About llvm-project patch.

[gtt1995]

Hello, I don’t know why, my RR (request for comment) on llvm's phabricator has never received a response, so with the help of clusterfuzz, please help me to pay attention. @Dor1s @oliverchang @jonathanmetzman Thanks very much!

[gtt1995]

In this patch, I found that libFuzzer’s fork mode has certain flaws. I proposed some strategies to solve these problems. Because it is an improvement of parallel fuzzing, I cannot evaluate it on fuzzbench. Finally, I uploaded my own local experiment data, and test targets are real-world programs from the OSS-Fuzz project.

[jonathanmetzman]

@gtt1995 Sorry for the delayed reply on phabricator. I think it was probably due to the holidays in the United States, where most libFuzzer's reviewers are located.

[gtt1995]

Thank you for your reply, you are also a member of the reviewer, can you check it for me? On the other hand, Can i evaluate this on Fuzzbench, because this is an improvement on parallel fuzzing? What should i do?

[jonathanmetzman]

Thank you for your reply, you are also a member of the reviewer, can you check it for me?

I think @kcc doesn't want me to land things if they aren't vetted by @morehouse first.

On the other hand, Can i evaluate this on Fuzzbench, because this is an improvement on parallel fuzzing? What should i do?

I would just submit a new fuzzer that is libfuzzer with your patch and then submit an experiment request. You can follow our guide: https://google.github.io/fuzzbench/getting-started/adding-a-new-fuzzer/

[gtt1995]

Thanks very much , i will try it .