Hi,
On 12-04-18 10:18, Ognjen Galic wrote:
On April 12, 2018 10:14:01 AM GMT+02:00, Hans de Goede <hdegoede@xxxxxxxxxx> wrote:
Hi,
On 11-04-18 19:47, Daniel Drake wrote:
On Wed, Apr 11, 2018 at 11:42 AM, Hans de Goede <hdegoede@xxxxxxxxxx>
wrote:
Either way userspace regressions are a big no-no and there are
enough
people running new kernels with an old(ish) userspace for this to be
a real problem, so we really need to go with FULL for now.
Would it be a regression though? According to your analysis userspace
is already reporting an inaccurate state.
Hmm, you are right, so if we do the right thing and resort
NOT_CHARGING then upower will end up reporting UNKNOWN to
the desktop environment instead of the current DISCHARGING,
which is a small improvement already, although some may see
this as a regression.
My patch was reverted because Rafael treated that as regression. A lot of people see that as regression.
In the case of your patch the /sys/class/power_supply device
when from reporting "unknown", which then usually correctly
gets translated to "full" by upower (*), to reporting "not charging"
which upower translates ironically enough to "unknown".
So with your patch what the (typical) desktop-environment sees
when the device is on AC and the charge is above the start-charging
threshold went from "full" to "unknown" which clearly is a regression
from the users pov.
In the case of the devices my patch applies to, these were wrongly
reporting "discharging" before, so what the DE sees goes from:
"discharging" (blatantly wrong) to "unknown" (not ideal, but at least
also not wrong).
Regards,
Hans
*) unless people have changed the start charging threshold, throwing
of upower's heuristics
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
