Nicholas Moore Bianca Curutan Pooya Samizadeh McMaster University March 30, 2012
Introduction TAL and PCC TAL-0 TAL-1 Real World Application
Typed Assembly Language (TAL) extends traditional untyped assembly languages with typing annotations, memory management primitives, and a sound set of typing rules These typing rules guarantee the memory safety, control flow safety, and type safety of TAL programs
Principle of Proof-Carrying Code (PCC): Eliminate the need to trust code by requiring a formal, machine-checkable proof that the code has some desired properties 1. What properties should we require of the code? 2. How do code producers construct a formal proof that their code has the desired properties?
Control-flow Safety: Ensure that a program jumps only to a well-defined subset of possible entry points A focus on control-flow safety allows us to develop a simple abstract machine and demonstrate the key ideas of adapting a type system to machine code
Instructions and Operands
Abstract Machine Syntax
Syntax additions
Limitations: ◦ Only supports simple tuple or record-like data structures ◦ Insufficient for compiling real-world high-level languages which provide data abstraction mechanisms ◦ Can only allocate objects whose size is known at compile time
Examples to make type system more useful: ◦ Annotate primitive memory type components with flags to control whether that component supports read-only, write-only, or read-write access ◦ Add support for subtyping ◦ Consider a read-write component to be a subtype of a read-only or a write-only component
Introduction Touchstone PCC Architecture Advantages Over Related Techniques Technical Difficulties to Overcome
Proof-Carrying Code (PCC): General framework that allows the host to verify properties about an agent via a formal proof that accompanies the executable code The host system can quickly verify the validity of the proof and compare the conclusions to its own security policy to determine if the agent has certain safety properties
1. Operates at load time before the agent code is installed in the host system 2. Trusted computing base is small 3. Can operate even on agents expressed in native-code form 4. General – all PCC has to do is verify safety explanations and match them with the code and safety policy
How to encode the formal proof? How to check the proof? How to relate the proof with the program?
Introduction Recall TAL-0 Example (Quotient) Description and Demo
Task ◦ Build an interpreter for the TAL-0 abstract machine in Haskell (i.e., functional programming language). Extra ◦ Literate programming using LaTeX
Instructions and Operands Abstract Machine Syntax
# reduce register r1 to 0 init|r1 := 10; r2:=11; => loop; loop|? r1 => done; r1 := r1 + -1; r1:=8; => loop; done|
Jones, Mark. Functional Programming with Overloading and Higher-Order Polymorphism. Diss. University of Nottingham. Nottingham. Print. Liang, Sheng, Paul Hudak, and Mark Jones. Monad Transformers and Modular Interpreters. Diss. Yale University. New Haven. Print. Necula, George. "Proof-Carrying Code." Computer Science Division, EECS at UC Berkeley. 22 July Web. 21 Mar Pierce, Benjamin C. Advanced Topics in Types and Programming Languages. Cambridge, MA: MIT, Print. "Typed Assembly Language Compiler." Cornell University Department of Computer Science. Web. 21 Mar