About this list Date view Thread view Subject view Author view Attachment view

From: Enrico Scholz (enrico.scholz_at_sigma-chemnitz.de)
Date: Wed 08 Oct 2003 - 12:40:07 BST


herbert_at_13thfloor.at (Herbert Poetzl) writes:

>> > > Using this new system call, chmod 000 is not needed anymore
>> > > and we can support vservers inside vservers.
>> > why don`t use private namespace ?
>>
>> How does it work ?
>
> you should join a discussion with Enrico and me on IRC
> .. basically it is based on CLONE_NEWNS() and relatives ...

IMO, it is not doable with current technology:

* CLONE_NEWNS has strange behavior[1]; this will be fixed[2] in
  2.4.23 probably

* CLONE_NEWNS + pivot_root are requiring CAP_SYS_ADMIN (which
  is not acceptably for vservers); using a new capability for
  CLONE_NEWNS seems to be possible, but pivot_root(2) needs
  additional logic. Else, when executed in root-namespace,
  pivot_root(2) can do really bad things with your system.

* joining foreign namespaces (e.g. for 'vserver ... enter') is
  not implemented in current kernel; I saw patches but AFAIS,
  they are missing important logic (e.g. no capability-check).
  This functionality will need hierarchical contextes also
  (e.g. parent-vserver can enter namespace of child-vservers,
  but not this of if siblings or parents).

Enrico

Footnotes:
[1] http://www.tu-chemnitz.de/~ensc/nst.c

[2] http://linux.bkbits.net:8080/linux-2.4/diffs/fs/namespace.c_at_1.24?nav=index.html|ChangeSet_at_-1d|cset_at_1.1182|hist/fs/namespace.c


About this list Date view Thread view Subject view Author view Attachment view
[Next/Previous Months] [Main vserver Project Homepage] [Howto Subscribe/Unsubscribe] [Paul Sladen's vserver stuff]
Generated on Wed 08 Oct 2003 - 14:50:28 BST by hypermail 2.1.3