Re: [PATCH] namei: permit linking with CAP_FOWNER in userns

From: ebiederm@xmission.com (Eric W. Biederman)
To: Seth Forshee <seth.forshee@canonical.com>
Cc: Dirk Steinmetz <public@rsjtdrjgfuzkfg.com>,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	Andy Lutomirski <luto@amacapital.net>,
	Kees Cook <keescook@chromium.org>,
	Serge Hallyn <serge.hallyn@canonical.com>
Subject: Re: [PATCH] namei: permit linking with CAP_FOWNER in userns
Date: Tue, 27 Oct 2015 16:04:25 -0500	[thread overview]
Message-ID: <87bnbkxbty.fsf@x220.int.ebiederm.org> (raw)
In-Reply-To: <20151027143344.GB132460@ubuntu-hedt> (Seth Forshee's message of "Tue, 27 Oct 2015 09:33:44 -0500")

Seth Forshee <seth.forshee@canonical.com> writes:

> On Tue, Oct 20, 2015 at 04:09:19PM +0200, Dirk Steinmetz wrote:
>> Attempting to hardlink to an unsafe file (e.g. a setuid binary) from
>> within an unprivileged user namespace fails, even if CAP_FOWNER is held
>> within the namespace. This may cause various failures, such as a gentoo
>> installation within a lxc container failing to build and install specific
>> packages.
>> 
>> This change permits hardlinking of files owned by mapped uids, if
>> CAP_FOWNER is held for that namespace. Furthermore, it improves consistency
>> by using the existing inode_owner_or_capable(), which is aware of
>> namespaced capabilities as of 23adbe12ef7d3 ("fs,userns: Change
>> inode_capable to capable_wrt_inode_uidgid").
>> 
>> Signed-off-by: Dirk Steinmetz <public@rsjtdrjgfuzkfg.com>
>
> Tested-by: Seth Forshee <seth.forshee@canonical.com>

If multiple groups are hitting this issue for different reasons
I am applying the supplied patch.

> This is hitting us in Ubuntu during some dpkg upgrades in containers.
> When upgrading a file dpkg creates a hard link to the old file to back
> it up before overwriting it. When packages upgrade suid files owned by a
> non-root user the link isn't permitted, and the package upgrade fails.
> This patch fixes our problem.
>
> I did want to point what seems to be an inconsistency in how
> capabilities in user namespaces are handled with respect to inodes. When
> I started looking at this my initial thought was to replace
> capable(CAP_FOWNER) with capable_wrt_inode_uidgid(inode, CAP_FOWNER). On
> the face of it this should be equivalent to what's done here, but it
> turns out that capable_wrt_inode_uidgid requires that the inode's uid
> and gid are both mapped into the namespace whereas
> inode_owner_or_capable only requires the uid be mapped. I'm not sure how
> significant that is, but it seems a bit odd.

It is a bit odd.

inode_owner_or_capable in this context is a gimme, as only being the
owner of the file in question is enough to create a hard link, and root
(in the user namespace) can become that user.

That said I think there have been some legitimate questions about setgid
executables in may_linkat (raised down thread), as well as legitimate
questions about capable_wrt_uidgid.  I will add the additional question
is it sane for us to ignore the acls in capable_wrt_uidgid.

All of this appears to be an area that no one except bad actors cares
about so I expect we can change things without causing regressions, and
on that note I encourage the conversation on the oddness to continue.

Eric