Bogus (380 mW) value for power consumption

[patrakov]

I have a HP EliteBook 735 G6 laptop. It is AMD-based. On it, powertop reports completely bogus values for power consumption and battery time remaining. Something like:

The battery reports a discharge rate of 368 mW
The energy consumed was 7.38 J
The estimated remaining time is 50 hours, 6 minutes

At the same time, Cinnamon battery applet correctly reports that the battery is good only for 2 hours and 20 minutes.

Presumably, this is because none of the methods that powertop can use for determining the battery discharge rate actually works:

20:06:58.602628 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/present", O_RDONLY) = 123
20:06:58.602687 read(123, "1\n", 8191)  = 2
20:06:58.603210 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/status", O_RDONLY) = 123
20:06:58.603286 read(123, "Discharging\n", 8191) = 12
20:06:58.604335 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/power_now", O_RDONLY) = -1 ENOENT (No such file or directory)
20:06:58.604398 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/energy_now", O_RDONLY) = -1 ENOENT (No such file or directory)
20:06:58.604449 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/voltage_now", O_RDONLY) = 123
20:06:58.604500 read(123, "11140000\n", 8191) = 9
20:06:58.604571 openat(AT_FDCWD, "/sys/class/power_supply/BAT0/current_now", O_RDONLY) = 123
20:06:58.604620 read(123, 0x5630a5a7e730, 8191) = -1 ENODEV (No such device)

The only usable attribute is charge_now, but, alas, powertop does not know how to use it.

$ LANG=en_US.UTF-8 grep -rH .
uevent:POWER_SUPPLY_NAME=BAT0
uevent:POWER_SUPPLY_TYPE=Battery
uevent:POWER_SUPPLY_STATUS=Discharging
uevent:POWER_SUPPLY_PRESENT=1
uevent:POWER_SUPPLY_TECHNOLOGY=Li-ion
uevent:POWER_SUPPLY_CYCLE_COUNT=35
uevent:POWER_SUPPLY_VOLTAGE_MIN_DESIGN=11550000
uevent:POWER_SUPPLY_VOLTAGE_NOW=11166000
uevent:POWER_SUPPLY_CHARGE_FULL_DESIGN=4330000
uevent:POWER_SUPPLY_CHARGE_FULL=3009000
uevent:POWER_SUPPLY_CHARGE_NOW=1459000
uevent:POWER_SUPPLY_CAPACITY=48
uevent:POWER_SUPPLY_CAPACITY_LEVEL=Normal
uevent:POWER_SUPPLY_MODEL_NAME=Primary
uevent:POWER_SUPPLY_MANUFACTURER=Hewlett-Packard
uevent:POWER_SUPPLY_SERIAL_NUMBER=13856 2020/04/22
charge_full_design:4330000
serial_number:13856 2020/04/22
technology:Li-ion
grep: current_now: No such device
charge_now:1459000
present:1
power/runtime_active_time:0
power/runtime_status:unsupported
grep: power/autosuspend_delay_ms: Input/output error
power/runtime_suspended_time:0
power/control:auto
manufacturer:Hewlett-Packard
hwmon2/in0_input:11166
hwmon2/power/runtime_active_time:0
hwmon2/power/runtime_status:unsupported
grep: hwmon2/power/autosuspend_delay_ms: Input/output error
hwmon2/power/runtime_suspended_time:0
hwmon2/power/control:auto
grep: hwmon2/curr1_input: No such device
hwmon2/name:BAT0
type:Battery
charge_full:3009000
capacity:48
cycle_count:35
voltage_now:11166000
status:Discharging
alarm:0
model_name:Primary
voltage_min_design:11550000
capacity_level:Normal

[spbisc97]

Same problem here with Fedora 37 and hp Elitebook 830 g5 , measuring around 360mW

[fenrus75]

this data comes from your bios ... not a lot powertop can do if the bios lies to it

On Wed, Nov 30, 2022 at 2:11 PM Simone Rotondi @.***> wrote:

Same problem here with Fedora 37 and hp Elitebook 830 g5 , measuring around 360mW

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>

[spbisc97]

It was strange for me because other utilities like the simple "power statistics" by gnome take the correct discharge rate. Anyway thanks for your help and for the excellent utility!

[paulmenzel]

Over what interface are these values read from the system firmware (“BIOS”)?