RFC: storvsp: new fuzzer #612
Conversation
from other test code to get something that compiles and runs. Coverage shows that the fuzzer (at this point) does hit much of the core storvsp packet processing, _except_ for the IO path. That remains a work in progress.
|
Overall this seems like a good starting point. Has this found anything interesting in its current state? |
Thanks! And thanks for the feedback - for what it's worth, this is the biggest chunk of rust code I've written (so I'm appreciative and very open to any stylistic, idiomatic, and general feedback). I was able to get the fuzzer to hang. It's not clear to me if this is a fuzzer bug or a bug in storvsp. I'm debugging. I've changed the fuzzer some since then, so I don't know if I'll be able to repro or not. I'll keep you posted. |
* Don't promote as many things to `pub`, and make it explicit when that happens. * .div_ceil rather than doing it manually * Arbitrary fields in enums. * cargo xtask fmt fix.
update: this found the issue fixed in #621 . That same test case now seems to hang. I'm going to hold off completing this until I get some more insight into that hang (or run out of time ...). |
First commit of a storvsp fuzzer. This begs, borrows, and steals from other test code to get something that compiles and runs.
Coverage shows that the fuzzer (at this point) does hit much of the core storvsp packet processing, except for the IO path. That remains a work in progress.