From: Aleksa Sarai <firstname.lastname@example.org>
To: Al Viro <email@example.com>,
Jeff Layton <firstname.lastname@example.org>,
"J. Bruce Fields" <email@example.com>,
Arnd Bergmann <firstname.lastname@example.org>,
David Howells <email@example.com>
Cc: Aleksa Sarai <firstname.lastname@example.org>,
Eric Biederman <email@example.com>,
Andy Lutomirski <firstname.lastname@example.org>, Jann Horn <email@example.com>,
Christian Brauner <firstname.lastname@example.org>,
David Drysdale <email@example.com>,
Tycho Andersen <firstname.lastname@example.org>,
Kees Cook <email@example.com>,
Andrew Morton <firstname.lastname@example.org>,
Alexei Starovoitov <email@example.com>,
Chanho Min <firstname.lastname@example.org>, Oleg Nesterov <email@example.com>,
Aleksa Sarai <firstname.lastname@example.org>,
Linus Torvalds <email@example.com>,
Subject: [PATCH RESEND v5 0/5] namei: vfs flags to restrict path resolution
Date: Thu, 21 Mar 2019 01:37:12 +1100 [thread overview]
Message-ID: <firstname.lastname@example.org> (raw)
Now that the holiday break is over, it's time to re-send this patch
series (with a few additions, due to new information we got from
CVE-2019-5736 -- which this patchset mostly protected against but had
some holes with regards to #!-style scripts).
* In response to CVE-2019-5736 (one of the vectors showed that
open(2)+fexec(3) cannot be used to scope binfmt_script's implicit
open_exec()), AT_* flags have been re-added and are now piped
through to binfmt_script (and other binfmt_* that use open_exec)
but are only supported for execveat(2) for now.
* Remove AT_* flag reservations, as they require more discussion.
* Switch to path_is_under() over __d_path() for breakout checking.
* Make O_XDEV no longer block openat("/tmp", "/", O_XDEV) -- dirfd
is now ignored for absolute paths to match other flags.
* Improve the dirfd_path_init() refactor and move it to a separate
* Remove reference to Linux-capsicum.
* Switch "proclink" name to "magic link".
* Made ".." resolution with AT_THIS_ROOT and AT_BENEATH safe(r) with
some semi-aggressive __d_path checking (see patch 3).
* Disallowed "proclinks" with AT_THIS_ROOT and AT_BENEATH, in the
hopes they can be re-enabled once safe.
* Removed the selftests as they will be reimplemented as xfstests.
* Removed stat(2) support, since you can already get it through
O_PATH and fstatat(2).
The need for some sort of control over VFS's path resolution (to avoid
malicious paths resulting in inadvertent breakouts) has been a very
long-standing desire of many userspace applications. This patchset is a
revival of Al Viro's old AT_NO_JUMPS[1,2] patchset (which was a variant
of David Drysdale's O_BENEATH patchset which was a spin-off of the
Capsicum project) with a few additions and changes made based on the
previous discussion within  as well as others I felt were useful.
In line with the conclusions of the original discussion of AT_NO_JUMPS,
the flag has been split up into separate flags:
* O_XDEV blocks all mountpoint crossings (upwards, downwards, or
through absolute links). Absolute pathnames alone in openat(2) do
not trigger this.
* O_NOMAGICLINKS blocks resolution through /proc/$pid/fd-style links.
This is done by blocking the usage of nd_jump_link() during
resolution in a filesystem. The term "magic links" is used to match
with the only reference to these links in Documentation/, but I'm
happy to change the name.
It should be noted that this is different to the scope of O_NOFOLLOW
in that it applies to all path components. However, you can do
open(O_NOFOLLOW|O_NOMAGICLINKS|O_PATH) on a "magic link" and it will
*not* fail (assuming that no parent component was a "magic link"),
and you will have an fd for the "magic link".
* O_BENEATH disallows escapes to outside the starting dirfd's tree,
using techniques such as ".." or absolute links. Absolute paths in
openat(2) are also disallowed. Conceptually this flag is to ensure
you "stay below" a certain point in the filesystem tree -- but this
requires some additional to protect against various races that would
allow escape using ".." (see patch 4 for more detail).
Currently O_BENEATH implies O_NOMAGICLINKS, because it can trivially
beam you around the filesystem (breaking the protection). In future,
there might be similar safety checks as in patch 4, but that
requires more discussion.
In addition, two new flags were added that expand on the above ideas:
* O_NOSYMLINKS does what it says on the tin. No symlink resolution is
allowed at all, including "magic links". Just as with O_NOMAGICLINKS
this can still be used with (O_PATH|O_NOFOLLOW) to open an fd for
the symlink as long as no parent path had a symlink component.
* O_THISROOT is an extension of O_BENEATH that, rather than blocking
attempts to move past the root, forces all such movements to be
scoped to the starting point. This provides chroot(2)-like
protection but without the cost of a chroot(2) for each filesystem
operation, as well as being safe against race attacks that chroot(2)
If a race is detected (as with O_BENEATH) then an error is
generated, and similar to O_BENEATH it is not permitted to cross
"magic links" with O_THISROOT.
The primary need for this is from container runtimes, which
currently need to do symlink scoping in userspace when opening
paths in a potentially malicious container. There is a long list of
CVEs that could have bene mitigated by having O_THISROOT (such as
CVE-2017-1002101, CVE-2017-1002102, CVE-2018-15664, and
CVE-2019-5736, just to name a few).
In addition, a mirror set of AT_* flags have been added (though
currently these are only supported for execveat(2) -- and not for any
other syscall). The need for these is explained in the final patch in
the series (it's motivated by CVE-2019-5736).
Cc: Al Viro <email@example.com>
Cc: Eric Biederman <firstname.lastname@example.org>
Cc: Andy Lutomirski <email@example.com>
Cc: David Howells <firstname.lastname@example.org>
Cc: Jann Horn <email@example.com>
Cc: Christian Brauner <firstname.lastname@example.org>
Cc: David Drysdale <email@example.com>
Cc: Tycho Andersen <firstname.lastname@example.org>
Cc: Kees Cook <email@example.com>
Aleksa Sarai (5):
namei: split out nd->dfd handling to dirfd_path_init
namei: O_BENEATH-style path resolution flags
namei: O_THISROOT: chroot-like path resolution
namei: aggressively check for nd->root escape on ".." resolution
binfmt_*: scope path resolution of interpreters
fs/binfmt_elf.c | 2 +-
fs/binfmt_elf_fdpic.c | 2 +-
fs/binfmt_em86.c | 4 +-
fs/binfmt_misc.c | 2 +-
fs/binfmt_script.c | 2 +-
fs/exec.c | 26 +++-
fs/fcntl.c | 2 +-
fs/namei.c | 205 ++++++++++++++++++++++---------
fs/open.c | 13 +-
include/linux/binfmts.h | 1 +
include/linux/fcntl.h | 3 +-
include/linux/fs.h | 9 +-
include/linux/namei.h | 8 ++
include/uapi/asm-generic/fcntl.h | 17 +++
include/uapi/linux/fcntl.h | 6 +
15 files changed, 228 insertions(+), 74 deletions(-)
next reply other threads:[~2019-03-20 14:38 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-20 14:37 Aleksa Sarai [this message]
2019-03-20 14:37 ` [PATCH RESEND v5 1/5] namei: split out nd->dfd handling to dirfd_path_init Aleksa Sarai
2019-03-20 14:37 ` [PATCH RESEND v5 2/5] namei: O_BENEATH-style path resolution flags Aleksa Sarai
2019-03-20 14:37 ` [PATCH RESEND v5 3/5] namei: O_THISROOT: chroot-like path resolution Aleksa Sarai
2019-03-20 14:37 ` [PATCH RESEND v5 4/5] namei: aggressively check for nd->root escape on ".." resolution Aleksa Sarai
2019-03-20 14:37 ` [PATCH RESEND v5 5/5] binfmt_*: scope path resolution of interpreters Aleksa Sarai
2019-03-21 17:06 ` [PATCH RESEND v5 0/5] namei: vfs flags to restrict path resolution Andy Lutomirski
2019-03-25 13:04 ` Aleksa Sarai
2019-04-23 20:13 ` Kees Cook
2019-04-23 20:24 ` Christian Brauner
2019-04-24 15:38 ` Aleksa Sarai
2019-04-25 13:22 ` Adam Borowski
2019-04-25 19:45 ` Aleksa Sarai
-- strict thread matches above, loose matches on Subject: below --
2019-03-06 19:12 Aleksa Sarai
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:
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
* 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).