You can also look into the signal() and alarm() C-functions
-S
--- In "Charles" <> wrote:
>
> Hi --
>
> --- In "ignac.kolenko" <ignac.kolenko@> wrote:
> >
> > i'm new to the system, and for a project i'm working on, i'm in need of a
> > task to wake up every 500usec to 1msec to monitor digital I/O lines on the
> > TS7500 unit. There may be a need to read/write data on I2C devices as well
> > at this time. essentially a 1kHz to 2kHz sampling interval would best sum
> > up the timing requirements.
> >
> > what is the best way to manage this on this unit. doing some tests i found
> > that having a task usleep(1000) to generate a 1msec wakeup didn't keep up.
> > getting closer to 1/4 second (usleep(250000)) seemed to create accurate 1/4
> > second timing.
> >
> > is there a better way of doing this on the TS7500?? is there a better task
> > delay function than usleep()??
>
> You are running into the granularity of the timer. Search the lists for info
> on 'HRT' high resolution timer option when building your kernel and for
> '10ms'. It comes up quite often here.
>
> The thread here is a good start:
> http://tech.groups.yahoo.com/group/ts-7000/message/16963
> and a patch/summary here:
> http://tech.groups.yahoo.com/group/ts-7000/message/17024
>
> Note: you have to check that the Cavium chip has HRT support in the kernel.
> I believe it does, but if not, you have to add it, using the ep93xx patches
> from the thread above as a guide; or check the LAKML.
>
> If you get it to work, please post a summary for others.
>
> regards, ........ Charlie
>
> >
> > can a user task configure a timer that fires off an interrupt to task at
> > this rate??
> >
> > running a task that does nothing but I/O work generates obviously 99% cpu
> > utilization, but the console prompt, etc., appear to be quite responsive.
> > busybox "top" command reports overall CPU load in this situation as 34%
> > which makes me think that there's plenty of horsepower to do a 2kHz to 1kHz
> > sampling rate.
> >
> >
> > or will I need to implement a kernel module to manage this, as I would
> > assume kernel timing would be more accurate than userland delay timing??
> >
> > any thoughts would be valuable.
> >
> > -ig
> >
>
------------------------------------
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/ts-7000/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/ts-7000/join
(Yahoo! ID required)
<*> To change settings via email:
<*> To unsubscribe from this group, send an email to:
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|