From: Seth Forshee <seth.forshee@canonical.com>
To: Dirk Steinmetz <public@rsjtdrjgfuzkfg.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
"Eric W. Biederman" <ebiederm@xmission.com>,
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 09:33:44 -0500 [thread overview]
Message-ID: <20151027143344.GB132460@ubuntu-hedt> (raw)
In-Reply-To: <1445350159-5489-1-git-send-email-public@rsjtdrjgfuzkfg.com>
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>
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.
Seth
> ---
> This is the third time I'm sending the patch, as the first two attempts did
> not provoke a reply. Feel free to point out any issues you see with it --
> including formal requirements, as this is the first patch I'm submitting.
> I'd really appreciate your time.
>
> Maybe a bit of rationale behind it would be helpful as well: some linux
> distributions, especially gentoo in which I discovered the behaviour,
> rely on root being able to hardlink arbitrary files. In the case of gentoo,
> this happens when building and installing 'man': the built binary has the
> suid-flag set and is owned by a user 'man'. The installation script
> (running as root) then attempts to insert a hardlink towards that binary.
>
> Thanks to user namespaces, a regular user can use subuids to create a user
> namespace, and acquire root-like capabilities within said namespace. It is
> then possible to install and use arbitrary linux distributions within such
> namespaces. When installing gentoo in that manner, building and installing
> 'man' fails, as may_linkat checks the capabilities in the init namespace,
> where the installation process is owned by a regular user.
>
> In my opinion may_linkat should permit linking in this case, as the file to
> link to is owned by one of the regular user's mapped subids. Note that, in
> the scenario described above, it is already possible to create the hardlink
> through other means (the following listing is from an unprivileged user
> namespace):
> > # cat /proc/$$/status | grep CapEff
> > CapEff: 0000003cfdfeffff
> > # ls -l
> > total 0
> > -rwSr--r-- 1 nobody nobody 0 Oct 20 15:40 file
> > # ln file link
> > ln: failed to create hard link 'link' => 'file': Operation not permitted
> > # su nobody -s /bin/bash -c "ln file link"
> > # ls -l
> > total 0
> > -rwSr--r-- 2 nobody nobody 0 Oct 20 15:40 file
> > -rwSr--r-- 2 nobody nobody 0 Oct 20 15:40 link
> As you can see, the process has CAP_FOWNER in the namespace, but cannot
> hardlink the file owned by 'nobody'. It can, however, use su to switch to
> 'nobody' and then create the link. After applying this patch, linking
> works as expected.
>
> Diffstat:
> fs/namei.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/fs/namei.c b/fs/namei.c
> index 33e9495..0d3340b 100644
> --- a/fs/namei.c
> +++ b/fs/namei.c
> @@ -955,26 +955,23 @@ static bool safe_hardlink_source(struct inode *inode)
> * - sysctl_protected_hardlinks enabled
> * - fsuid does not match inode
> * - hardlink source is unsafe (see safe_hardlink_source() above)
> - * - not CAP_FOWNER
> + * - not CAP_FOWNER in a namespace with the inode owner uid mapped
> *
> * Returns 0 if successful, -ve on error.
> */
> static int may_linkat(struct path *link)
> {
> - const struct cred *cred;
> struct inode *inode;
>
> if (!sysctl_protected_hardlinks)
> return 0;
>
> - cred = current_cred();
> inode = link->dentry->d_inode;
>
> /* Source inode owner (or CAP_FOWNER) can hardlink all they like,
> * otherwise, it must be a safe source.
> */
> - if (uid_eq(cred->fsuid, inode->i_uid) || safe_hardlink_source(inode) ||
> - capable(CAP_FOWNER))
> + if (inode_owner_or_capable(inode) || safe_hardlink_source(inode))
> return 0;
>
> audit_log_link_denied("linkat", link);
> --
> 2.1.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
next prev parent reply other threads:[~2015-10-27 14:34 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-10 14:59 [PATCH] namei: permit linking with CAP_FOWNER in userns Dirk Steinmetz
2015-10-20 14:09 ` Dirk Steinmetz
2015-10-27 14:33 ` Seth Forshee [this message]
2015-10-27 18:08 ` Dirk Steinmetz
2015-10-27 20:28 ` Serge Hallyn
2015-10-28 15:07 ` Dirk Steinmetz
2015-10-28 17:33 ` Serge Hallyn
2015-11-02 15:10 ` Dirk Steinmetz
2015-11-02 18:02 ` Serge Hallyn
2015-11-02 19:57 ` Andy Lutomirski
2015-11-03 0:39 ` [RFC] namei: prevent sgid-hardlinks for unmapped gids Dirk Steinmetz
2015-11-03 15:44 ` Serge Hallyn
2015-11-03 18:20 ` Kees Cook
2015-11-03 23:21 ` Dirk Steinmetz
2015-11-03 23:29 ` Kees Cook
2015-11-04 6:58 ` Willy Tarreau
2015-11-04 17:59 ` Andy Lutomirski
2015-11-04 18:15 ` Willy Tarreau
2015-11-04 18:17 ` Andy Lutomirski
2015-11-04 18:28 ` Willy Tarreau
2015-11-06 21:59 ` Kees Cook
2015-11-06 22:30 ` Andy Lutomirski
2015-11-07 0:11 ` Kees Cook
2015-11-07 0:16 ` Kees Cook
2015-11-07 0:48 ` Andy Lutomirski
2015-11-07 5:05 ` Kees Cook
2015-11-08 2:02 ` Theodore Ts'o
2015-11-10 15:08 ` Jan Kara
2015-11-19 20:11 ` Kees Cook
2015-11-19 21:57 ` Andy Lutomirski
2015-11-19 22:02 ` Dave Chinner
2015-11-20 0:11 ` Kees Cook
2015-11-04 14:46 ` Serge E. Hallyn
2015-10-27 21:04 ` [PATCH] namei: permit linking with CAP_FOWNER in userns Eric W. Biederman
2015-11-03 17:51 ` Kees Cook
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20151027143344.GB132460@ubuntu-hedt \
--to=seth.forshee@canonical.com \
--cc=ebiederm@xmission.com \
--cc=keescook@chromium.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=public@rsjtdrjgfuzkfg.com \
--cc=serge.hallyn@canonical.com \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).