From: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
To: Martin Shepherd <mcs@domain.hid>
Cc: xenomai@xenomai.org
Subject: Re: [Xenomai-help] Running xeno-test
Date: Sun, 12 Apr 2009 11:11:54 +0200 [thread overview]
Message-ID: <49E1B05A.4040705@domain.hid> (raw)
In-Reply-To: <Pine.LNX.4.64.0904111820330.6444@domain.hid>
Martin Shepherd wrote:
> On Sat, 11 Apr 2009, Gilles Chanteperdrix wrote:
>> We have just discovered that FPU handling on x86 had a few bugs, one of
>> them involving CPUs without fxsr (check /proc/cpuinfo).
>
> The AMD Athlon that I am using, does have fxsr, according to
> /proc/cpuinfo. Regardless, now that I have seen lockups during the
> latency tests as well, I don't think that the problem can be related
> to floating point. It looks more like an interrupt problem.
Ok. But the patch mentioned previously also fixes kernels using FPU in
kernel-space, which the Athlon does according to Konfig. So, please try
the FPU patch or select K6 instead of Athlon.
>
> Xeno-test sporadically locks up the computer during both the latency
> tests and the switching tests. It just happens more often during the
> switching test. When the computer is in the locked-up state, the clock
> stops updating, any ssh sessions to the computer hang, and I can't
> either start a new ssh session, or ping the computer from an external
> computer. This persists indefinitely. However all that I have to do to
> unhang the computer, is move the mouse, or hit any keyboard key on the
> offending PC. Then xeno-test resumes, pings get through etc, until the
> next hang. Also, occasionally, moving the mouse or hitting any key
> when xeno-test is running, will cause xeno-test to freeze for a
> moment, and then report a multi-millisecond latency when I resume it
> by moving the mouse.
>
> My guess is that when it hangs, it is waiting for a timer interrupt
> that for some reason never arives, and that moving the mouse or
> hitting a key on the keyboard generates an interrupt that wakes it
> up. Both the keyboard and mouse are PS/2 devices, which presumably go
> through the i8042 driver. The contents of /proc/interrupts, shown
> below, don't indicate anything attempting to share interrupts with
> this.
> CPU0
> 0: 2545784 XT-PIC-XT timer
> 1: 9149 XT-PIC-XT i8042
> 2: 0 XT-PIC-XT cascade
> 3: 3 XT-PIC-XT
> 4: 3 XT-PIC-XT
> 5: 2755 XT-PIC-XT VIA686A
> 7: 1 XT-PIC-XT parport0
> 8: 1 XT-PIC-XT rtc0
> 9: 0 XT-PIC-XT acpi, uhci_hcd:usb1, uhci_hcd:usb2
> 11: 8410900 XT-PIC-XT eth0, radeon@domain.hid
> 12: 70922 XT-PIC-XT i8042
> 14: 847875 XT-PIC-XT pata_via
> 15: 0 XT-PIC-XT pata_via
> NMI: 0 Non-maskable interrupts
> ERR: 0
>
> Any ideas?
Could you tell us the versions of Linux and Xenomai you are using and
post your .config ?
--
Gilles.
next prev parent reply other threads:[~2009-04-12 9:11 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-04-08 23:08 [Xenomai-help] Selecting the TSC Martin Shepherd
2009-04-09 0:11 ` Martin Shepherd
2009-04-09 2:49 ` Steven Seeger
2009-04-10 22:15 ` [Xenomai-help] Xeno-test bug Martin Shepherd
2009-04-11 0:20 ` [Xenomai-help] Running xeno-test Martin Shepherd
2009-04-11 0:24 ` Steven Seeger
2009-04-11 0:35 ` Martin Shepherd
2009-04-11 12:39 ` Gilles Chanteperdrix
2009-04-11 20:00 ` Martin Shepherd
2009-04-11 20:07 ` Gilles Chanteperdrix
2009-04-12 2:02 ` Martin Shepherd
2009-04-12 9:11 ` Gilles Chanteperdrix [this message]
2009-04-12 17:40 ` Gilles Chanteperdrix
2009-04-12 19:10 ` Martin Shepherd
2009-04-12 19:23 ` Gilles Chanteperdrix
2009-04-16 3:20 ` Martin Shepherd
[not found] ` <4F0B1E98-862D-4031-B22E-D7DAEA554EEB@comcast.net>
[not found] ` <Pine.LNX.4.64.0904161059220.24221@domain.hid>
2009-04-16 18:43 ` Steven Seeger
2009-04-16 22:18 ` Martin Shepherd
2009-04-17 2:50 ` Martin Shepherd
[not found] ` <1239872431.5845.25.camel@domain.hid>
[not found] ` <Pine.LNX.4.64.0904161537400.26955@domain.hid>
2009-04-17 8:01 ` Vikesh Rambaran
2009-04-17 12:57 ` Gilles Chanteperdrix
2009-04-17 23:34 ` Martin Shepherd
2009-04-19 5:15 ` Martin Shepherd
2009-04-12 10:31 ` [Xenomai-help] Xeno-test bug Gilles Chanteperdrix
2009-04-12 10:36 ` Gilles Chanteperdrix
2009-04-12 18:52 ` Martin Shepherd
2009-04-12 19:25 ` Gilles Chanteperdrix
2009-04-13 0:42 ` Martin Shepherd
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=49E1B05A.4040705@domain.hid \
--to=gilles.chanteperdrix@xenomai.org \
--cc=mcs@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.