Re: [patch] sharpsl_pm refactor

[Pavel Machek <pavel@ucw.cz>]

Hi!

> > > I'm unconvinced as to why collie needs an ifdef in there and looking at
> > > what I think you're leading to, its ugly. Perhaps you could change the 2
> > > to a variable set by the machine instead or something, depending upon
> > > your intention.
> > 
> > Well, I hate the if/else maze -- IMO returns are more readable. Anyway
> > collie needs both count and time checks disabled, AFAICT.
> 
> To me it looks much worse after you changed it as I can understand it at
> the moment and afterwards with the ifdefs in, I can't.

Well, maybe you'll not get the ifdefs after all... They were just
handy in my tree and code with returns (vs. code with if/else maze)
looked better to me.

> Ignoring that issue, why does collie need them disabled? Do they break
> collie somehow or is this just because the sharp driver didn't do
> it?

Sharp driver did not do it, and forcing charge 3 times when charger
tells me that it is done seems a bit cruel. What is worse, if I get
charge-too-fast timeout too long, it will keep charging battery
over-and-over-and-over-and-over. 

> I'd prefer to keep the charging techniques the same across as many of
> the devices as we can and I can't see how this technique causes a
> problem. The charging hardware and the battery is very similar across
> the models (although you wouldn't believe it looking at the charging
> driver).

Okay, you may be right here. I am just trying to be careful -- hot
lithium scares me a bit ;-).
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html