Re: [PATCH 1/2 v3][RFC] trace: profile likely and unlikelyannotations

From: Steven Rostedt
Date: Thu Oct 30 2008 - 11:10:54 EST



On Thu, 30 Oct 2008, Theodore Tso wrote:

> On Thu, Oct 30, 2008 at 03:32:39PM +0100, J?rn Engel wrote:
> > On Wed, 29 October 2008 18:39:55 -0400, Steven Rostedt wrote:
> > >
> > > 69768 61064 87 __switch_to process_64.c 624
> > > 15557 115251 100 __switch_to process_64.c 594
> > > 15555 115227 100 __switch_to process_64.c 590
> >
> > I may be out of school for a while, but that math sure looks odd.
> >
> > > + if (p->correct) {
> > > + percent = p->incorrect * 100;
> > > + percent /= p->correct;
> > percent /= p->correct + p->incorect;
>
> And once you do the above fix, I don't think what is below is
> necessary any more. :-)

Bah! Total brain-fart. Thanks ;-)

>
> > > + /* No need to see huge numbers */
> > > + if (percent > 100)
> > > + percent = 100;
>
> I would also calculate the percent correct rather than the percent
> incorrect, on the general theory that 100% good, 0% bad is easier for
> my little brain to understand, but that's just a minor thing...

I thought about whether or not to show percent correct or precent
incorrect. I chose, incorrect, just because it is easier to spot the
trouble makers. You can say, lets allow 25% incorrect, better than saying
lets have a minimum of 75%. I think the mind can pick out those that go
over a number better than it can see those that are under a number.

Thanks,

-- Steve

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/