linux-kernel - Re: [PATCH] SELinux: Measure state and hash of policy using IMA

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 8 Sep 2020 07:58:07 -0400
From:   Stephen Smalley <stephen.smalley.work@...il.com>
To:     Lakshmi Ramasubramanian <nramas@...ux.microsoft.com>
Cc:     Mimi Zohar <zohar@...ux.ibm.com>, Paul Moore <paul@...l-moore.com>,
        Ondrej Mosnacek <omosnace@...hat.com>,
        Casey Schaufler <casey@...aufler-ca.com>,
        Tyler Hicks <tyhicks@...ux.microsoft.com>,
        tusharsu@...ux.microsoft.com, Sasha Levin <sashal@...nel.org>,
        James Morris <jmorris@...ei.org>,
        linux-integrity@...r.kernel.org,
        SElinux list <selinux@...r.kernel.org>,
        LSM List <linux-security-module@...r.kernel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] SELinux: Measure state and hash of policy using IMA

On Tue, Sep 8, 2020 at 12:44 AM Lakshmi Ramasubramanian
<nramas@...ux.microsoft.com> wrote:
>
> On 9/7/20 3:32 PM, Stephen Smalley wrote:
>
> >> Signed-off-by: Lakshmi Ramasubramanian <nramas@...ux.microsoft.com>
> >> Suggested-by: Stephen Smalley <stephen.smalley.work@...il.com>
> >> Reported-by: kernel test robot <lkp@...el.com> # error: implicit declaration of function 'vfree'
> >> Reported-by: kernel test robot <lkp@...el.com> # error: implicit declaration of function 'crypto_alloc_shash'
> >> Reported-by: kernel test robot <lkp@...el.com> # sparse: symbol 'security_read_selinux_policy' was not declared. Should it be static?
> >
> > Not sure these Reported-by lines are useful since they were just on
> > submitted versions of the patch not on an actual merged commit.
>
> I'll remove them when I update the patch.
>
> >
> >> diff --git a/security/selinux/measure.c b/security/selinux/measure.c
> >> new file mode 100644
> >> index 000000000000..caf9107937d9
> >> --- /dev/null
> >> +++ b/security/selinux/measure.c
> > <snip>
> >> +void selinux_measure_state(struct selinux_state *state, bool policy_mutex_held)
> >> +{
> > <snip>
> >> +
> >> +       if (!policy_mutex_held)
> >> +               mutex_lock(&state->policy_mutex);
> >> +
> >> +       rc = security_read_policy_kernel(state, &policy, &policy_len);
> >> +
> >> +       if (!policy_mutex_held)
> >> +               mutex_unlock(&state->policy_mutex);
> >
> > This kind of conditional taking of a mutex is generally frowned upon
> > in my experience.
> > You should likely just always take the mutex in the callers of
> > selinux_measure_state() instead.
> > In some cases, it may be the caller of the caller.  Arguably selinuxfs
> > could be taking it around all state modifying operations (e.g.
> > enforce, checkreqprot) not just policy modifying ones although it
> > isn't strictly for that purpose.
>
> Since currently policy_mutex is not used to synchronize access to state
> variables (enforce, checkreqprot, etc.) I am wondering if
> selinux_measure_state() should measure only state if policy_mutex is not
> held by the caller - similar to how we skip measuring policy if
> initialization is not yet completed.

No, we want to measure policy whenever there is a policy to measure.
Just move the taking of the mutex to the callers of
selinux_measure_state() so that it can be unconditional.