How To Write A Linux Security Module That Makes Sense For You
Casey Schaufler
https://linux.conf.au/schedule/30136/...
The traditional Linux security model traces it's fundamental concepts to the mini-computers of the 1970's. It makes a lot of sense when you consider a machine without a network connection that is being shared by a handful of friendly collaborators. Linux security modules (LSM) were introduced at the turn of the century to address the needs of high security environments. A recent and ongoing rework of the LSM infrastructure makes it significantly easier to implement and upstream small, targeted security features.
This talk will teach you want you can accomplish with a Linux security module, and what you can't. You'll learn the difference between a major module and a minor one. Techniques for implementing access controls on files, inter process communications and sockets will be covered, as will the underlying mechanisms required to maintain the data needed. You'll find out what a secid is, how it relates to a security context. and why it matters to CIPSO. The difference between inode based schemes and path name based ones will be made clear.
When the talk is over you'll have the tools you need to create a security module that protects what you care about instead of what seemed like a good idea to a government researcher during the Cold War.