Prev: Adding LogicPD OMAP 3530 LV SOM AND OMAP 35X
Next: [PATCH] ima: appraise measurement required
From: Joe Perches on 2 Jun 2010 19:40
On Thu, 2010-06-03 at 01:23 +0200, Sven Eckelmann wrote:
> You've also changed the output. So you may partly broke batctl too (have to
> check that first).

There was no change in output.
MAC_FMT output is the same as %pM

cheers, Joe

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Joe Perches on 2 Jun 2010 19:50
On Thu, 2010-06-03 at 01:33 +0200, Sven Eckelmann wrote:
> Joe Perches wrote:
> > On Thu, 2010-06-03 at 01:23 +0200, Sven Eckelmann wrote:
> > > You've also changed the output. So you may partly broke batctl too (have
> > > to check that first).
> > There was no change in output.
> > MAC_FMT output is the same as %pM
> I meant the other patch :)

batctl is parsing dmesg or equivalent? ouch.

If so, may I suggest you consider using something other than
a message logging parser for batctl?

I glanced at the source and don't see any such use.
It looks like it's only /sys, /proc and debugfs,
but I only spent a few seconds at it.

http://www.open-mesh.net/changeset/1682/trunk?old_path=%2F&format=zip

cheers, Joe

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Joe Perches on 2 Jun 2010 20:30
On Thu, 2010-06-03 at 01:56 +0200, Sven Eckelmann wrote:
> see sys.c -> log_print
> and read_file -> read_file

Yuck.

The patch changes the prefix from "batman-adv:" to "batman_adv: "
so yes, it would break as-is.

I think the concept is broken though, I believe dmesg output
specifically is not guaranteed to remain stable, and batman should
use some other, perhaps private, logger based on ethtool events.

cheers, Joe

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Greg KH on 3 Jun 2010 11:10
On Thu, Jun 03, 2010 at 05:15:53PM +0800, Marek Lindner wrote:
> On Thursday 03 June 2010 08:20:01 Joe Perches wrote:
> > The patch changes the prefix from "batman-adv:" to "batman_adv: "
> > so yes, it would break as-is.
> >
> > I think the concept is broken though, I believe dmesg output
> > specifically is not guaranteed to remain stable, and batman should
> > use some other, perhaps private, logger based on ethtool events.
>
> I think changing a dash to an underscore is not such a big deal (I did not
> check the rest of the patch yet). But I'm interested to hear more about your
> "private logger" idea because the current solution is far from being perfect.
> As we have to debug the routing protocol every now and then it would be very
> helpful to get direct access to some internal logging facility. In fact, that
> existed before (inside of /proc) but was removed to be more compliant with the
> linux kernel and the existing log facilities.

You can always use debugfs if you like for something like this.

Or tie into the profile/perf subsystem, that might be even easier.

thanks,

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 | 
Pages: 1
Prev: Adding LogicPD OMAP 3530 LV SOM AND OMAP 35X
Next: [PATCH] ima: appraise measurement required