From: Tahera Fahimi <fahimitahera@gmail.com>
To: Jann Horn <jannh@google.com>
Cc: outreachy@lists.linux.dev, mic@digikod.net, gnoack@google.com,
paul@paul-moore.com, jmorris@namei.org, serge@hallyn.com,
linux-security-module@vger.kernel.org,
linux-kernel@vger.kernel.org, bjorn3_gh@protonmail.com,
netdev@vger.kernel.org
Subject: Re: [PATCH v2 1/4] Landlock: Add signal control
Date: Tue, 6 Aug 2024 16:00:02 -0600 [thread overview]
Message-ID: <ZrKc4i1PhdMwA77h@tahera-OptiPlex-5000> (raw)
In-Reply-To: <CAG48ez3o9fmqz5FkFh3YoJs_jMdtDq=Jjj-qMj7v=CxFROq+Ew@mail.gmail.com>
On Tue, Aug 06, 2024 at 08:56:15PM +0200, Jann Horn wrote:
> On Tue, Aug 6, 2024 at 8:11 PM Tahera Fahimi <fahimitahera@gmail.com> wrote:
> > Currently, a sandbox process is not restricted to send a signal
> > (e.g. SIGKILL) to a process outside of the sandbox environment.
> > Ability to sending a signal for a sandboxed process should be
> > scoped the same way abstract unix sockets are scoped. Therefore,
> > we extend "scoped" field in a ruleset with
> > "LANDLOCK_SCOPED_SIGNAL" to specify that a ruleset will deny
> > sending any signal from within a sandbox process to its
> > parent(i.e. any parent sandbox or non-sandboxed procsses).
> [...]
> > diff --git a/security/landlock/task.c b/security/landlock/task.c
> > index 7e8579ebae83..a73cff27bb91 100644
> > --- a/security/landlock/task.c
> > +++ b/security/landlock/task.c
> > @@ -261,11 +261,54 @@ static int hook_unix_may_send(struct socket *const sock,
> > return -EPERM;
> > }
> >
> > +static int hook_task_kill(struct task_struct *const p,
> > + struct kernel_siginfo *const info, const int sig,
> > + const struct cred *const cred)
> > +{
> > + bool is_scoped;
> > + const struct landlock_ruleset *target_dom;
> > +
> > + /* rcu is already locked */
> > + target_dom = landlock_get_task_domain(p);
> > + if (cred)
> > + /* dealing with USB IO */
> > + is_scoped = domain_IPC_scope(landlock_cred(cred)->domain,
> > + target_dom,
> > + LANDLOCK_SCOPED_SIGNAL);
> > + else
> > + is_scoped = domain_IPC_scope(landlock_get_current_domain(),
> > + target_dom,
> > + LANDLOCK_SCOPED_SIGNAL);
>
> This might be a bit more concise if you turn it into something like:
>
> /* only USB IO supplies creds */
> cred = cred ?: current_cred();
> is_scoped = domain_IPC_scope(landlock_cred(cred)->domain,
> target_dom, LANDLOCK_SCOPED_SIGNAL);
>
> but that's just a question of style, feel free to keep it as-is
> depending on what you prefer.
Hi Jann,
Thanks for the feedback:)
> > + if (is_scoped)
> > + return 0;
> > +
> > + return -EPERM;
> > +}
> > +
> > +static int hook_file_send_sigiotask(struct task_struct *tsk,
> > + struct fown_struct *fown, int signum)
> > +{
> > + bool is_scoped;
> > + const struct landlock_ruleset *dom, *target_dom;
> > + struct task_struct *result = get_pid_task(fown->pid, fown->pid_type);
>
> I'm not an expert on how the fowner stuff works, but I think this will
> probably give you "result = NULL" if the file owner PID has already
> exited, and then the following landlock_get_task_domain() would
> probably crash? But I'm not entirely sure about how this works.
I considered since the file structure can always be obtained, then the
file owner PID always exist. I can check if we can use the credentials
stored in struct file instead.
> I think the intended way to use this hook would be to instead use the
> "file_set_fowner" hook to record the owning domain (though the setup
> for that is going to be kind of a pain...), see the Smack and SELinux
> definitions of that hook. Or alternatively maybe it would be even
> nicer to change the fown_struct to record a cred* instead of a uid and
> euid and then use the domain from those credentials for this hook...
> I'm not sure which of those would be easier.
Because Landlock does not use any security blob for this purpose, I am
not sure how to record the owner's doamin.
The alternative way looks nice.
> > + /* rcu is already locked! */
> > + dom = landlock_get_task_domain(result);
> > + target_dom = landlock_get_task_domain(tsk);
> > + is_scoped = domain_IPC_scope(dom, target_dom, LANDLOCK_SCOPED_SIGNAL);
> > + put_task_struct(result);
> > + if (is_scoped)
> > + return 0;
> > + return -EPERM;
> > +}
next prev parent reply other threads:[~2024-08-06 22:00 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-06 18:10 [PATCH v2 0/4] Landlock: Signal Scoping Support Tahera Fahimi
2024-08-06 18:10 ` [PATCH v2 1/4] Landlock: Add signal control Tahera Fahimi
2024-08-06 18:56 ` Jann Horn
2024-08-06 21:55 ` Jann Horn
2024-08-07 18:16 ` Mickaël Salaün
2024-08-07 23:36 ` Tahera Fahimi
2024-08-08 1:10 ` Jann Horn
2024-08-08 14:09 ` Mickaël Salaün
2024-08-08 14:42 ` Jann Horn
2024-08-09 10:59 ` Mickaël Salaün
2024-08-09 12:40 ` Mickaël Salaün
2024-08-09 12:45 ` Mickaël Salaün
2024-08-09 12:44 ` Jann Horn
2024-08-09 13:17 ` f_modown and LSM inconsistency (was [PATCH v2 1/4] Landlock: Add signal control) Mickaël Salaün
2024-08-09 14:00 ` Jann Horn
2024-08-09 14:44 ` Christian Brauner
2024-08-11 22:04 ` Paul Moore
2024-08-12 13:09 ` Jann Horn
2024-08-12 14:55 ` Mickaël Salaün
2024-08-12 14:57 ` Paul Moore
2024-08-12 15:06 ` Jann Horn
2024-08-12 16:30 ` Paul Moore
2024-08-12 17:27 ` Mickaël Salaün
2024-08-12 18:17 ` Paul Moore
2024-08-12 14:49 ` [PATCH] fs,security: Fix file_set_fowner LSM hook inconsistencies Mickaël Salaün
2024-08-12 15:00 ` Paul Moore
2024-08-13 1:32 ` kernel test robot
2024-08-13 1:42 ` kernel test robot
2024-08-13 1:42 ` kernel test robot
2024-08-13 11:44 ` kernel test robot
2024-08-09 13:37 ` [PATCH v2 1/4] Landlock: Add signal control Mickaël Salaün
2024-08-09 13:57 ` Jann Horn
2024-08-06 22:00 ` Tahera Fahimi [this message]
2024-08-06 22:55 ` Jann Horn
2024-08-06 18:10 ` [PATCH v2 2/4] selftest/Landlock: Signal restriction tests Tahera Fahimi
2024-08-06 18:10 ` [PATCH v2 3/4] sample/Landlock: Support signal scoping restriction Tahera Fahimi
2024-08-06 18:10 ` [PATCH v2 4/4] Landlock: Document LANDLOCK_SCOPED_SIGNAL Tahera Fahimi
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=ZrKc4i1PhdMwA77h@tahera-OptiPlex-5000 \
--to=fahimitahera@gmail.com \
--cc=bjorn3_gh@protonmail.com \
--cc=gnoack@google.com \
--cc=jannh@google.com \
--cc=jmorris@namei.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-security-module@vger.kernel.org \
--cc=mic@digikod.net \
--cc=netdev@vger.kernel.org \
--cc=outreachy@lists.linux.dev \
--cc=paul@paul-moore.com \
--cc=serge@hallyn.com \
/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