[RFC][v8][PATCH 3/10]: Make pid_max a pid_ns property

[Serge E. Hallyn]

Quoting Pavel Emelyanov (xemul at openvz.org):
> Sukadev Bhattiprolu wrote:
> > 
> > From: Serge Hallyn <serue at us.ibm.com>
> > Subject: [RFC][v8][PATCH 3/10]: Make pid_max a pid_ns property
> > 
> > Remove the pid_max global, and make it a property of the
> > pid_namespace.  When a pid_ns is created, it inherits
> > the parent's pid_ns.
> > 
> > Fixing up sysctl (trivial akin to ipc version, but
> > potentially tedious to get right for all CONFIG*
> > combinations) is left for later.
> > 
> > Changelog[v2]:
> > 	- Port to newer kernel
> > 	- Make pid_max a local variable in alloc_pidmap() to simplify code/patch
> > 
> > Signed-off-by: Serge Hallyn <serue at us.ibm.com>
> > Signed-off-by: Sukadev Bhattiprolu <sukadev at us.ibm.com>
> 
> Not that I'm about to slow down or block the process, but...

This patch isn't a core part of the clone_with_pid functionality,
just something Eric has asked for.  So I don't object to dropping
it.  But I disagree with Alexey's claim that this isn't a namespace
property.  It should be.

> frankly I don't see the reason for doing so. Why should we?
> Especially taking into account, that we essentially cannot
> change thin in the namespace level 3 and deeper?

What do you mean by that?  With this patchset we're not, it's
true, but we trivially can - even now, userspace can simply not
give the container CAP_SYS_ADMIN or write access to the sysctl
so they can't do any more CLONE_NEWPIDS or change the sysctl.

-serge