From: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
To: Roberto Bielli <roberto.bielli@axelsw.it>
Cc: "xenomai@xenomai.org" <xenomai@xenomai.org>
Subject: Re: [Xenomai] high latency with imx6q without smp
Date: Tue, 01 Oct 2013 21:10:10 +0200 [thread overview]
Message-ID: <524B1E12.2010909@xenomai.org> (raw)
In-Reply-To: <524AC518.2040901@axelsw.it>
On 10/01/2013 02:50 PM, Roberto Bielli wrote:
> Hi,
>
> excuse me for the previous question i have not seen your email.
> i have tried the kernel with the ipipe patch 3.0.43 on xenomai 2.6.x
> branch without smp and has very high worst latency:
> RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat
> best|--lat worst
> RTD| 5.999| 11.333| 25.333| 0| 0| 5.666| 328.999
> RTD| 10.999| 11.333| 19.666| 0| 0| 5.666| 328.999
> RTD| 5.999| 10.999| 17.999| 0| 0| 5.666| 328.999
> RTD| 10.999| 11.333| 17.999| 0| 0| 5.666| 328.999
> RTD| 5.999| 11.333| 26.999| 0| 0| 5.666| 328.999
>
> This behaviour isn't in smp version with 2 processors.
It may be here in SMP but you do not see it. You should run the latency
test under load to try and trigger the worst case latency.
>
> So i read your email and if i understand correctly could be the L2 cache
> spinlock that create large latencies (?328.999).
The L2 cache spinlock section are alreay reduced, the maximum was rather
around 40us anyway.
> I'm not using framebuffer directly.
> You told me that there is a patch that can fix this behaviour in the
> last ipipe patch.
> is there a commit for fixing only this behaviour that i can move on my
> kernel with patch 3.0.43 ?
Please try disabling the framebuffer driver completely, and see if it
changes anything. If it does not change anything, you have to use the
I-pipe tracer.
--
Gilles.
next prev parent reply other threads:[~2013-10-01 19:10 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-10-01 12:24 [Xenomai] high latency with imx6q without smp Roberto Bielli
2013-10-01 12:26 ` Gilles Chanteperdrix
2013-10-01 12:50 ` Roberto Bielli
2013-10-01 19:10 ` Gilles Chanteperdrix [this message]
2013-10-03 8:36 ` Roberto Bielli
2013-10-03 13:27 ` Gilles Chanteperdrix
2013-10-03 13:33 ` Drasko DRASKOVIC
2013-10-03 13:51 ` Roberto Bielli
2013-10-03 14:07 ` 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=524B1E12.2010909@xenomai.org \
--to=gilles.chanteperdrix@xenomai.org \
--cc=roberto.bielli@axelsw.it \
--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.