Re: [PATCH] Combine instrumentation menus in kernel/Kconfig.instrumentation

From: Mathieu Desnoyers
Date: Wed Sep 19 2007 - 07:20:48 EST


* Valdis.Kletnieks@xxxxxx (Valdis.Kletnieks@xxxxxx) wrote:
> On Tue, 18 Sep 2007 21:06:28 EDT, Mathieu Desnoyers said:
>
> > * Valdis.Kletnieks@xxxxxx (Valdis.Kletnieks@xxxxxx) wrote:
> > > OK, I'll bite - given the mention of 'debugging' there, do we want to go for
> > > broke and *also* suck in the 'Kernel Hacking' menu as well?
> >
> > Instrumentation primarity aims at debugging user-space applications by
> > giving the ability to extract information across execution layers, hence
> > being a feature useful to users, not only kernel hackers.
>
> Ahh.. Where I come from, "Instrumentation" includes *all* the code that doesn't
> actually *do* the work, but exists solely so you can tell what the code is
> doing - so almost everything on the i386/x86_64 "Kernel Hacking" menu counts
> as "instrumentation" in my book.
>
> Now if we're talking about the features provided by one specific *instance*
> of instrumentation code, I can buy that too - but I think that help text
> needs to be clear on what is included under "debugging"
>
> (I'd settle for "If you're trying to debug the kernel itself, go see the
> Kernel Hacking menu" or similar.. ;)
>

Ok, let's add that.

> > to many, not only kernel developers. Please have a look at the
> > papers (especially the OLS2007 paper) linked on http://ltt.polymtl.ca as a
> > starting point if you are intereted in the question.
>
> Ahh, LTT. *that* I recognize. Yeah, I count that as *one flavor* of
> instrumentation. :)
>

Systemtap, kprobes and oprofile would also fall into the same category.

Mathieu


--
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68

Attachment: signature.asc
Description: Digital signature