-
Reducers are Fuzzers
A test case isn’t just a test case: it lives in the (generally extremely large) space of inputs for the software system you are testing. If we have a test case that triggers a bug, here’s one way we can look at it: The set of test cases triggering a bug is a useful notion…
-
Multi-Version Execution Defeats a Compiler-Bug-Based Backdoor
[This piece is jointly authored by Cristian Cadar, Luís Pina, and John Regehr] What should you do if you’re worried that someone might have exploited a compiler bug to introduce a backdoor into code that you are running? One option is to find a bug-free compiler. Another is to run versions of the code produced…
-
Testcase Reduction for Non-Preprocessed C and C++
C-Reduce takes a C or C++ file that triggers a bug in a compiler (or other tool that processes source code) and turns it into the smallest possible test case that still triggers the bug. Most often, we try to reduce code that has already been preprocessed. This post is about how to reduce non-preprocessed…
-
A Few Synthesizing Superoptimizer Results
For this post, I crippled Souper by disabling its path conditions and limiting the depth of harvested expressions to two LLVM instructions. The first goal was to create a nice easy burn-in test for Souper’s instruction synthesizer, which uses a variant of this method; the second goal was to see if depth-limited, path-condition-free expressions would…
-
Nibble Sort Denouement
Back in January my nibble sort contest resulted in entries that dramatically exceeded my expectations. Since then I’ve been trying to write up a post explaining the various strategies that people used and since you don’t care about my excuses I won’t tell you them, but I never got it written. However! I want to…
-
Defending Against Compiler-Based Backdoors
Scotty Bauer (a Utah grad student), Pascal Cuoq, and I have an article in the latest PoC||GTFO about introducing a backdoor into sudo using a compiler bug. In other words, the C code implementing sudo does not contain a backdoor, but a backdoor appears when sudo is built using a particular compiler version (clang 3.3,…
-
Generating a Random Program vs. Generating All Programs
Generating all possible inputs — up to some maximum length — to a software system is one way of creating test cases, and this technique even has a name: bounded exhaustive testing. Back when we were doing Csmith, my then-student Yang Chen spent a while on a bounded exhaustive C program generator which was in…
-
UB Canaries
If you report an undefined behavior bug, a common reaction from software developers is “So what? Our code works just fine.” As a random example, here is a discussion I had with Rasmus Lerdorf about five years ago about some UBs in the PHP interpreter. One might point out that it wasn’t a very mature…
-
Instruction Synthesis is Fun and Weird
Synthesis is sort of a hot topic in PL research lately. It basically means “implement a specification automatically.” Of course, at some level this isn’t very different from what compilers have been doing for ages, if we consider the source language program to be the specification, but when we’re doing synthesis the connotation is that…
-
Souper Results 2
The Souper superoptimizer has made some progress since my last post about it. We wrote compiler drivers that usually reduce the problem of building a project with Souper to make CC=sclang CXX=sclang++. Souper now uses Redis to cache optimizations so that even if the initial build of a program using Souper is slow, subsequent builds…