to go to 11.01? Anything in the kernel, U-Boot, ATF?
Any help is appreciated. Thanks
Matthias

