How to evaluate my fuzzer with Magma? #171
Comments
|
Almost all of the fuzzers integrated into magma pull local files into the
docker container during building. Eg angora has local files under the `src`
directory and these are available inside the container under `$FUZZER/src`
(see angoras fetch.sh). You can just put your entire code base inside a sec
directory and do the same thing; no need to git clone anything.
…On Tue, 10 Sep 2024 at 6:59 PM, 123huahua ***@***.***> wrote:
I want to evaluate my fuzzer with Magma, but I do not have a public git
repo. How can I modify the Docker image to use my local code?Could you
please provide detailed implementation steps?
—
Reply to this email directly, view it on GitHub
<#171>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ACB2DESTNI26YBKS2GO6KJTZV2YGRAVCNFSM6AAAAABN6IKA5GVHI2DSMVQWIX3LMV43ASLTON2WKOZSGUYTKOBTHE2DSMQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
How should I modify the script files in the fuzzer folder and what should I pay attention to? The script copied from other fuzzers cannot run my fuzzer properly. |
|
Because my fuzzer is implemented based on afl++, but it is different from the version of afl++ in magma. I copied those script files to my fuzzer and encountered the following error。Using the same script to run AFL++, everything is OK。 
Dockerfile:8684 | ENV LDFLAGS -L"${OUT}" -g
|
|
You need to build the fuzzer in the container; don’t copy across prebuilt
binaries. It looks like you have a gcc mismatch
…On Tue, 10 Sep 2024 at 10:29 PM, 123huahua ***@***.***> wrote:
Because my fuzzer is implemented based on afl++, but it is different from
the version of afl++ in magma. I copied those script files to my fuzzer and
encountered the following error。Using the same script to run AFL++,
everything is OK。
20240910202710.png (view on web)
<https://github.com/user-attachments/assets/cf5b0bbb-4690-450f-bcf5-45f90dea63a0>
[20/20] RUN /magma/fuzzers/mesfuzz/instrument.sh --progress=plain
--no-cache:
0.439 + CXXFLAGS='-include /magma/magma/src/canary.h
-DMAGMA_ENABLE_CANARIES -DMAGMA_FATAL_CANARIES -g -O0 -stdlib=libc++'
0.439 + export OUT=/magma_out/afl
0.439 + OUT=/magma_out/afl
0.439 + export 'LDFLAGS=-L/magma_out -g -L/magma_out/afl'
0.439 + LDFLAGS='-L/magma_out -g -L/magma_out/afl'
0.439 + /magma/magma/build.sh
0.449 + MAGMA_STORAGE=/magma_shared/canaries.raw
0.449 + /magma/fuzzers/mesfuzz/repo/afl-clang-fast -include
/magma/magma/src/canary.h -DMAGMA_ENABLE_CANARIES -DMAGMA_FATAL_CANARIES -g
-O0 '-DMAGMA_STORAGE="/magma_shared/canaries.raw"' -c
/magma/magma/src/canary.c -fPIC -I /magma/magma/src/ -o
/magma_out/afl/canary.o -L/magma_out -g -L/magma_out/afl
0.456 /magma/fuzzers/mesfuzz/repo/afl-clang-fast:
/lib/x86_64-linux-gnu/libc.so.6: version GLIBC_2.33' not found (required
by /magma/fuzzers/mesfuzz/repo/afl-clang-fast) 0.456
/magma/fuzzers/mesfuzz/repo/afl-clang-fast:
/lib/x86_64-linux-gnu/libc.so.6: version GLIBC_2.34' not found (required
by /magma/fuzzers/mesfuzz/repo/afl-clang-fast)
------------------------------
Dockerfile:86 84 | ENV LDFLAGS -L"${OUT}" -g
85 |
86 | >>> RUN ${FUZZER}/instrument.sh --progress=plain --no-cache
87 |
88 | ENTRYPOINT "${MAGMA}/run.sh"
ERROR: failed to solve: process "/bin/sh -c ${FUZZER}/instrument.sh
--progress=plain --no-cache" did not complete successfully: exit code: 1
—
Reply to this email directly, view it on GitHub
<#171 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ACB2DEV46DXGTHAWKH7327DZV3Q4DAVCNFSM6AAAAABN6IKA5GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBQGU4DGOJVGE>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
Why does the fuzzing process end in just a few seconds when the TIMEOUT is set to 1h?
|
|
It is impossible to debug this based on the information given. I would look
through the magma logs produced in the out directory and see if you can
find any errors. I suspect the fuzzer is crashing.
…On Wed, 11 Sep 2024 at 10:54 PM, 123huahua ***@***.***> wrote:
Why does the fuzzing process end in just a few seconds when the TIMEOUT is
set to 1h?
1.png (view on web)
<https://github.com/user-attachments/assets/6543b868-6fc4-4fbd-a1ab-414094ccd865>
2.png (view on web)
<https://github.com/user-attachments/assets/5a413ab7-5bb4-4fdd-8e8c-63d6c9ee4f97>
3.png (view on web)
<https://github.com/user-attachments/assets/fa5bbf23-d3c6-4ad8-ba44-62439c496b13>
—
Reply to this email directly, view it on GitHub
<#171 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ACB2DES4K7LDUK6WJBBTRIDZWA4PPAVCNFSM6AAAAABN6IKA5GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNBTGU4TAMZVGE>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
I want to evaluate my fuzzer with Magma, but I do not have a public git repo. How can I modify the Docker image to use my local code?Could you please provide detailed implementation steps?
The text was updated successfully, but these errors were encountered: