Conversation
The paper is here: cc.gatech.edu/~mxu80/pubs/xu
BTW, the leader of this project, Meng Xu(cc.gatech.edu/~mxu80/), will be joining the University of Waterloo as an assistant professor in 2021. You can contact him for research opportunities in security, system, and program analysis.
1
3
11
Have you looked at KTSAN? It uses a happens-before approach to kernel data-race detection. The paper suspiciously doesn't mention it.
github.com/google/ktsan/w
1
2
2
Replying to
KTSAN’s instrumentation has conflicts with our instrumentation. So we only tried our lockset+ happens before detector.
1
Show replies
Do you deal with disjoined atomics and memory barriers? (E.g. wmb() + WRITE_ONCE() forms a happens-before relation with the matching READ_ONCE() + rmb().) We had some troubles with those in KTSAN.
1
No, we do not. I am curious to know how did you deal with those in KTSAN.
1
Show replies
KCSAN is not a module of syzkaller, it's a kernel data-race detector on its own.
Have you looked at KernelStrider? It uses instrumentation (binary though) to collect event logs from kernel modules and then analyzes them in userspace. The paper doesn't mention it either.
2