public inbox for linux-arm-kernel@lists.infradead.org
 help / color / mirror / Atom feed
From: fweisbec@gmail.com (Frederic Weisbecker)
To: linux-arm-kernel@lists.infradead.org
Subject: Question about ARM function graph tracing
Date: Wed, 16 Feb 2011 05:39:43 +0100	[thread overview]
Message-ID: <20110216043941.GA21346@nowhere> (raw)
In-Reply-To: <1297821945.23343.136.camel@gandalf.stny.rr.com>

On Tue, Feb 15, 2011 at 09:05:45PM -0500, Steven Rostedt wrote:
> On Tue, 2011-02-15 at 17:39 -0800, Tim Bird wrote:
> I've had to debug things that broke in the function graph tracer like
> this before. Is dynamic ftrace available in arm?
> 
> Hmm, from arch/arm/Kconfig:
> 
> 	select HAVE_DYNAMIC_FTRACE if (!XIP_KERNEL)
> 
> 
> I guess so. If you are running something other than this XIP_KERNEL, I
> would suggest using the dynamic functions to binary search the issue.
> 
> cat available_filter_functions > /tmp/funcs
> 
> Then copy half the funcs to /tmp/funcs1 and the other half
> to /tmp/funcs2
> 
> cat /tmp/funcs1 > set_ftrace_filter
> 
> This may take a while, as setting the filter one by one is slow.
> 
> then run the function graph tracer. If it locks up, try it again
> with /tmp/funcs2. If it locks up, you may have two issues or it could be
> something wrong with the function graph tracer itself. But I'm betting
> one will work and one wont.
> 
> Then just keep chopping these funcs files until you get down to a set of
> functions that are causing issues. Then it should point out the problem.
> 
> This is the procedure I used to find out why a kvm guest was locking up
> with function graph tracer.
> 
> See commit:
> 
> 258af47479980d8238a04568b94a4e55aa1cb537
> 
> Something tells me you may have a similar issue ;)


That's a very nice idea to debug ftrace recursions!

As a last resort, in case the binary search doesn't work, you can try:

1) do a local_irq_save() in asm (to avoid possible recursions with common funcs)
2) test recursion, if so, printk_once() the traced func
3) exit recursion
3) local_irq_restore() in asm

      reply	other threads:[~2011-02-16  4:39 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-02-03 23:42 Question about ARM function graph tracing Tim Bird
2011-02-04  4:07 ` Rabin Vincent
2011-02-04 20:29   ` Tim Bird
2011-02-16  1:39     ` Tim Bird
2011-02-16  2:05       ` Steven Rostedt
2011-02-16  4:39         ` Frederic Weisbecker [this message]

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=20110216043941.GA21346@nowhere \
    --to=fweisbec@gmail.com \
    --cc=linux-arm-kernel@lists.infradead.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox