Some issues regarding battery status reading

Hi,

I have seen this in my dmesg:

call to w1_ds2760_status failed (0xc38f9e58)
Stopping tasks:
================================================================call to
w1_ds2760_status failed (0xc38f9e58) ==== stopping tasks failed (1
tasks remaining) Restarting tasks...<6> Strange, w1_bus_master1 not
stopped

This was produced using an 2.6.15-hh2 kernel with h2200_leds module enabled,
apmd and mb-applet-battery were not running.
I just did:
cd "/sys/class/battery/h2200 battery"
while true ; do sleep 1 ; cat charge current status ; done
I started that after some time of cycling (max. 2 minutes).

The ipaq was powered by my hub dynamo (of course with a
rectifier, a capacitor and a voltage converter. When cycling slowly,
the frequency is low, so the ipaq sees power interruptions
and generates lots of "AC plug" interrupts.
The ds2760 handler does not seem to like it.

After stopping, I tried to suspend the ipaq with apm --suspend

The yellow led still reacts on plugging AC. So I guess
the h2200_ac_plug_isr() is still called.
But I think this might be evil:

        /* Force the data to be re-read. */
        battery_status.update_time = 0;

        /* Wait a few seconds to let the DS2760 update its current
reading. */ queue_delayed_work(monitor_q, &monitor_work, HZ * 7);

It will load the queue with useless stuff.

What other side effects might th power plugging have? Is it only this interrupt
and the fuel gauge seeing a different current direction? I guess I can
cure symptoms by fiddling around with that h2200_ac_plug_isr() but I
would prefer to solve the problem.
Any hints? I have seen the problem also without that hub dynamo torture but very
rarely.

Greetings
Andreas Kemnade
Received on Sun Dec 30 2007 - 14:04:48 EST