Re: [RFC] create instrumentation/ directory

[Randy Dunlap <rdunlap@xenotime.net>]

On Wed, 24 Oct 2007 13:31:02 -0400 Mathieu Desnoyers wrote:

> * Randy Dunlap (rdunlap@xenotime.net) wrote:
> > On Wed, 24 Oct 2007 09:19:05 -0400 Jeff Garzik wrote:
> > 
> > > Honestly, I don't care at all about building the code.  If that's what 
> > > you want, great.
> > 
> > Yes, that is wanted.  They bitrot too easily -- not good.
> > 
> > > My objection is more to adding a samples/ directory, which is contra to 
> > > past experience:
> > > 
> > > A new net driver sample should go in drivers/net/ like the existing 
> > > skeleton files I already listed, not samples/net/.  A top-level 
> > 
> > Can't disagree with that.
> > 
> > > "samples/" seems like it exists only duplicate the rest of the tree 
> > > hierarchy.
> > 
> > It seems odd to be building code in Documentation/, but I can live
> > with that.  It was primarily Christoph who was opposed to that.
> > He suggested samples/ and I went along with it just to break the
> > impasse (since no one else was making any comments on it at that
> > time).
> > 
> > Sam, would building code in Documentation/ cause problems for
> > kbuild?
> 
> Since we already have the Instrumentation menu in
> kernel/Kconfig.instrumentation and instrumentation code all over the
> kernel tree:
> 
> arch/*/oprofile/*.c
> kernel/kprobes.c
> arch/*/kernel/kprobes.c
> kernel/marker.c
> kernel/profile.c
> kernel/profile.S
> kernel/lockdep.c
> vm/vmstat.c
> block/blktrace.c
> 
> We could move them to 
> 
> instrumentation/
> arch/*/instrumentation/
> 
> Therefore, we could also move the kprobes and marker samples under
> 
> instrumentation/samples/
> 
> My main concern is that 15 characters long directory name might be
> inelegant (however, it only beats Documentation by 2).

Ack, instrumentation deserves that much (in the new world order).


---
~Randy