fuzz: Mutate -max_len= during generation phase #30371

1. 
   maflcko commented at 3:16 pm on July 1, 2024: member

   This revives #28178 (comment), because it helps to find #30367, which failed to be found for more than a year on any of the existing fuzz servers.

   Locally, I can find the bug with -max_len=84 -use_value_profile=1 on a single thread on a laptop. The reason is likely that a smaller max_len results in a faster fuzzing speed (iterations). As a side-effect it may also be effective at reducing the size of existing inputs (without losing coverage), but I haven’t benchmarked this.

2. fuzz: Mutate -max_len= during generation phase 9999b60298
3. 
   DrahtBot commented at 3:16 pm on July 1, 2024: contributor

   The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

   Code Coverage & Benchmarks

   For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/30371.

   Reviews

   See the guideline for information on the review process. A summary of reviews will appear here.

4. DrahtBot added the label Tests on Jul 1, 2024
5. 
   dergoegge commented at 3:57 pm on July 1, 2024: member

   I don’t know how to evaluate if this is a good idea but I’m also not using the test runner, so wouldn’t mind. Also, I don’t think oss-fuzz is doing this, so how come they found the bug but we didn’t?
6. 
   maflcko commented at 10:56 am on July 2, 2024: member

   I don’t know how to evaluate if this is a good idea but I’m also not using the test runner, so wouldn’t mind. Also, I don’t think oss-fuzz is doing this, so how come they found the bug but we didn’t?

   Good points. Possibly, the single run on my laptop was just extremely lucky and it is highly variant to find the bug regardless of max_len. The high variance will probably make this hard to benchmark.

7. 
   TheCharlatan commented at 7:09 am on July 3, 2024: contributor

   Re #30371#issue-2384114654

   Locally, I can find the bug with -max_len=84 -use_value_profile=1 on a single thread on a laptop.

   I’ve been running this for a day and did not find the crash. How long was this running for?

8. 
   maflcko commented at 7:20 am on July 3, 2024: member

   16kk iterations; N=1. Again, I guess my Laptop just got extremely lucky.
9. maflcko marked this as a draft on Jul 3, 2024
10. 
    maflcko commented at 7:22 am on July 3, 2024: member

    Turning into draft for now, to allow for more time to benchmark and evaluate this.
11. 
    DrahtBot commented at 1:25 am on December 30, 2024: contributor

    There hasn’t been much activity lately. What is the status here?

    Finding reviewers may take time. However, if the patch is no longer relevant, please close this pull request. If the author lost interest or time to work on this, please close it and mark it ‘Up for grabs’ with the label, so that it can be picked up in the future.

12. DrahtBot added the label CI failed on Feb 6, 2025
13. 
    fanquake commented at 7:15 pm on February 20, 2025: member

    Turning into draft for now, to allow for more time to benchmark and evaluate this.

    Any more thoughts on this?

14. maflcko closed this on Feb 20, 2025

    ---

Contributors
maflcko DrahtBot dergoegge TheCharlatan fanquake

Labels
Tests CI failed