I3status-rust: Battery: sysfs driver can't get the power draw

[travankor]

When I use the sysfs driver, I can't get the power consumption to show on the bar. However, the upower driver does work.

Worth mentioning is that my laptop lacks a /sys/class/power_supply/BAT0/power_now file but it does have current_now and voltage_now.

[ammgws]

@travankor Do you still experience this?

[travankor]

@ammgws Yes, and I don't see any commits that fixed it.

Basically, this function has to be translated to Rust. So instead of not using current here, use the calculate_rate() function.

Alternatively, you can just use P=IV to get a quick and dirty result.

I've been meaning to learn Rust enough to fix this but I didn't have the time yet :/

[GladOSkar]

I assume that just taking current * voltage is probably closer to the way the power sysfs value would behave.

IMO, actually taking a linear regression over time would probably be neat and more accurate (even for the power case), but also quite a bit more complicated

[ammgws]

For now I just implemented the P=IV method, but need someone to test it.

Also, I found this. @travankor does your system have this powercap device?

[travankor]

Also, I found this. @travankor does your system have this powercap device?

Yes, it does. Thanks for looking into this.

[ammgws]

Can you check what values it outputs? Then we can compare it to the current * voltage method and see if maybe we want to prefer powercap before falling back to P=VI.

[GladOSkar]

As mentioned in the Stackexchange post you linked and on my system as well, it only provides a microjoule counter, which is probably just as difficult to get a power draw from as doing it via the other charge values

[ammgws]

Yeah it would need to be sampled over time. Was just wondering how its values correlate with the other method to see if it's worth looking into in the future.

[GladOSkar]

Yeah maybe. I put some thought and research into calculations like this (averaging intervals and such) in #214 , might be worth a look