Re: [PATCH][DOCUMENTATION] Minimal controller code for a quick start

["Serge E. Hallyn" <serue@us.ibm.com>]

Quoting Peter Zijlstra (a.p.zijlstra@chello.nl):
> 
> On Thu, 2008-02-07 at 18:37 +0300, Pavel Emelyanov wrote:
> > The Documentation/cgroups.txt file contains the info on how
> > to write some controller for cgroups subsystem, but even with
> > this, one need to write quite a lot of code before developing
> > the core (or copy-n-paste it from some other place).
> > 
> > I propose to put this minimal controller into Documentation
> > directory to let people copy-n-paste a) from a known place and 
> > b) a small piece of code.
> > 
> > Besides, many people learn better reading an example rather
> > than/along with a document.
> 
> 
> While on the subject, could someone document struct cgroup_subsys. In
> particular, I've wondered why we have: cgroup_subsys::can_attach() and

The point of can_attach() is to decide whether an attach should be
permitted.  kernel/ns_cgroup.c defines it.

> not use a return value in cgroup_subsys::attach()?

IIUC, by the point ->attach() is called, the task has already been placed
in the new cgroup, and we're just asking each subsystem to update any
relevant accounting.

-serge