ltib
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Ltib] Kernel Configuration - Timing HZ on ARM - NXP LPC3250


From: Kevin Wells
Subject: RE: [Ltib] Kernel Configuration - Timing HZ on ARM - NXP LPC3250
Date: Fri, 10 Sep 2010 00:08:49 +0200

Hi Tim,

I've always wondered how tick-less systems handle drivers that use cpu_relax() 
with jiffie counters.
I use that approach in several drivers. It should be ok, but I'm not sure 
myself.

I used to have support for changing the tick rate in the kernel.
I took this out for mainline during review and in turn removed it from the 
older releases.
The new drivers now have no idea what the tick rate is going to be as the 
configuration info comes from the kernel.
(I don't even know anymore what the default rate is!)

Have you tried turn off the tickles support in the kernel to see if it helps?
What types of issues are you having? Are they specific drivers or interfaces?

Kevin

> -----Original Message-----
> From: address@hidden [mailto:ltib-
> address@hidden On Behalf Of Tim Nelson
> Sent: Thursday, September 09, 2010 12:15 PM
> To: ltib
> Subject: [Ltib] Kernel Configuration - Timing HZ on ARM - NXP LPC3250
> 
> Greetings-
> 
> While working with some communications sensitive applications on the NXP
> LPC3250 platform, I've started wondering if kernel frequency may be the
> culprit to the problems experienced.
> 
> In the kernel config, the CONFIG_HZ is set to 250. I've changed this on
> x86 systems for various reasons, including running a tickless kernel. Are
> these options available with the LPC3250 SoC? In some of my reading I've
> found changing the CONFIG_HZ value on the ARM platform may not work as
> expected, if at all.
> 
> Thoughts?
> 
> Thank you!
> 
> --Tim
> 
> _______________________________________________
> LTIB home page: http://ltib.org
> 
> Ltib mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/ltib

reply via email to

[Prev in Thread] Current Thread [Next in Thread]