Re: [PATCH] selinux: reorder hooks to make runtime disable less broken

[Casey Schaufler <casey@schaufler-ca.com>]

On 12/10/2019 11:50 AM, Stephen Smalley wrote:
> On 12/10/19 2:43 PM, Casey Schaufler wrote:
>> On 12/10/2019 11:29 AM, Paul Moore wrote:
>>> On Tue, Dec 10, 2019 at 6:19 AM Ondrej Mosnacek <omosnace@redhat.com> wrote:
>>>> On Mon, Dec 9, 2019 at 2:21 PM Stephen Smalley <sds@tycho.nsa.gov> wrote:
>>>>> On 12/9/19 2:57 AM, Ondrej Mosnacek wrote:
>>>>>> Commit b1d9e6b0646d ("LSM: Switch to lists of hooks") switched the LSM
>>>>>> infrastructure to use per-hook lists, which meant that removing the
>>>>>> hooks for a given module was no longer atomic. Even though the commit
>>>>>> clearly documents that modules implementing runtime revmoval of hooks
>>>>>> (only SELinux attempts this madness) need to take special precautions to
>>>>>> avoid race conditions, SELinux has never addressed this.
>>>>>>
>>>>>> By inserting an artificial delay between the loop iterations of
>>>>>> security_delete_hooks() (I used 100 ms), booting to a state where
>>>>>> SELinux is enabled, but policy is not yet loaded, and running these
>>>>>> commands:
>>>>>>
>>>>>>       while true; do ping -c 1 <some IP>; done &
>>>>>>       echo -n 1 >/sys/fs/selinux/disable
>>>>>>       kill %1
>>>>>>       wait
>>>>>>
>>>>>> ...I was able to trigger NULL pointer dereferences in various places. I
>>>>>> also have a report of someone getting panics on a stock RHEL-8 kernel
>>>>>> after setting SELINUX=disabled in /etc/selinux/config and rebooting
>>>>>> (without adding "selinux=0" to kernel command-line).
>>>>>>
>>>>>> Reordering the SELinux hooks such that those that allocate structures
>>>>>> are removed last seems to prevent these panics. It is very much possible
>>>>>> that this doesn't make the runtime disable completely race-free, but at
>>>>>> least it makes the operation much less fragile.
>>>>>>
>>>>>> An alternative (and safer) solution would be to add NULL checks to each
>>>>>> hook, but doing this just to support the runtime disable hack doesn't
>>>>>> seem to be worth the effort...
>>>>> Personally, I would prefer to just get rid of runtime disable
>>>>> altogether; it also precludes making the hooks read-only after
>>>>> initialization.  IMHO, selinux=0 is the proper way to disable SELinux if
>>>>> necessary.  I believe there is an open bugzilla on Fedora related to
>>>>> this issue, since runtime disable was originally introduced for Fedora.
>>>> I, too, would like to see it gone, but removing it immediately would
>>>> likely cause issues for existing users (see [1]) ...
>>>>
>>>> [1] https://bugzilla.redhat.com/show_bug.cgi?id=1430944#c2
>>> For the record, and for those who didn't click on the RHBZ link above,
>>> I'm a big fan of getting rid of SELinux's runtime disable but concede
>>> that it must be done in such a way to as not horribly break userspace.
>>
>> Is there some reason that changing the "disable SELinux" option
>> has to remove the hooks? Why can't it set selinux_enabled to 0
>> and be done with it?
>
> selinux_enabled is only used during initialization to deal with selinux=0 across the different components of SELinux.  It isn't checked by the hooks themselves.  And if we were to add a if (!selinux_enabled) return 0 to the start of every hook, then that's just another easy target for kernel exploits to leverage.

That's what I expected, but I wanted to see it explicitly stated. Thanks.