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

[TNKS]

Shem Multinymous wrote:

>> > There's just one more test that would be useful, if you get a chance:
>> > unload tp_smapi+thinkpad_ec+hdaps, load the vanilla hdaps module (not
>> > the one patched by tp_smapi) and see if this causes the hang:
>> > # while true; do cat /sys/devices/platform/hdaps/position; done
>>
>> my quick check of the old hdaps seemed to indicate that it is stable with
>> respect to this problem.
> 
> With the aforementioned tight polling loop of .../position? Can you do
> a more prologned check please?
> 

So I did a prolonged test on the unpatched hdaps position interface.  I just
ran it for a few minutes.  I know that doesn't sound long, but in my
experience of this problem, that's enough to convince me that hdaps is far
less fragile than tp_smapi.

Right after my multi-minute test of the unpatched hdaps, I polled dump, and
within seconds the EC hung.

I've been using tp_smapi for a while now and never had an issue with my
normal use.  I guess it really showed itself pretty much when I made that
silly watch script, just to monitor the tp_smapi interface.  

It's a little frightening that my EC can crap out like this.  So no one else
has verified this behavior?  It would be really helpful if there were
another system out there that could replicate this problem, then we could
look for some commonality between the two (relative to systems that seem
stable)

- Sukant