search

epi052 / fuzzing-101-solutions

Companion repository to the Fuzzing101 with LibAFL series of blog posts.
https://epi052.gitlab.io/notes-to-self/blog/2021-11-01-fuzzing-101-with-libafl/
141 stars 26 forks source link

Can't find crashes in exercise one. #3

Closed lentikr closed 2 years ago

lentikr commented 2 years ago

Hello. I'm following your code to solve exercise one with LibAFL(version 0.8.1). I found in your blog that you used 10 mins to generate 600+ inputs. But my corpus grow slowly after the size comes to 140. And I can't find any timeout in hours. I don't know where the problem is. Can anyone help me? OS: Ubuntu20.04 on VMware main(rs format).txt image

epi052 commented 2 years ago

hey there, sorry, i spaced on replying to this; did you ever get it solved?

lentikr commented 2 years ago

hey there, sorry, i spaced on replying to this; did you ever get it solved?

Thanks, I solved the problem. Just del the shemem_provider function when building the fork_server. The function means the fuzzer get the input from memory, that's not what I want. For detailed information, we can check the document of libafl's API.

let fork_server = ForkserverExecutor::builder()
       .program("./xpdf/install/bin/pdftotext".to_string())
       .parse_afl_cmdline(&[String::from("@@")])
       .shmem_provider(&mut shmem_provider)// delete this line
       .build(tuple_list!(time_observer, edges_observer))
       .unwrap();