AW: duration of udelay differs with activated realtime-preempt patch?

["Schultheiss, Christoph" <Christoph.Schultheiss@eurocopter.com>]

> > > after measuring the duration of the function udelay (with
oscilloscope
> > > at parallel port), I figured out that udelay (5usec) with
activated
> > > realtime- preempt patch lasts a little bit longer. Without the
patch the
> > > time is exact.
> > > All kernel debug switches are turned off at compile time.
> > > Can anyone suggest why this happens?
> > 
> > Well, the -rt patch, has changed the udelay function.  BTW, you are 
> > using the constant udelay, right?  Maybe an example of the code you 
> > used to test might be useful.
> > 
> > Ingo or John?
>
> yes, i found this problem too, a week ago or so. It's due to the GTOD 
> changes to i386's __delay() function. Does it still occur with -rt15 
> [which has the -B11 GTOD patchset] ?

I've done the first measurements only with rt5 and rt13. 
Now with rt15 it looks fine (nice 5usec!)

fine patch, thanks!

christoph
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/