Re:BSD vs. Linux (2005)
by Anthony-G on Tuesday July 05, 2016 @04:51AM
(#12035163)
Back in 2011, security researcher, Joanna Rutkowska wrote a very enlightening article on the inherent security deficiencies of the X server architecture which allows allows any windowed application to control any other one: http://blog.invisiblethings.org/2011/04/23/linux-security-ci...
[ Parent
| on HN ]
Re:BSD vs. Linux (2005)
by nickpsecurity on Tuesday July 05, 2016 @09:10AM
(#12036136)
Back in the 90's, the high-assurance community fleshed out the issues and solved many of them on top of strong endpoints. Here's an example:
https://www.acsac.org/2006/papers/epstein-paper.pdf
One from 2005 that I sent Joanna in gripe about Qubes not having a trusted path:
https://www.acsac.org/2005/papers/54.pdf
Good to see she added a better graphics system at some point. I suggest looking at prior work, though, as it's by security engineers with knowledge of adding it to every aspect of lifecycle and intent to minimize TCB. For example, GenodeOS is using Nitpicker + a microkernel. The abandoned EROS OS, whose site is still up, combined a different windowing system with a capability kernel and new networking stack. All to minimize risk in system in fine-grained way.