Re: ENE KB3924 CIR driver

[Maxim L. <max...@gm...>]

On Sun, 2009-08-09 at 15:47 +0200, Christoph Bartelmus wrote:
> Hi!
> 
> Maxim Levitsky "max...@gm..." wrote:
> [...]
> >> Because LIRC supports gaps up to something around 16 s. You are truncating
> >> the gap at maximum_gap. Your driver should support the maximum supported
> >> by the interface. That user space applications stop working when you
> >> change maximum_gap should have been hint enough that something is wrong.
> > Indeed, you are right.
> >
> > I misunderstood the kernel->userspace interface.
> >
> > First of all I now assume that I can report several gaps one after
> > another, and even several pulses one after another?
> 
> No, you always have to report gaps and pulses by turns.
> The first data reported should be a long gap.
> 
> > Second, I assume you want me to do the gap handling in following way:
> >
> > 1 - first data is recieved, and reported as is.
> > 2 - after a fixed timeout (you say 150000 - 250000) I turn off the
> > sampler, but I continue to report gaps of arbitrarily lengths, till I
> > receive new interrupt from the hardware (and this is done using a timer)
> 
> After let's say 150000 us you turn off the sampler to reduce the interrupt  
> load. At that time the last data reported was the last pulse, you don't  
> report the gap yet because you don't know yet how long it will take. When  
> the next pulse is reported by the hardware you calculate the gap by  
> measuring the time since the last pulse with do_gettimeofday. Now you know  
> the final gap value and can report it to user space.

Ah, got it now, thanks for clear explanation!


Best regards,
	Maxim Levitsky


> 
> Christoph
> 
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
> trial. Simplify your report design, integration and deployment - and focus on 
> what you do best, core application coding. Discover what's new with 
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july