For the general instruction manual, see docs/README.md.
The code in this directory allows you to build a standalone feature that leverages the QEMU "user emulation" mode and allows callers to obtain instrumentation output for black-box, closed-source binaries. This mechanism can be then used by afl-fuzz to stress-test targets that couldn't be built with afl-cc.
The usual performance cost is 2-5x, which is considerably better than seen so far in experiments with tools such as DynamoRIO and PIN.
The idea and much of the initial implementation comes from Andrew Griffiths. The actual implementation on current QEMU (shipped as qemuafl) is from Andrea Fioraldi. Special thanks to abiondo that re-enabled TCG chaining.
The feature is implemented with a patched QEMU. The simplest way to build it is to run ./build_qemu_support.sh. The script will download, configure, and compile the QEMU binary for you.
QEMU is a big project, so this will take a while, and you may have to resolve a couple of dependencies (most notably, you will definitely need libtool and glib2-devel).
Once the binaries are compiled, you can leverage the QEMU tool by calling
afl-fuzz and all the related utilities with -Q in the command line.
Note that QEMU requires a generous memory limit to run; somewhere around 200 MB
is a good starting point, but considerably more may be needed for more complex
programs. The default -m limit will be automatically bumped up to 200 MB when
specifying -Q to afl-fuzz; be careful when overriding this.
In principle, if you set CPU_TARGET before calling ./build_qemu_support.sh,
you should get a build capable of running non-native binaries (say, you can try
CPU_TARGET=arm). This is also necessary for running 32-bit binaries on a
64-bit system (CPU_TARGET=i386). If you're trying to run QEMU on a different
architecture, you can also set HOST to the cross-compiler prefix to use (for
example HOST=arm-linux-gnueabi to use arm-linux-gnueabi-gcc).
You can also compile statically-linked binaries by setting STATIC=1. This can
be useful when compiling QEMU on a different system than the one you're planning
to run the fuzzer on and is most often used with the HOST variable.
Note: when targeting the i386 architecture, on some binaries the forkserver handshake may fail due to the lack of reserved memory. Fix it with:
export QEMU_RESERVED_VA=0x1000000
Note: if you want the QEMU helper to be installed on your system for all users,
you need to build it before issuing make install in the parent directory.
If you want to specify a different path for libraries (e.g., to run an arm64
binary on x86_64) use QEMU_LD_PREFIX.
As for LLVM mode (refer to instrumentation/README.llvm.md for mode details), QEMU mode supports the deferred initialization.
This can be enabled by setting the environment variable AFL_ENTRYPOINT which
allows to move the forkserver to a different part, e.g., just before the file is
opened (e.g., way after command line parsing and config file loading, etc.)
which can be a huge speed improvement.
AFL++'s QEMU mode now supports also persistent mode for x86, x86_64, arm, and aarch64 targets. This increases the speed by several factors, however, it is a bit of work to set up - but worth the effort.
For more information, see README.persistent.md.
As an extension to persistent mode, qemuafl can snapshot and restore the memory state and brk(). For details, see README.persistent.md.
The environment variable that enables the ready to use snapshot mode is
AFL_QEMU_SNAPSHOT and takes a hex address as a value that is the snapshot
entry point.
Snapshot mode can work restoring all the writeable pages, that is typically slower than fork() mode but, on the other hand, it can scale better with multicore. If the AFL++ snapshot kernel module is loaded, qemuafl will use it and, in this case, the speed is better than fork() and also the scaling capabilities.
You can tell QEMU to instrument only a part of the address space.
Just set AFL_QEMU_INST_RANGES=A,B,C....
The format of the items in the list is either a range of addresses like 0x123-0x321 or a module name like module.so (that is matched in the mapped object filename).
Alternatively, you can tell QEMU to ignore part of an address space for instrumentation.
Just set AFL_QEMU_EXCLUDE_RANGES=A,B,C....
The format of the items on the list is the same as for AFL_QEMU_INST_RANGES
and excluding ranges takes priority over any included ranges or AFL_INST_LIBS.
CompareCoverage is a sub-instrumentation with effects similar to laf-intel.
You have to set AFL_PRELOAD=/path/to/libcompcov.so together with setting the
AFL_COMPCOV_LEVEL you want to enable it.
AFL_COMPCOV_LEVEL=1 is to instrument comparisons with only immediate
values/read-only memory.
AFL_COMPCOV_LEVEL=2 instruments all comparison instructions and memory
comparison functions when libcompcov is preloaded.
AFL_COMPCOV_LEVEL=3 has the same effects of AFL_COMPCOV_LEVEL=2 but enables
also the instrumentation of the floating-point comparisons on x86 and x86_64
(experimental).
Integer comparison instructions are currently instrumented only on the x86, x86_64, arm, and aarch64 targets.
Recommended, but not as good as CMPLOG mode (see below).
Another new feature is CMPLOG, which is based on the Redqueen project. Here all immediates in CMP instructions are learned and put into a dynamic dictionary and applied to all locations in the input that reached that CMP, trying to solve and pass it. This is a very effective feature and it is available for x86, x86_64, arm, and aarch64.
To enable it, you must pass on the command line of afl-fuzz:
-c /path/to/your/target
AFL++ QEMU can use Wine to fuzz Win32 PE binaries. Use the -W flag of
afl-fuzz.
Note that some binaries require user interaction with the GUI and must be patched.
For examples, look here.
The feature is supported only on Linux. Supporting BSD may amount to porting the changes made to linux-user/elfload.c and applying them to bsd-user/elfload.c, but I have not looked into this yet.
The instrumentation follows only the .text section of the first ELF binary encountered in the linking process. It does not trace shared libraries. In practice, this means two things:
-
Any libraries you want to analyze must be linked statically into the executed ELF file (this will usually be the case for closed-source apps).
-
Standard C libraries and other stuff that is wasteful to instrument should be linked dynamically - otherwise, AFL++ will have no way to avoid peeking into them.
Setting AFL_INST_LIBS=1 can be used to circumvent the .text detection logic
and instrument every basic block encountered.
If you want to compare the performance of the QEMU instrumentation with that of afl-clang-fast compiled code against the same target, you need to build the non-instrumented binary with the same optimization flags that are normally injected by afl-clang-fast, and make sure that the bits to be tested are statically linked into the binary. A common way to do this would be:
CFLAGS="-O3 -funroll-loops" ./configure --disable-shared
make clean all
Comparative measurements of execution speed or instrumentation coverage will be fairly meaningless if the optimization levels or instrumentation scopes don't match.
With AFL_QEMU_FORCE_DFL, you force QEMU to ignore the registered signal
handlers of the target.
If you need to fix up checksums or do other cleanups on mutated test cases, see
afl_custom_post_process in custom_mutators/examples/example.c for a viable
solution.
Do not mix QEMU mode with ASAN, MSAN, or the likes; QEMU doesn't appreciate the "shadow VM" trick employed by the sanitizers and will probably just run out of memory.
Compared to fully-fledged virtualization, the user emulation mode is NOT a security boundary. The binaries can freely interact with the host OS. If you somehow need to fuzz an untrusted binary, put everything in a sandbox first.
QEMU does not necessarily support all CPU or hardware features that your target
program may be utilizing. In particular, it does not appear to have full support
for AVX2/FMA3. Using binaries for older CPUs or recompiling them with
-march=core2, can help.
Beyond that, this is an early-stage mechanism, so fields reports are welcome. You can send them to [email protected].
Statically rewriting binaries just once, instead of attempting to translate them at run time, can be a faster alternative. That said, static rewriting is fraught with peril, because it depends on being able to properly and fully model program control flow without actually executing each and every code path.
For more information and hints, check out docs/fuzzing_binary-only_targets.md.