On 2020-01-23 00:43, Burn Alting wrote:
> Well, there is one thing we may need to change.
> If USBGuard is configured to sent it's logs to auditd generating a 
> USER_DEVICE event
> type, we are going to have to add some new keys to the typetab.h table so 
> that one
> can interpret some of the USBGuard keys. Unfortunately, USBGuard appears to 
> decide
> to encode a key's value dependant on it's content (e.g. embedded spaces or 
> quotes)
> not the key. I am not sure if ausearch's interpret code can decide on the fly
> whether to unescape a possible encoded key's value or that will need to be 
> built in.
> Also any down-stream processing of interpreted values may be challenging 
> given an
> example key value pair
> likeencoded:device_rule=616C6C6F7720696420316436623A303030322073657269616C202230303
> 0303A30303A31612E3022206E616D6520224548434920486F737420436F6E74726F6C6C6572222068617
> 3682022656A3157566564794C79554D4C6951787A45637277625934357A436F64775638354B7A7937686
> D324776343D2220706172656E742D686173682022652F5257306D4D624D2B54534651787052694D45664
> C372F33524A664B5664716666426D39463571412B453D22207669612D706F72742022757362312220776
> 974682D696E746572666163652030393A30303A3030decoded:device_rule=allow id 
> 1d6b:0002
> serial "0000:00:1a.0" name "EHCI Host Controller" hash
> "ej1WVedyLyUMLiQxzEcrwbY45zCodwV85Kzy7hm2Gv4=" parent-hash
> "e/RW0mMbM+TSFQxpRiMEfL7/3RJfKVdqffBm9F5qA+E=" via-port "usb1" with-interface
> 09:00:00 

Yes, fields need to be consistently encoded since the parser isn't able
to detect that.

> On Wed, 2020-01-22 at 20:27 +1100, Burn Alting wrote:
> > Richard,
> > 
> > On the surface, it appears to have value, but as you say it would need to be
> > extended to other traditional, and non-traditional, removable media. 
> > Further, the
> > initial appeal in having the capability directly within the kernel was to 
> > make it
> > a little more difficult to subvert, centralise auditing policy/monitoring 
> > and, if
> > frame-worked appropriately, easily extensible to other than USB media types 
> > (which
> > was the basis for the Proof of Concept developed by RedHat back in 2016).
> > 
> > I have not used USBGuard myself, so will do some experimentation and report 
> > back.
> > 
> > Regards
> > 
> > On Tue, 2020-01-21 at 15:16 -0500, Richard Guy Briggs wrote:
> > > Hi Burn, and all,
> > > I've been aware of this issue for a while now, but wasn't directlyworking 
> > > on
> > > it.  Now that I'm taking a closer look at this issue, I amwondering how 
> > > much
> > > USBGuard changes the equation?
> > > https://www.kernel.org/doc/Documentation/usb/authorization.txt
> > > https://usbguard.github.io/
> > >   https://github.com/USBGuard/usbguard
> > > 
> > > https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/security_guide/sec-using-usbguard
> > > 
> > > It has tools to generate baseline lists of devices, but this is only
> > > forusb.  Other interfaces would need to be appropriately instrumented.
> > > - RGB

- RGB

--
Richard Guy Briggs <[email protected]>
Sr. S/W Engineer, Kernel Security, Base Operating Systems
Remote, Ottawa, Red Hat Canada
IRC: rgb, SunRaycer
Voice: +1.647.777.2635, Internal: (81) 32635

--
Linux-audit mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/linux-audit

Reply via email to