This is very irritating for people using dual-cell batteries and charging with higher currents (1250 mA), as it limits charging to <500 mA, resulting in getting less-than-half-charged device, after usual "full charging" time passes. Not to mention that device "thinks" it's connected as USB client indefinitely, even after you unplug charger.

One can workaround it without rebooting device, by (ab)using hostmode applet - just select mode "peripheral with charging (it doesn't matter that you're in this mode already), then, issuing (as root):
...then plugging and unplugging charger, then issuing above command again, and finally, connecting charger back. At this point, it should charge as it should (in my case, full 1250 mA), without any illusion of being connected in USB data mode, nor side effects. Sometimes it need to repeat those steps 2 or 3 times, but will eventually get unstuck.

As you have seen in my post from CSSU thread, new version of applet may be PITA, in case it decides to use "design capacity" as "full". I always end up with nonsenses like 3300 current mAh, 1200 mAh max. Also, as shadowjk noted, even for people with "regular" batteries, it (percents, current charge, max charge) doesn't add up arithmetically, if you spend a minute and calculate it by hand.

Solution - modify one line in applet's source code, to make it use bq27x00_battery reported capacity with precedence over rx51_battery/bme reported max capacity, as the latter is plain BS. For kernel-power users (everyone and his uncle) it will work OK, for non-kp users, it will work just like to does now (aka report idiotic values, but still looking pretty, in case of regular batteries).

/Estel

// Edit

I hope CSSU guys will fix it or did so already - if it won't get fixed "upstream", I'll bloody do that one liner, and upload fork to extras-*. It would be, well, silly situation, but Meh.