DGFuzz Testing #2000

DanBlackwell · 2024-07-07T19:45:45Z

This is a new fuzzer that I'd like to test out.

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-07-dgfuzz --fuzzers dgfuzz libafl

DonggeLiu · 2024-07-07T23:38:57Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-08-dgfuzz --fuzzers dgfuzz

DonggeLiu · 2024-07-07T23:46:51Z

Ops:
Would you mind making a trivial modification to service/gcbrun_experiment.py?
This will allow PR experiments. Here is an example to add a dummy comment : )
Thanks!

DanBlackwell · 2024-07-08T06:56:09Z

Done! Forgot about that sorry

DonggeLiu · 2024-07-08T10:01:44Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-08-dgfuzz --fuzzers dgfuzz

DanBlackwell · 2024-07-08T11:37:48Z

experiment name: 2024-07-08-dgfuzz
Report: here
experiment data: here

DanBlackwell · 2024-07-08T12:41:14Z

@DonggeLiu Agh, forgot to run make presubmit and the linter had taken issue with my whitespace. I think it needs restarting? The experiment report didn't populate, so I assume it's down to the CI checks failing.

DonggeLiu · 2024-07-08T12:56:04Z

@DonggeLiu Agh, forgot to run make presubmit and the linter had taken issue with my whitespace. I think it needs restarting?

Hi @DanBlackwell
That experiment has launched: The data directory is not empty.
The report normally takes longer to show after FB finishes building the benchmarks and fuzzers.

I've manually terminated all VM instances of that experiment and will relaunch a new one to use the latest code.

The experiment report didn't populate, so I assume it's down to the CI checks failing.

The CI checks won't block PR experiments, they are separated.
But it's good that you double-checked CIs, as sometimes the fuzzer might not be compatible with some benchmarks : )

DonggeLiu · 2024-07-08T12:56:27Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-08-dgfuzz-1 --fuzzers dgfuzz

DanBlackwell · 2024-07-08T13:03:25Z

Hi @DanBlackwell
That experiment has launched: The data directory is not empty.
The report normally takes longer to show after FB finishes building the benchmarks and fuzzers.

My apologies, I thought the experiment-folders and coverage dirs in the data directory normally showed up after 1-2 hours; then when I saw the CI checks failed I jumped to the conclusion that it only ran the build and bailed.

But it's good that you double-checked CIs, as sometimes the fuzzer might not be compatible with some benchmarks : )

This fuzzer will only work on a subset for now, as DFSan requires all dependencies to be compiled with the -fsanitize=dataflow flag set (including the C++ standard library, OpenSSL etc.). If it looks promising then I'll put in some time to figure out how that can be done.

DanBlackwell · 2024-07-08T13:03:51Z

experiment name: `2024-07-08-dgfuzz-1`

Report: here
experiment data: here

DanBlackwell · 2024-07-08T19:28:54Z

@DonggeLiu Hmm, seems like it’s not started the experiment again (no experiment-folders created again). Is it possible that because it’s the only fuzzer in the experiment - and some of the targets did not build?

If so, maybe it could be run with just the 8 targets that do build. I've now added dgfuzz to all the unsupported_fuzzers fields for those that do not build, so in theory it can be run with the same gcb command. (from what I saw here, it looks like gcbrun only lets you do one benchmark at a time, is that correct?)

Additionally, I’ve set the merge_with_nonprivate flag to false just in case running less programs causes any weirdness; it would be best if we could also run libafl as a baseline / control given that it won't merge now.

If it helps, these are the 8 that build ok:

harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

PS, if the experiment did start at some point (it's been 5 hours without starting at the time of writing), then ignore all this and let it run through :)

DonggeLiu · 2024-07-09T04:24:33Z

PS, if the experiment did start at some point (it's been 5 hours without starting at the time of writing), then ignore all this and let it run through :)

I think it did : )
https://storage.googleapis.com/www.fuzzbench.com/reports/experimental/2024-07-08-dgfuzz-1/index.html

Is it possible that because it’s the only fuzzer in the experiment - and some of the targets did not build?

I don't think this will block the experiment. If this happens, then that benchmark will be excluded from the exp without affecting others.

DanBlackwell · 2024-07-10T15:06:34Z

Hi @DonggeLiu , I've pushed up some updates - I think that last run was getting killed due to shared memory not getting freed correctly. Can you run the following:

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-10-dgfuzz --fuzzers dgfuzz

DonggeLiu · 2024-07-11T00:22:57Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-11-dgfuzz --fuzzers dgfuzz

DanBlackwell · 2024-07-11T08:37:46Z

Experiment 2024-07-11-dgfuzz data and results will be available later at:
The experiment data.
The experiment report(experimental).

DanBlackwell · 2024-07-11T12:37:29Z

@DonggeLiu Just a note that most of these did not build with some apt errors it seems. The following was taken from libpcap_fuzz_both:

Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.257 Get:240 http://archive.ubuntu.com/ubuntu focal/main amd64 python3-dev amd64 3.8.2-0ubuntu2 [1212 B]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.447 Err:233 http://security.ubuntu.com/ubuntu focal-updates/main amd64 libpython3.8 amd64 3.8.10-0ubuntu1~20.04.9
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.447   404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.555 Err:234 http://security.ubuntu.com/ubuntu focal-updates/main amd64 libpython3.8-dev amd64 3.8.10-0ubuntu1~20.04.9
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.555   404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.555 Err:239 http://security.ubuntu.com/ubuntu focal-updates/main amd64 python3.8-dev amd64 3.8.10-0ubuntu1~20.04.9
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.555   404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.562 E: Failed to fetch http://security.ubuntu.com/ubuntu/pool/main/p/python3.8/libpython3.8_3.8.10-0ubuntu1~20.04.9_amd64.deb  404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.562 E: Failed to fetch http://security.ubuntu.com/ubuntu/pool/main/p/python3.8/libpython3.8-dev_3.8.10-0ubuntu1~20.04.9_amd64.deb  404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.562 E: Failed to fetch http://security.ubuntu.com/ubuntu/pool/main/p/python3.8/python3.8-dev_3.8.10-0ubuntu1~20.04.9_amd64.deb  404  Not Found [IP: 91.189.91.81 80]
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.562 E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 7.562 Fetched 124 MB in 4s (28.4 MB/s)
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": #12 ERROR: executor failed running [/bin/sh -c apt-get install -y         build-essential         python3-dev         python3-setuptools         automake         cmake         git         flex         bison         libglib2.0-dev         libpixman-1-dev         cargo         libgtk-3-dev         ninja-build         gcc-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-plugin-dev         libstdc++-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-dev]: exit code: 100
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": ------
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate":  > [8/9] RUN apt-get install -y         build-essential         python3-dev         python3-setuptools         automake         cmake         git         flex         bison         libglib2.0-dev         libpixman-1-dev         cargo         libgtk-3-dev         ninja-build         gcc-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-plugin-dev         libstdc++-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-dev:
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": ------
Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate": executor failed running [/bin/sh -c apt-get install -y         build-essential         python3-dev         python3-setuptools         automake         cmake         git         flex         bison         libglib2.0-dev         libpixman-1-dev         cargo         libgtk-3-dev         ninja-build         gcc-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-plugin-dev         libstdc++-$(gcc --version|head -n1|sed 's/\..*//'|sed 's/.* //')-dev]: exit code: 100
Finished Step #2 - "dgfuzz-libpcap_fuzz_both-builder-intermediate"
ERROR
ERROR: build step 2 "gcr.io/cloud-builders/docker" failed: step exited with non-zero status: 1

From what I see it looks like http://archive.ubuntu.com/ goes through fine but http://security.ubuntu.com fails. Strangely it did work for mbed_fuzz_dtlsclient.

DonggeLiu · 2024-07-11T13:22:03Z

From what I see it looks like http://archive.ubuntu.com/ goes through fine but http://security.ubuntu.com fails. Strangely it did work for mbed_fuzz_dtlsclient.

Yep, this should be a flaky internet connection failure.
We sometimes see this but cannot do much other than re-run the exp.

DonggeLiu · 2024-07-11T13:22:18Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-07-11-dgfuzz-1 --fuzzers dgfuzz

DanBlackwell · 2024-07-11T13:41:22Z

Experiment 2024-07-11-dgfuzz-1 data and results will be available later at:
The experiment data.
The experiment report(experimental).

DanBlackwell · 2024-07-11T18:02:06Z

Apt is still not playing ball it seems. It's trying to fetch:

http://security.ubuntu.com/ubuntu/pool/main/p/python3.8/libpython3.8_3.8.10-0ubuntu1~20.04.9_amd64.deb

Looking here it seems there is no 20.04.9, but there is 20.04.10 (python3.8-dbg_3.8.10-0ubuntu1~20.04.10_amd64.deb). ~~As to why it is trying to fetch that particular version, I don't know. The particular command that fails is copied from AFL++ builder.Dockerfile, so maybe that will have the same issue?~~

EDIT: it was missing apt-get update; I guess so far I'd got lucky that the base image was new enough that nothing had been replaced yet.

DanBlackwell · 2024-09-03T07:43:33Z

Hi @DonggeLiu are you logged off for the day yet? Can see you pulled a big stint for oss-fuzz-gen so not sure if you saw this yet

DonggeLiu · 2024-09-03T12:32:31Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-03-dgfuzz --fuzzers dgfuzz_e0d055 --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer

DanBlackwell · 2024-09-03T15:20:30Z

Experiment 2024-09-03-dgfuzz data and results will be available later at:
The experiment data.
The experiment report(experimental).

…tracking logs

DanBlackwell · 2024-09-04T10:38:13Z

Hi @DonggeLiu any chance you could run the following for me please?

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-04-dgfuzz --fuzzers dgfuzz_22977c --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer

DonggeLiu · 2024-09-04T12:13:29Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-04-dgfuzz --fuzzers dgfuzz_22977c --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer

DanBlackwell · 2024-09-04T16:27:06Z

Hmm @DonggeLiu the experiment failed to build (see here) for some reason - but I can't access the logs to see what went wrong, as they're on GCP.

Note that the CI always fails for this PR, but usually the experiment builds fine.

DanBlackwell · 2024-09-05T07:55:34Z

Ok @DonggeLiu , I added the log for retcode in the measurer so I think it's worth adding zlib back in and see if we can catch what was killing it. That last run yesterday failed for some reason as mentioned in the comment above; if there's nothing suspicious in the logs and it was just flakiness then could we run the following instead? (Otherwise I'm happy to try debug if you can send me the logs)

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-05-dgfuzz --fuzzers dgfuzz_0ebf1a9 --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-05T23:54:48Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-06-dgfuzz --fuzzers dgfuzz_0ebf1a9 --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-05T23:58:35Z

That last run yesterday failed for some reason as mentioned in the comment above

I think that was because the fuzzer dgfuzz_22977c does not exist.
Did you push it to the fuzzers dir?

"Step #1: gcbrun_experiment.py: error: argument -f/--fuzzers: invalid choice: 'dgfuzz_22977c' (choose from 'hastefuzz', 'aflfast', 'symcc_afl_single', 'libafl_libfuzzer', 'dgfuzz_4aacb8', 'dgfuzz_e0d055', 'darwin', 'symcc_aflplusplus_single', 'tortoisefuzz', 'libfuzzer', 'learnperffuzz', 'aflplusplus_um_prioritize_75', 'aflplusplus_um_random', 'dgfuzz_93227b', 'dgfuzz_22997c', 'klee', 'lafintel', 'afl_2_52_b', 'aflplusplus_um_random_75', 'honggfuzz_um_parallel', 'pastis', 'nautilus', 'honggfuzz_um_prioritize_75', 'aflpp_random_wrs_rp', 'mopt', 'glibfuzzer', 'gramatron', 'afl_qemu', 'aflcc', 'eclipser_new', 'aflplusplus_frida', 'fuzzolic_aflplusplus_z3', 'aflpp_random_wrs', 'fafuzz', 'aflpp_random_no_favs', 'dgfuzz_c286e5', 'aflplusplus_um_prioritize', 'dgfuzz_653cc9', 'dgfuzz_3dfa24', 'eclipser_aflplusplus', 'dgfuzz_135c90', 'dgfuzz_d3c951', 'symcc_afl', 'dgfuzz_8d1cc7', 'manul', 'token_level', 'symcc_aflplusplus', 'aflpp_random_wrs_rf_rp', 'aflsmart', 'wingfuzz', 'honggfuzz', 'centipede', 'afl_virginmap', 'honggfuzz_qemu', 'symsan', 'honggfuzz_um_prioritize', 'aflpp_random_default', 'fairfuzz', 'dgfuzz_3f8f81', 'aflplusplus', 'aflsmart_plusplus', 'pythia_effect_bb', 'dgfuzz_065684_depth_3', 'aflrustrust', 'dgfuzz_control', 'weizz_qemu', 'eclipser', 'aflplusplus_qemu', 'dgfuzz_065684_depth_1', 'libafl_text', 'dgfuzz_065684_depth_2', 'afl', 'dgfuzz_dc6e84', 'pythia_bb', 'dgfuzz_0e010d', 'dgfuzz_dc6e84_rand_sched', 'ecofuzz', 'neuzz', 'libafl', 'aflplusplus_um_parallel', 'grimoire', 'symqemu_aflplusplus', 'dgfuzz_83ca1e', 'aflpp_random_wrs_rf', 'fuzzolic_aflplusplus_fuzzy', 'afl_random_favored', 'dgfuzz_ab0800', 'centipede_function_filter', 'honggfuzz_um_random', 'honggfuzz_um_random_75')"

DanBlackwell · 2024-09-06T06:20:33Z

That last run yesterday failed for some reason as mentioned in the comment above

I think that was because the fuzzer dgfuzz_22977c does not exist. Did you push it to the fuzzers dir?

"Step #1: gcbrun_experiment.py: error: argument -f/--fuzzers: invalid choice: 'dgfuzz_22977c' (choose from 'hastefuzz', 'aflfast', 'symcc_afl_single', 'libafl_libfuzzer', 'dgfuzz_4aacb8', 'dgfuzz_e0d055', 'darwin', 'symcc_aflplusplus_single', 'tortoisefuzz', 'libfuzzer', 'learnperffuzz', 'aflplusplus_um_prioritize_75', 'aflplusplus_um_random', 'dgfuzz_93227b', 'dgfuzz_22997c', 'klee', 'lafintel', 'afl_2_52_b', 'aflplusplus_um_random_75', 'honggfuzz_um_parallel', 'pastis', 'nautilus', 'honggfuzz_um_prioritize_75', 'aflpp_random_wrs_rp', 'mopt', 'glibfuzzer', 'gramatron', 'afl_qemu', 'aflcc', 'eclipser_new', 'aflplusplus_frida', 'fuzzolic_aflplusplus_z3', 'aflpp_random_wrs', 'fafuzz', 'aflpp_random_no_favs', 'dgfuzz_c286e5', 'aflplusplus_um_prioritize', 'dgfuzz_653cc9', 'dgfuzz_3dfa24', 'eclipser_aflplusplus', 'dgfuzz_135c90', 'dgfuzz_d3c951', 'symcc_afl', 'dgfuzz_8d1cc7', 'manul', 'token_level', 'symcc_aflplusplus', 'aflpp_random_wrs_rf_rp', 'aflsmart', 'wingfuzz', 'honggfuzz', 'centipede', 'afl_virginmap', 'honggfuzz_qemu', 'symsan', 'honggfuzz_um_prioritize', 'aflpp_random_default', 'fairfuzz', 'dgfuzz_3f8f81', 'aflplusplus', 'aflsmart_plusplus', 'pythia_effect_bb', 'dgfuzz_065684_depth_3', 'aflrustrust', 'dgfuzz_control', 'weizz_qemu', 'eclipser', 'aflplusplus_qemu', 'dgfuzz_065684_depth_1', 'libafl_text', 'dgfuzz_065684_depth_2', 'afl', 'dgfuzz_dc6e84', 'pythia_bb', 'dgfuzz_0e010d', 'dgfuzz_dc6e84_rand_sched', 'ecofuzz', 'neuzz', 'libafl', 'aflplusplus_um_parallel', 'grimoire', 'symqemu_aflplusplus', 'dgfuzz_83ca1e', 'aflpp_random_wrs_rf', 'fuzzolic_aflplusplus_fuzzy', 'afl_random_favored', 'dgfuzz_ab0800', 'centipede_function_filter', 'honggfuzz_um_random', 'honggfuzz_um_random_75')"

You're right, I made a typo in the commit hash - should have been dgfuzz_22997c rather than dgfuzz_22977c. No worries, the one running now is newer (and hopefully better).

DanBlackwell · 2024-09-08T21:12:14Z

Hi @DonggeLiu, could you please run the following:

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-09-dgfuzz --fuzzers dgfuzz_7fa3ff --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

n.b. this is doing something quite memory heavy so I won't be surprised if there's a bunch of OOMs.

DonggeLiu · 2024-09-09T03:52:01Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-09-dgfuzz --fuzzers dgfuzz_7fa3ff --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DanBlackwell · 2024-09-09T19:51:46Z

Hi @DonggeLiu , thanks for that! Could you please run the following? PS looks like whatever the issue was with zlib last month, it is no longer occurring

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-10-dgfuzz --fuzzers dgfuzz_e989ba --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-09T21:43:44Z

PS looks like whatever the issue was with zlib last month, it is no longer occurring

Good to know!
I am really busy with other things recently and does not have the time to debug FB.

DonggeLiu · 2024-09-09T21:43:52Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-10-dgfuzz --fuzzers dgfuzz_e989ba --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DanBlackwell · 2024-09-11T20:55:03Z

Hi @DonggeLiu , could you run the following please?

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-12-dgfuzz --fuzzers dgfuzz_1d7283 dgfuzz_32d973 --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-12T03:45:45Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-12-dgfuzz --fuzzers dgfuzz_1d7283 dgfuzz_32d973 --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DanBlackwell · 2024-09-12T08:24:20Z

Experiment 2024-09-12-dgfuzz data and results will be available later at:
The experiment data.
The experiment report(experimental).

DanBlackwell · 2024-09-12T21:58:34Z

Hi @DonggeLiu , any chance you can run the following please?

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-13-dgfuzz --fuzzers dgfuzz_49a0ef --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-12T22:30:49Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-13-dgfuzz --fuzzers dgfuzz_49a0ef --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

…isabled

DanBlackwell · 2024-09-15T17:30:30Z

Hi @DonggeLiu, could you run the following for me please?

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-16-dgfuzz --fuzzers dgfuzz_23de6c --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DonggeLiu · 2024-09-16T22:18:20Z

/gcbrun run_experiment.py -a --experiment-config /opt/fuzzbench/service/experiment-config.yaml --experiment-name 2024-09-16-dgfuzz --fuzzers dgfuzz_23de6c --benchmarks harfbuzz_hb-shape-fuzzer lcms_cms_transform_fuzzer libpcap_fuzz_both mbedtls_fuzz_dtlsclient openthread_ot-ip6-send-fuzzer stb_stbi_read_fuzzer vorbis_decode_fuzzer zlib_zlib_uncompress_fuzzer

DanBlackwell added 3 commits July 7, 2024 20:29

First attempt at integrating dgfuzz

566edb3

Fixed DGFuzz fuzzer.py (tested on libpcap only)

910f093

clone DGFuzz from Github

8589be6

Add dummy comment to make gcbrun happy

836f2ee

Make that linter happy

18b6d84

DanBlackwell added 2 commits July 8, 2024 20:00

disable merge with main experiment

73a85d5

Set unsupported_fuzzers correctly for dgfuzz

15fc38d

Update DGFuzz to weight target bytes

7b9d958

Bump DGFuzz version with tweaked parameters

51dcfde

Fix for apt related build error (missing apt-get update)

7b34460

Added dgfuzz_22997c with updated DFSan havoc logic and some mutation …

784244f

…tracking logs

Add extra log for retcode upon coverage measurer failure

c7221b4

Added dgfuzz_0ebf1a9 with bug fixes

7040c8d

Added dgfuzz_7fa3ff with Bloom filtering

611e73c

Added dgfuzz_e989ba with small improvements

3edc8c0

Added setups dgfuzz_1d7283 dgfuzz_32d973 with OOM fixes

b4d4c20

Added dgfuzz_49a0ef with small improvements and libafl cmplog

95b7793

Added dgfuzz_23de6c with const cmplog built in and prescient cmplog d…

09646ec

…isabled

DGFuzz Testing #2000

Are you sure you want to change the base?

DGFuzz Testing #2000

Conversation

DanBlackwell commented Jul 7, 2024 • edited Loading

DonggeLiu commented Jul 7, 2024

DonggeLiu commented Jul 7, 2024

DanBlackwell commented Jul 8, 2024

DonggeLiu commented Jul 8, 2024

DanBlackwell commented Jul 8, 2024

DanBlackwell commented Jul 8, 2024

DonggeLiu commented Jul 8, 2024

DonggeLiu commented Jul 8, 2024

DanBlackwell commented Jul 8, 2024 • edited Loading

DanBlackwell commented Jul 8, 2024

experiment name: 2024-07-08-dgfuzz-1

DanBlackwell commented Jul 8, 2024

DonggeLiu commented Jul 9, 2024

DanBlackwell commented Jul 10, 2024

DonggeLiu commented Jul 11, 2024

DanBlackwell commented Jul 11, 2024

DanBlackwell commented Jul 11, 2024

DonggeLiu commented Jul 11, 2024

DonggeLiu commented Jul 11, 2024

DanBlackwell commented Jul 11, 2024

DanBlackwell commented Jul 11, 2024 • edited Loading

DanBlackwell commented Sep 3, 2024

DonggeLiu commented Sep 3, 2024

DanBlackwell commented Sep 3, 2024

DanBlackwell commented Sep 4, 2024

DonggeLiu commented Sep 4, 2024

DanBlackwell commented Sep 4, 2024 • edited Loading

DanBlackwell commented Sep 5, 2024

DonggeLiu commented Sep 5, 2024

DonggeLiu commented Sep 5, 2024 • edited Loading

DanBlackwell commented Sep 6, 2024

DanBlackwell commented Sep 8, 2024

DonggeLiu commented Sep 9, 2024

DanBlackwell commented Sep 9, 2024

DonggeLiu commented Sep 9, 2024

DonggeLiu commented Sep 9, 2024

DanBlackwell commented Sep 11, 2024

DonggeLiu commented Sep 12, 2024

DanBlackwell commented Sep 12, 2024

DanBlackwell commented Sep 12, 2024

DonggeLiu commented Sep 12, 2024

DanBlackwell commented Sep 15, 2024

DonggeLiu commented Sep 16, 2024

DanBlackwell commented Jul 7, 2024 •

edited

Loading

DanBlackwell commented Jul 8, 2024 •

edited

Loading

experiment name: `2024-07-08-dgfuzz-1`

DanBlackwell commented Jul 11, 2024 •

edited

Loading

DanBlackwell commented Sep 4, 2024 •

edited

Loading

DonggeLiu commented Sep 5, 2024 •

edited

Loading