[ltp] Re: Re: Re: Re: EC Hanging with latest tp-smapi (0.29)

[TNKS]

Shem Multinymous wrote:

> Whereas hdaps 'position' queries the hardware just once per read, the
> 'dump' attribute makes 24 back-to-back queries. So if both trigger the
> same firmware bug, I expect 'dump' to do it much faster.
> 
> Is 'dump' a lot more problematic than other battery status attributes
> (e.g., current_now)?

Shem, 

You'll probably be slightly relieved to know that I can only make the EC
hang with rapid calls to dump.  I wonder if this is a reason that few
people have found this problem (dump is such a generally unaccessed
interface).  I don't know if it's possible to hang the EC using the other
interfaces or with hdaps; spin-loop polling didn't do it, though.

The "real" monitors I use only use the non-dump interfaces, so I think about
as stable as I've always been. 

If anyone at all is reading this post, can you please close down all your
important programs (prepare your system for a EC hang) and do a:

    while true; do cat /sys/devices/platform/smapi/BAT0/dump ; done

Unless Shem deems this unnecessary, I'm really curious whether this problem
is somehow unique to my system.  I'm particularly interested in testing out
a T42-esque laptop running a current kernel and a tp_smapi module. 
Both "have the problem" and "don't have it" responses are helpful, I think.

- Sukant