ts-7000
[Top] [All Lists]

[ts-7000] Re: high resolution timers and real time

To:
Subject: [ts-7000] Re: high resolution timers and real time
From: "Charles" <>
Date: Tue, 19 Jan 2010 01:37:38 -0000
Hi Peter --

--- In  Peter Gammie <> wrote:
>
> Charles:
> 
> On 18/01/2010, at 4:20 PM, Charles wrote:
> 
> > Any chance you could pass the clockevents patch on to Matthieu to include 
> > in his patchset? HRT would help lots of users, I think, even without the 
> > -rt support.
> 
> No promises, but I could try to do that if anyone else shows an interest. :-)
> 
> Are Matthieu's patches hosted anywhere except in the Yahoo file 

Sure, they are at http://mcrapet.free.fr/
He's up to 2.6.32.3, but doesn't have HRT yet AFAIK.  Search this group archive 
for '10ms' to gauge interest, it's come up many times before :)

area? Can he (or someone) tell me what the status wrt the linux mainline is? 
I'm not a kernel hacker, but I surmise that RT got merged sometime around 
2.6.31, and there seem to be have been a steady stream of ts72xx/ep93xx patches 
going in.

Only some of his patches have gone in, and some will never make it (like the 
force_machine_id patch).  Matt's patchset has never included the preempt-rt 
code from
http://www.kernel.org/pub/linux/kernel/projects/rt/
which is now up to 2.6.31.6-rt19.

> 
> The 2.6.29.6 kernel I built does sub-10ms delays, dare I say even sub-100ns 
> delays using clock_nanosleep, but the RT scheduler (FIFO/highest priority) is 
> not keeping up its part of the bargain: running top glitches the multiplexed 
> display scanning loop. This loop is only twiddling some GPIO lines in the 
> manner of Jim Jackson's ts_arm_util (i.e. reading and writing to the mmap'd 
> GPIO addresses) - I can't make it much tighter. The CPU load due to my 
> program is something like 0.01% all up.

There are some cool tests here:
http://www.kernel.org/pub/linux//kernel/people/clrkwllms/rt-tests

And a good explanation of how to trace latencies in this thread:
http://marc.info/?l=linux-rt-users&m=126142802114499&w=2

You might have better luck *without* the -rt patch, ironically.  That intends 
to have deterministic latencies, but not necessarily the lowest.  Try just the 
normal config-preempt from the vanilla kernel (but with HRT of course).  You 
don't have the PM stuff enabled in Matt's patches do you?  If there's anything 
you don't need (USB, ethernet, serial, etc), take it out.

regards, .......... Charlie

> 
> I'll try to look into it sometime soon.
> 
> cheers
> peter
> 
> -- 
> http://peteg.org/
>




------------------------------------

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/

<Prev in Thread] Current Thread [Next in Thread>
Admin

Disclaimer: Neither Andrew Taylor nor the University of NSW School of Computer and Engineering take any responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the birding-aus mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU