Documenting the requirement of CAP_SETFCAP to map UID 0
Serge E. Hallyn
serge at hallyn.com
Tue Aug 10 23:58:38 UTC 2021
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;
> * 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
> * 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).
> Michael Kerrisk
> Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
> Linux/UNIX System Programming Training: http://man7.org/training/
More information about the Linux-security-module-archive