All of lore.kernel.org
 help / color / mirror / Atom feed
From: Benjamin ZORES <benjamin.zores@domain.hid>
To: xenomai-core <xenomai@xenomai.org>
Subject: [Xenomai-core] Timing Issues on x86_32 SMP
Date: Tue, 06 May 2008 10:11:02 +0200	[thread overview]
Message-ID: <48201296.50905@domain.hid> (raw)

Hi,

I'm currently running an x86_32 SMP system and facing some issues with 
periodic tasks.
I'd like to get a bit more information on a few assumptions I've made.

Quick sum-up of my setup:
- adeos-ipipe-2.6.23-i386-1.12-03
- xenomai-2.4.3.patch
- Core 2 Duo x86_32 running in SMP
- Linux 2.6.23.17
- Timer frequency: 1000 Hz
- Tick-less mode activated (CONFIG_NO_HZ)
- Xenomai Periodic Timing enabled (CONFIG_XENO_OPT_TIMING_PERIODIC)
- tasks are kernel-based.

I'm trying to schedule a periodic task each 8ms and I'm running into 
period miss
(loosing from 5 up to 75, which is quite A LOT). The problem however 
doesn't appear
on UP system (or at least when booting the kernel with maxcpus=1 ).

I'd like to have more information regarding timer handling.

First, I create my 8ms periodic task, the following way:

    SRTIME ticks;
    int period = 8;

    /* need to feed with nano-seconds, then convert to CPU ticks */
    ticks = rt_timer_ns2ticks(1000000 * period);
    rt_task_set_periodic (task, TM_NOW, ticks);

Is it the correct way to do ?
It seems the rt_timer_ns2ticks() function is useless, i.e. it seems that 
1 tick = 1 ns,
which sees weird to me.

Also, what does the rt_timer_read() function really return in my context ?
Jiffies ? Nanoseconds ? Clock ticks ?
 From the different headers and documentation pieces I've read, none 
seem to be really
explicit regarding the returned value.

Any help, is really appreciated.

Ben





             reply	other threads:[~2008-05-06  8:11 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-05-06  8:11 Benjamin ZORES [this message]
2008-05-06  9:13 ` [Xenomai-core] Timing Issues on x86_32 SMP Gilles Chanteperdrix
2008-05-06  9:27   ` Benjamin ZORES
2008-05-06  9:31     ` Gilles Chanteperdrix

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=48201296.50905@domain.hid \
    --to=benjamin.zores@domain.hid \
    --cc=xenomai@xenomai.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.