Filesystem fuzzing

Posted Mar 19, 2015 6:01 UTC (Thu) by eru (subscriber, #2753)
Parent article: Filesystem fuzzing

For one thing, there is no pass/fail criteria.

I once tested some aspects of a C compiler being developed by a kind of fuzzer that produced semi-randomly valid program snippets that were then compiled and run both by the tested compiler and a reference compiler, and the outputs compared. Differences (or a crash) indicated there was something requiring a closer look. Unfortunately some differences were benign, resulting from different expression evaluation order, but it did help find many problems.

I wonder if some of the file system fuzzing could employ the same idea, by doing identical operations to two different file systems and comparing the effect.

Filesystem fuzzing

Posted Mar 20, 2015 10:00 UTC (Fri) by hkario (subscriber, #94864) [Link]

problem is that many file systems have explicitly different behaviour between them