Ingo Molnar writes:
> On Tue, 5 Sep 2000, Richard Gooch wrote:
> > everyone has to start from nothing. But if the learning/development
> > curve is too steep, or the process is too frustrating, you are going
> > to lose a proportion of the potential gurus. You can't push people in
> > a direction they don't want to go. [...]
> nobody wants to force anything on anyone - you are free to apply KDB
> and the other debugging patches. But we can certainly help the usage
> of the right kind of debugging methodology.
You might think you're encouraging people to think instead of
monkey-tap, but consider the people you're putting off. It's the other
side of that coin.
> > [...] From what I've seen, most of the
> > active developers are chronically overworked. Many projects compete
> > for your attention. Why hit your head against a purposely erected
> > brick wall?
> active developers certainly have the skills to apply debugging
> facilities they need.
It's not just the skills. It's the time and energy. It's more effort,
because you have to download it (and hope it's current and kernel
drift hasn't rooted it), and you have to maintain more kernel trees
(otherwise you send out patches with IKD mistakenly included (and
don't tell me it doesn't happen, even by the most esteemed
Skilled people are usually busy people (they have their fingers in too
many pies). While they are better able to download and apply IKD, by
the same token they are less inclined to, because they are busy, and
besides, they know that they *shouldn't have to*.
> The reference kernel should be IMO 'untainted' though. Believe me,
> during the 2.3.2x pagecache rewrite my kernel was hacked with ad-hoc
> debugging code beyond recognition - eg. automatic checksumming of
> every physical page in the system to detect stray DMA related memory
> corruption. No rocket science, but ugly enough to 'taint' the
> kernel proper. Would any of the debugging facilities advocated in
> this thread have helped in the bugs we were chasing at that time?
> Nope. Do i want such debugging code to ever show up in the mainsteam
> kernel? Hell no.
Would you classify IKD as a pile of warts you wouldn't want to see in
Surely there must be some useful features that can be included in the
kernel without uglyfing it or slowing it down (configed out)? Leaving
aside the social engineering attempts, of course :-)
> you dont want to carry a whole car-construction factory along with
> you when you are rally-racing. You probably want to carry a spare
> tire and a mobile phone - anything else would be a slowdown.
We're already carrying a small city. One extra factory won't
hurt. It's just source bloat, not object or data bloat.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to firstname.lastname@example.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Sep 07 2000 - 21:00:23 EST