From: Willy Tarreau <w@1wt.eu>
To: Andy Lutomirski <luto@amacapital.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
"security@kernel.org" <security@kernel.org>,
Ingo Molnar <mingo@kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Oleg Nesterov <oleg@redhat.com>,
Al Viro <viro@zeniv.linux.org.uk>,
Linux FS Devel <linux-fsdevel@vger.kernel.org>,
Brad Spengler <spender@grsecurity.net>
Subject: Re: [PATCH v2] vfs: Tighten up linkat(..., AT_EMPTY_PATH)
Date: Thu, 22 Aug 2013 21:39:10 +0200 [thread overview]
Message-ID: <20130822193910.GJ31117@1wt.eu> (raw)
In-Reply-To: <CALCETrWYeNXhEDY7S9XLD6L0wOZu8iPVO6LN3A_797hLHfBzAw@mail.gmail.com>
On Thu, Aug 22, 2013 at 12:05:50PM -0700, Andy Lutomirski wrote:
> On Thu, Aug 22, 2013 at 11:53 AM, Willy Tarreau <w@1wt.eu> wrote:
> > On Thu, Aug 22, 2013 at 11:48:10AM -0700, Linus Torvalds wrote:
> >> And I'm wondering if we shouldn't actually do that at "path_init"
> >> time. Right now the code says:
> >>
> >> /* Caller must check execute permissions on the
> >> starting path component */
> >> struct fd f = fdget_raw(dfd);
> >>
> >> and then uses the struct file mindlessly.
> >>
> >> I'm wondering if we should just do some validation in that place, and say:
> >>
> >> - for directories, we require exec permissions here
> >> - for everything else, we require that f->f_cred == current->cred check.
>
> Does this work for the procfs case? As far as I understand it (which
> isn't saying much), it goes through the symlink-following path.
Indeed I checked yesterday that it seems to use proc_pid_follow_link() for
fd/, cwd, root and exe, which means the same tests are used everywhere.
> >> I dunno. But that I_LINKABLE thing just bothers me. It screams "I'm
> >> hacky" to me.
> >
> > I agreed, simply because the condition here is different from the one in /proc.
> >
> > I have read some code last evening to try to understand how /proc/pid/fd
> > entries were granted access to various processes, because I would love to
> > see the same condition being used in both places. Unfortunately, it's beyond
> > my skills, and I stopped after my random attempts gave me some panics.
>
> What if we added another field to struct nameidata that's indicates
> what restrictions need to be enforced when following magical symlinks
> and then enforcing them when nd_jump_link gets used. (There are only
> two of these, both in procfs.)
I tried to add a test based on a mount option before this call to
nd_jump_link() when I realized my attempt was a total disaster because
I'm a noob. But what I think would be nice (at least as an opt-in) would
be :
- processes which don't share the same root should not be allowed to
access files through /proc/pid/{root,cwd,exe,fd/*}
- keep the current restrictions as well of course
- the exact same restrictions should apply to AT_EMPTY_PATH
I might be totally wrong, but as a user that's what I find natural and
what I tend to expect.
> Then open could check that the original fd was opened for a superset
> of the requested permissions (or that the caller has
> CAP_DAC_OVERRIDE), linkat could check whatever it feels like checking,
> etc.
Do not forget that 2 other syscalls seem to support AT_EMPTH_PATH as
well if that makes a difference.
> This would allow all of these issues to be fixed for real (controlled
> by sysctl, presumably).
If needed for backwards compatibility, possibly, though I doubt that
there are apps that *rely* on the current lack of isolation between
chroots. But at the same time I hate to break existing setups :-)
> --Andy
Willy
prev parent reply other threads:[~2013-08-22 19:39 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-08-21 19:14 [PATCH v2] vfs: Tighten up linkat(..., AT_EMPTY_PATH) Andy Lutomirski
[not found] ` <CA+55aFxi-ps2f2M8xPhfbuQ0pToqupPrDsLi2+GPUK2sqdYfUw@mail.gmail.com>
[not found] ` <CALCETrW7+LcexA6v6RQDKhni_yJAduOmiSDneCpq3v8sPDvwUQ@mail.gmail.com>
2013-08-21 20:16 ` Willy Tarreau
2013-08-22 18:48 ` Linus Torvalds
2013-08-22 18:53 ` Willy Tarreau
2013-08-22 19:05 ` Andy Lutomirski
2013-08-22 19:23 ` Linus Torvalds
2013-08-22 20:10 ` Andy Lutomirski
2013-08-22 20:15 ` Willy Tarreau
2013-08-22 20:22 ` Andy Lutomirski
2013-08-22 20:44 ` Linus Torvalds
2013-08-22 20:48 ` Andy Lutomirski
2013-08-22 20:54 ` Linus Torvalds
2013-08-22 20:58 ` Andy Lutomirski
2013-08-23 1:07 ` Al Viro
2013-08-25 3:37 ` Al Viro
2013-08-25 7:26 ` Andy Lutomirski
2013-08-25 14:23 ` Al Viro
2013-08-25 17:04 ` Andy Lutomirski
2013-08-25 19:57 ` Linus Torvalds
2013-08-25 20:06 ` Al Viro
2013-08-25 20:23 ` Linus Torvalds
2013-08-26 17:37 ` Linus Torvalds
2013-08-26 18:07 ` Linus Torvalds
2013-08-26 18:11 ` Andy Lutomirski
2013-08-27 19:16 ` [RFC PATCH] fs: Add user_file_or_path_at and use it for truncate Andy Lutomirski
2013-08-27 19:32 ` Linus Torvalds
2013-08-27 20:28 ` Andy Lutomirski
2013-08-28 6:16 ` Al Viro
2013-08-28 16:24 ` Linus Torvalds
2013-08-28 19:04 ` Andy Lutomirski
2013-08-28 19:59 ` Al Viro
2013-08-28 21:07 ` Andy Lutomirski
2013-08-27 23:08 ` Al Viro
2013-08-27 23:13 ` Andy Lutomirski
2013-08-24 18:29 ` /proc/pid/fd && anon_inode_fops Oleg Nesterov
2013-08-24 21:24 ` Willy Tarreau
2013-08-25 5:23 ` Al Viro
2013-08-25 6:50 ` Willy Tarreau
2013-08-25 18:51 ` Linus Torvalds
2013-08-25 19:48 ` Oleg Nesterov
2013-08-25 20:05 ` Linus Torvalds
2013-08-26 15:33 ` Oleg Nesterov
2013-08-26 16:37 ` Oleg Nesterov
2013-08-26 17:54 ` [PATCH] proc: make proc_fd_permission() thread-friendly Oleg Nesterov
2013-08-26 18:09 ` Linus Torvalds
2013-08-26 19:35 ` Linus Torvalds
2013-08-26 20:20 ` Willy Tarreau
2013-08-27 15:05 ` Oleg Nesterov
2013-08-27 14:39 ` [PATCH 0/1] proc: make /proc/self point to thread Oleg Nesterov
2013-08-27 14:40 ` [PATCH 1/1] " Oleg Nesterov
2013-08-27 16:39 ` Linus Torvalds
2013-08-27 17:49 ` Oleg Nesterov
2013-08-27 18:15 ` Linus Torvalds
2013-08-27 18:28 ` Oleg Nesterov
[not found] ` <CALCETrXP-mYBPRon=0NzexW1FK1Qxz2+Bwv7-WeHBQpvW7ywRg@mail.gmail.com>
2013-08-27 15:45 ` [PATCH] proc: make proc_fd_permission() thread-friendly Oleg Nesterov
2013-08-26 18:32 ` Eric W. Biederman
2013-08-26 18:46 ` Oleg Nesterov
2013-08-26 18:56 ` Oleg Nesterov
2013-08-26 19:10 ` Eric W. Biederman
2013-08-27 14:53 ` Oleg Nesterov
2013-08-25 18:32 ` /proc/pid/fd && anon_inode_fops Linus Torvalds
2013-08-25 19:11 ` Al Viro
2013-08-25 19:17 ` Andy Lutomirski
2013-09-03 15:58 ` Pavel Machek
2013-08-25 15:45 ` Oleg Nesterov
2013-08-22 19:39 ` Willy Tarreau [this message]
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=20130822193910.GJ31117@1wt.eu \
--to=w@1wt.eu \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=mingo@kernel.org \
--cc=oleg@redhat.com \
--cc=security@kernel.org \
--cc=spender@grsecurity.net \
--cc=torvalds@linux-foundation.org \
--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).