linux-rt-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chirag Jog <chirag@linux.vnet.ibm.com>
To: Sven-Thorsten Dietrich <sven@thebigcorporation.com>
Cc: Sripathi Kodi <sripathik@in.ibm.com>, linux-rt-users@vger.kernel.org
Subject: Re: Fwd: -rt and PowerPC
Date: Mon, 31 Dec 2007 09:56:23 +0530	[thread overview]
Message-ID: <20071231042623.GA8421@linux.vnet.ibm.com> (raw)
In-Reply-To: <1198859012.12074.0.camel@sven.thebigcorporation.com>

Hi,
* Sven-Thorsten Dietrich <sven@thebigcorporation.com> [2007-12-28 08:23:32]:

> On Wed, 2007-12-26 at 09:53 +0530, Chirag Jog wrote:
> > * sven@thebigcorporation.com <sven@thebigcorporation.com> [2007-12-24 13:00:30]:
> > 
> > Hi,
> 
> Sorry for top-posting the last email. Have not learned how to inline on
> the Blackberry.
> 
> Comments below.
> 
> > > Do you have DEBUG and LOCKDEP configured?
> > CONFIG_PREEMPT_DEBUG is enabled
> > but LOCKDEP is not.
> > > 
> > > -----Original Message-----
> > > From:	Sripathi Kodi <sripathik@in.ibm.com>
> > > 
> > >     We are trying to get a ppc64 box booted with the -rt kernel.
> > > 
> > >     Tried the latest 2.6.24-rc5-rt1 kernel.
> > > 
> > >     Everything goes well [ kernel boots up etc] until the services
> > >     start coming up.
> > >     It takes a lot of time at Starting udev.
> > >     After which either services take too long to start or we get
> > >     SEGFAULT.
> 
> I have seen this exact behavior with lockdep enabled on earlier x86-SMP
> Kernels. I saw the issue predominantly on larger SMP machines, but it
> behaved much like a race. This was observed on 2.6.21 Kernels, I will
> try and verify whether .23 has the same behavior.
> 
> We eventually bisected it down to CONFIG_PROVE_LOCKING, but I wasn't
> totally convinced whethere the root cause was in lockdep, or
> elsewhere.  
> 
> > > 
> > >     This is easily reproducible. Also if we get a chance to login;
> > >     simple commands like ls, vi etc take either too long or get
> > >     SEGFAULT.
> 
> > > 
> > >     Booting the kernel with maxcpus=1 or turning off SMP, doesn't
> > >     solve the problem.
> > >
> 
> Hmmm, so have you COMPILED-out SMP, or just booted with nosmp?
> 
> It might help narrowing this down to compile a UP Kernel.
We have tried both options, compiled out SMP and boot with maxcpus=1.
Similar results.
Sripathi also tried with turning off HIGH_RES_TIMERS but didn't seem
to help.
The system took forever to come up.
> 
> Thanks
> 
> Sven
 
<snip>
-- 
Cheers,
Chirag Jog

  reply	other threads:[~2007-12-31  4:28 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-12-24  6:10 Fwd: -rt and PowerPC Sripathi Kodi
     [not found] ` <562556514-1198501255-cardhu_decombobulator_blackberry.rim.net-1487612076-@bxe137.bisx.prod.on.blackberry>
2007-12-26  4:23   ` Chirag Jog
2007-12-28 16:23     ` Sven-Thorsten Dietrich
2007-12-31  4:26       ` Chirag Jog [this message]
2007-12-31  2:57     ` Paul E. McKenney
2007-12-31 15:44       ` Sripathi Kodi
2007-12-31 16:43         ` Paul E. McKenney

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=20071231042623.GA8421@linux.vnet.ibm.com \
    --to=chirag@linux.vnet.ibm.com \
    --cc=linux-rt-users@vger.kernel.org \
    --cc=sripathik@in.ibm.com \
    --cc=sven@thebigcorporation.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).