On Wed, 2012-04-18 at 12:43 -0400, Steven Rostedt wrote:
> On Wed, 2012-04-18 at 11:30 -0400, Dave Jones wrote:
> > On Wed, Apr 18, 2012 at 10:56:05AM -0400, Steven Rostedt wrote:
> > > On Wed, 2012-04-18 at 09:58 -0400, Dave Jones wrote:
> > >
> > > > > OK, function tracing is broken on your box. Could you send me your full
> > > > > config and I'll take a look at it tomorrow. Bed time for me now ;-)
> > > >
> > > > http://fpaste.org/b8PN/raw/
> > >
> > > Hmm, I had no problems with this config. But that wasn't on my Xeon. I'm
> > > compiling for that box now.
> > >
> > > Could you also paste your full dmesg.
> >
> > Oh, I guess it's this..
> >
> > "PEBS disabled due to CPU errata."
> >
> > Crap. Never even noticed that before.
>
> Yeah, but that shouldn't stop the function tracer from working.
>
Hi Dave,
Have we ever figured out why this wasn't working? Did it start working
on newer kernels?
-- Steve
On Wed, May 23, 2012 at 10:37:08AM -0400, Steven Rostedt wrote:
> > > Oh, I guess it's this..
> > >
> > > "PEBS disabled due to CPU errata."
> > >
> > > Crap. Never even noticed that before.
> >
> > Yeah, but that shouldn't stop the function tracer from working.
>
> Hi Dave,
>
> Have we ever figured out why this wasn't working? Did it start working
> on newer kernels?
I think that was where we left it.
But! Linus' current tree seems to have working tracing on that box.
Don't break it again! ;-)
Dave
On Wed, 2012-05-23 at 11:05 -0400, Dave Jones wrote:
> But! Linus' current tree seems to have working tracing on that box.
> Don't break it again! ;-)
You could install ktest and run a reverse bisect to find out what fixed
it ;-)
-- Steve