Cannot combine afl.rs with asan when testing real world programs

[ydongyeon]

Hi, I'v stuck with a problem that I cannot combine afl.rs with asan when it deals with real world programs without any use of 'fuzz' macro that afl.rs serves.

let mut rustflags = format!(
    "--cfg fuzzing \
     -C debug-assertions \
     -C overflow_checks \
     -C passes={} \
     -C codegen-units=1 \
     -C llvm-args=-sanitizer-coverage-level=3 \
     -C llvm-args=-sanitizer-coverage-trace-pc-guard \
     -C llvm-args=-sanitizer-coverage-prune-blocks=0 \
     -C opt-level=3 \
     -C target-cpu=native \
     -C debuginfo=0 \
     -l afl-llvm-rt \
     -L {} ",
    passes,
    common::afl_llvm_rt_dir(None).display()
);

This is the flags which are afl.rs is using. And if I build the program using "RUSTFLAGS="-Zsanitizer=address" cargo afl build" and run the afl fuzzing by "cargo afl fuzz -i in -o out target/debug/exectuable" it shows a problem like below

I think the pass (sancov-module) that afl.rs use is not compatible with asan with those flag options

let mut rustflags = format!(
    "--cfg fuzzing \
     -C debug-assertions \
     -C overflow_checks \
     -C passes={} \
     -C codegen-units=1 \
     -C llvm-args=-sanitizer-coverage-level=3 \
     -C llvm-args=-sanitizer-coverage-trace-pc-guard \
     -C llvm-args=-sanitizer-coverage-prune-blocks=0 \
     -C opt-level=3 \
     -C target-cpu=native \
     -C debuginfo=0 \
     -l afl-llvm-rt \
     -L {} ",

I've tried changing opt-level to 0, but it still does not solve the problem. Is there any way to use afl.rs with asan ?