Documenting the requirement of CAP_SETFCAP to map UID 0

[Serge E. Hallyn]

On Sun, Aug 08, 2021 at 11:09:30AM +0200, Michael Kerrisk (man-pages) wrote:
> Hello Serge,
> 
> Your commit:
> 
> [[
> commit db2e718a47984b9d71ed890eb2ea36ecf150de18
> Author: Serge E. Hallyn <serge at hallyn.com>
> Date:   Tue Apr 20 08:43:34 2021 -0500
> 
>     capabilities: require CAP_SETFCAP to map uid 0
> ]]
> 
> added a new requirement when updating a UID map a user namespace
> with a value of '0 0 *'.
> 
> Kir sent a patch to briefly document this change, but I think much more
> should be written. I've attempted to do so. Could you tell me whether the
> following text (to be added in user_namespaces(7)) is accurate please:

Sorry for the delay - this did not go into my main mailbox.

The text looks good.  Thanks!

> [[
>       In  order  for  a  process  to  write  to  the /proc/[pid]/uid_map
>        (/proc/[pid]/gid_map) file, all of the following requirements must
>        be met:
> 
>        [...]
> 
>        4. If  updating  /proc/[pid]/uid_map to create a mapping that maps
>           UID 0 in the parent namespace, then one of the  following  must
>           be true:
> 
>           *  if  writing process is in the parent user namespace, then it
>              must have the CAP_SETFCAP capability in that user namespace;
>              or
> 
>           *  if  the writing process is in the child user namespace, then
>              the process that created the user namespace  must  have  had
>              the CAP_SETFCAP capability when the namespace was created.
> 
>           This rule has been in place since Linux 5.12.  It eliminates an
>           earlier security bug whereby a UID 0  process  that  lacks  the
>           CAP_SETFCAP capability, which is needed to create a binary with
>           namespaced file capabilities (as described in capabilities(7)),
>           could  nevertheless  create  such  a  binary,  by the following
>           steps:
> 
>           *  Create a new user namespace with the identity mapping (i.e.,
>              UID  0 in the new user namespace maps to UID 0 in the parent
>              namespace), so that UID 0 in both namespaces  is  equivalent
>              to the same root user ID.
> 
>           *  Since  the  child process has the CAP_SETFCAP capability, it
>              could create a binary with namespaced file capabilities that
>              would  then  be  effective in the parent user namespace (be‐
>              cause the root user IDs are the same in the two namespaces).
> 
>        [...]
> ]]
> 
> Thanks,
> 
> Michael
> 
> -- 
> Michael Kerrisk
> Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
> Linux/UNIX System Programming Training: http://man7.org/training/