diff options
103 files changed, 6182 insertions, 1098 deletions
diff --git a/.custom-format.py b/.custom-format.py index 6f1b0bfa..60f6d9c3 100755 --- a/.custom-format.py +++ b/.custom-format.py @@ -32,7 +32,8 @@ if CLANG_FORMAT_BIN is None: p = subprocess.Popen(["clang-format-10", "--version"], stdout=subprocess.PIPE) o, _ = p.communicate() o = str(o, "utf-8") - o = o[len("clang-format version "):].strip() + o = re.sub(r".*ersion ", "", o) + #o = o[len("clang-format version "):].strip() o = o[:o.find(".")] o = int(o) except: diff --git a/.dockerignore b/.dockerignore new file mode 100644 index 00000000..d05bf1c6 --- /dev/null +++ b/.dockerignore @@ -0,0 +1,65 @@ +.test +.test2 +.sync_tmp +*.o +*.so +*.pyc +*.dSYM +as +ld +in +out +core* +afl-analyze +afl-as +afl-clang +afl-clang\+\+ +afl-clang-fast +afl-clang-fast\+\+ +afl-clang-lto +afl-clang-lto\+\+ +afl-fuzz +afl-g\+\+ +afl-gcc +afl-gcc-fast +afl-g\+\+-fast +afl-gotcpu +afl-ld +afl-ld-lto +afl-qemu-trace +afl-showmap +afl-tmin +afl-analyze.8 +afl-as.8 +afl-clang-fast\+\+.8 +afl-clang-fast.8 +afl-clang-lto.8 +afl-clang-lto\+\+.8 +afl-cmin.8 +afl-cmin.bash.8 +afl-fuzz.8 +afl-gcc.8 +afl-gcc-fast.8 +afl-g\+\+-fast.8 +afl-gotcpu.8 +afl-plot.8 +afl-showmap.8 +afl-system-config.8 +afl-tmin.8 +afl-whatsup.8 +qemu_mode/libcompcov/compcovtest +qemu_mode/qemu-* +unicorn_mode/samples/*/\.test-* +unicorn_mode/samples/*/output +unicorn_mode/unicornafl +test/unittests/unit_maybe_alloc +test/unittests/unit_preallocable +test/unittests/unit_list +test/unittests/unit_rand +test/unittests/unit_hash +examples/afl_network_proxy/afl-network-server +examples/afl_network_proxy/afl-network-client +examples/afl_frida/afl-frida +examples/afl_frida/libtestinstr.so +examples/afl_frida/frida-gum-example.c +examples/afl_frida/frida-gum.h \ No newline at end of file diff --git a/.gitignore b/.gitignore index 1000cc6f..b2c2fc62 100644 --- a/.gitignore +++ b/.gitignore @@ -1,8 +1,15 @@ .test .test2 +.sync_tmp *.o *.so *.pyc +*.dSYM +as +ld +in +out +core* afl-analyze afl-as afl-clang @@ -41,13 +48,10 @@ afl-system-config.8 afl-tmin.8 afl-whatsup.8 qemu_mode/libcompcov/compcovtest -as -ld qemu_mode/qemu-* unicorn_mode/samples/*/\.test-* unicorn_mode/samples/*/output/ unicorn_mode/unicornafl -core\.* test/unittests/unit_maybe_alloc test/unittests/unit_preallocable test/unittests/unit_list @@ -55,3 +59,7 @@ test/unittests/unit_rand test/unittests/unit_hash examples/afl_network_proxy/afl-network-server examples/afl_network_proxy/afl-network-client +examples/afl_frida/afl-frida +examples/afl_frida/libtestinstr.so +examples/afl_frida/frida-gum-example.c +examples/afl_frida/frida-gum.h diff --git a/.travis.yml b/.travis.yml index 4271b821..96cdc86c 100644 --- a/.travis.yml +++ b/.travis.yml @@ -51,6 +51,7 @@ script: - gcc -v - clang -v - sudo -E ./afl-system-config + - sudo sysctl -w kernel.shmmax=10000000000 - if [ "$TRAVIS_OS_NAME" = "osx" ]; then export LLVM_CONFIG=`pwd`/"$NAME" ; make source-only ASAN_BUILD=1 ; fi - if [ "$TRAVIS_OS_NAME" = "linux" -a "$TRAVIS_CPU_ARCH" = "amd64" ]; then make distrib ASAN_BUILD=1 ; fi - if [ "$TRAVIS_CPU_ARCH" = "arm64" ] ; then echo DEBUG ; find / -name llvm-config.h 2>/dev/null; apt-cache search clang | grep clang- ; apt-cache search llvm | grep llvm- ; dpkg -l | egrep 'clang|llvm'; echo DEBUG ; export LLVM_CONFIG=llvm-config-6.0 ; make ASAN_BUILD=1 ; cd qemu_mode && sh ./build_qemu_support.sh ; cd .. ; fi diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 90049432..ccacef5f 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -1,5 +1,7 @@ # How to submit a Pull Request to AFLplusplus +All contributions (pull requests) must be made against our `dev` branch. + Each modified source file, before merging, must be formatted. ``` @@ -18,5 +20,5 @@ No camel case at all and use the AFL's macros wherever possible (e.g. WARNF, FATAL, MAP_SIZE, ...). Remember that AFLplusplus has to build and run on many platforms, so -generalize your Makefiles (or your patches to our pre-existing Makefiles) -to be as much generic as possible. +generalize your Makefiles/GNUmakefile (or your patches to our pre-existing +Makefiles) to be as much generic as possible. diff --git a/Dockerfile b/Dockerfile index 905e8265..64b04ba6 100644 --- a/Dockerfile +++ b/Dockerfile @@ -5,7 +5,7 @@ # has focal has gcc-10 but not g++-10 ... # -FROM ubuntu:20.04 +FROM ubuntu:20.04 AS aflplusplus MAINTAINER afl++ team <afl@aflplus.plus> LABEL "about"="AFLplusplus docker image" @@ -20,11 +20,11 @@ RUN apt-get update && apt-get upgrade -y && \ python3 python3-dev python3-setuptools python-is-python3 \ libtool libtool-bin \ libglib2.0-dev \ - wget vim jupp nano \ + wget vim jupp nano bash-completion \ apt-utils apt-transport-https ca-certificates gnupg dialog \ libpixman-1-dev -RUN echo deb http://apt.llvm.org/focal/ llvm-toolchain-focal main >> /etc/apt/sources.list && \ +RUN echo deb http://apt.llvm.org/focal/ llvm-toolchain-focal-11 main >> /etc/apt/sources.list && \ wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add - RUN echo deb http://ppa.launchpad.net/ubuntu-toolchain-r/test/ubuntu focal main >> /etc/apt/sources.list && \ @@ -46,17 +46,19 @@ RUN update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-10 0 RUN rm -rf /var/cache/apt/archives/* -ARG CC=gcc-10 -ARG CXX=g++-10 -ARG LLVM_CONFIG=llvm-config-11 +ENV LLVM_CONFIG=llvm-config-11 +ENV AFL_SKIP_CPUFREQ=1 -RUN git clone https://github.com/AFLplusplus/AFLplusplus -RUN cd AFLplusplus && export REAL_CXX=g++-10 && make distrib && \ - make install && make clean +RUN git clone https://github.com/vanhauser-thc/afl-cov /afl-cov +RUN cd /afl-cov && make install && cd .. -RUN git clone https://github.com/vanhauser-thc/afl-cov afl-cov -RUN cd afl-cov && make install +COPY . /AFLplusplus +WORKDIR /AFLplusplus -RUN echo 'alias joe="jupp --wordwrap"' >> ~/.bashrc +RUN export REAL_CXX=g++-10 && export CC=gcc-10 && \ + export CXX=g++-10 && make clean && \ + make distrib && make install && make clean -ENV AFL_SKIP_CPUFREQ=1 +RUN echo 'alias joe="jupp --wordwrap"' >> ~/.bashrc +RUN echo 'export PS1="[afl++]$PS1"' >> ~/.bashrc +ENV IS_DOCKER="1" diff --git a/GNUmakefile b/GNUmakefile index 748cd73c..e0d89274 100644 --- a/GNUmakefile +++ b/GNUmakefile @@ -56,8 +56,10 @@ endif ifneq "$(shell uname)" "Darwin" ifeq "$(shell echo 'int main() {return 0; }' | $(CC) $(CFLAGS) -Werror -x c - -march=native -o .test 2>/dev/null && echo 1 || echo 0 ; rm -f .test )" "1" + ifndef SOURCE_DATE_EPOCH #CFLAGS_OPT += -march=native SPECIAL_PERFORMANCE += -march=native + endif endif # OS X does not like _FORTIFY_SOURCE=2 CFLAGS_OPT += -D_FORTIFY_SOURCE=2 @@ -65,7 +67,7 @@ endif ifeq "$(shell uname)" "SunOS" CFLAGS_OPT += -Wno-format-truncation - LDFLAGS=-lkstat + LDFLAGS=-lkstat -lrt endif ifdef STATIC @@ -96,7 +98,7 @@ ifneq "$(shell uname -m)" "x86_64" endif CFLAGS ?= -O3 -funroll-loops $(CFLAGS_OPT) -override CFLAGS += -Wall -g -Wno-pointer-sign -Wmissing-declarations\ +override CFLAGS += -Wall -g -Wno-pointer-sign \ -I include/ -DAFL_PATH=\"$(HELPER_PATH)\" \ -DBIN_PATH=\"$(BIN_PATH)\" -DDOC_PATH=\"$(DOC_PATH)\" @@ -196,7 +198,7 @@ else endif ifneq "$(filter Linux GNU%,$(shell uname))" "" - LDFLAGS += -ldl + LDFLAGS += -ldl -lrt endif ifneq "$(findstring FreeBSD, $(shell uname))" "" @@ -254,13 +256,13 @@ ifeq "$(shell echo '$(HASH)include <sys/ipc.h>@$(HASH)include <sys/shm.h>@int ma else SHMAT_OK=0 override CFLAGS+=-DUSEMMAP=1 - LDFLAGS += -Wno-deprecated-declarations -lrt + LDFLAGS += -Wno-deprecated-declarations endif ifdef TEST_MMAP SHMAT_OK=0 override CFLAGS += -DUSEMMAP=1 - LDFLAGS += -Wno-deprecated-declarations -lrt + LDFLAGS += -Wno-deprecated-declarations endif all: test_x86 test_shm test_python ready $(PROGS) afl-as test_build all_done @@ -455,10 +457,10 @@ code-format: ./.custom-format.py -i llvm_mode/*.h ./.custom-format.py -i llvm_mode/*.cc ./.custom-format.py -i gcc_plugin/*.c - #./.custom-format.py -i gcc_plugin/*.h + @#./.custom-format.py -i gcc_plugin/*.h ./.custom-format.py -i gcc_plugin/*.cc ./.custom-format.py -i custom_mutators/*/*.c - ./.custom-format.py -i custom_mutators/*/*.h + @#./.custom-format.py -i custom_mutators/*/*.h # destroys input.h :-( ./.custom-format.py -i examples/*/*.c ./.custom-format.py -i examples/*/*.h ./.custom-format.py -i test/*.c diff --git a/README.md b/README.md index 14a42b7e..bd2784ae 100644 --- a/README.md +++ b/README.md @@ -1,4 +1,4 @@ -# american fuzzy lop plus plus (afl++) +# American Fuzzy Lop plus plus (afl++) <img align="right" src="https://raw.githubusercontent.com/andreafioraldi/AFLplusplus-website/master/static/logo_256x256.png" alt="AFL++ Logo"> @@ -8,61 +8,36 @@ Github Version: 2.66d - includes all necessary/interesting changes from Google's afl 2.56b - - Originally developed by Michal "lcamtuf" Zalewski. - Repository: [https://github.com/AFLplusplus/AFLplusplus](https://github.com/AFLplusplus/AFLplusplus) afl++ is maintained by: - * Marc "van Hauser" Heuse <mh@mh-sec.de>, - * Heiko "hexcoder-" Eißfeldt <heiko.eissfeldt@hexco.de>, - * Andrea Fioraldi <andreafioraldi@gmail.com> and - * Dominik Maier <mail@dmnk.co>. - - Note that although afl now has a Google afl repository [https://github.com/Google/afl](https://github.com/Google/afl), - it is unlikely to receive any notable enhancements: [https://twitter.com/Dor3s/status/1154737061787660288](https://twitter.com/Dor3s/status/1154737061787660288) - -## The enhancements compared to the original stock afl - - Many improvements were made over the official afl release - which did not - get any feature improvements since November 2017. - - Among other changes afl++ has a more performant llvm_mode, supports - llvm up to version 11, QEMU 3.1, more speed and crashfixes for QEMU, - better *BSD and Android support and much, much more. - - Additionally the following features and patches have been integrated: - - * AFLfast's power schedules by Marcel Böhme: [https://github.com/mboehme/aflfast](https://github.com/mboehme/aflfast) - - * The new excellent MOpt mutator: [https://github.com/puppet-meteor/MOpt-AFL](https://github.com/puppet-meteor/MOpt-AFL) - - * InsTrim, a very effective CFG llvm_mode instrumentation implementation for large targets: [https://github.com/csienslab/instrim](https://github.com/csienslab/instrim) - - * C. Holler's afl-fuzz Python mutator module and llvm_mode instrument file support: [https://github.com/choller/afl](https://github.com/choller/afl) - - * Custom mutator by a library (instead of Python) by kyakdan - * Unicorn mode which allows fuzzing of binaries from completely different platforms (integration provided by domenukk) - - * LAF-Intel or CompCov support for llvm_mode, qemu_mode and unicorn_mode + * Marc "van Hauser" Heuse <mh@mh-sec.de>, + * Heiko "hexcoder-" Eißfeldt <heiko.eissfeldt@hexco.de>, + * Andrea Fioraldi <andreafioraldi@gmail.com> and + * Dominik Maier <mail@dmnk.co>. - * NeverZero patch for afl-gcc, llvm_mode, qemu_mode and unicorn_mode which prevents a wrapping map value to zero, increases coverage - - * Persistent mode and deferred forkserver for qemu_mode - - * Win32 PE binary-only fuzzing with QEMU and Wine + Originally developed by Michał "lcamtuf" Zalewski. - * Radamsa mutator (as a custom mutator). + afl++ is a superiour fork to Google's afl - more speed, more and better + mutations, more and better instrumentation, custom module support, etc. - * QBDI mode to fuzz android native libraries via QBDI framework +## Contents - * The new CmpLog instrumentation for LLVM and QEMU inspired by [Redqueen](https://www.syssec.ruhr-uni-bochum.de/media/emma/veroeffentlichungen/2018/12/17/NDSS19-Redqueen.pdf) + 1. [Features](#important-features-of-afl) + 2. [How to compile and install afl++](#building-and-installing-afl) + 3. [How to fuzz a target](#how-to-fuzz-with-afl) + 4. [Fuzzing binary-only targets](#fuzzing-binary-only-targets) + 5. [Good examples and writeups of afl++ usages](#good-examples-and-writeups) + 6. [Branches](#branches) + 7. [Want to help?](#help-wanted) + 8. [Detailed help and description of afl++](#challenges-of-guided-fuzzing) - * LLVM mode Ngram coverage by Adrian Herrera [https://github.com/adrianherrera/afl-ngram-pass](https://github.com/adrianherrera/afl-ngram-pass) +## Important features of afl++ - A more thorough list is available in the [PATCHES](docs/PATCHES.md) file. + afl++ supports llvm up to version 12, very fast binary fuzzing with QEMU 3.1 + with laf-intel and redqueen, unicorn mode, gcc plugin, full *BSD, Solaris and + Android support and much, much, much more. | Feature/Instrumentation | afl-gcc | llvm_mode | gcc_plugin | qemu_mode | unicorn_mode | | ----------------------- |:-------:|:---------:|:----------:|:----------------:|:------------:| @@ -75,23 +50,37 @@ | InsTrim | | x | | | | | Ngram prev_loc coverage | | x(6) | | | | | Context coverage | | x | | | | + | Auto dictionary | | x(7) | | | | | Snapshot LKM support | | x | | (x)(5) | | - neverZero: + 1. default for LLVM >= 9.0, env var for older version due an efficiency bug in llvm <= 8 + 2. GCC creates non-performant code, hence it is disabled in gcc_plugin + 3. partially via AFL_CODE_START/AFL_CODE_END + 4. with pcguard mode and LTO mode for LLVM >= 11 + 5. upcoming, development in the branch + 6. not compatible with LTO instrumentation and needs at least LLVM >= 4.1 + 7. only in LTO mode with LLVM >= 11 - (1) default for LLVM >= 9.0, env var for older version due an efficiency bug in llvm <= 8 + Among others, the following features and patches have been integrated: - (2) GCC creates non-performant code, hence it is disabled in gcc_plugin - - (3) partially via AFL_CODE_START/AFL_CODE_END - - (4) Only for LLVM >= 11 and not all targets compile - - (5) upcoming, development in the branch + * NeverZero patch for afl-gcc, llvm_mode, qemu_mode and unicorn_mode which prevents a wrapping map value to zero, increases coverage + * Persistent mode and deferred forkserver for qemu_mode + * Unicorn mode which allows fuzzing of binaries from completely different platforms (integration provided by domenukk) + * The new CmpLog instrumentation for LLVM and QEMU inspired by [Redqueen](https://www.syssec.ruhr-uni-bochum.de/media/emma/veroeffentlichungen/2018/12/17/NDSS19-Redqueen.pdf) + * Win32 PE binary-only fuzzing with QEMU and Wine + * AFLfast's power schedules by Marcel Böhme: [https://github.com/mboehme/aflfast](https://github.com/mboehme/aflfast) + * The MOpt mutator: [https://github.com/puppet-meteor/MOpt-AFL](https://github.com/puppet-meteor/MOpt-AFL) + * LLVM mode Ngram coverage by Adrian Herrera [https://github.com/adrianherrera/afl-ngram-pass](https://github.com/adrianherrera/afl-ngram-pass) + * InsTrim, an effective CFG llvm_mode instrumentation implementation for large targets: [https://github.com/csienslab/instrim](https://github.com/csienslab/instrim) + * C. Holler's afl-fuzz Python mutator module and llvm_mode instrument file support: [https://github.com/choller/afl](https://github.com/choller/afl) + * Custom mutator by a library (instead of Python) by kyakdan + * LAF-Intel/CompCov support for llvm_mode, qemu_mode and unicorn_mode (with enhanced capabilities) + * Radamsa and hongfuzz mutators (as custom mutators). + * QBDI mode to fuzz android native libraries via Quarkslab's [QBDI](https://github.com/QBDI/QBDI) framework - (6) not compatible with LTO instrumentation and needs at least LLVM >= 4.1 + A more thorough list is available in the [PATCHES](docs/PATCHES.md) file. - So all in all this is the best-of afl that is currently out there :-) + So all in all this is the best-of afl that is out there :-) For new versions and additional information, check out: [https://github.com/AFLplusplus/AFLplusplus](https://github.com/AFLplusplus/AFLplusplus) @@ -106,7 +95,7 @@ The following branches exist: - * [master/trunk](https://github.com/AFLplusplus/AFLplusplus/) : stable state of afl++ - it is synced from dev from time to + * [stable/trunk](https://github.com/AFLplusplus/AFLplusplus/) : stable state of afl++ - it is synced from dev from time to time when we are satisfied with it's stability * [dev](https://github.com/AFLplusplus/AFLplusplus/tree/dev) : development state of afl++ - bleeding edge and you might catch a checkout which does not compile or has a bug. *We only accept PRs in dev!!* @@ -115,7 +104,7 @@ For releases, please see the [Releases](https://github.com/AFLplusplus/AFLplusplus/releases) tab. -## Google Summer of Code 2020 (and any other students and enthusiast developers) +## Help wanted We are happy to be part of [Google Summer of Code 2020](https://summerofcode.withgoogle.com/organizations/5100744400699392/)! :-) @@ -140,7 +129,7 @@ hence afl-clang-lto is available!) or just pull directly from the docker hub: docker pull aflplusplus/aflplusplus docker run -ti -v /location/of/your/target:/src aflplusplus/aflplusplus ``` -This container is automatically generated when a push to master happens. +This image is automatically generated when a push to the stable repo happens. You will find your target source code in /src in the container. If you want to build afl++ yourself you have many options. @@ -151,11 +140,11 @@ sudo apt install build-essential libtool-bin python3-dev automake flex bison lib make distrib sudo make install ``` -It is recommended to install the newest available gcc and clang and llvm-dev +It is recommended to install the newest available gcc, clang and llvm-dev possible in your distribution! Note that "make distrib" also builds llvm_mode, qemu_mode, unicorn_mode and -more. If you just want plain afl then do "make all", however compiling and +more. If you just want plain afl++ then do "make all", however compiling and using at least llvm_mode is highly recommended for much better results - hence in this case @@ -197,205 +186,552 @@ These build options exist: e.g.: make ASAN_BUILD=1 -## Challenges of guided fuzzing +## Good examples and writeups -Fuzzing is one of the most powerful and proven strategies for identifying -security issues in real-world software; it is responsible for the vast -majority of remote code execution and privilege escalation bugs found to date -in security-critical software. +Here are some good writeups to show how to effectively use AFL++: -Unfortunately, fuzzing is also relatively shallow; blind, random mutations -make it very unlikely to reach certain code paths in the tested code, leaving -some vulnerabilities firmly outside the reach of this technique. + * [https://aflplus.plus/docs/tutorials/libxml2_tutorial/](https://aflplus.plus/docs/tutorials/libxml2_tutorial/) + * [https://bananamafia.dev/post/gb-fuzz/](https://bananamafia.dev/post/gb-fuzz/) + * [https://securitylab.github.com/research/fuzzing-challenges-solutions-1](https://securitylab.github.com/research/fuzzing-challenges-solutions-1) + * [https://securitylab.github.com/research/fuzzing-software-2](https://securitylab.github.com/research/fuzzing-software-2) + * [https://securitylab.github.com/research/fuzzing-sockets-FTP](https://securitylab.github.com/research/fuzzing-sockets-FTP) -There have been numerous attempts to solve this problem. One of the early -approaches - pioneered by Tavis Ormandy - is corpus distillation. The method -relies on coverage signals to select a subset of interesting seeds from a -massive, high-quality corpus of candidate files, and then fuzz them by -traditional means. The approach works exceptionally well but requires such -a corpus to be readily available. In addition, block coverage measurements -provide only a very simplistic understanding of the program state and are less -useful for guiding the fuzzing effort in the long haul. +If you are interested in fuzzing structured data (where you define what the +structure is), these links have you covered: + * Superion for afl++: [https://github.com/adrian-rt/superion-mutator](https://github.com/adrian-rt/superion-mutator) + * libprotobuf raw: [https://github.com/bruce30262/libprotobuf-mutator_fuzzing_learning/tree/master/4_libprotobuf_aflpp_custom_mutator](https://github.com/bruce30262/libprotobuf-mutator_fuzzing_learning/tree/master/4_libprotobuf_aflpp_custom_mutator) + * libprotobuf for old afl++ API: [https://github.com/thebabush/afl-libprotobuf-mutator](https://github.com/thebabush/afl-libprotobuf-mutator) -Other, more sophisticated research has focused on techniques such as program -flow analysis ("concolic execution"), symbolic execution, or static analysis. -All these methods are extremely promising in experimental settings, but tend -to suffer from reliability and performance problems in practical uses - and -currently do not offer a viable alternative to "dumb" fuzzing techniques. +If you find other good ones, please send them to us :-) +## How to fuzz with afl++ -## The afl-fuzz approach +The following describes how to fuzz with a target if source code is available. +If you have a binary-only target please skip to [#Instrumenting binary-only apps](#Instrumenting binary-only apps) -American Fuzzy Lop is a brute-force fuzzer coupled with an exceedingly simple -but rock-solid instrumentation-guided genetic algorithm. It uses a modified -form of edge coverage to effortlessly pick up subtle, local-scale changes to -program control flow. +Fuzzing source code is a three step process. -Simplifying a bit, the overall algorithm can be summed up as: +1. compile the target with a special compiler that prepares the target to be + fuzzed efficiently. This step is called "instrumenting a target". +2. Prepare the fuzzing by selecting and optimizing the input corpus for the + target. +3. perform the fuzzing of the target by randomly mutating input and assessing + if a generated input was processed in a new path in the target binary. - 1) Load user-supplied initial test cases into the queue, +### 1. Instrumenting that target - 2) Take the next input file from the queue, +#### a) Selecting the best afl++ compiler for instrumenting the target - 3) Attempt to trim the test case to the smallest size that doesn't alter - the measured behavior of the program, +afl++ comes with different compilers and instrumentation options. +The following evaluation flow will help you to select the best possible. - 4) Repeatedly mutate the file using a balanced and well-researched variety - of traditional fuzzing strategies, +It is highly recommended to have the newest llvm version possible installed, +anything below 9 is not recommended. - 5) If any of the generated mutations resulted in a new state transition - recorded by the instrumentation, add mutated output as a new entry in the - queue. +``` ++--------------------------------+ +| clang/clang++ 11+ is available | --> use afl-clang-lto and afl-clang-lto++ ++--------------------------------+ see [llvm/README.lto.md](llvm/README.lto.md) + | + | if not, or if the target fails with afl-clang-lto/++ + | + v ++---------------------------------+ +| clang/clang++ 3.3+ is available | --> use afl-clang-fast and afl-clang-fast++ ++---------------------------------+ see [llvm/README.md](llvm/README.md) + | + | if not, or if the target fails with afl-clang-fast/++ + | + v + +--------------------------------+ + | if you want to instrument only | -> use afl-gcc-fast and afl-gcc-fast++ + | parts of the target | see [gcc_plugin/README.md](gcc_plugin/README.md) and + +--------------------------------+ [gcc_plugin/README.instrument_file.md](gcc_plugin/README.instrument_file.md) + | + | if not, or if you do not have a gcc with plugin support + | + v + use afl-gcc and afl-g++ (or afl-clang and afl-clang++) +``` - 6) Go to 2. +Clickable README links for the chosen compiler: + + * [afl-clang-lto](llvm/README.lto.md) + * [afl-clang-fast](llvm/README.md) + * [afl-gcc-fast](gcc_plugin/README.md) + * afl-gcc has no README as it has no features + +#### b) Selecting instrumentation options + +The following options are available when you instrument with afl-clang-fast or +afl-clang-lto: + + * Splitting integer, string, float and switch comparisons so afl++ can easier + solve these. This is an important option if you do not have a very good + and large input corpus. This technique is called laf-intel or COMPCOV. + To use this set the following environment variable before compiling the + target: `export AFL_LLVM_LAF_ALL=1` + You can read more about this in [llvm/README.laf-intel.md](llvm/README.laf-intel.md) + * A different technique (and usually a bit better than laf-intel) is to + instrument the target so that any compare values in the target are sent to + afl++ which then tries to put this value into the fuzzing data at different + locations. This technique is very fast and good - if the target does not + transform input data before comparison. Therefore this technique is called + `input to state` or `redqueen`. + If you want to use this technique, then you have to compile the target + twice, once specifically with/for this mode, and pass this binary to afl-fuzz + via the `-c` parameter. + You can read more about this in [llvm_mode/README.cmplog.md](llvm_mode/README.cmplog.md) + +If you use afl-clang-fast, afl-clang-lto or afl-gcc-fast you have the option to +selectively only instrument parts of the target that you are interested in: + + * To instrument only those parts of the target that you are interested in + create a file with all the filenames of the source code that should be + instrumented. + For afl-clang-lto and afl-gcc-fast - or afl-clang-fast if either the clang + version is < 7 or the CLASSIC instrumentation is used - just put one + filename per line, no directory information necessary, and set + `export AFL_LLVM_INSTRUMENT_FILE=yourfile.txt` + see [llvm_mode/README.instrument_file.md](llvm_mode/README.instrument_file.md) + For afl-clang-fast > 6.0 or if PCGUARD instrumentation is used then use the + llvm sancov allow-list feature: [http://clang.llvm.org/docs/SanitizerCoverage.html](http://clang.llvm.org/docs/SanitizerCoverage.html) + +There are many more options and modes available however these are most of the +time less effective. See: + * [llvm_mode/README.ctx.md](llvm_mode/README.ctx.md) + * [llvm_mode/README.ngram.md](llvm_mode/README.ngram.md) + * [llvm_mode/README.instrim.md](llvm_mode/README.instrim.md) + * [llvm_mode/README.neverzero.md](llvm_mode/README.neverzero.md) + +#### c) Modify the target + +If the target has features that makes fuzzing more difficult, e.g. +checksums, HMAC etc. then modify the source code so that this is +removed. +This can even be done for productional source code be eliminating +these checks within this specific defines: -The discovered test cases are also periodically culled to eliminate ones that -have been obsoleted by newer, higher-coverage finds; and undergo several other -instrumentation-driven effort minimization steps. +``` +#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION + // say that the checksum or HMAC was fine - or whatever is required + // to eliminate the need for the fuzzer to guess the right checksum + return 0; +#endif +``` -As a side result of the fuzzing process, the tool creates a small, -self-contained corpus of interesting test cases. These are extremely useful -for seeding other, labor- or resource-intensive testing regimes - for example, -for stress-testing browsers, office applications, graphics suites, or -closed-source tools. +#### d) Instrument the target -The fuzzer is thoroughly tested to deliver out-of-the-box performance far -superior to blind fuzzing or coverage-only tools. +In this step the target source code is compiled so that it can be fuzzed. +Basically you have to tell the target build system that the selected afl++ +compiler is used. Also - if possible - you should always configure the +build system that the target is compiled statically and not dynamically. +How to do this is described below. -## Instrumenting programs for use with AFL +Then build the target. (Usually with `make`) -PLEASE NOTE: llvm_mode compilation with afl-clang-fast/afl-clang-fast++ -instead of afl-gcc/afl-g++ is much faster and has many cool features. -See llvm_mode/ - however few code does not compile with llvm. -We support llvm versions 3.4 to 11. +##### configure -When source code is available, instrumentation can be injected by a companion -tool that works as a drop-in replacement for gcc or clang in any standard build -process for third-party code. +For `configure` build systems this is usually done by: +`CC=afl-clang-fast CXX=afl-clang-fast++ ./configure --disable-shared` -The instrumentation has a fairly modest performance impact; in conjunction with -other optimizations implemented by afl-fuzz, most programs can be fuzzed as fast -or even faster than possible with traditional tools. +Note that if you are using the (better) afl-clang-lto compiler you also have to +set AR to llvm-ar[-VERSION] and RANLIB to llvm-ranlib[-VERSION] - as it is +described in [llvm/README.lto.md](llvm/README.lto.md) -The correct way to recompile the target program may vary depending on the -specifics of the build process, but a nearly-universal approach would be: +##### cmake -```shell -CC=/path/to/afl/afl-gcc ./configure -make clean all -``` +For `configure` build systems this is usually done by: +`mkdir build; cd build; CC=afl-clang-fast CXX=afl-clang-fast++ cmake ..` -For C++ programs, you'd would also want to set `CXX=/path/to/afl/afl-g++`. +Some cmake scripts require something like `-DCMAKE_CC=... -DCMAKE_CXX=...` +or `-DCMAKE_C_COMPILER=... DCMAKE_CPP_COMPILER=...` instead. -The clang wrappers (afl-clang and afl-clang++) can be used in the same way; -clang users may also opt to leverage a higher-performance instrumentation mode, -as described in [llvm_mode/README.md](llvm_mode/README.md). -Clang/LLVM has a much better performance and works with LLVM version 3.4 to 11. +Note that if you are using the (better) afl-clang-lto compiler you also have to +set AR to llvm-ar[-VERSION] and RANLIB to llvm-ranlib[-VERSION] - as it is +described in [llvm/README.lto.md](llvm/README.lto.md) -Using the LAF Intel performance enhancements are also recommended, see -[llvm_mode/README.laf-intel.md](llvm_mode/README.laf-intel.md) +##### other build systems or if configure/cmake didn't work -Using partial instrumentation is also recommended, see -[llvm_mode/README.instrument_file.md](llvm_mode/README.instrument_file.md) +Sometimes cmake and configure do not pick up the afl++ compiler, or the +ranlib/ar that is needed - because this was just not foreseen by the developer +of the target. Or they have non-standard options. Figure out if there is a +non-standard way to set this, otherwise set up the build normally and edit the +generated build environment afterwards manually to point to the right compiler +(and/or ranlib and ar). -When testing libraries, you need to find or write a simple program that reads -data from stdin or from a file and passes it to the tested library. In such a -case, it is essential to link this executable against a static version of the -instrumented library or to make sure that the correct .so file is loaded at -runtime (usually by setting `LD_LIBRARY_PATH`). The simplest option is a static -build, usually possible via: +#### d) Better instrumentation + +If you just fuzz a target program as-is you are wasting a great opportunity for +much more fuzzing speed. + +This requires the usage of afl-clang-lto or afl-clang-fast. + +This is the so-called `persistent mode`, which is much, much faster but +requires that you code a source file that is specifically calling the target +functions that you want to fuzz, plus a few specific afl++ functions around +it. See [llvm_mode/README.persistent_mode.md](llvm_mode/README.persistent_mode.md) for details. + +Basically if you do not fuzz a target in persistent mode then you are just +doing it for a hobby and not professionally :-) + +### 2. Preparing the fuzzing + +As you fuzz the target with mutated input, having as diverse inputs for the +target as possible improves the efficiency a lot. + +#### a) Collect inputs +Try to gather valid inputs for the target from wherever you can. E.g. if it is +the PNG picture format try to find as many png files as possible, e.g. from +reported bugs, test suites, random downloads from the internet, unit test +case data - from all kind of PNG software. + +If the input format is not known, you can also modify a target program to write +away normal data it receives and processes to a file and use these. + +#### b) Making the input corpus unique + +Use the afl++ tool `afl-cmin` to remove inputs from the corpus that do not +produce a new path in the target. + +Put all files from step a) into one directory, e.g. INPUTS. + +If the target program is to be called by fuzzing as `bin/target -d INPUTFILE` +the run afl-cmin like this: +`afl-cmin -i INPUTS -o INPUTS_UNIQUE -- bin/target -d @@` +Note that the INPUTFILE argument that the target program would read from has to be set as `@@`. + +If the target reads from stdin instead, just omit the `@@` as this is the +default. + +#### c) Minimizing all corpus files + +The shorter the input files that still traverse the same path +within the target, the better the fuzzing will be. This is done with `afl-tmin` +however it is a long process as this has to be done for every file: -```shell -CC=/path/to/afl/afl-gcc ./configure --disable-shared ``` +mkdir input +cd INPUTS_UNIQUE +for i in *; do + afl-tmin -i "$i" -o "../input/$i" -- bin/target -d @@ +done +``` + +This can also be parallelized, e.g. with `parallel` + +#### Done! + +The INPUTS_UNIQUE/ directory from step b) - or even better the directory input/ +if you minimized the corpus in step c) - is the resulting input corpus directory +to be used in fuzzing! :-) + +### 3. Fuzzing the target + +In this final step we fuzz the target. +There are not that many useful options to run the target - unless you want to +use many CPU cores/threads for the fuzzing, which will make the fuzzing much +more useful. + +If you just use one CPU for fuzzing, then you are fuzzing just for fun and not +seriously :-) + +Pro tip: load the [afl++ snapshot module](https://github.com/AFLplusplus/AFL-Snapshot-LKM) +before the start of afl-fuzz as this improves performance by a x2 speed increase! -Setting `AFL_HARDEN=1` when calling 'make' will cause the CC wrapper to -automatically enable code hardening options that make it easier to detect -simple memory bugs. Libdislocator, a helper library included with AFL (see -[libdislocator/README.md](libdislocator/README.md)) can help uncover heap corruption issues, too. +#### a) Running afl-fuzz -PS. ASAN users are advised to review [docs/notes_for_asan.md](docs/notes_for_asan.md) -file for important caveats. +Before to do even a test run of afl-fuzz execute `sudo afl-system-config` (on +the host if you execute afl-fuzz in a docker container). This reconfigures the +system for optimal speed - which afl-fuzz checks and bails otherwise. +Set `export AFL_SKIP_CPUFREQ=1` for afl-fuzz to skip this check if you cannot run +afl-system-config with root privileges on the host for whatever reason. +If you have an input corpus from step 2 then specify this directory with the `-i` +option. Otherwise create a new directory and create a file with any content +as test data in there. -## Instrumenting binary-only apps +If you do not want anything special, the defaults are already usually best, +hence all you need is to specify the seed input directory with the result of +step [2. Collect inputs](#a)a-collect-inputs)): +`afl-fuzz -i input -o output -- bin/target -d @@` +Note that the directory specified with -o will be created if it does not exist. -When source code is *NOT* available, the fuzzer offers experimental support for -fast, on-the-fly instrumentation of black-box binaries. This is accomplished -with a version of QEMU running in the lesser-known "user space emulation" mode. +If you need to stop and re-start the fuzzing, use the same command line options +(or even change them by selecting a different power schedule or another +mutation mode!) and switch the input directory with a dash (`-`): +`afl-fuzz -i - -o output -- bin/target -d @@` +Note that afl-fuzz enforces memory limits to prevent the system to run out +of memory. By default this is 50MB for a process. If this is too little for +the target (which you can usually see by afl-fuzz bailing with the message +that it could not connect to the forkserver), then you can increase this +with the `-m` option, the value is in MB. To disable any memory limits +(beware!) set `-m 0` - which is usually required for ASAN compiled targets. + +Adding a dictionary is helpful. See the [dictionaries/](dictionaries/) if +something is already included for your data format, and tell afl-fuzz to load +that dictionary by adding `-x dictionaries/FORMAT.dict`. With afl-clang-lto +you have an autodictionary generation for which you need to do nothing except +to use afl-clang-lto as the compiler. You also have the option to generate +a dictionary yourself, see [libtokencap/README.md](libtokencap/README.md). + +afl-fuzz has a variety of options that help to workaround target quirks like +specific locations for the input file (`-f`), not performing deterministic +fuzzing (`-d`) and many more. Check out `afl-fuzz -h`. + +afl-fuzz never stops fuzzing. To terminate afl++ simply press Control-C. + +When you start afl-fuzz you will see a user interface that shows what the status +is: + + +All labels are explained in [docs/status_screen.md](docs/status_screen.md). + +#### b) Using multiple cores/threads + +If you want to seriously fuzz then use as many cores/threads as possible to +fuzz your target. + +On the same machine - due to the design of how afl++ works - there is a maximum +number of CPU cores/threads that are useful, use more and the overall performance +degrades instead. This value depends on the target and the limit is between 48 +and 96 cores/threads per machine. + +There should be one main fuzzer (`-M main` option) and as many secondary +fuzzers (eg `-S variant1`) as you have cores that you use. +Every -M/-S entry needs a unique name (that can be whatever), however the same +-o output directory location has to be used for all instances. + +For every secondary fuzzer there should be a variation, e.g.: + * one should fuzz the target that was compiled differently: with sanitizers + activated (`export AFL_USE_ASAN=1 ; export AFL_USE_UBSAN=1 ; + export AFL_USE_CFISAN=1 ; ` + * one should fuzz the target with CMPLOG/redqueen (see above) + * one to three should fuzz a target compiled with laf-intel/COMPCOV (see above). + +All other secondaries should be used like this: + * A third to a half with the MOpt mutator enabled: `-L 0` + * run with a different power schedule, available are: + `explore (default), fast, coe, lin, quad, exploit, mmopt, rare, seek` + which you can set with e.g. `-p seek` + +You can also use different fuzzers. +If you are using afl spinoffs or afl conforming fuzzers, then just use the +same -o directory and give it a unique `-S` name. +Examples are: + * [Angora](https://github.com/AngoraFuzzer/Angora) + * [Untracer](https://github.com/FoRTE-Research/UnTracer-AFL) + * [AFLsmart](https://github.com/aflsmart/aflsmart) + * [FairFuzz](https://github.com/carolemieux/afl-rb) + * [Neuzz](https://github.com/Dongdongshe/neuzz) + +A long list can be found at [https://github.com/Microsvuln/Awesome-AFL](https://github.com/Microsvuln/Awesome-AFL) + +However you can also sync afl++ with honggfuzz, libfuzzer, entropic, etc. +Just show the main fuzzer (-M) with the `-F` option where the queue +directory of a different fuzzer is, e.g. `-F /src/target/honggfuzz`. + +#### c) The status of the fuzz campaign + +afl++ comes with the `afl-whatsup` script to show the status of the fuzzing +campaign. + +Just supply the directory that afl-fuzz is given with the -o option and +you will see a detailed status of every fuzzer in that campaign plus +a summary. + +To have only the summary use the `-s` switch e.g.: `afl-whatsup -s output/` + +#### d) Checking the coverage of the fuzzing + +The `paths found` value is a bad indicator how good the coverage is. +It is better to check out the exact lines of code that have been reached - +and which have not been found so far. + +An "easy" helper script for this is [https://github.com/vanhauser-thc/afl-cov](https://github.com/vanhauser-thc/afl-cov), +just follow the README of that seperate project. + +If you see that an important area or a feature has not been covered so far then +try to find an input that is able to reach that and start a new secondary in +that fuzzing campaign with that seed as input, let it run for a few minutes, +then terminate it. The main node will pick it up and make it available to the +other secondary nodes over time. Set `export AFL_NO_AFFINITY=1` if you have no +free core. + +#### e) How long to fuzz a target? + +This is a difficult question. +Basically if no new path is found for a long time (e.g. for a day or a week) +then you can expect that your fuzzing won't be fruitful anymore. +However often this just means that you should switch out secondaries for +others, e.g. custom mutator modules, sync to very different fuzzers, etc. + +#### f) Improve the speed! + + * Use [persistent mode](llvm_mode/README.persistent_mode.md) (x2-x20 speed increase) + * If you do not use shmem persistent mode, use `AFL_TMPDIR` to point the input file on a tempfs location, see [docs/env_variables.md](docs/env_variables.md) + * Linux: Use the [afl++ snapshot module](https://github.com/AFLplusplus/AFL-Snapshot-LKM) (x2 speed increase) + * Linux: Improve kernel performance: modify `/etc/default/grub`, set `GRUB_CMDLINE_LINUX_DEFAULT="ibpb=off ibrs=off kpti=off l1tf=off mds=off mitigations=off no_stf_barrier noibpb noibrs nopcid nopti nospec_store_bypass_disable nospectre_v1 nospectre_v2 pcid=off pti=off spec_store_bypass_disable=off spectre_v2=off stf_barrier=off"`; then `update-grub` and `reboot` (warning: makes the system more insecure) + * Linux: Running on an `ext2` filesystem with `noatime` mount option will be a bit faster than on any other journaling filesystem + * Use your cores! [3.b) Using multiple cores/threads](#b-using-multiple-coresthreads) + +### The End + +Check out the [docs/FAQ](docs/FAQ.md) if it maybe answers your question (that +you might not even have known you had ;-) ). + +This is basically all you need to know to professionally run fuzzing campaigns. +If you want to know more, the rest of this README and the tons of texts in +[docs/](docs/) will have you covered. + +Note that there are also a lot of tools out there that help fuzzing with afl++ +(some might be deprecated or unsupported): + +Minimization of test cases: + * [afl-pytmin](https://github.com/ilsani/afl-pytmin) - a wrapper for afl-tmin that tries to speed up the process of the minimization of test case by using many CPU cores. + * [afl-ddmin-mod](https://github.com/MarkusTeufelberger/afl-ddmin-mod) - a variation of afl-tmin based on the ddmin algorithm. + * [halfempty](https://github.com/googleprojectzero/halfempty) - is a fast utility for minimizing test cases by Tavis Ormandy based on parallelization. + +Distributed execution: + * [disfuzz-afl](https://github.com/MartijnB/disfuzz-afl) - distributed fuzzing for AFL. + * [AFLDFF](https://github.com/quantumvm/AFLDFF) - AFL distributed fuzzing framework. + * [afl-launch](https://github.com/bnagy/afl-launch) - a tool for the execution of many AFL instances. + * [afl-mothership](https://github.com/afl-mothership/afl-mothership) - management and execution of many synchronized AFL fuzzers on AWS cloud. + * [afl-in-the-cloud](https://github.com/abhisek/afl-in-the-cloud) - another script for running AFL in AWS. + +Deployment, management, monitoring, reporting + * [afl-other-arch](https://github.com/shellphish/afl-other-arch) - is a set of patches and scripts for easily adding support for various non-x86 architectures for AFL. + * [afl-trivia](https://github.com/bnagy/afl-trivia) - a few small scripts to simplify the management of AFL. + * [afl-monitor](https://github.com/reflare/afl-monitor) - a script for monitoring AFL. + * [afl-manager](https://github.com/zx1340/afl-manager) - a web server on Python for managing multi-afl. + * [afl-remote](https://github.com/block8437/afl-remote) - a web server for the remote management of AFL instances. + +Crash processing + * [afl-utils](https://gitlab.com/rc0r/afl-utils) - a set of utilities for automatic processing/analysis of crashes and reducing the number of test cases. + * [afl-crash-analyzer](https://github.com/floyd-fuh/afl-crash-analyzer) - another crash analyzer for AFL. + * [fuzzer-utils](https://github.com/ThePatrickStar/fuzzer-utils) - a set of scripts for the analysis of results. + * [atriage](https://github.com/Ayrx/atriage) - a simple triage tool. + * [afl-kit](https://github.com/kcwu/afl-kit) - afl-cmin on Python. + * [AFLize](https://github.com/d33tah/aflize) - a tool that automatically generates builds of debian packages suitable for AFL. + * [afl-fid](https://github.com/FoRTE-Research/afl-fid) - a set of tools for working with input data. + +## Fuzzing binary-only targets + +When source code is *NOT* available, afl++ offers various support for fast, +on-the-fly instrumentation of black-box binaries. + +### QEMU + +For linux programs and it's libraries this is accomplished with a version of +QEMU running in the lesser-known "user space emulation" mode. QEMU is a project separate from AFL, but you can conveniently build the feature by doing: - ```shell cd qemu_mode ./build_qemu_support.sh ``` - For additional instructions and caveats, see [qemu_mode/README.md](qemu_mode/README.md). - If possible you should use the persistent mode, see [qemu_mode/README.persistent.md](qemu_mode/README.persistent.md). - -The mode is approximately 2-5x slower than compile-time instrumentation, is -less conducive to parallelization, and may have some other quirks. +The mode is approximately 2-5x slower than compile-time instrumentation, and is +less conducive to parallelization. If [afl-dyninst](https://github.com/vanhauser-thc/afl-dyninst) works for your binary, then you can use afl-fuzz normally and it will have twice -the speed compared to qemu_mode. +the speed compared to qemu_mode (but slower than persistent mode). + +### Unicorn + +For non-Linux binaries you can use afl++'s unicorn mode which can emulate +anything you want - for the price of speed and the user writing scripts. +See [unicorn_mode](unicorn_mode/README.md). + +It can be easily build by: +```shell +cd unicorn_mode +./build_unicorn_support.sh +``` + +### Shared libraries + +If the goal is to fuzz a dynamic library then there are two options available. +For both you need to write a small hardness that loads and calls the library. +Faster is the frida solution: [examples/afl_frida/README.md](examples/afl_frida/README.md) + +Another, less precise and slower option is using ptrace with debugger interrupt +instrumentation: [examples/afl_untracer/README.md](examples/afl_untracer/README.md) + +### More A more comprehensive description of these and other options can be found in [docs/binaryonly_fuzzing.md](docs/binaryonly_fuzzing.md) -## Good examples and writeups +## Challenges of guided fuzzing -Here are some good writeups to show how to effectively use AFL++: +Fuzzing is one of the most powerful and proven strategies for identifying +security issues in real-world software; it is responsible for the vast +majority of remote code execution and privilege escalation bugs found to date +in security-critical software. - * [https://aflplus.plus/docs/tutorials/libxml2_tutorial/](https://aflplus.plus/docs/tutorials/libxml2_tutorial/) - * [https://bananamafia.dev/post/gb-fuzz/](https://bananamafia.dev/post/gb-fuzz/) - * [https://securitylab.github.com/research/fuzzing-challenges-solutions-1](https://securitylab.github.com/research/fuzzing-challenges-solutions-1) - * [https://securitylab.github.com/research/fuzzing-sockets-FTP](https://securitylab.github.com/research/fuzzing-sockets-FTP) +Unfortunately, fuzzing is also relatively shallow; blind, random mutations +make it very unlikely to reach certain code paths in the tested code, leaving +some vulnerabilities firmly outside the reach of this technique. -If you are interested in fuzzing structured data (where you define what the -structure is), these two links have you covered: - * [https://github.com/bruce30262/libprotobuf-mutator_fuzzing_learning/tree/master/4_libprotobuf_aflpp_custom_mutator](https://github.com/bruce30262/libprotobuf-mutator_fuzzing_learning/tree/master/4_libprotobuf_aflpp_custom_mutator) - * [https://github.com/thebabush/afl-libprotobuf-mutator](https://github.com/thebabush/afl-libprotobuf-mutator) +There have been numerous attempts to solve this problem. One of the early +approaches - pioneered by Tavis Ormandy - is corpus distillation. The method +relies on coverage signals to select a subset of interesting seeds from a +massive, high-quality corpus of candidate files, and then fuzz them by +traditional means. The approach works exceptionally well but requires such +a corpus to be readily available. In addition, block coverage measurements +provide only a very simplistic understanding of the program state and are less +useful for guiding the fuzzing effort in the long haul. -If you find other good ones, please send them to us :-) +Other, more sophisticated research has focused on techniques such as program +flow analysis ("concolic execution"), symbolic execution, or static analysis. +All these methods are extremely promising in experimental settings, but tend +to suffer from reliability and performance problems in practical uses - and +currently do not offer a viable alternative to "dumb" fuzzing techniques. -## Power schedules +## Background: The afl-fuzz approach -The power schedules were copied from Marcel Böhme's excellent AFLfast -implementation and expand on the ability to discover new paths and -therefore may increase the code coverage. +American Fuzzy Lop is a brute-force fuzzer coupled with an exceedingly simple +but rock-solid instrumentation-guided genetic algorithm. It uses a modified +form of edge coverage to effortlessly pick up subtle, local-scale changes to +program control flow. -The available schedules are: - - - explore (default, original AFL) - - exploit (original AFL) - - fast (AFLfast) - - coe (AFLfast) - - quad (AFLfast) - - lin (AFLfast) - - rare (afl++ experimental) - - mmopt (afl++ experimental) - - seek (afl++ experimental) +Simplifying a bit, the overall algorithm can be summed up as: -In parallel mode (-M/-S, several instances with the shared queue), we suggest to -run the main node using the explore or fast schedule (-p explore) and the secondary -nodes with a combination of cut-off-exponential (-p coe), exponential (-p fast), -explore (-p explore) and mmopt (-p mmopt) schedules. If a schedule does -not perform well for a target, restart the secondary nodes with a different schedule. + 1) Load user-supplied initial test cases into the queue, -In single mode, using -p fast is usually slightly more beneficial than the -default explore mode. -(We don't want to change the default behavior of afl, so "fast" has not been -made the default mode). + 2) Take the next input file from the queue, -More details can be found in the paper published at the 23rd ACM Conference on -Computer and Communications Security [CCS'16](https://www.sigsac.org/ccs/CCS2016/accepted-papers/) + 3) Attempt to trim the test case to the smallest size that doesn't alter + the measured behavior of the program, -## Choosing initial test cases + 4) Repeatedly mutate the file using a balanced and well-researched variety + of traditional fuzzing strategies, + + 5) If any of the generated mutations resulted in a new state transition + recorded by the instrumentation, add mutated output as a new entry in the + queue. + + 6) Go to 2. + +The discovered test cases are also periodically culled to eliminate ones that +have been obsoleted by newer, higher-coverage finds; and undergo several other +instrumentation-driven effort minimization steps. + +As a side result of the fuzzing process, the tool creates a small, +self-contained corpus of interesting test cases. These are extremely useful +for seeding other, labor- or resource-intensive testing regimes - for example, +for stress-testing browsers, office applications, graphics suites, or +closed-source tools. + +The fuzzer is thoroughly tested to deliver out-of-the-box performance far +superior to blind fuzzing or coverage-only tools. + +## Help: Choosing initial test cases To operate correctly, the fuzzer requires one or more starting file that contains a good example of the input data normally expected by the targeted @@ -415,45 +751,7 @@ PS. If a large corpus of data is available for screening, you may want to use the afl-cmin utility to identify a subset of functionally distinct files that exercise different code paths in the target binary. - -## Fuzzing binaries - -The fuzzing process itself is carried out by the afl-fuzz utility. This program -requires a read-only directory with initial test cases, a separate place to -store its findings, plus a path to the binary to test. - -For target binaries that accept input directly from stdin, the usual syntax is: - -```shell -./afl-fuzz -i testcase_dir -o findings_dir /path/to/program [...params...] -``` - -For programs that take input from a file, use '@@' to mark the location in -the target's command line where the input file name should be placed. The -fuzzer will substitute this for you: - -```shell -./afl-fuzz -i testcase_dir -o findings_dir /path/to/program @@ -``` - -You can also use the -f option to have the mutated data written to a specific -file. This is useful if the program expects a particular file extension or so. - -Non-instrumented binaries can be fuzzed in the QEMU mode (add -Q in the command -line) or in a traditional, blind-fuzzer mode (specify -n). - -You can use -t and -m to override the default timeout and memory limit for the -executed process; rare examples of targets that may need these settings touched -include compilers and video decoders. - -Tips for optimizing fuzzing performance are discussed in [perf_tips.md](docs/perf_tips.md). - -Note that afl-fuzz starts by performing an array of deterministic fuzzing -steps, which can take several days, but tend to produce neat test cases. If you -want quick & dirty results right away - akin to zzuf and other traditional -fuzzers - add the -d option to the command line. - -## Interpreting output +## Help: Interpreting output See the [docs/status_screen.md](docs/status_screen.md) file for information on how to interpret the displayed stats and monitor the health of the process. Be @@ -513,53 +811,7 @@ If you have gnuplot installed, you can also generate some pretty graphs for any active fuzzing task using afl-plot. For an example of how this looks like, see [http://lcamtuf.coredump.cx/afl/plot/](http://lcamtuf.coredump.cx/afl/plot/). -## Parallelized fuzzing - -Every instance of afl-fuzz takes up roughly one core. This means that on -multi-core systems, parallelization is necessary to fully utilize the hardware. -For tips on how to fuzz a common target on multiple cores or multiple networked -machines, please refer to [docs/parallel_fuzzing.md](docs/parallel_fuzzing.md). - -The parallel fuzzing mode also offers a simple way for interfacing AFL to other -fuzzers, to symbolic or concolic execution engines, and so forth; again, see the -last section of [docs/parallel_fuzzing.md](docs/parallel_fuzzing.md) for tips. - -## Fuzzer dictionaries - -By default, afl-fuzz mutation engine is optimized for compact data formats - -say, images, multimedia, compressed data, regular expression syntax, or shell -scripts. It is somewhat less suited for languages with particularly verbose and -redundant verbiage - notably including HTML, SQL, or JavaScript. - -To avoid the hassle of building syntax-aware tools, afl-fuzz provides a way to -seed the fuzzing process with an optional dictionary of language keywords, -magic headers, or other special tokens associated with the targeted data type --- and use that to reconstruct the underlying grammar on the go: - - [http://lcamtuf.blogspot.com/2015/01/afl-fuzz-making-up-grammar-with.html](http://lcamtuf.blogspot.com/2015/01/afl-fuzz-making-up-grammar-with.html) - -To use this feature, you first need to create a dictionary in one of the two -formats discussed in [dictionaries/README.md](dictionaries/README.md); -and then point the fuzzer to it via the -x option in the command line. - -(Several common dictionaries are already provided in that subdirectory, too.) - -There is no way to provide more structured descriptions of the underlying -syntax, but the fuzzer will likely figure out some of this based on the -instrumentation feedback alone. This actually works in practice, say: - - [http://lcamtuf.blogspot.com/2015/04/finding-bugs-in-sqlite-easy-way.html](http://lcamtuf.blogspot.com/2015/04/finding-bugs-in-sqlite-easy-way.html) - -PS. Even when no explicit dictionary is given, afl-fuzz will try to extract -existing syntax tokens in the input corpus by watching the instrumentation -very closely during deterministic byte flips. This works for some types of -parsers and grammars but isn't nearly as good as the -x mode. - -If a dictionary is really hard to come by, another option is to let AFL run -for a while and then use the token capture library that comes as a companion -utility with AFL. For that, see [libtokencap/README.md](libtokencap/README.tokencap.md). - -## Crash triage +## Help: Crash triage The coverage-based grouping of crashes usually produces a small data set that can be quickly triaged manually or with a very simple GDB or Valgrind script. @@ -593,13 +845,13 @@ can be operated in a very simple way: The tool works with crashing and non-crashing test cases alike. In the crash mode, it will happily accept instrumented and non-instrumented binaries. In the -non-crashing mode, the minimizer relies on standard AFL instrumentation to make +non-crashing mode, the minimizer relies on standard afl++ instrumentation to make the file simpler without altering the execution path. The minimizer accepts the -m, -t, -f and @@ syntax in a manner compatible with afl-fuzz. -Another recent addition to AFL is the afl-analyze tool. It takes an input +Another tool in afl++ is the afl-analyze tool. It takes an input file, attempts to sequentially flip bytes, and observes the behavior of the tested program. It then color-codes the input based on which sections appear to be critical, and which are not; while not bulletproof, it can often offer quick @@ -627,7 +879,8 @@ found by modifying the target programs to call abort() when say: Implementing these or similar sanity checks usually takes very little time; if you are the maintainer of a particular package, you can make this code conditional with `#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION` (a flag also -shared with libfuzzer) or `#ifdef __AFL_COMPILER` (this one is just for AFL). +shared with libfuzzer and honggfuzz) or `#ifdef __AFL_COMPILER` (this one is +just for AFL). ## Common-sense risks @@ -641,7 +894,7 @@ tasks, fuzzing may put a strain on your hardware and on the OS. In particular: for something to blow up. - Targeted programs may end up erratically grabbing gigabytes of memory or - filling up disk space with junk files. AFL tries to enforce basic memory + filling up disk space with junk files. afl++ tries to enforce basic memory limits, but can't prevent each and every possible mishap. The bottom line is that you shouldn't be fuzzing on systems where the prospect of data loss is not an acceptable risk. @@ -662,7 +915,7 @@ tasks, fuzzing may put a strain on your hardware and on the OS. In particular: Here are some of the most important caveats for AFL: - - AFL detects faults by checking for the first spawned process dying due to + - afl++ detects faults by checking for the first spawned process dying due to a signal (SIGSEGV, SIGABRT, etc). Programs that install custom handlers for these signals may need to have the relevant code commented out. In the same vein, faults in child processes spawned by the fuzzed target may evade @@ -691,9 +944,6 @@ Here are some of the most important caveats for AFL: Some useful tips for modifying network-based services can be also found at: [https://www.fastly.com/blog/how-to-fuzz-server-american-fuzzy-lop](https://www.fastly.com/blog/how-to-fuzz-server-american-fuzzy-lop) - - AFL doesn't output human-readable coverage data. If you want to monitor - coverage, use afl-cov from Michael Rash: [https://github.com/mrash/afl-cov](https://github.com/mrash/afl-cov) - - Occasionally, sentient machines rise against their creators. If this happens to you, please consult [http://lcamtuf.coredump.cx/prep/](http://lcamtuf.coredump.cx/prep/). @@ -758,6 +1008,6 @@ Thank you! Questions? Concerns? Bug reports? The contributors can be reached via [https://github.com/AFLplusplus/AFLplusplus](https://github.com/AFLplusplus/AFLplusplus) -There is also a mailing list for the afl project; to join, send a mail to -<afl-users+subscribe@googlegroups.com>. Or, if you prefer to browse -archives first, try: [https://groups.google.com/group/afl-users](https://groups.google.com/group/afl-users) +There is also a mailing list for the afl/afl++ project; to join, send a mail to +<afl-users+subscribe@googlegroups.com>. Or, if you prefer to browse archives +first, try: [https://groups.google.com/group/afl-users](https://groups.google.com/group/afl-users) diff --git a/TODO.md b/TODO.md index 8085bc07..999cb9d3 100644 --- a/TODO.md +++ b/TODO.md @@ -1,20 +1,20 @@ # TODO list for AFL++ -## Roadmap 2.66+ +## Roadmap 2.67+ + - expand on AFL_LLVM_INSTRUMENT_FILE to also support sancov allowlist format - AFL_MAP_SIZE for qemu_mode and unicorn_mode - - namespace for targets? e.g. network - - learn from honggfuzz (mutations, maybe ptrace?) - CPU affinity for many cores? There seems to be an issue > 96 cores ## Further down the road afl-fuzz: - - ascii_only mode for mutation output - or use a custom mutator for this? - setting min_len/max_len/start_offset/end_offset limits for mutation output + - add __sanitizer_cov_trace_cmp* support via shmem llvm_mode: - LTO - imitate sancov + - add __sanitizer_cov_trace_cmp* support gcc_plugin: - (wait for submission then decide) @@ -22,7 +22,7 @@ gcc_plugin: - better instrumentation (seems to be better with gcc-9+) qemu_mode: - - update to 5.x (if the performance bug if gone) + - update to 5.x (if the performance bug is gone) - non colliding instrumentation - rename qemu specific envs to AFL_QEMU (AFL_ENTRYPOINT, AFL_CODE_START/END, AFL_COMPCOV_LEVEL?) @@ -30,3 +30,15 @@ qemu_mode: persistent mode - add/implement AFL_QEMU_INST_LIBLIST and AFL_QEMU_NOINST_PROGRAM - add/implement AFL_QEMU_INST_REGIONS as a list of _START/_END addresses + +## Ideas + + - LTO/sancov: write current edge to prev_loc and use that information when + using cmplog or __sanitizer_cov_trace_cmp*. maybe we can deduct by follow + up edge numbers that both following cmp paths have been found and then + disable working on this edge id + + - new tancov: use some lightweight taint analysis to see which parts of a + new queue entry is accessed and only fuzz these bytes - or better, only + fuzz those bytes that are newly in coverage compared to the queue entry + the new one is based on diff --git a/afl-plot b/afl-plot index 1074552a..0faed0ec 100755 --- a/afl-plot +++ b/afl-plot @@ -68,6 +68,15 @@ if [ ! -f "$inputdir/plot_data" ]; then fi +LINES=`cat "$inputdir/plot_data" | wc -l` + +if [ "$LINES" -lt 3 ]; then + + echo "[-] Error: plot_data carries too little data, let it run longer." 1>&2 + exit 1 + +fi + BANNER="`cat "$inputdir/fuzzer_stats" 2> /dev/null | grep '^afl_banner ' | cut -d: -f2- | cut -b2-`" test "$BANNER" = "" && BANNER="(none)" @@ -118,6 +127,9 @@ set key outside set autoscale xfixmin set autoscale xfixmax +set xlabel "all times in UTC" font "small" + +set ytics auto plot '$inputdir/plot_data' using 1:4 with filledcurve x1 title 'total paths' linecolor rgb '#000000' fillstyle transparent solid 0.2 noborder, \\ '' using 1:3 with filledcurve x1 title 'current path' linecolor rgb '#f0f0f0' fillstyle transparent solid 0.5 noborder, \\ '' using 1:5 with lines title 'pending paths' linecolor rgb '#0090ff' linewidth 3, \\ @@ -127,6 +139,7 @@ plot '$inputdir/plot_data' using 1:4 with filledcurve x1 title 'total paths' lin set terminal png truecolor enhanced size 1000,200 butt set output '$outputdir/low_freq.png' +set ytics 1 plot '$inputdir/plot_data' using 1:8 with filledcurve x1 title '' linecolor rgb '#c00080' fillstyle transparent solid 0.2 noborder, \\ '' using 1:8 with lines title ' uniq crashes' linecolor rgb '#c00080' linewidth 3, \\ '' using 1:9 with lines title 'uniq hangs' linecolor rgb '#c000f0' linewidth 3, \\ @@ -135,6 +148,7 @@ plot '$inputdir/plot_data' using 1:8 with filledcurve x1 title '' linecolor rgb set terminal png truecolor enhanced size 1000,200 butt set output '$outputdir/exec_speed.png' +set ytics auto plot '$inputdir/plot_data' using 1:11 with filledcurve x1 title '' linecolor rgb '#0090ff' fillstyle transparent solid 0.2 noborder, \\ '$inputdir/plot_data' using 1:11 with lines title ' execs/sec' linecolor rgb '#0090ff' linewidth 3 smooth bezier; diff --git a/afl-system-config b/afl-system-config index 1e180d8b..34db61aa 100755 --- a/afl-system-config +++ b/afl-system-config @@ -1,5 +1,5 @@ #!/bin/sh -test "$1" = "-h" && { +test "$1" = "-h" -o "$1" = "-hh" && { echo 'afl-system-config by Marc Heuse <mh@mh-sec.de>' echo echo $0 diff --git a/afl-whatsup b/afl-whatsup index 1a276964..abcddbf1 100755 --- a/afl-whatsup +++ b/afl-whatsup @@ -20,7 +20,7 @@ echo "$0 status check tool for afl-fuzz by Michal Zalewski" echo -test "$1" = "-h" && { +test "$1" = "-h" -o "$1" = "-hh" && { echo $0 [-s] output_directory echo echo Options: @@ -162,7 +162,8 @@ for i in `find . -maxdepth 2 -iname fuzzer_stats | sort`; do ALIVE_CNT=$((ALIVE_CNT + 1)) - EXEC_SEC=$((execs_done / RUN_UNIX)) + EXEC_SEC=0 + test -z "$RUN_UNIX" -o "$RUN_UNIX" = 0 || EXEC_SEC=$((execs_done / RUN_UNIX)) PATH_PERC=$((cur_path * 100 / paths_total)) TOTAL_TIME=$((TOTAL_TIME + RUN_UNIX)) @@ -184,7 +185,9 @@ for i in `find . -maxdepth 2 -iname fuzzer_stats | sort`; do echo " ${RED}timeout_ratio $TIMEOUT_PERC%${NC}" fi - if [ $EXEC_SEC -lt 100 ]; then + if [ $EXEC_SEC -eq 0 ]; then + echo " ${YELLOW}no data yet, 0 execs/sec${NC}" + elif [ $EXEC_SEC -lt 100 ]; then echo " ${RED}slow execution, $EXEC_SEC execs/sec${NC}" fi diff --git a/afl-wine-trace b/afl-wine-trace index 65525a33..8853a757 100755 --- a/afl-wine-trace +++ b/afl-wine-trace @@ -68,7 +68,12 @@ else: argv = sys.argv[1:] for i in range(len(argv)): if ".cur_input" in argv[i]: - argv[i] = subprocess.run([os.path.join(os.path.dirname(wine_path), "winepath"), "--windows", argv[i]], universal_newlines=True, stdout=subprocess.PIPE).stdout + # Get the Wine translated path using the winepath tool + arg_translated = subprocess.run([os.path.join(os.path.dirname(wine_path), "winepath"), "--windows", argv[i]], universal_newlines=True, stdout=subprocess.PIPE).stdout + # Remove the spurious LF at the end of the path + if len(arg_translated) > 0 and arg_translated[-1] == '\n': + arg_translated = arg_translated[:-1] + argv[i] = arg_translated break print("[afl-wine-trace] exec:", " ".join([qemu_path, wine_path] + argv)) diff --git a/custom_mutators/README.md b/custom_mutators/README.md index 0abce32f..a3b164be 100644 --- a/custom_mutators/README.md +++ b/custom_mutators/README.md @@ -10,3 +10,22 @@ Use with e.g. and add `AFL_CUSTOM_MUTATOR_ONLY=1` if you only want to use the custom mutator. Multiple custom mutators can be used by separating their paths with `:` in the environment variable. + +# Other custom mutators + +## Superion port + +Adrian Tiron ported the Superion grammar fuzzer to afl++, it is WIP and +requires cmake (among other things): +[https://github.com/adrian-rt/superion-mutator](https://github.com/adrian-rt/superion-mutator) + +## Protobuf + +There are two WIP protobuf projects, that require work to be working though: + +transforms protobuf raw: +https://github.com/bruce30262/libprotobuf-mutator_fuzzing_learning/tree/master/4_libprotobuf_aflpp_custom_mutator + +has a transform function you need to fill for your protobuf format, however +needs to be ported to the updated afl++ custom mutator API (not much work): +https://github.com/thebabush/afl-libprotobuf-mutator diff --git a/custom_mutators/honggfuzz/Makefile b/custom_mutators/honggfuzz/Makefile new file mode 100644 index 00000000..2f46d0e7 --- /dev/null +++ b/custom_mutators/honggfuzz/Makefile @@ -0,0 +1,15 @@ + +CFLAGS = -O3 -funroll-loops -fPIC -Wl,-Bsymbolic + +all: honggfuzz.so + +honggfuzz.so: honggfuzz.c input.h mangle.c ../../src/afl-performance.c + $(CC) $(CFLAGS) -I../../include -I. -shared -o honggfuzz.so honggfuzz.c mangle.c ../../src/afl-performance.c + +update: + wget --unlink https://github.com/google/honggfuzz/raw/master/mangle.c + wget --unlink https://github.com/google/honggfuzz/raw/master/mangle.h + wget --unlink https://github.com/google/honggfuzz/raw/master/honggfuzz.h + +clean: + rm -f *.o *~ *.so core diff --git a/custom_mutators/honggfuzz/README.md b/custom_mutators/honggfuzz/README.md new file mode 100644 index 00000000..8824976f --- /dev/null +++ b/custom_mutators/honggfuzz/README.md @@ -0,0 +1,12 @@ +# custum mutator: honggfuzz mangle + +this is the very good honggfuzz mutator in mangle.c as a custom mutator +module for afl++. It is the original mangle.c, mangle.h and honggfuzz.h +with a lot of mocking around it :-) + +just type `make` to build + +```AFL_CUSTOM_MUTATOR_LIBRARY=custom_mutators/honggfuzz/honggfuzz.so afl-fuzz ...``` + +> Original repository: https://github.com/google/honggfuzz +> Source commit: d0fbcb0373c32436b8fb922e6937da93b17291f5 diff --git a/custom_mutators/honggfuzz/common.h b/custom_mutators/honggfuzz/common.h new file mode 100644 index 00000000..e69de29b --- /dev/null +++ b/custom_mutators/honggfuzz/common.h diff --git a/custom_mutators/honggfuzz/custom_mutator_helpers.h b/custom_mutators/honggfuzz/custom_mutator_helpers.h new file mode 100644 index 00000000..57754697 --- /dev/null +++ b/custom_mutators/honggfuzz/custom_mutator_helpers.h @@ -0,0 +1,22 @@ +#ifndef CUSTOM_MUTATOR_HELPERS +#define CUSTOM_MUTATOR_HELPERS + +#include "config.h" +#include "types.h" +#include "afl-fuzz.h" +#include <stdlib.h> + +#define INITIAL_GROWTH_SIZE (64) + +/* Use in a struct: creates a name_buf and a name_size variable. */ +#define BUF_VAR(type, name) \ + type * name##_buf; \ + size_t name##_size; +/* this filles in `&structptr->something_buf, &structptr->something_size`. */ +#define BUF_PARAMS(struct, name) \ + (void **)&struct->name##_buf, &struct->name##_size + +#undef INITIAL_GROWTH_SIZE + +#endif + diff --git a/custom_mutators/honggfuzz/honggfuzz.c b/custom_mutators/honggfuzz/honggfuzz.c new file mode 100644 index 00000000..bde922c6 --- /dev/null +++ b/custom_mutators/honggfuzz/honggfuzz.c @@ -0,0 +1,143 @@ +#include <stdio.h> +#include <stdint.h> +#include <stdlib.h> +#include <string.h> + +#include "custom_mutator_helpers.h" +#include "mangle.h" + +#define NUMBER_OF_MUTATIONS 5 + +uint8_t * queue_input; +size_t queue_input_size; +afl_state_t * afl_struct; +run_t run; +honggfuzz_t global; +struct _dynfile_t dynfile; + +typedef struct my_mutator { + + afl_state_t *afl; + run_t * run; + u8 * mutator_buf; + unsigned int seed; + unsigned int extras_cnt, a_extras_cnt; + +} my_mutator_t; + +my_mutator_t *afl_custom_init(afl_state_t *afl, unsigned int seed) { + + my_mutator_t *data = calloc(1, sizeof(my_mutator_t)); + if (!data) { + + perror("afl_custom_init alloc"); + return NULL; + + } + + if ((data->mutator_buf = malloc(MAX_FILE)) == NULL) { + + perror("mutator_buf alloc"); + return NULL; + + } + + run.dynfile = &dynfile; + run.global = &global; + data->afl = afl; + data->seed = seed; + data->run = &run; + afl_struct = afl; + + run.global->mutate.maxInputSz = MAX_FILE; + run.global->mutate.mutationsPerRun = NUMBER_OF_MUTATIONS; + run.mutationsPerRun = NUMBER_OF_MUTATIONS; + run.global->timing.lastCovUpdate = 6; + + // global->feedback.cmpFeedback + // global->feedback.cmpFeedbackMap + + return data; + +} + +/* When a new queue entry is added we check if there are new dictionary + entries to add to honggfuzz structure */ + +void afl_custom_queue_new_entry(my_mutator_t * data, + const uint8_t *filename_new_queue, + const uint8_t *filename_orig_queue) { + + if (run.global->mutate.dictionaryCnt >= 1024) return; + + while (data->extras_cnt < data->afl->extras_cnt && + run.global->mutate.dictionaryCnt < 1024) { + + memcpy(run.global->mutate.dictionary[run.global->mutate.dictionaryCnt].val, + data->afl->extras[data->extras_cnt].data, + data->afl->extras[data->extras_cnt].len); + run.global->mutate.dictionary[run.global->mutate.dictionaryCnt].len = + data->afl->extras[data->extras_cnt].len; + run.global->mutate.dictionaryCnt++; + data->extras_cnt++; + + } + + while (data->a_extras_cnt < data->afl->a_extras_cnt && + run.global->mutate.dictionaryCnt < 1024) { + + memcpy(run.global->mutate.dictionary[run.global->mutate.dictionaryCnt].val, + data->afl->a_extras[data->a_extras_cnt].data, + data->afl->a_extras[data->a_extras_cnt].len); + run.global->mutate.dictionary[run.global->mutate.dictionaryCnt].len = + data->afl->a_extras[data->a_extras_cnt].len; + run.global->mutate.dictionaryCnt++; + data->a_extras_cnt++; + + } + +} + +/* we could set only_printable if is_ascii is set ... let's see +uint8_t afl_custom_queue_get(void *data, const uint8_t *filename) { + + //run.global->cfg.only_printable = ... + +} + +*/ + +/* here we run the honggfuzz mutator, which is really good */ + +size_t afl_custom_fuzz(my_mutator_t *data, uint8_t *buf, size_t buf_size, + u8 **out_buf, uint8_t *add_buf, size_t add_buf_size, + size_t max_size) { + + /* set everything up, costly ... :( */ + memcpy(data->mutator_buf, buf, buf_size); + queue_input = data->mutator_buf; + run.dynfile->data = data->mutator_buf; + queue_input_size = buf_size; + run.dynfile->size = buf_size; + *out_buf = data->mutator_buf; + + /* the mutation */ + mangle_mangleContent(&run, NUMBER_OF_MUTATIONS); + + /* return size of mutated data */ + return run.dynfile->size; + +} + +/** + * Deinitialize everything + * + * @param data The data ptr from afl_custom_init + */ +void afl_custom_deinit(my_mutator_t *data) { + + free(data->mutator_buf); + free(data); + +} + diff --git a/custom_mutators/honggfuzz/honggfuzz.h b/custom_mutators/honggfuzz/honggfuzz.h new file mode 100644 index 00000000..4e045272 --- /dev/null +++ b/custom_mutators/honggfuzz/honggfuzz.h @@ -0,0 +1,460 @@ +/* + * + * honggfuzz - core structures and macros + * ----------------------------------------- + * + * Author: Robert Swiecki <swiecki@google.com> + * + * Copyright 2010-2018 by Google Inc. All Rights Reserved. + * + * Licensed under the Apache License, Version 2.0 (the "License"); you may + * not use this file except in compliance with the License. You may obtain + * a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + * implied. See the License for the specific language governing + * permissions and limitations under the License. + * + */ + +#ifndef _HF_HONGGFUZZ_H_ +#define _HF_HONGGFUZZ_H_ + +#include <dirent.h> +#include <inttypes.h> +#include <limits.h> +#include <pthread.h> +#include <signal.h> +#include <stdbool.h> +#include <stdint.h> +#include <sys/param.h> +#include <sys/queue.h> +#include <sys/types.h> +#include <time.h> + +#include "libhfcommon/util.h" + +#define PROG_NAME "honggfuzz" +#define PROG_VERSION "2.2" + +/* Name of the template which will be replaced with the proper name of the file + */ +#define _HF_FILE_PLACEHOLDER "___FILE___" + +/* Default name of the report created with some architectures */ +#define _HF_REPORT_FILE "HONGGFUZZ.REPORT.TXT" + +/* Default stack-size of created threads. */ +#define _HF_PTHREAD_STACKSIZE (1024ULL * 1024ULL * 2ULL) /* 2MB */ + +/* Name of envvar which indicates sequential number of fuzzer */ +#define _HF_THREAD_NO_ENV "HFUZZ_THREAD_NO" + +/* Name of envvar which indicates that the netDriver should be used */ +#define _HF_THREAD_NETDRIVER_ENV "HFUZZ_USE_NETDRIVER" + +/* Name of envvar which indicates honggfuzz's log level in use */ +#define _HF_LOG_LEVEL_ENV "HFUZZ_LOG_LEVEL" + +/* Number of crash verifier iterations before tag crash as stable */ +#define _HF_VERIFIER_ITER 5 + +/* Size (in bytes) for report data to be stored in stack before written to file + */ +#define _HF_REPORT_SIZE 32768 + +/* Perf bitmap size */ +#define _HF_PERF_BITMAP_SIZE_16M (1024U * 1024U * 16U) +#define _HF_PERF_BITMAP_BITSZ_MASK 0x7FFFFFFULL +/* Maximum number of PC guards (=trace-pc-guard) we support */ +#define _HF_PC_GUARD_MAX (1024ULL * 1024ULL * 64ULL) + +/* Maximum size of the input file in bytes (1 MiB) */ +#define _HF_INPUT_MAX_SIZE (1024ULL * 1024ULL) + +/* Default maximum size of produced inputs */ +#define _HF_INPUT_DEFAULT_SIZE (1024ULL * 8) + +/* Per-thread bitmap */ +#define _HF_PERTHREAD_BITMAP_FD 1018 +/* FD used to report back used int/str constants from the fuzzed process */ +#define _HF_CMP_BITMAP_FD 1019 +/* FD used to log inside the child process */ +#define _HF_LOG_FD 1020 +/* FD used to represent the input file */ +#define _HF_INPUT_FD 1021 +/* FD used to pass coverage feedback from the fuzzed process */ +#define _HF_COV_BITMAP_FD 1022 +#define _HF_BITMAP_FD _HF_COV_BITMAP_FD /* Old name for _HF_COV_BITMAP_FD */ +/* FD used to pass data to a persistent process */ +#define _HF_PERSISTENT_FD 1023 + +/* Input file as a string */ +#define _HF_INPUT_FILE_PATH "/dev/fd/" HF_XSTR(_HF_INPUT_FD) + +/* Maximum number of supported execve() args */ +#define _HF_ARGS_MAX 2048 + +/* Message indicating that the fuzzed process is ready for new data */ +static const uint8_t HFReadyTag = 'R'; + +/* Maximum number of active fuzzing threads */ +#define _HF_THREAD_MAX 1024U + +/* Persistent-binary signature - if found within file, it means it's a + * persistent mode binary */ +#define _HF_PERSISTENT_SIG "\x01_LIBHFUZZ_PERSISTENT_BINARY_SIGNATURE_\x02\xFF" +/* HF NetDriver signature - if found within file, it means it's a + * NetDriver-based binary */ +#define _HF_NETDRIVER_SIG "\x01_LIBHFUZZ_NETDRIVER_BINARY_SIGNATURE_\x02\xFF" + +/* printf() nonmonetary separator. According to MacOSX's man it's supported + * there as well */ +#define _HF_NONMON_SEP "'" + +typedef enum { + + _HF_DYNFILE_NONE = 0x0, + _HF_DYNFILE_INSTR_COUNT = 0x1, + _HF_DYNFILE_BRANCH_COUNT = 0x2, + _HF_DYNFILE_BTS_EDGE = 0x10, + _HF_DYNFILE_IPT_BLOCK = 0x20, + _HF_DYNFILE_SOFT = 0x40, + +} dynFileMethod_t; + +typedef struct { + + uint64_t cpuInstrCnt; + uint64_t cpuBranchCnt; + uint64_t bbCnt; + uint64_t newBBCnt; + uint64_t softCntPc; + uint64_t softCntEdge; + uint64_t softCntCmp; + +} hwcnt_t; + +typedef enum { + + _HF_STATE_UNSET = 0, + _HF_STATE_STATIC, + _HF_STATE_DYNAMIC_DRY_RUN, + _HF_STATE_DYNAMIC_MAIN, + _HF_STATE_DYNAMIC_MINIMIZE, + +} fuzzState_t; + +typedef enum { + + HF_MAYBE = -1, + HF_NO = 0, + HF_YES = 1, + +} tristate_t; + +struct _dynfile_t { + + size_t size; + uint64_t cov[4]; + size_t idx; + int fd; + uint64_t timeExecUSecs; + char path[PATH_MAX]; + struct _dynfile_t *src; + uint32_t refs; + uint8_t * data; + TAILQ_ENTRY(_dynfile_t) pointers; + +}; + +typedef struct _dynfile_t dynfile_t; + +struct strings_t { + + size_t len; + TAILQ_ENTRY(strings_t) pointers; + char s[]; + +}; + +typedef struct { + + uint8_t pcGuardMap[_HF_PC_GUARD_MAX]; + uint8_t bbMapPc[_HF_PERF_BITMAP_SIZE_16M]; + uint32_t bbMapCmp[_HF_PERF_BITMAP_SIZE_16M]; + uint64_t pidNewPC[_HF_THREAD_MAX]; + uint64_t pidNewEdge[_HF_THREAD_MAX]; + uint64_t pidNewCmp[_HF_THREAD_MAX]; + uint64_t guardNb; + uint64_t pidTotalPC[_HF_THREAD_MAX]; + uint64_t pidTotalEdge[_HF_THREAD_MAX]; + uint64_t pidTotalCmp[_HF_THREAD_MAX]; + +} feedback_t; + +typedef struct { + + uint32_t cnt; + struct { + + uint8_t val[32]; + uint32_t len; + + } valArr[1024 * 16]; + +} cmpfeedback_t; + +typedef struct { + + struct { + + size_t threadsMax; + size_t threadsFinished; + uint32_t threadsActiveCnt; + pthread_t mainThread; + pid_t mainPid; + pthread_t threads[_HF_THREAD_MAX]; + + } threads; + + struct { + + const char *inputDir; + const char *outputDir; + DIR * inputDirPtr; + size_t fileCnt; + size_t testedFileCnt; + const char *fileExtn; + size_t maxFileSz; + size_t newUnitsAdded; + char workDir[PATH_MAX]; + const char *crashDir; + const char *covDirNew; + bool saveUnique; + size_t dynfileqMaxSz; + size_t dynfileqCnt; + dynfile_t * dynfileqCurrent; + dynfile_t * dynfileq2Current; + TAILQ_HEAD(dyns_t, _dynfile_t) dynfileq; + bool exportFeedback; + + } io; + + struct { + + int argc; + const char *const *cmdline; + bool nullifyStdio; + bool fuzzStdin; + const char * externalCommand; + const char * postExternalCommand; + const char * feedbackMutateCommand; + bool netDriver; + bool persistent; + uint64_t asLimit; + uint64_t rssLimit; + uint64_t dataLimit; + uint64_t coreLimit; + uint64_t stackLimit; + bool clearEnv; + char * env_ptrs[128]; + char env_vals[128][4096]; + sigset_t waitSigSet; + + } exe; + + struct { + + time_t timeStart; + time_t runEndTime; + time_t tmOut; + time_t lastCovUpdate; + int64_t timeOfLongestUnitUSecs; + bool tmoutVTALRM; + + } timing; + + struct { + + struct { + + uint8_t val[256]; + size_t len; + + } dictionary[1024]; + + size_t dictionaryCnt; + const char *dictionaryFile; + size_t mutationsMax; + unsigned mutationsPerRun; + size_t maxInputSz; + + } mutate; + + struct { + + bool useScreen; + char cmdline_txt[65]; + int64_t lastDisplayUSecs; + + } display; + + struct { + + bool useVerifier; + bool exitUponCrash; + const char *reportFile; + size_t dynFileIterExpire; + bool only_printable; + bool minimize; + bool switchingToFDM; + + } cfg; + + struct { + + bool enable; + bool del_report; + + } sanitizer; + + struct { + + fuzzState_t state; + feedback_t * covFeedbackMap; + int covFeedbackFd; + cmpfeedback_t * cmpFeedbackMap; + int cmpFeedbackFd; + bool cmpFeedback; + const char * blacklistFile; + uint64_t * blacklist; + size_t blacklistCnt; + bool skipFeedbackOnTimeout; + uint64_t maxCov[4]; + dynFileMethod_t dynFileMethod; + hwcnt_t hwCnts; + + } feedback; + + struct { + + size_t mutationsCnt; + size_t crashesCnt; + size_t uniqueCrashesCnt; + size_t verifiedCrashesCnt; + size_t blCrashesCnt; + size_t timeoutedCnt; + + } cnts; + + struct { + + bool enabled; + int serverSocket; + int clientSocket; + + } socketFuzzer; + + struct { + + pthread_rwlock_t dynfileq; + pthread_mutex_t feedback; + pthread_mutex_t report; + pthread_mutex_t state; + pthread_mutex_t input; + pthread_mutex_t timing; + + } mutex; + + /* For the Linux code */ + struct { + + int exeFd; + uint64_t dynamicCutOffAddr; + bool disableRandomization; + void * ignoreAddr; + const char *symsBlFile; + char ** symsBl; + size_t symsBlCnt; + const char *symsWlFile; + char ** symsWl; + size_t symsWlCnt; + uintptr_t cloneFlags; + tristate_t useNetNs; + bool kernelOnly; + bool useClone; + + } arch_linux; + + /* For the NetBSD code */ + struct { + + void * ignoreAddr; + const char *symsBlFile; + char ** symsBl; + size_t symsBlCnt; + const char *symsWlFile; + char ** symsWl; + size_t symsWlCnt; + + } arch_netbsd; + +} honggfuzz_t; + +typedef enum { + + _HF_RS_UNKNOWN = 0, + _HF_RS_WAITING_FOR_INITIAL_READY = 1, + _HF_RS_WAITING_FOR_READY = 2, + _HF_RS_SEND_DATA = 3, + +} runState_t; + +typedef struct { + + honggfuzz_t *global; + pid_t pid; + int64_t timeStartedUSecs; + char crashFileName[PATH_MAX]; + uint64_t pc; + uint64_t backtrace; + uint64_t access; + int exception; + char report[_HF_REPORT_SIZE]; + bool mainWorker; + unsigned mutationsPerRun; + dynfile_t * dynfile; + bool staticFileTryMore; + uint32_t fuzzNo; + int persistentSock; + runState_t runState; + bool tmOutSignaled; + char * args[_HF_ARGS_MAX + 1]; + int perThreadCovFeedbackFd; + unsigned triesLeft; + dynfile_t * current; +#if !defined(_HF_ARCH_DARWIN) + timer_t timerId; +#endif // !defined(_HF_ARCH_DARWIN) + hwcnt_t hwCnts; + + struct { + + /* For Linux code */ + uint8_t *perfMmapBuf; + uint8_t *perfMmapAux; + int cpuInstrFd; + int cpuBranchFd; + int cpuIptBtsFd; + + } arch_linux; + +} run_t; + +#endif + diff --git a/custom_mutators/honggfuzz/input.h b/custom_mutators/honggfuzz/input.h new file mode 100644 index 00000000..7b0c55ae --- /dev/null +++ b/custom_mutators/honggfuzz/input.h @@ -0,0 +1,106 @@ +#ifndef _HG_INPUT_ +#define _HG_INPUT_ + +#include <stdarg.h> +#ifdef __clang__ +#include <stdatomic.h> +#endif +#include <stdbool.h> +#include <stdint.h> +#include <time.h> + +#include "honggfuzz.h" +#include "afl-fuzz.h" + +/* + * Go-style defer scoped implementation + * + * If compiled with clang, use: -fblocks -lBlocksRuntime + * + * Example of use: + * + * { + * int fd = open(fname, O_RDONLY); + * if (fd == -1) { + * error(....); + * return; + * } + * defer { close(fd); }; + * ssize_t sz = read(fd, buf, sizeof(buf)); + * ... + * ... + * } + * + */ + +#define __STRMERGE(a, b) a##b +#define _STRMERGE(a, b) __STRMERGE(a, b) +#ifdef __clang__ +#if __has_extension(blocks) +static void __attribute__((unused)) __clang_cleanup_func(void (^*dfunc)(void)) { + (*dfunc)(); +} + +#define defer \ + void (^_STRMERGE(__defer_f_, __COUNTER__))(void) \ + __attribute__((cleanup(__clang_cleanup_func))) __attribute__((unused)) = ^ + +#else /* __has_extension(blocks) */ +#define defer UNIMPLEMENTED - NO - SUPPORT - FOR - BLOCKS - IN - YOUR - CLANG - ENABLED +#endif /* __has_extension(blocks) */ +#else /* !__clang__, e.g.: gcc */ + +#define __block +#define _DEFER(a, count) \ + auto void _STRMERGE(__defer_f_, count)(void* _defer_arg __attribute__((unused))); \ + int _STRMERGE(__defer_var_, count) __attribute__((cleanup(_STRMERGE(__defer_f_, count)))) \ + __attribute__((unused)); \ + void _STRMERGE(__defer_f_, count)(void* _defer_arg __attribute__((unused))) +#define defer _DEFER(a, __COUNTER__) +#endif /* ifdef __clang__ */ + +#define HF_MIN(x, y) (x <= y ? x : y) +#define HF_MAX(x, y) (x >= y ? x : y) +#define ATOMIC_GET +#define ARRAYSIZE(x) (sizeof(x) / sizeof(*x)) +#define HF_ATTR_UNUSED __attribute__((unused)) +#define util_Malloc(x) malloc(x) + +extern uint8_t * queue_input; +extern size_t queue_input_size; +extern afl_state_t * afl_struct; + +inline void wmb() { } +inline void LOG_F(const char *format, ...) { } +static inline uint64_t util_rndGet(uint64_t min, uint64_t max) { + return min + rand_below(afl_struct, max - min + 1); +} +static inline uint64_t util_rnd64() { return rand_below(afl_struct, 1 << 30); } + +static inline size_t input_getRandomInputAsBuf(run_t *run, const uint8_t **buf) { + *buf = queue_input; + run->dynfile->data = queue_input; + run->dynfile->size = queue_input_size; + return queue_input_size; +} +static inline void input_setSize(run_t* run, size_t sz) { + run->dynfile->size = sz; +} +static inline void util_turnToPrintable(uint8_t* buf, size_t sz) { + for (size_t i = 0; i < sz; i++) + buf[i] = buf[i] % 95 + 32; +} +static inline void util_rndBuf(uint8_t* buf, size_t sz) { + if (sz == 0) return; + for (size_t i = 0; i < sz; i++) + buf[i] = (uint8_t)rand_below(afl_struct, 256); +} +static inline uint8_t util_rndPrintable() { + return 32 + rand_below(afl_struct, 127 - 32); +} +static inline void util_rndBufPrintable(uint8_t* buf, size_t sz) { + for (size_t i = 0; i < sz; i++) + buf[i] = util_rndPrintable(); +} + +#endif diff --git a/custom_mutators/honggfuzz/libhfcommon b/custom_mutators/honggfuzz/libhfcommon new file mode 120000 index 00000000..945c9b46 --- /dev/null +++ b/custom_mutators/honggfuzz/libhfcommon @@ -0,0 +1 @@ +. \ No newline at end of file diff --git a/custom_mutators/honggfuzz/log.h b/custom_mutators/honggfuzz/log.h new file mode 120000 index 00000000..51e19654 --- /dev/null +++ b/custom_mutators/honggfuzz/log.h @@ -0,0 +1 @@ +common.h \ No newline at end of file diff --git a/custom_mutators/honggfuzz/mangle.c b/custom_mutators/honggfuzz/mangle.c new file mode 100644 index 00000000..05e0dcfa --- /dev/null +++ b/custom_mutators/honggfuzz/mangle.c @@ -0,0 +1,1039 @@ +/* + * + * honggfuzz - run->dynfile->datafer mangling routines + * ----------------------------------------- + * + * Author: + * Robert Swiecki <swiecki@google.com> + * + * Copyright 2010-2018 by Google Inc. All Rights Reserved. + * + * Licensed under the Apache License, Version 2.0 (the "License"); you may + * not use this file except in compliance with the License. You may obtain + * a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + * implied. See the License for the specific language governing + * permissions and limitations under the License. + * + */ + +#include "mangle.h" + +#include <ctype.h> +#include <inttypes.h> +#include <stdbool.h> +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <sys/types.h> +#include <time.h> + +#include "input.h" +#include "libhfcommon/common.h" +#include "libhfcommon/log.h" +#include "libhfcommon/util.h" + +static inline size_t mangle_LenLeft(run_t *run, size_t off) { + + if (off >= run->dynfile->size) { + + LOG_F("Offset is too large: off:%zu >= len:%zu", off, run->dynfile->size); + + } + + return (run->dynfile->size - off - 1); + +} + +/* Get a random value between <1:max> with x^2 distribution */ +static inline size_t mangle_getLen(size_t max) { + + if (max > _HF_INPUT_MAX_SIZE) { + + LOG_F("max (%zu) > _HF_INPUT_MAX_SIZE (%zu)", max, + (size_t)_HF_INPUT_MAX_SIZE); + + } + + if (max == 0) { LOG_F("max == 0"); } + if (max == 1) { return 1; } + + const uint64_t max2 = (uint64_t)max * max; + const uint64_t max3 = (uint64_t)max * max * max; + const uint64_t rnd = util_rndGet(1, max2 - 1); + + uint64_t ret = rnd * rnd; + ret /= max3; + ret += 1; + + if (ret < 1) { + + LOG_F("ret (%" PRIu64 ") < 1, max:%zu, rnd:%" PRIu64, ret, max, rnd); + + } + + if (ret > max) { + + LOG_F("ret (%" PRIu64 ") > max (%zu), rnd:%" PRIu64, ret, max, rnd); + + } + + return (size_t)ret; + +} + +/* Prefer smaller values here, so use mangle_getLen() */ +static inline size_t mangle_getOffSet(run_t *run) { + + return mangle_getLen(run->dynfile->size) - 1; + +} + +/* Offset which can be equal to the file size */ +static inline size_t mangle_getOffSetPlus1(run_t *run) { + + size_t reqlen = HF_MIN(run->dynfile->size + 1, _HF_INPUT_MAX_SIZE); + return mangle_getLen(reqlen) - 1; + +} + +static inline void mangle_Move(run_t *run, size_t off_from, size_t off_to, + size_t len) { + + if (off_from >= run->dynfile->size) { return; } + if (off_to >= run->dynfile->size) { return; } + if (off_from == off_to) { return; } + + size_t len_from = run->dynfile->size - off_from; + len = HF_MIN(len, len_from); + + size_t len_to = run->dynfile->size - off_to; + len = HF_MIN(len, len_to); + + memmove(&run->dynfile->data[off_to], &run->dynfile->data[off_from], len); + +} + +static inline void mangle_Overwrite(run_t *run, size_t off, const uint8_t *src, + size_t len, bool printable) { + + if (len == 0) { return; } + size_t maxToCopy = run->dynfile->size - off; + if (len > maxToCopy) { len = maxToCopy; } + + memmove(&run->dynfile->data[off], src, len); + if (printable) { util_turnToPrintable(&run->dynfile->data[off], len); } + +} + +static inline size_t mangle_Inflate(run_t *run, size_t off, size_t len, + bool printable) { + + if (run->dynfile->size >= run->global->mutate.maxInputSz) { return 0; } + if (len > (run->global->mutate.maxInputSz - run->dynfile->size)) { + + len = run->global->mutate.maxInputSz - run->dynfile->size; + + } + + input_setSize(run, run->dynfile->size + len); + mangle_Move(run, off, off + len, run->dynfile->size); + if (printable) { memset(&run->dynfile->data[off], ' ', len); } + + return len; + +} + +static inline void mangle_Insert(run_t *run, size_t off, const uint8_t *val, + size_t len, bool printable) { + + len = mangle_Inflate(run, off, len, printable); + mangle_Overwrite(run, off, val, len, printable); + +} + +static inline void mangle_UseValue(run_t *run, const uint8_t *val, size_t len, + bool printable) { + + if (util_rnd64() % 2) { + + mangle_Insert(run, mangle_getOffSetPlus1(run), val, len, printable); + + } else { + + mangle_Overwrite(run, mangle_getOffSet(run), val, len, printable); + + } + +} + +static void mangle_MemSwap(run_t *run, bool printable HF_ATTR_UNUSED) { + + size_t off1 = mangle_getOffSet(run); + size_t maxlen1 = run->dynfile->size - off1; + + size_t off2 = mangle_getOffSet(run); + size_t maxlen2 = run->dynfile->size - off2; + + size_t len = mangle_getLen(HF_MIN(maxlen1, maxlen2)); + uint8_t *tmpbuf = (uint8_t *)util_Malloc(len); + defer { + + free(tmpbuf); + + }; + + memcpy(tmpbuf, &run->dynfile->data[off1], len); + memmove(&run->dynfile->data[off1], &run->dynfile->data[off2], len); + memcpy(&run->dynfile->data[off2], tmpbuf, len); + +} + +static void mangle_MemCopy(run_t *run, bool printable HF_ATTR_UNUSED) { + + size_t off = mangle_getOffSet(run); + size_t len = mangle_getLen(run->dynfile->size - off); + + /* Use a temp buf, as Insert/Inflate can change source bytes */ + uint8_t *tmpbuf = (uint8_t *)util_Malloc(len); + defer { + + free(tmpbuf); + + }; + + memcpy(tmpbuf, &run->dynfile->data[off], len); + + mangle_UseValue(run, tmpbuf, len, printable); + +} + +static void mangle_Bytes(run_t *run, bool printable) { + + uint16_t buf; + if (printable) { + + util_rndBufPrintable((uint8_t *)&buf, sizeof(buf)); + + } else { + + buf = util_rnd64(); + + } + + /* Overwrite with random 1-2-byte values */ + size_t toCopy = util_rndGet(1, 2); + mangle_UseValue(run, (const uint8_t *)&buf, toCopy, printable); + +} + +static void mangle_ByteRepeatOverwrite(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t destOff = off + 1; + size_t maxSz = run->dynfile->size - destOff; + + /* No space to repeat */ + if (!maxSz) { + + mangle_Bytes(run, printable); + return; + + } + + size_t len = mangle_getLen(maxSz); + memset(&run->dynfile->data[destOff], run->dynfile->data[off], len); + +} + +static void mangle_ByteRepeatInsert(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t destOff = off + 1; + size_t maxSz = run->dynfile->size - destOff; + + /* No space to repeat */ + if (!maxSz) { + + mangle_Bytes(run, printable); + return; + + } + + size_t len = mangle_getLen(maxSz); + len = mangle_Inflate(run, destOff, len, printable); + memset(&run->dynfile->data[destOff], run->dynfile->data[off], len); + +} + +static void mangle_Bit(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + run->dynfile->data[off] ^= (uint8_t)(1U << util_rndGet(0, 7)); + if (printable) { util_turnToPrintable(&(run->dynfile->data[off]), 1); } + +} + +static const struct { + + const uint8_t val[8]; + const size_t size; + +} mangleMagicVals[] = { + + /* 1B - No endianness */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x01\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x02\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x03\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x04\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x05\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x06\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x07\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x08\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x09\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0A\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0B\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0C\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0D\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0E\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x0F\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x10\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x20\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x40\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x7E\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x7F\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\x81\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\xC0\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\xFE\x00\x00\x00\x00\x00\x00\x00", 1}, + {"\xFF\x00\x00\x00\x00\x00\x00\x00", 1}, + /* 2B - NE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x01\x01\x00\x00\x00\x00\x00\x00", 2}, + {"\x80\x80\x00\x00\x00\x00\x00\x00", 2}, + {"\xFF\xFF\x00\x00\x00\x00\x00\x00", 2}, + /* 2B - BE */ + {"\x00\x01\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x02\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x03\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x04\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x05\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x06\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x07\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x08\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x09\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0A\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0B\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0C\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0D\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0E\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x0F\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x10\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x20\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x40\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x7E\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x7F\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x80\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x81\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\xC0\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\xFE\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\xFF\x00\x00\x00\x00\x00\x00", 2}, + {"\x7E\xFF\x00\x00\x00\x00\x00\x00", 2}, + {"\x7F\xFF\x00\x00\x00\x00\x00\x00", 2}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x80\x01\x00\x00\x00\x00\x00\x00", 2}, + {"\xFF\xFE\x00\x00\x00\x00\x00\x00", 2}, + /* 2B - LE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x01\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x02\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x03\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x04\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x05\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x06\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x07\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x08\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x09\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0A\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0B\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0C\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0D\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0E\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x0F\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x10\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x20\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x40\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x7E\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x7F\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\x81\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\xC0\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\xFE\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\xFF\x00\x00\x00\x00\x00\x00\x00", 2}, + {"\xFF\x7E\x00\x00\x00\x00\x00\x00", 2}, + {"\xFF\x7F\x00\x00\x00\x00\x00\x00", 2}, + {"\x00\x80\x00\x00\x00\x00\x00\x00", 2}, + {"\x01\x80\x00\x00\x00\x00\x00\x00", 2}, + {"\xFE\xFF\x00\x00\x00\x00\x00\x00", 2}, + /* 4B - NE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x01\x01\x01\x01\x00\x00\x00\x00", 4}, + {"\x80\x80\x80\x80\x00\x00\x00\x00", 4}, + {"\xFF\xFF\xFF\xFF\x00\x00\x00\x00", 4}, + /* 4B - BE */ + {"\x00\x00\x00\x01\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x02\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x03\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x04\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x05\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x06\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x07\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x08\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x09\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0A\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0B\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0C\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0D\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0E\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x0F\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x10\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x20\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x40\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x7E\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x7F\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x80\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x81\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\xC0\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\xFE\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\xFF\x00\x00\x00\x00", 4}, + {"\x7E\xFF\xFF\xFF\x00\x00\x00\x00", 4}, + {"\x7F\xFF\xFF\xFF\x00\x00\x00\x00", 4}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x80\x00\x00\x01\x00\x00\x00\x00", 4}, + {"\xFF\xFF\xFF\xFE\x00\x00\x00\x00", 4}, + /* 4B - LE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x01\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x02\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x03\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x04\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x05\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x06\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x07\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x08\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x09\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0A\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0B\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0C\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0D\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0E\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x0F\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x10\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x20\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x40\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x7E\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x7F\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\x81\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\xC0\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\xFE\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\xFF\x00\x00\x00\x00\x00\x00\x00", 4}, + {"\xFF\xFF\xFF\x7E\x00\x00\x00\x00", 4}, + {"\xFF\xFF\xFF\x7F\x00\x00\x00\x00", 4}, + {"\x00\x00\x00\x80\x00\x00\x00\x00", 4}, + {"\x01\x00\x00\x80\x00\x00\x00\x00", 4}, + {"\xFE\xFF\xFF\xFF\x00\x00\x00\x00", 4}, + /* 8B - NE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x01\x01\x01\x01\x01\x01\x01\x01", 8}, + {"\x80\x80\x80\x80\x80\x80\x80\x80", 8}, + {"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", 8}, + /* 8B - BE */ + {"\x00\x00\x00\x00\x00\x00\x00\x01", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x02", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x03", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x04", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x05", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x06", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x07", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x08", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x09", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0A", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0B", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0C", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0D", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0E", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x0F", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x10", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x20", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x40", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x7E", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x7F", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x80", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x81", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\xC0", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\xFE", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\xFF", 8}, + {"\x7E\xFF\xFF\xFF\xFF\xFF\xFF\xFF", 8}, + {"\x7F\xFF\xFF\xFF\xFF\xFF\xFF\xFF", 8}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x80\x00\x00\x00\x00\x00\x00\x01", 8}, + {"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFE", 8}, + /* 8B - LE */ + {"\x00\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x01\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x02\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x03\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x04\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x05\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x06\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x07\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x08\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x09\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0A\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0B\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0C\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0D\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0E\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x0F\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x10\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x20\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x40\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x7E\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x7F\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x80\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\x81\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\xC0\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\xFE\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\xFF\x00\x00\x00\x00\x00\x00\x00", 8}, + {"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\x7E", 8}, + {"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\x7F", 8}, + {"\x00\x00\x00\x00\x00\x00\x00\x80", 8}, + {"\x01\x00\x00\x00\x00\x00\x00\x80", 8}, + {"\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF", 8}, + +}; + +static void mangle_Magic(run_t *run, bool printable) { + + uint64_t choice = util_rndGet(0, ARRAYSIZE(mangleMagicVals) - 1); + mangle_UseValue(run, mangleMagicVals[choice].val, + mangleMagicVals[choice].size, printable); + +} + +static void mangle_StaticDict(run_t *run, bool printable) { + + if (run->global->mutate.dictionaryCnt == 0) { + + mangle_Bytes(run, printable); + return; + + } + + uint64_t choice = util_rndGet(0, run->global->mutate.dictionaryCnt - 1); + mangle_UseValue(run, run->global->mutate.dictionary[choice].val, + run->global->mutate.dictionary[choice].len, printable); + +} + +static inline const uint8_t *mangle_FeedbackDict(run_t *run, size_t *len) { + + if (!run->global->feedback.cmpFeedback) { return NULL; } + cmpfeedback_t *cmpf = run->global->feedback.cmpFeedbackMap; + uint32_t cnt = ATOMIC_GET(cmpf->cnt); + if (cnt == 0) { return NULL; } + if (cnt > ARRAYSIZE(cmpf->valArr)) { cnt = ARRAYSIZE(cmpf->valArr); } + uint32_t choice = util_rndGet(0, cnt - 1); + *len = (size_t)ATOMIC_GET(cmpf->valArr[choice].len); + if (*len == 0) { return NULL; } + return cmpf->valArr[choice].val; + +} + +static void mangle_ConstFeedbackDict(run_t *run, bool printable) { + + size_t len; + const uint8_t *val = mangle_FeedbackDict(run, &len); + if (val == NULL) { + + mangle_Bytes(run, printable); + return; + + } + + mangle_UseValue(run, val, len, printable); + +} + +static void mangle_MemSet(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t len = mangle_getLen(run->dynfile->size - off); + int val = + printable ? (int)util_rndPrintable() : (int)util_rndGet(0, UINT8_MAX); + + memset(&run->dynfile->data[off], val, len); + +} + +static void mangle_RandomOverwrite(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t len = mangle_getLen(run->dynfile->size - off); + if (printable) { + + util_rndBufPrintable(&run->dynfile->data[off], len); + + } else { + + util_rndBuf(&run->dynfile->data[off], len); + + } + +} + +static void mangle_RandomInsert(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t len = mangle_getLen(run->dynfile->size - off); + + len = mangle_Inflate(run, off, len, printable); + + if (printable) { + + util_rndBufPrintable(&run->dynfile->data[off], len); + + } else { + + util_rndBuf(&run->dynfile->data[off], len); + + } + +} + +static inline void mangle_AddSubWithRange(run_t *run, size_t off, size_t varLen, + uint64_t range, bool printable) { + + int64_t delta = (int64_t)util_rndGet(0, range * 2) - (int64_t)range; + + switch (varLen) { + + case 1: { + + run->dynfile->data[off] += delta; + break; + + } + + case 2: { + + int16_t val; + memcpy(&val, &run->dynfile->data[off], sizeof(val)); + if (util_rnd64() & 0x1) { + + val += delta; + + } else { + + /* Foreign endianess */ + val = __builtin_bswap16(val); + val += delta; + val = __builtin_bswap16(val); + + } + + mangle_Overwrite(run, off, (uint8_t *)&val, varLen, printable); + break; + + } + + case 4: { + + int32_t val; + memcpy(&val, &run->dynfile->data[off], sizeof(val)); + if (util_rnd64() & 0x1) { + + val += delta; + + } else { + + /* Foreign endianess */ + val = __builtin_bswap32(val); + val += delta; + val = __builtin_bswap32(val); + + } + + mangle_Overwrite(run, off, (uint8_t *)&val, varLen, printable); + break; + + } + + case 8: { + + int64_t val; + memcpy(&val, &run->dynfile->data[off], sizeof(val)); + if (util_rnd64() & 0x1) { + + val += delta; + + } else { + + /* Foreign endianess */ + val = __builtin_bswap64(val); + val += delta; + val = __builtin_bswap64(val); + + } + + mangle_Overwrite(run, off, (uint8_t *)&val, varLen, printable); + break; + + } + + default: { + + LOG_F("Unknown variable length size: %zu", varLen); + + } + + } + +} + +static void mangle_AddSub(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + + /* 1,2,4,8 */ + size_t varLen = 1U << util_rndGet(0, 3); + if ((run->dynfile->size - off) < varLen) { varLen = 1; } + + uint64_t range; + switch (varLen) { + + case 1: + range = 16; + break; + case 2: + range = 4096; + break; + case 4: + range = 1048576; + break; + case 8: + range = 268435456; + break; + default: + LOG_F("Invalid operand size: %zu", varLen); + + } + + mangle_AddSubWithRange(run, off, varLen, range, printable); + +} + +static void mangle_IncByte(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + if (printable) { + + run->dynfile->data[off] = (run->dynfile->data[off] - 32 + 1) % 95 + 32; + + } else { + + run->dynfile->data[off] += (uint8_t)1UL; + + } + +} + +static void mangle_DecByte(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + if (printable) { + + run->dynfile->data[off] = (run->dynfile->data[off] - 32 + 94) % 95 + 32; + + } else { + + run->dynfile->data[off] -= (uint8_t)1UL; + + } + +} + +static void mangle_NegByte(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + if (printable) { + + run->dynfile->data[off] = 94 - (run->dynfile->data[off] - 32) + 32; + + } else { + + run->dynfile->data[off] = ~(run->dynfile->data[off]); + + } + +} + +static void mangle_Expand(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + size_t len; + if (util_rnd64() % 16) { + + len = mangle_getLen(HF_MIN(16, run->global->mutate.maxInputSz - off)); + + } else { + + len = mangle_getLen(run->global->mutate.maxInputSz - off); + + } + + mangle_Inflate(run, off, len, printable); + +} + +static void mangle_Shrink(run_t *run, bool printable HF_ATTR_UNUSED) { + + if (run->dynfile->size <= 2U) { return; } + + size_t off_start = mangle_getOffSet(run); + size_t len = mangle_LenLeft(run, off_start); + if (len == 0) { return; } + if (util_rnd64() % 16) { + + len = mangle_getLen(HF_MIN(16, len)); + + } else { + + len = mangle_getLen(len); + + } + + size_t off_end = off_start + len; + size_t len_to_move = run->dynfile->size - off_end; + + mangle_Move(run, off_end, off_start, len_to_move); + input_setSize(run, run->dynfile->size - len); + +} + +static void mangle_ASCIINum(run_t *run, bool printable) { + + size_t len = util_rndGet(2, 8); + + char buf[20]; + snprintf(buf, sizeof(buf), "%-19" PRId64, (int64_t)util_rnd64()); + + mangle_UseValue(run, (const uint8_t *)buf, len, printable); + +} + +static void mangle_ASCIINumChange(run_t *run, bool printable) { + + size_t off = mangle_getOffSet(run); + + /* Find a digit */ + for (; off < run->dynfile->size; off++) { + + if (isdigit(run->dynfile->data[off])) { break; } + + } + + if (off == run->dynfile->size) { + + mangle_Bytes(run, printable); + return; + + } + + size_t len = HF_MIN(20, run->dynfile->size - off); + char numbuf[21] = {}; + strncpy(numbuf, (const char *)&run->dynfile->data[off], len); + uint64_t val = (uint64_t)strtoull(numbuf, NULL, 10); + + switch (util_rndGet(0, 5)) { + + case 0: + val += util_rndGet(1, 256); + break; + case 1: + val -= util_rndGet(1, 256); + break; + case 2: + val *= util_rndGet(1, 256); + break; + case 3: + val /= util_rndGet(1, 256); + break; + case 4: + val = ~(val); + break; + case 5: + val = util_rnd64(); + break; + default: + LOG_F("Invalid choice"); + + }; + + len = HF_MIN((size_t)snprintf(numbuf, sizeof(numbuf), "%" PRIu64, val), len); + mangle_Overwrite(run, off, (const uint8_t *)numbuf, len, printable); + +} + +static void mangle_Splice(run_t *run, bool printable) { + + const uint8_t *buf; + size_t sz = input_getRandomInputAsBuf(run, &buf); + if (!sz) { + + mangle_Bytes(run, printable); + return; + + } + + size_t remoteOff = mangle_getLen(sz) - 1; + size_t len = mangle_getLen(sz - remoteOff); + mangle_UseValue(run, &buf[remoteOff], len, printable); + +} + +static void mangle_Resize(run_t *run, bool printable) { + + ssize_t oldsz = run->dynfile->size; + ssize_t newsz = 0; + + uint64_t choice = util_rndGet(0, 32); + switch (choice) { + + case 0: /* Set new size arbitrarily */ + newsz = (ssize_t)util_rndGet(1, run->global->mutate.maxInputSz); + break; + case 1 ... 4: /* Increase size by a small value */ + newsz = oldsz + (ssize_t)util_rndGet(0, 8); + break; + case 5: /* Increase size by a larger value */ + newsz = oldsz + (ssize_t)util_rndGet(9, 128); + break; + case 6 ... 9: /* Decrease size by a small value */ + newsz = oldsz - (ssize_t)util_rndGet(0, 8); + break; + case 10: /* Decrease size by a larger value */ + newsz = oldsz - (ssize_t)util_rndGet(9, 128); + break; + case 11 ... 32: /* Do nothing */ + newsz = oldsz; + break; + default: + LOG_F("Illegal value from util_rndGet: %" PRIu64, choice); + break; + + } + + if (newsz < 1) { newsz = 1; } + if (newsz > (ssize_t)run->global->mutate.maxInputSz) { + + newsz = run->global->mutate.maxInputSz; + + } + + input_setSize(run, (size_t)newsz); + if (newsz > oldsz) { + + if (printable) { memset(&run->dynfile->data[oldsz], ' ', newsz - oldsz); } + + } + +} + +void mangle_mangleContent(run_t *run, int speed_factor) { + + static void (*const mangleFuncs[])(run_t * run, bool printable) = { + + /* Every *Insert or Expand expands file, so add more Shrink's */ + mangle_Shrink, + mangle_Shrink, + mangle_Shrink, + mangle_Shrink, + mangle_Expand, + mangle_Bit, + mangle_IncByte, + mangle_DecByte, + mangle_NegByte, + mangle_AddSub, + mangle_MemSet, + mangle_MemSwap, + mangle_MemCopy, + mangle_Bytes, + mangle_ASCIINum, + mangle_ASCIINumChange, + mangle_ByteRepeatOverwrite, + mangle_ByteRepeatInsert, + mangle_Magic, + mangle_StaticDict, + mangle_ConstFeedbackDict, + mangle_RandomOverwrite, + mangle_RandomInsert, + mangle_Splice, + + }; + + if (run->mutationsPerRun == 0U) { return; } + if (run->dynfile->size == 0U) { + + mangle_Resize(run, /* printable= */ run->global->cfg.only_printable); + + } + + uint64_t changesCnt = run->global->mutate.mutationsPerRun; + + if (speed_factor < 5) { + + changesCnt = util_rndGet(1, run->global->mutate.mutationsPerRun); + + } else if (speed_factor < 10) { + + changesCnt = run->global->mutate.mutationsPerRun; + + } else { + + changesCnt = HF_MIN(speed_factor, 12); + changesCnt = HF_MAX(changesCnt, run->global->mutate.mutationsPerRun); + + } + + /* If last coverage acquisition was more than 5 secs ago, use splicing more + * frequently */ + if ((time(NULL) - ATOMIC_GET(run->global->timing.lastCovUpdate)) > 5) { + + if (util_rnd64() % 2) { + + mangle_Splice(run, run->global->cfg.only_printable); + + } + + } + + for (uint64_t x = 0; x < changesCnt; x++) { + + uint64_t choice = util_rndGet(0, ARRAYSIZE(mangleFuncs) - 1); + mangleFuncs[choice](run, /* printable= */ run->global->cfg.only_printable); + + } + + wmb(); + +} + diff --git a/custom_mutators/honggfuzz/mangle.h b/custom_mutators/honggfuzz/mangle.h new file mode 100644 index 00000000..1b6a4943 --- /dev/null +++ b/custom_mutators/honggfuzz/mangle.h @@ -0,0 +1,32 @@ +/* + * + * honggfuzz - buffer mangling routines + * ----------------------------------------- + * + * Author: Robert Swiecki <swiecki@google.com> + * + * Copyright 2010-2018 by Google Inc. All Rights Reserved. + * + * Licensed under the Apache License, Version 2.0 (the "License"); you may + * not use this file except in compliance with the License. You may obtain + * a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + * implied. See the License for the specific language governing + * permissions and limitations under the License. + * + */ + +#ifndef _HF_MANGLE_H_ +#define _HF_MANGLE_H_ + +#include "honggfuzz.h" + +extern void mangle_mangleContent(run_t *run, int speed_factor); + +#endif + diff --git a/custom_mutators/honggfuzz/util.h b/custom_mutators/honggfuzz/util.h new file mode 120000 index 00000000..51e19654 --- /dev/null +++ b/custom_mutators/honggfuzz/util.h @@ -0,0 +1 @@ +common.h \ No newline at end of file diff --git a/dictionaries/rtf.dict b/dictionaries/rtf.dict index 8b8f0cad..2d5019ef 100644 --- a/dictionaries/rtf.dict +++ b/dictionaries/rtf.dict @@ -3,10 +3,12 @@ # charset "\\ansi" "\\mac" +"\\pc" "\\pca" # font table "\\fnil" +"\\fRoman" "\\fswiss" "\\fmodern" "\\fscript" @@ -28,6 +30,7 @@ "\\cb" # pictures +"\\pict" "\\macpict" "\\pmmetafile" "\\wmetafile" @@ -40,36 +43,15 @@ "\\pich" "\\picwgoal" "\\pichgoal" +"\\picscalex" "\\picscaley" "\\picscaled" "\\piccropt" "\\piccropb" "\\piccropl" "\\piccropr" -"\\brdrs" -"\\brdrdb" -"\\brdrth" -"\\brdrsh" -"\\brdrdot" -"\\brdrhair" -"\\brdrw" -"\\brdrcf" -"\\shading" -"\\bghoriz" -"\\bgvert" -"\\bgfdiag" -"\\bgbdiag" -"\\bgcross" -"\\bgdcross" -"\\bgdkhoriz" -"\\bgdkvert" -"\\bgdkfdiag" -"\\bgdkbdiag" -"\\bgdkcross" -"\\bgdkdcross" -"\\cfpat" -"\\cbpat" "\\bin" +# these strings are probably not necessary "MM_TEXT" "MM_LOMETRIC" "MM_HIMETRIC" @@ -86,25 +68,31 @@ "PU_HIENGLISH" "PU_TWIPS" -# headers and gooters -"\headerr" -"\headerf" -"\footerl" -"\footerr" -"\footerf" +# headers and footers +"\\headerl" +"\\headerr" +"\\headerf" +"\\footerl" +"\\footerr" +"\\footerf" # misc -"\\chftn" "\\*\\footnote" "\\*\\annotation" +"\\xe" +"\\txe" +"\\rxe" "\\bxe" "\\ixe" "\\tcf" "\\tcl" "\\*\\bkmkstart" "\\*\\bkmkend" +"\\bkmkcolf" +"\\bkmkcoll" # metadata +"\\info" "\\title" "\\subject" "\\author" @@ -128,10 +116,13 @@ "\\nofwords" "\\nofchars" "\\id" +"\\field" "\\flddirty" "\\fldedit" "\\fldlock" "\\fldpriv" +"\\*\\fldinst" +"\\fldrslt" # objects "\\objemb" @@ -166,16 +157,14 @@ # macintosh editor "\\bkmkpub" "\\pubauto" -"\\objalias" -"\\objsect" # formating "\\deftab" "\\hyphhotz" "\\linestart" "\\fracwidth" -"\\*\nextfile" -"\\*\template" +"\\*\\nextfile" +"\\*\\template" "\\makebackup" "\\defformat" "\\psover" diff --git a/docs/Changelog.md b/docs/Changelog.md index afb9dea6..ae7377f2 100644 --- a/docs/Changelog.md +++ b/docs/Changelog.md @@ -9,6 +9,37 @@ Want to stay in the loop on major new features? Join our mailing list by sending a mail to <afl-users+subscribe@googlegroups.com>. +### Version ++2.66d (devel) + - Support for improved afl++ snapshot module: + https://github.com/AFLplusplus/AFL-Snapshot-LKM + - afl-fuzz: + - added -F option to allow -M main fuzzers to sync to foreign fuzzers, + e.g. honggfuzz or libfuzzer + - added -b option to bind to a specific CPU + - eliminated CPU affinity race condition for -S/-M runs + - expanded havoc mode added, on no cycle finds add extra splicing and + MOpt into the mix + - fixed a bug in redqueen for strings + - llvm_mode: + - now supports llvm 12! + - fixes for laf-intel float splitting (thanks to mark-griffin for + reporting) + - LTO: autodictionary mode is a default + - LTO: instrim instrumentation disabled, only classic support used + as it is always better + - LTO: env var AFL_LLVM_DOCUMENT_IDS=file will document which edge ID + was given to which function during compilation + - LTO: single block functions were not implemented by default, fixed + - LTO: AFL_LLVM_SKIP_NEVERZERO behaviour was inversed, fixed + - setting AFL_LLVM_LAF_SPLIT_FLOATS now activates + AFL_LLVM_LAF_SPLIT_COMPARES + - added honggfuzz mangle as a custom mutator in custom_mutators/honggfuzz + - added afl-frida gum solution to examples/afl_frida (mostly imported + from https://github.com/meme/hotwax/) + - small fixes to afl-plot, afl-whatsup and man page creation + - new README, added FAQ + + ### Version ++2.66c (release) - renamed the main branch on Github to "stable" - renamed master/slave to main/secondary diff --git a/docs/FAQ.md b/docs/FAQ.md new file mode 100644 index 00000000..c15cd484 --- /dev/null +++ b/docs/FAQ.md @@ -0,0 +1,126 @@ +# Frequently asked questions about afl++ + +## Contents + + 1. [How to improve the fuzzing speed?](#how-to-improve-the-fuzzing-speed) + 2. [What is an edge?](#what-is-an-edge) + 3. [Why is my stability below 100%?](#why-is-my-stability-below-100) + 4. [How can I improve the stability value](#how-can-i-improve-the-stability-value) + +If you find an interesting or important question missing, submit it via +[https://github.com/AFLplusplus/AFLplusplus/issues](https://github.com/AFLplusplus/AFLplusplus/issues) + +## How to improve the fuzzing speed + + 1. use [llvm_mode](docs/llvm_mode/README.md): afl-clang-lto (llvm >= 11) or afl-clang-fast (llvm >= 9 recommended) + 2. Use [persistent mode](llvm_mode/README.persistent_mode.md) (x2-x20 speed increase) + 3. Use the [afl++ snapshot module](https://github.com/AFLplusplus/AFL-Snapshot-LKM) (x2 speed increase) + 4. If you do not use shmem persistent mode, use `AFL_TMPDIR` to point the input file on a tempfs location, see [docs/env_variables.md](docs/env_variables.md) + 5. Improve kernel performance: modify `/etc/default/grub`, set `GRUB_CMDLINE_LINUX_DEFAULT="ibpb=off ibrs=off kpti=off l1tf=off mds=off mitigations=off no_stf_barrier noibpb noibrs nopcid nopti nospec_store_bypass_disable nospectre_v1 nospectre_v2 pcid=off pti=off spec_store_bypass_disable=off spectre_v2=off stf_barrier=off"`; then `update-grub` and `reboot` (warning: makes the system more insecure) + 6. Running on an `ext2` filesystem with `noatime` mount option will be a bit faster than on any other journaling filesystem + 7. Use your cores! [README.md:3.b) Using multiple cores/threads](../README.md#b-using-multiple-coresthreads) + +## What is an "edge" + +A program contains `functions`, `functions` contain the compiled machine code. +The compiled machine code in a `function` can be in a single or many `basic blocks`. +A `basic block` is the largest possible number of subsequent machine code +instructions that runs independent, meaning it does not split up to different +locations nor is it jumped into it from a different location: +``` +function() { + A: + some + code + B: + if (x) goto C; else goto D; + C: + some code + goto D + D: + some code + goto B + E: + return +} +``` +Every code block between two jump locations is a `basic block`. + +An `edge` is then the unique relationship between two `basic blocks` (from the +code example above): +``` + Block A + | + v + Block B <------+ + / \ | + v v | + Block C Block D --+ + \ + v + Block E +``` +Every line between two blocks is an `edge`. + +## Why is my stability below 100 + +Stability is measured by how many percent of the edges in the target are +"stable". Sending the same input again and again should take the exact same +path through the target every time. If that is the case, the stability is 100%. + +If however randomness happens, e.g. a thread reading from shared memory, +reaction to timing, etc. then in some of the re-executions with the same data +will result in the edge information being different accross runs. +Those edges that change are then flagged "unstable". + +The more "unstable" edges, the more difficult for afl++ to identify valid new +paths. + +A value above 90% is usually fine and a value above 80% is also still ok, and +even above 20% can still result in successful finds of bugs. +However, it is recommended that below 90% or 80% you should take measures to +improve the stability. + +## How can I improve the stability value + +Four steps are required to do this and requires quite some knowledge of +coding and/or disassembly and it is only effectively possible with +afl-clang-fast PCGUARD and afl-clang-lto LTO instrumentation! + + 1. First step: Identify which edge ID numbers are unstable + + run the target with `export AFL_DEBUG=1` for a few minutes then terminate. + The out/fuzzer_stats file will then show the edge IDs that were identified + as unstable. + + 2. Second step: Find the responsible function. + + a) For LTO instrumented binaries this can be documented during compile + time, just set `export AFL_LLVM_DOCUMENT_IDS=/path/to/afile`. + This file will have one assigned edge ID and the corresponding function + per line. + + b) For PCGUARD instrumented binaries it is much more difficult. Here you + can either modify the __sanitizer_cov_trace_pc_guard function in + llvm_mode/afl-llvm-rt.o.c to write a backtrace to a file if the ID in + __afl_area_ptr[*guard] is one of the unstable edge IDs. + (Example code is already there). + Then recompile and reinstall llvm_mode and rebuild your target. + Run the recompiled target with afl-fuzz for a while and then check the + file that you wrote with the backtrace information. + Alternatively you can use `gdb` to hook __sanitizer_cov_trace_pc_guard_init + on start, check to which memory address the edge ID value is written + and set a write breakpoint to that address (`watch 0x.....`). + + 3. Third step: create a text file with the filenames + + Identify which source code files contain the functions that you need to + remove from instrumentation. + + Simply follow this document on how to do this: [llvm_mode/README.instrument_file.md](llvm_mode/README.instrument_file.md) + If PCGUARD is used, then you need to follow this guide (needs llvm 12+!): + [http://clang.llvm.org/docs/SanitizerCoverage.html#partially-disabling-instrumentation](http://clang.llvm.org/docs/SanitizerCoverage.html#partially-disabling-instrumentation) + + 4. Fourth step: recompile the target + + Recompile, fuzz it, be happy :) diff --git a/docs/binaryonly_fuzzing.md b/docs/binaryonly_fuzzing.md index 7c9be418..a3d3330f 100644 --- a/docs/binaryonly_fuzzing.md +++ b/docs/binaryonly_fuzzing.md @@ -8,12 +8,17 @@ The following is a description of how these binaries can be fuzzed with afl++ + ## TL;DR: qemu_mode in persistent mode is the fastest - if the stability is high enough. Otherwise try retrowrite, afl-dyninst and if these fail too then standard qemu_mode with AFL_ENTRYPOINT to where you need it. + If your target is a library use examples/afl_frida/. + + If your target is non-linux then use unicorn_mode/. + ## QEMU @@ -57,6 +62,20 @@ As it is included in afl++ this needs no URL. +## AFL FRIDA + + If you want to fuzz a binary-only shared library then you can fuzz it with + frida-gum via examples/afl_frida/, you will have to write a harness to + call the target function in the library, use afl-frida.c as a template. + + +## AFL UNTRACER + + If you want to fuzz a binary-only shared library then you can fuzz it with + examples/afl_untracer/, use afl-untracer.c as a template. + It is slower than AFL FRIDA (see above). + + ## DYNINST Dyninst is a binary instrumentation framework similar to Pintool and diff --git a/docs/custom_mutators.md b/docs/custom_mutators.md index 464acbee..a22c809b 100644 --- a/docs/custom_mutators.md +++ b/docs/custom_mutators.md @@ -36,7 +36,7 @@ size_t afl_custom_fuzz(void *data, uint8_t *buf, size_t buf_size, u8 **out_buf, size_t afl_custom_post_process(void *data, uint8_t *buf, size_t buf_size, uint8_t **out_buf); int32_t afl_custom_init_trim(void *data, uint8_t *buf, size_t buf_size); size_t afl_custom_trim(void *data, uint8_t **out_buf); -int32_t afl_custom_post_trim(void *data, int success) { +int32_t afl_custom_post_trim(void *data, int success); size_t afl_custom_havoc_mutation(void *data, u8 *buf, size_t buf_size, u8 **out_buf, size_t max_size); uint8_t afl_custom_havoc_mutation_probability(void *data); uint8_t afl_custom_queue_get(void *data, const uint8_t *filename); diff --git a/docs/env_variables.md b/docs/env_variables.md index 87344331..811c5658 100644 --- a/docs/env_variables.md +++ b/docs/env_variables.md @@ -121,18 +121,16 @@ Then there are a few specific features that are only available in llvm_mode: built if LLVM 11 or newer is used. - AFL_LLVM_INSTRUMENT=CFG will use Control Flow Graph instrumentation. - (recommended) - - - AFL_LLVM_LTO_AUTODICTIONARY will generate a dictionary in the target - binary based on string compare and memory compare functions. - afl-fuzz will automatically get these transmitted when starting to - fuzz. + (not recommended!) None of the following options are necessary to be used and are rather for manual use (which only ever the author of this LTO implementation will use). These are used if several seperated instrumentation are performed which are then later combined. + - AFL_LLVM_DOCUMENT_IDS=file will document to a file which edge ID was given + to which function. This helps to identify functions with variable bytes + or which functions were touched by an input. - AFL_LLVM_MAP_ADDR sets the fixed map address to a different address than the default 0x10000. A value of 0 or empty sets the map address to be dynamic (the original afl way, which is slower) @@ -254,15 +252,6 @@ checks or alter some of the more exotic semantics of the tool: useful if you can't change the defaults (e.g., no root access to the system) and are OK with some performance loss. - - Setting AFL_NO_FORKSRV disables the forkserver optimization, reverting to - fork + execve() call for every tested input. This is useful mostly when - working with unruly libraries that create threads or do other crazy - things when initializing (before the instrumentation has a chance to run). - - Note that this setting inhibits some of the user-friendly diagnostics - normally done when starting up the forkserver and causes a pretty - significant performance drop. - - AFL_EXIT_WHEN_DONE causes afl-fuzz to terminate when all existing paths have been fuzzed and there were no new finds for a while. This would be normally indicated by the cycle counter in the UI turning green. May be @@ -273,6 +262,9 @@ checks or alter some of the more exotic semantics of the tool: the target. This must be equal or larger than the size the target was compiled with. + - Setting AFL_DISABLE_TRIM tells afl-fuzz to no trim test cases. This is + usually a bad idea! + - Setting AFL_NO_AFFINITY disables attempts to bind to a specific CPU core on Linux systems. This slows things down, but lets you run more instances of afl-fuzz than would be prudent (if you really want to). @@ -338,6 +330,13 @@ checks or alter some of the more exotic semantics of the tool: - In QEMU mode (-Q), AFL_PATH will be searched for afl-qemu-trace. + - Setting AFL_CYCLE_SCHEDULES will switch to a different schedule everytime + a cycle is finished. + + - Setting AFL_EXPAND_HAVOC_NOW will start in the extended havoc mode that + includes costly mutations. afl-fuzz automatically enables this mode when + deemed useful otherwise. + - Setting AFL_PRELOAD causes AFL to set LD_PRELOAD for the target binary without disrupting the afl-fuzz process itself. This is useful, among other things, for bootstrapping libdislocator.so. @@ -365,6 +364,15 @@ checks or alter some of the more exotic semantics of the tool: for an existing out folder, even if a different `-i` was provided. Without this setting, afl-fuzz will refuse execution for a long-fuzzed out dir. + - Setting AFL_NO_FORKSRV disables the forkserver optimization, reverting to + fork + execve() call for every tested input. This is useful mostly when + working with unruly libraries that create threads or do other crazy + things when initializing (before the instrumentation has a chance to run). + + Note that this setting inhibits some of the user-friendly diagnostics + normally done when starting up the forkserver and causes a pretty + significant performance drop. + - Outdated environment variables that are that not supported anymore: AFL_DEFER_FORKSRV AFL_PERSISTENT diff --git a/docs/parallel_fuzzing.md b/docs/parallel_fuzzing.md index 271f8369..2ab1466c 100644 --- a/docs/parallel_fuzzing.md +++ b/docs/parallel_fuzzing.md @@ -99,7 +99,15 @@ example may be: This is not a concern if you use @@ without -f and let afl-fuzz come up with the file name. -## 3) Multi-system parallelization +## 3) Syncing with non-afl fuzzers or independant instances + +A -M main node can be told with the `-F other_fuzzer_queue_directory` option +to sync results from other fuzzers, e.g. libfuzzer or honggfuzz. + +Only the specified directory will by synced into afl, not subdirectories. +The specified directories do not need to exist yet at the start of afl. + +## 4) Multi-system parallelization The basic operating principle for multi-system parallelization is similar to the mechanism explained in section 2. The key difference is that you need to @@ -176,7 +184,7 @@ It is *not* advisable to skip the synchronization script and run the fuzzers directly on a network filesystem; unexpected latency and unkillable processes in I/O wait state can mess things up. -## 4) Remote monitoring and data collection +## 5) Remote monitoring and data collection You can use screen, nohup, tmux, or something equivalent to run remote instances of afl-fuzz. If you redirect the program's output to a file, it will @@ -200,7 +208,7 @@ Keep in mind that crashing inputs are *not* automatically propagated to the main instance, so you may still want to monitor for crashes fleet-wide from within your synchronization or health checking scripts (see afl-whatsup). -## 5) Asymmetric setups +## 6) Asymmetric setups It is perhaps worth noting that all of the following is permitted: diff --git a/docs/screenshot.png b/docs/screenshot.png new file mode 100644 index 00000000..7b4dd7e4 --- /dev/null +++ b/docs/screenshot.png Binary files differdiff --git a/examples/afl_frida/GNUmakefile b/examples/afl_frida/GNUmakefile new file mode 100644 index 00000000..c154f3a4 --- /dev/null +++ b/examples/afl_frida/GNUmakefile @@ -0,0 +1,23 @@ +ifdef DEBUG + OPT=-O0 -D_DEBUG=\"1\" +else + OPT=-O3 -funroll-loops +endif + +all: afl-frida libtestinstr.so + +libfrida-gum.a: + @echo Download and extract frida-gum-devkit-VERSION-PLATFORM.tar.xz for your platform from https://github.com/frida/frida/releases/latest + @exit 1 + +afl-frida: afl-frida.c libfrida-gum.a + $(CC) -g $(OPT) -o afl-frida -Wno-format -Wno-pointer-sign -I. -fpermissive -fPIC afl-frida.c ../../afl-llvm-rt.o libfrida-gum.a -ldl -lresolv -pthread + +libtestinstr.so: libtestinstr.c + $(CC) -g -O0 -fPIC -o libtestinstr.so -shared libtestinstr.c + +clean: + rm -f afl-frida *~ core *.o libtestinstr.so + +deepclean: clean + rm -f libfrida-gum.a frida-gum* diff --git a/examples/afl_frida/Makefile b/examples/afl_frida/Makefile new file mode 100644 index 00000000..0b306dde --- /dev/null +++ b/examples/afl_frida/Makefile @@ -0,0 +1,2 @@ +all: + @echo please use GNU make, thanks! diff --git a/examples/afl_frida/README.md b/examples/afl_frida/README.md new file mode 100644 index 00000000..7743479b --- /dev/null +++ b/examples/afl_frida/README.md @@ -0,0 +1,34 @@ +# afl-frida - faster fuzzing of binary-only libraries + +## Introduction + +afl-frida is an example skeleton file which can easily be used to fuzz +a closed source library. + +It requires less memory and is x5-10 faster than qemu_mode but does not +provide interesting features like compcov or cmplog. + +## How-to + +### Modify afl-frida.c + +Read and modify afl-frida.c then `make`. +To adapt afl-frida.c to your needs, read the header of the file and then +search and edit the `STEP 1`, `STEP 2` and `STEP 3` locations. + +### Fuzzing + +Example (after modifying afl-frida.c to your needs and compile it): +``` +LD_LIBRARY_PATH=/path/to/the/target/library afl-fuzz -i in -o out -- ./afl-frida +``` +(or even remote via afl-network-proxy). + +# Speed and stability + +The speed is very good, about x12 of fork() qemu_mode. +However the stability is low. Reason is currently unknown. + +# Background + +This code is copied for a larger part from https://github.com/meme/hotwax diff --git a/examples/afl_frida/afl-frida.c b/examples/afl_frida/afl-frida.c new file mode 100644 index 00000000..2ad5a72a --- /dev/null +++ b/examples/afl_frida/afl-frida.c @@ -0,0 +1,541 @@ +/* + american fuzzy lop++ - afl-frida skeleton example + ------------------------------------------------- + + Copyright 2020 AFLplusplus Project. All rights reserved. + + Written mostly by meme -> https://github.com/meme/hotwax + + Modificationy by Marc Heuse <mh@mh-sec.de> + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at: + + http://www.apache.org/licenses/LICENSE-2.0 + + HOW-TO + ====== + + You only need to change the following: + + 1. set the defines and function call parameters. + 2. dl load the library you want to fuzz, lookup the functions you need + and setup the calls to these. + 3. in the while loop you call the functions in the necessary order - + incl the cleanup. the cleanup is important! + + Just look these steps up in the code, look for "// STEP x:" + +*/ + +#include <stdio.h> +#include <stdint.h> +#include <unistd.h> +#include <stdint.h> +#include <stddef.h> +#include <sys/shm.h> +#include <dlfcn.h> + +#ifndef __APPLE__ + #include <sys/wait.h> + #include <sys/personality.h> +#endif + +int debug = 0; + +// STEP 1: + +// The presets are for the example libtestinstr.so: + +/* What is the name of the library to fuzz */ +#define TARGET_LIBRARY "libtestinstr.so" + +/* What is the name of the function to fuzz */ +#define TARGET_FUNCTION "testinstr" + +/* here you need to specify the parameter for the target function */ +static void *(*o_function)(uint8_t *, int); + +// END STEP 1 + +#include "frida-gum.h" + +G_BEGIN_DECLS + +#define GUM_TYPE_FAKE_EVENT_SINK (gum_fake_event_sink_get_type()) +G_DECLARE_FINAL_TYPE(GumFakeEventSink, gum_fake_event_sink, GUM, + FAKE_EVENT_SINK, GObject) + +struct _GumFakeEventSink { + + GObject parent; + GumEventType mask; + +}; + +GumEventSink *gum_fake_event_sink_new(void); +void gum_fake_event_sink_reset(GumFakeEventSink *self); + +G_END_DECLS + +static void gum_fake_event_sink_iface_init(gpointer g_iface, + gpointer iface_data); +static void gum_fake_event_sink_finalize(GObject *obj); +static GumEventType gum_fake_event_sink_query_mask(GumEventSink *sink); +static void gum_fake_event_sink_process(GumEventSink *sink, const GumEvent *ev); +void instr_basic_block(GumStalkerIterator *iterator, GumStalkerOutput *output, + gpointer user_data); +void afl_setup(void); +void afl_start_forkserver(void); +int __afl_persistent_loop(unsigned int max_cnt); + +static void gum_fake_event_sink_class_init(GumFakeEventSinkClass *klass) { + + GObjectClass *object_class = G_OBJECT_CLASS(klass); + object_class->finalize = gum_fake_event_sink_finalize; + +} + +static void gum_fake_event_sink_iface_init(gpointer g_iface, + gpointer iface_data) { + + GumEventSinkInterface *iface = (GumEventSinkInterface *)g_iface; + iface->query_mask = gum_fake_event_sink_query_mask; + iface->process = gum_fake_event_sink_process; + +} + +G_DEFINE_TYPE_EXTENDED(GumFakeEventSink, gum_fake_event_sink, G_TYPE_OBJECT, 0, + G_IMPLEMENT_INTERFACE(GUM_TYPE_EVENT_SINK, + gum_fake_event_sink_iface_init)) + +#include "../../config.h" + +// Shared memory fuzzing. +int __afl_sharedmem_fuzzing = 1; +extern unsigned int * __afl_fuzz_len; +extern unsigned char *__afl_fuzz_ptr; + +// Notify AFL about persistent mode. +static volatile char AFL_PERSISTENT[] = "##SIG_AFL_PERSISTENT##"; +int __afl_persistent_loop(unsigned int); + +// Notify AFL about deferred forkserver. +static volatile char AFL_DEFER_FORKSVR[] = "##SIG_AFL_DEFER_FORKSRV##"; +void __afl_manual_init(); + +// Because we do our own logging. +extern uint8_t * __afl_area_ptr; +static __thread guint64 previous_pc; + +// Frida stuff below. +typedef struct { + + GumAddress base_address; + guint64 code_start, code_end; + +} range_t; + +inline static void afl_maybe_log(guint64 current_pc) { + + // fprintf(stderr, "PC: %p ^ %p\n", current_pc, previous_pc); + + current_pc = (current_pc >> 4) ^ (current_pc << 8); + current_pc &= MAP_SIZE - 1; + + __afl_area_ptr[current_pc ^ previous_pc]++; + previous_pc = current_pc >> 1; + +} + +static void on_basic_block(GumCpuContext *context, gpointer user_data) { + + afl_maybe_log((guint64)user_data); + +} + +void instr_basic_block(GumStalkerIterator *iterator, GumStalkerOutput *output, + gpointer user_data) { + + range_t *range = (range_t *)user_data; + + const cs_insn *instr; + gboolean begin = TRUE; + while (gum_stalker_iterator_next(iterator, &instr)) { + + if (begin) { + + if (instr->address >= range->code_start && + instr->address <= range->code_end) { + + gum_stalker_iterator_put_callout(iterator, on_basic_block, + (gpointer)instr->address, NULL); + begin = FALSE; + + } + + } + + gum_stalker_iterator_keep(iterator); + + } + +} + +static void gum_fake_event_sink_init(GumFakeEventSink *self) { + +} + +static void gum_fake_event_sink_finalize(GObject *obj) { + + G_OBJECT_CLASS(gum_fake_event_sink_parent_class)->finalize(obj); + +} + +GumEventSink *gum_fake_event_sink_new(void) { + + GumFakeEventSink *sink; + sink = (GumFakeEventSink *)g_object_new(GUM_TYPE_FAKE_EVENT_SINK, NULL); + return GUM_EVENT_SINK(sink); + +} + +void gum_fake_event_sink_reset(GumFakeEventSink *self) { + +} + +static GumEventType gum_fake_event_sink_query_mask(GumEventSink *sink) { + + return 0; + +} + +typedef struct library_list { + + uint8_t *name; + uint64_t addr_start, addr_end; + +} library_list_t; + +#define MAX_LIB_COUNT 256 +static library_list_t liblist[MAX_LIB_COUNT]; +static u32 liblist_cnt; + +void read_library_information() { + +#if defined(__linux__) + FILE *f; + u8 buf[1024], *b, *m, *e, *n; + + if ((f = fopen("/proc/self/maps", "r")) == NULL) { + + fprintf(stderr, "Error: cannot open /proc/self/maps\n"); + exit(-1); + + } + + if (debug) fprintf(stderr, "Library list:\n"); + while (fgets(buf, sizeof(buf), f)) { + + if (strstr(buf, " r-x")) { + + if (liblist_cnt >= MAX_LIB_COUNT) { + + fprintf( + stderr, + "Warning: too many libraries to old, maximum count of %d reached\n", + liblist_cnt); + return; + + } + + b = buf; + m = index(buf, '-'); + e = index(buf, ' '); + if ((n = rindex(buf, '/')) == NULL) n = rindex(buf, ' '); + if (n && + ((*n >= '0' && *n <= '9') || *n == '[' || *n == '{' || *n == '(')) + n = NULL; + else + n++; + if (b && m && e && n && *n) { + + *m++ = 0; + *e = 0; + if (n[strlen(n) - 1] == '\n') n[strlen(n) - 1] = 0; + + if (rindex(n, '/') != NULL) { + + n = rindex(n, '/'); + n++; + + } + + liblist[liblist_cnt].name = strdup(n); + liblist[liblist_cnt].addr_start = strtoull(b, NULL, 16); + liblist[liblist_cnt].addr_end = strtoull(m, NULL, 16); + if (debug) + fprintf( + stderr, "%s:%llx (%llx-%llx)\n", liblist[liblist_cnt].name, + liblist[liblist_cnt].addr_end - liblist[liblist_cnt].addr_start, + liblist[liblist_cnt].addr_start, + liblist[liblist_cnt].addr_end - 1); + liblist_cnt++; + + } + + } + + } + + if (debug) fprintf(stderr, "\n"); + +#elif defined(__FreeBSD__) + int mib[] = {CTL_KERN, KERN_PROC, KERN_PROC_VMMAP, getpid()}; + char * buf, *start, *end; + size_t miblen = sizeof(mib) / sizeof(mib[0]); + size_t len; + + if (debug) fprintf(stderr, "Library list:\n"); + if (sysctl(mib, miblen, NULL, &len, NULL, 0) == -1) { return; } + + len = len * 4 / 3; + + buf = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_ANON, -1, 0); + if (buf == MAP_FAILED) { return; } + if (sysctl(mib, miblen, buf, &len, NULL, 0) == -1) { + + munmap(buf, len); + return; + + } + + start = buf; + end = buf + len; + + while (start < end) { + + struct kinfo_vmentry *region = (struct kinfo_vmentry *)start; + size_t size = region->kve_structsize; + + if (size == 0) { break; } + + if ((region->kve_protection & KVME_PROT_READ) && + !(region->kve_protection & KVME_PROT_EXEC)) { + + liblist[liblist_cnt].name = + region->kve_path[0] != '\0' ? strdup(region->kve_path) : 0; + liblist[liblist_cnt].addr_start = region->kve_start; + liblist[liblist_cnt].addr_end = region->kve_end; + + if (debug) { + + fprintf(stderr, "%s:%x (%lx-%lx)\n", liblist[liblist_cnt].name, + liblist[liblist_cnt].addr_end - liblist[liblist_cnt].addr_start, + liblist[liblist_cnt].addr_start, + liblist[liblist_cnt].addr_end - 1); + + } + + liblist_cnt++; + + } + + start += size; + + } + +#endif + +} + +library_list_t *find_library(char *name) { + + char *filename = rindex(name, '/'); + + if (filename) + filename++; + else + filename = name; + +#if defined(__linux__) + u32 i; + for (i = 0; i < liblist_cnt; i++) + if (strcmp(liblist[i].name, filename) == 0) return &liblist[i]; +#elif defined(__APPLE__) && defined(__LP64__) + kern_return_t err; + static library_list_t lib; + + // get the list of all loaded modules from dyld + // the task_info mach API will get the address of the dyld all_image_info + // struct for the given task from which we can get the names and load + // addresses of all modules + task_dyld_info_data_t task_dyld_info; + mach_msg_type_number_t count = TASK_DYLD_INFO_COUNT; + err = task_info(mach_task_self(), TASK_DYLD_INFO, + (task_info_t)&task_dyld_info, &count); + + const struct dyld_all_image_infos *all_image_infos = + (const struct dyld_all_image_infos *)task_dyld_info.all_image_info_addr; + const struct dyld_image_info *image_infos = all_image_infos->infoArray; + + for (size_t i = 0; i < all_image_infos->infoArrayCount; i++) { + + const char * image_name = image_infos[i].imageFilePath; + mach_vm_address_t image_load_address = + (mach_vm_address_t)image_infos[i].imageLoadAddress; + if (strstr(image_name, name)) { + + lib.name = name; + lib.addr_start = (u64)image_load_address; + lib.addr_end = 0; + return &lib; + + } + + } + +#endif + + return NULL; + +} + +static void gum_fake_event_sink_process(GumEventSink * sink, + const GumEvent *ev) { + +} + +/* Because this CAN be called more than once, it will return the LAST range */ +static int enumerate_ranges(const GumRangeDetails *details, + gpointer user_data) { + + GumMemoryRange *code_range = (GumMemoryRange *)user_data; + memcpy(code_range, details->range, sizeof(*code_range)); + return 0; + +} + +int main() { + +#ifndef __APPLE__ + (void)personality(ADDR_NO_RANDOMIZE); // disable ASLR +#endif + + // STEP 2: load the library you want to fuzz and lookup the functions, + // inclusive of the cleanup functions. + // If there is just one function, then there is nothing to change + // or add here. + + void *dl = dlopen(TARGET_LIBRARY, RTLD_LAZY); + if (!dl) { + + fprintf(stderr, "Could not load %s\n", TARGET_LIBRARY); + exit(-1); + + } + + if (!(o_function = dlsym(dl, TARGET_FUNCTION))) { + + fprintf(stderr, "Could not find function %s\n", TARGET_FUNCTION); + exit(-1); + + } + + // END STEP 2 + + read_library_information(); + library_list_t *lib = find_library(TARGET_LIBRARY); + + if (lib == NULL) { + + fprintf(stderr, "Could not find target library\n"); + exit(-1); + + } + + gum_init_embedded(); + if (!gum_stalker_is_supported()) { + + gum_deinit_embedded(); + return 1; + + } + + GumStalker *stalker = gum_stalker_new(); + + /* + This does not work here as we load a shared library. pretty sure this + would also be easily solvable with frida gum, but I already have all the + code I need from afl-untracer + + GumAddress base_address = gum_module_find_base_address(TARGET_LIBRARY); + GumMemoryRange code_range; + gum_module_enumerate_ranges(TARGET_LIBRARY, GUM_PAGE_RX, enumerate_ranges, + &code_range); + guint64 code_start = code_range.base_address - base_address; + guint64 code_end = (code_range.base_address + code_range.size) - base_address; + range_t instr_range = {base_address, code_start, code_end}; + */ + range_t instr_range = {0, lib->addr_start, lib->addr_end}; + + GumStalkerTransformer *transformer = + gum_stalker_transformer_make_from_callback(instr_basic_block, + &instr_range, NULL); + + GumEventSink *event_sink = gum_fake_event_sink_new(); + + // to ensure that the signatures are not optimized out + memcpy(__afl_area_ptr, (void *)AFL_PERSISTENT, sizeof(AFL_PERSISTENT) + 1); + memcpy(__afl_area_ptr + 32, (void *)AFL_DEFER_FORKSVR, + sizeof(AFL_DEFER_FORKSVR) + 1); + __afl_manual_init(); + + // + // any expensive target library initialization that has to be done just once + // - put that here + // + + gum_stalker_follow_me(stalker, transformer, event_sink); + + while (__afl_persistent_loop(UINT32_MAX) != 0) { + + previous_pc = 0; // Required! + +#ifdef _DEBUG + fprintf(stderr, "CLIENT crc: %016llx len: %u\n", hash64(__afl_fuzz_ptr, *__a + fprintf(stderr, "RECV:"); + for (int i = 0; i < *__afl_fuzz_len; i++) + fprintf(stderr, "%02x", __afl_fuzz_ptr[i]); + fprintf(stderr,"\n"); +#endif + + // STEP 3: ensure the minimum length is present and setup the target + // function to fuzz. + + if (*__afl_fuzz_len > 0) { + + __afl_fuzz_ptr[*__afl_fuzz_len] = 0; // if you need to null terminate + (*o_function)(__afl_fuzz_ptr, *__afl_fuzz_len); + + } + + // END STEP 3 + + } + + gum_stalker_unfollow_me(stalker); + + while (gum_stalker_garbage_collect(stalker)) + g_usleep(10000); + + g_object_unref(stalker); + g_object_unref(transformer); + g_object_unref(event_sink); + gum_deinit_embedded(); + + return 0; + +} + diff --git a/examples/afl_frida/afl-frida.h b/examples/afl_frida/afl-frida.h new file mode 100644 index 00000000..efa3440f --- /dev/null +++ b/examples/afl_frida/afl-frida.h @@ -0,0 +1,53 @@ +extern int is_persistent; + +G_BEGIN_DECLS + +#define GUM_TYPE_FAKE_EVENT_SINK (gum_fake_event_sink_get_type()) + +G_DECLARE_FINAL_TYPE(GumFakeEventSink, gum_fake_event_sink, GUM, + FAKE_EVENT_SINK, GObject) + +struct _GumFakeEventSink { + + GObject parent; + GumEventType mask; + +}; + +GumEventSink *gum_fake_event_sink_new(void); +void gum_fake_event_sink_reset(GumFakeEventSink *self); + +G_END_DECLS + +typedef struct { + + GumAddress base_address; + guint64 code_start, code_end; + +} range_t; + +void instr_basic_block(GumStalkerIterator *iterator, GumStalkerOutput *output, + gpointer user_data); +#pragma once + +void afl_setup(void); +void afl_start_forkserver(void); +int __afl_persistent_loop(unsigned int max_cnt); + +inline static inline void afl_maybe_log(guint64 current_pc) { + + extern unsigned int afl_instr_rms; + extern uint8_t * afl_area_ptr; + + static __thread guint64 previous_pc; + + current_pc = (current_pc >> 4) ^ (current_pc << 8); + current_pc &= MAP_SIZE - 1; + + if (current_pc >= afl_instr_rms) return; + + afl_area_ptr[current_pc ^ previous_pc]++; + previous_pc = current_pc >> 1; + +} + diff --git a/examples/afl_frida/libtestinstr.c b/examples/afl_frida/libtestinstr.c new file mode 100644 index 00000000..96b1cf21 --- /dev/null +++ b/examples/afl_frida/libtestinstr.c @@ -0,0 +1,35 @@ +/* + american fuzzy lop++ - a trivial program to test the build + -------------------------------------------------------- + Originally written by Michal Zalewski + Copyright 2014 Google Inc. All rights reserved. + Copyright 2019-2020 AFLplusplus Project. All rights reserved. + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at: + http://www.apache.org/licenses/LICENSE-2.0 + */ + +#include <stdio.h> +#include <stdlib.h> +#include <unistd.h> +#include <string.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <fcntl.h> + +void testinstr(char *buf, int len) { + + if (len < 1) return; + buf[len] = 0; + + // we support three input cases + if (buf[0] == '0') + printf("Looks like a zero to me!\n"); + else if (buf[0] == '1') + printf("Pretty sure that is a one!\n"); + else + printf("Neither one or zero? How quaint!\n"); + +} + diff --git a/examples/afl_network_proxy/afl-network-client.c b/examples/afl_network_proxy/afl-network-client.c index 5af41055..a2451fdc 100644 --- a/examples/afl_network_proxy/afl-network-client.c +++ b/examples/afl_network_proxy/afl-network-client.c @@ -34,7 +34,9 @@ #include <netinet/ip6.h> #include <arpa/inet.h> #include <sys/mman.h> -#include <sys/shm.h> +#ifndef USEMMAP + #include <sys/shm.h> +#endif #include <sys/wait.h> #include <sys/types.h> #include <sys/socket.h> diff --git a/examples/afl_untracer/README.md b/examples/afl_untracer/README.md index e59792cb..ada0c916 100644 --- a/examples/afl_untracer/README.md +++ b/examples/afl_untracer/README.md @@ -32,13 +32,14 @@ To easily run the scripts without needing to run the GUI with Ghidra: /opt/ghidra/support/analyzeHeadless /tmp/ tmp$$ -import libtestinstr.so -postscript ./ghidra_get_patchpoints.java rm -rf /tmp/tmp$$ ``` +The file is created at `~/Desktop/patches.txt` ### Fuzzing Example (after modifying afl-untracer.c to your needs, compiling and creating patches.txt): ``` -AFL_UNTRACER_FILE=./patches.txt afl-fuzz -i in -o out -- ./afl-untracer +LD_LIBRARY_PATH=/path/to/target/library AFL_UNTRACER_FILE=./patches.txt afl-fuzz -i in -o out -- ./afl-untracer ``` (or even remote via afl-network-proxy). diff --git a/examples/afl_untracer/TODO b/examples/afl_untracer/TODO new file mode 100644 index 00000000..fffffacf --- /dev/null +++ b/examples/afl_untracer/TODO @@ -0,0 +1,2 @@ + * add shmem fuzzing + * add snapshot feature? diff --git a/examples/afl_untracer/afl-untracer.c b/examples/afl_untracer/afl-untracer.c index 664e691c..77b15eb8 100644 --- a/examples/afl_untracer/afl-untracer.c +++ b/examples/afl_untracer/afl-untracer.c @@ -56,6 +56,7 @@ #include <sys/shm.h> #include <sys/wait.h> #include <sys/types.h> +#include <sys/personality.h> #if defined(__linux__) #include <sys/ucontext.h> @@ -73,6 +74,9 @@ // STEP 1: +/* here you need to specify the parameter for the target function */ +static void *(*o_function)(u8 *buf, int len); + /* use stdin (1) or a file on the commandline (0) */ static u32 use_stdin = 1; @@ -111,10 +115,10 @@ static library_list_t liblist[MAX_LIB_COUNT]; static u32 liblist_cnt; static void sigtrap_handler(int signum, siginfo_t *si, void *context); -static void fuzz(); +static void fuzz(void); /* read the library information */ -void read_library_information() { +void read_library_information(void) { #if defined(__linux__) FILE *f; @@ -280,7 +284,7 @@ library_list_t *find_library(char *name) { // this seems to work for clang too. nice :) requires gcc 4.4+ #pragma GCC push_options #pragma GCC optimize("O0") -void breakpoint() { +void breakpoint(void) { if (debug) fprintf(stderr, "Breakpoint function \"breakpoint\" reached.\n"); @@ -395,7 +399,7 @@ static void __afl_map_shm(void) { } /* Fork server logic. */ -static void __afl_start_forkserver(void) { +inline static void __afl_start_forkserver(void) { u8 tmp[4] = {0, 0, 0, 0}; u32 status = 0; @@ -411,7 +415,7 @@ static void __afl_start_forkserver(void) { } -static u32 __afl_next_testcase(u8 *buf, u32 max_len) { +inline static u32 __afl_next_testcase(u8 *buf, u32 max_len) { s32 status; @@ -437,7 +441,7 @@ static u32 __afl_next_testcase(u8 *buf, u32 max_len) { } -static void __afl_end_testcase(int status) { +inline static void __afl_end_testcase(int status) { if (write(FORKSRV_FD + 1, &status, 4) != 4) do_exit = 1; // fprintf(stderr, "write2 %d\n", do_exit); @@ -457,7 +461,7 @@ static void __afl_end_testcase(int status) { ((uintptr_t)addr & 0x3) * 0x10000000000)) #endif -void setup_trap_instrumentation() { +void setup_trap_instrumentation(void) { library_list_t *lib_base = NULL; size_t lib_size = 0; @@ -667,12 +671,11 @@ static void sigtrap_handler(int signum, siginfo_t *si, void *context) { } -/* here you need to specify the parameter for the target function */ -static void *(*o_function)(u8 *buf, int len); - /* the MAIN function */ int main(int argc, char *argv[]) { + (void)personality(ADDR_NO_RANDOMIZE); // disable ASLR + pid = getpid(); if (getenv("AFL_DEBUG")) debug = 1; @@ -706,6 +709,9 @@ int main(int argc, char *argv[]) { while (1) { + // instead of fork() we could also use the snapshot lkm or do our own mini + // snapshot feature like in https://github.com/marcinguy/fuzzer + // -> snapshot.c if ((pid = fork()) == -1) PFATAL("fork failed"); if (pid) { @@ -738,7 +744,11 @@ int main(int argc, char *argv[]) { } -static void fuzz() { +#ifndef _DEBUG +inline +#endif + static void + fuzz(void) { // STEP 3: call the function to fuzz, also the functions you might // need to call to prepare the function and - important! - diff --git a/examples/aflpp_driver/GNUmakefile b/examples/aflpp_driver/GNUmakefile index a993c8a9..b118a8b5 100644 --- a/examples/aflpp_driver/GNUmakefile +++ b/examples/aflpp_driver/GNUmakefile @@ -7,25 +7,25 @@ ifneq "" "$(LLVM_BINDIR)" LLVM_BINDIR := $(LLVM_BINDIR)/ endif -FLAGS=-O3 -funroll-loops -g +CFLAGS := -O3 -funroll-loops -g all: libAFLDriver.a libAFLQemuDriver.a aflpp_qemu_driver_hook.so -aflpp_driver.o: aflpp_driver.cpp - $(LLVM_BINDIR)clang++ $(FLAGS) -stdlib=libc++ -std=c++11 -c aflpp_driver.cpp +aflpp_driver.o: aflpp_driver.c + $(LLVM_BINDIR)clang -I. -I../../include $(CFLAGS) -c aflpp_driver.c libAFLDriver.a: aflpp_driver.o ar ru libAFLDriver.a aflpp_driver.o debug: - $(LLVM_BINDIR)clang++ -Wno-deprecated -I../../include $(FLAGS) -D_DEBUG=\"1\" -c -o afl-performance.o ../../src/afl-performance.c - $(LLVM_BINDIR)clang++ -I../../include -D_DEBUG=\"1\" -g -stdlib=libc++ -funroll-loops -std=c++11 -c aflpp_driver.cpp - #$(LLVM_BINDIR)clang++ -S -emit-llvm -Wno-deprecated -I../../include $(FLAGS) -D_DEBUG=\"1\" -c -o afl-performance.ll ../../src/afl-performance.c - #$(LLVM_BINDIR)clang++ -S -emit-llvm -I../../include -D_DEBUG=\"1\" -g -stdlib=libc++ -funroll-loops -std=c++11 -c aflpp_driver.cpp + $(LLVM_BINDIR)clang -Wno-deprecated -I../../include $(CFLAGS) -D_DEBUG=\"1\" -c -o afl-performance.o ../../src/afl-performance.c + $(LLVM_BINDIR)clang -I../../include -D_DEBUG=\"1\" -g -funroll-loops -c aflpp_driver.c + #$(LLVM_BINDIR)clang -S -emit-llvm -Wno-deprecated -I../../include $(CFLAGS) -D_DEBUG=\"1\" -c -o afl-performance.ll ../../src/afl-performance.c + #$(LLVM_BINDIR)clang -S -emit-llvm -I../../include -D_DEBUG=\"1\" -g -funroll-loops -c aflpp_driver.c ar ru libAFLDriver.a afl-performance.o aflpp_driver.o aflpp_qemu_driver.o: aflpp_qemu_driver.c - $(LLVM_BINDIR)clang $(FLAGS) -O0 -funroll-loops -c aflpp_qemu_driver.c + $(LLVM_BINDIR)clang $(CFLAGS) -O0 -funroll-loops -c aflpp_qemu_driver.c libAFLQemuDriver.a: aflpp_qemu_driver.o ar ru libAFLQemuDriver.a aflpp_qemu_driver.o @@ -34,11 +34,11 @@ aflpp_qemu_driver_hook.so: aflpp_qemu_driver_hook.o $(LLVM_BINDIR)clang -shared aflpp_qemu_driver_hook.o -o aflpp_qemu_driver_hook.so aflpp_qemu_driver_hook.o: aflpp_qemu_driver_hook.c - $(LLVM_BINDIR)clang -fPIC $(FLAGS) -funroll-loops -c aflpp_qemu_driver_hook.c + $(LLVM_BINDIR)clang -fPIC $(CFLAGS) -funroll-loops -c aflpp_qemu_driver_hook.c test: debug - #clang++ -S -emit-llvm -D_DEBUG=\"1\" -I../../include -Wl,--allow-multiple-definition -stdlib=libc++ -funroll-loops -std=c++11 -o aflpp_driver_test.ll aflpp_driver_test.cpp - afl-clang-fast++ -D_DEBUG=\"1\" -I../../include -Wl,--allow-multiple-definition -stdlib=libc++ -funroll-loops -std=c++11 -o aflpp_driver_test aflpp_driver_test.cpp libAFLDriver.a + #clang -S -emit-llvm -D_DEBUG=\"1\" -I../../include -Wl,--allow-multiple-definition -funroll-loops -o aflpp_driver_test.ll aflpp_driver_test.c + afl-clang-fast -D_DEBUG=\"1\" -I../../include -Wl,--allow-multiple-definition -funroll-loops -o aflpp_driver_test aflpp_driver_test.c libAFLDriver.a afl-performance.o clean: rm -f *.o libAFLDriver*.a libAFLQemuDriver.a aflpp_qemu_driver_hook.so *~ core aflpp_driver_test diff --git a/examples/aflpp_driver/aflpp_driver.cpp b/examples/aflpp_driver/aflpp_driver.c index d6163bdf..86c7a69f 100644 --- a/examples/aflpp_driver/aflpp_driver.cpp +++ b/examples/aflpp_driver/aflpp_driver.c @@ -14,12 +14,15 @@ cat << EOF > test_fuzzer.cc #include <stddef.h> #include <stdint.h> extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { + if (size > 0 && data[0] == 'H') if (size > 1 && data[1] == 'I') if (size > 2 && data[2] == '!') __builtin_trap(); return 0; + } + EOF # Build your target with -fsanitize-coverage=trace-pc-guard using fresh clang. clang -g -fsanitize-coverage=trace-pc-guard test_fuzzer.cc -c @@ -49,188 +52,195 @@ If 1, close stdout at startup. If 2 close stderr; if 3 close both. #include <stdlib.h> #include <string.h> #include <unistd.h> +#include <limits.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <fcntl.h> -#include <fstream> -#include <iostream> -#include <vector> +#include "config.h" #ifdef _DEBUG -#include "hash.h" + #include "hash.h" #endif // Platform detection. Copied from FuzzerInternal.h #ifdef __linux__ -#define LIBFUZZER_LINUX 1 -#define LIBFUZZER_APPLE 0 -#define LIBFUZZER_NETBSD 0 -#define LIBFUZZER_FREEBSD 0 -#define LIBFUZZER_OPENBSD 0 + #define LIBFUZZER_LINUX 1 + #define LIBFUZZER_APPLE 0 + #define LIBFUZZER_NETBSD 0 + #define LIBFUZZER_FREEBSD 0 + #define LIBFUZZER_OPENBSD 0 #elif __APPLE__ -#define LIBFUZZER_LINUX 0 -#define LIBFUZZER_APPLE 1 -#define LIBFUZZER_NETBSD 0 -#define LIBFUZZER_FREEBSD 0 -#define LIBFUZZER_OPENBSD 0 + #define LIBFUZZER_LINUX 0 + #define LIBFUZZER_APPLE 1 + #define LIBFUZZER_NETBSD 0 + #define LIBFUZZER_FREEBSD 0 + #define LIBFUZZER_OPENBSD 0 #elif __NetBSD__ -#define LIBFUZZER_LINUX 0 -#define LIBFUZZER_APPLE 0 -#define LIBFUZZER_NETBSD 1 -#define LIBFUZZER_FREEBSD 0 -#define LIBFUZZER_OPENBSD 0 + #define LIBFUZZER_LINUX 0 + #define LIBFUZZER_APPLE 0 + #define LIBFUZZER_NETBSD 1 + #define LIBFUZZER_FREEBSD 0 + #define LIBFUZZER_OPENBSD 0 #elif __FreeBSD__ -#define LIBFUZZER_LINUX 0 -#define LIBFUZZER_APPLE 0 -#define LIBFUZZER_NETBSD 0 -#define LIBFUZZER_FREEBSD 1 -#define LIBFUZZER_OPENBSD 0 + #define LIBFUZZER_LINUX 0 + #define LIBFUZZER_APPLE 0 + #define LIBFUZZER_NETBSD 0 + #define LIBFUZZER_FREEBSD 1 + #define LIBFUZZER_OPENBSD 0 #elif __OpenBSD__ -#define LIBFUZZER_LINUX 0 -#define LIBFUZZER_APPLE 0 -#define LIBFUZZER_NETBSD 0 -#define LIBFUZZER_FREEBSD 0 -#define LIBFUZZER_OPENBSD 1 + #define LIBFUZZER_LINUX 0 + #define LIBFUZZER_APPLE 0 + #define LIBFUZZER_NETBSD 0 + #define LIBFUZZER_FREEBSD 0 + #define LIBFUZZER_OPENBSD 1 #else -#error "Support for your platform has not been implemented" + #error "Support for your platform has not been implemented" #endif -int __afl_sharedmem_fuzzing = 1; -extern unsigned int *__afl_fuzz_len; +int __afl_sharedmem_fuzzing = 1; +extern unsigned int * __afl_fuzz_len; extern unsigned char *__afl_fuzz_ptr; // libFuzzer interface is thin, so we don't include any libFuzzer headers. -extern "C" { int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size); __attribute__((weak)) int LLVMFuzzerInitialize(int *argc, char ***argv); -} // Notify AFL about persistent mode. static volatile char AFL_PERSISTENT[] = "##SIG_AFL_PERSISTENT##"; -extern "C" int __afl_persistent_loop(unsigned int); -static volatile char suppress_warning2 = AFL_PERSISTENT[0]; +int __afl_persistent_loop(unsigned int); // Notify AFL about deferred forkserver. static volatile char AFL_DEFER_FORKSVR[] = "##SIG_AFL_DEFER_FORKSRV##"; -extern "C" void __afl_manual_init(); -static volatile char suppress_warning1 = AFL_DEFER_FORKSVR[0]; - -// Input buffer. -static const size_t kMaxAflInputSize = 1 << 20; -static uint8_t AflInputBuf[kMaxAflInputSize]; +void __afl_manual_init(); // Use this optionally defined function to output sanitizer messages even if // user asks to close stderr. -__attribute__((weak)) extern "C" void __sanitizer_set_report_fd(void *); +__attribute__((weak)) void __sanitizer_set_report_fd(void *); // Keep track of where stderr content is being written to, so that // dup_and_close_stderr can use the correct one. -static FILE *output_file = stderr; +static FILE *output_file; // Experimental feature to use afl_driver without AFL's deferred mode. // Needs to run before __afl_auto_init. __attribute__((constructor(0))) static void __decide_deferred_forkserver(void) { + if (getenv("AFL_DRIVER_DONT_DEFER")) { + if (unsetenv("__AFL_DEFER_FORKSRV")) { + perror("Failed to unset __AFL_DEFER_FORKSRV"); abort(); + } + } + } // If the user asks us to duplicate stderr, then do it. static void maybe_duplicate_stderr() { + char *stderr_duplicate_filename = getenv("AFL_DRIVER_STDERR_DUPLICATE_FILENAME"); - if (!stderr_duplicate_filename) - return; + if (!stderr_duplicate_filename) return; FILE *stderr_duplicate_stream = freopen(stderr_duplicate_filename, "a+", stderr); if (!stderr_duplicate_stream) { + fprintf( stderr, "Failed to duplicate stderr to AFL_DRIVER_STDERR_DUPLICATE_FILENAME"); abort(); + } + output_file = stderr_duplicate_stream; + } // Most of these I/O functions were inspired by/copied from libFuzzer's code. static void discard_output(int fd) { + FILE *temp = fopen("/dev/null", "w"); - if (!temp) - abort(); + if (!temp) abort(); dup2(fileno(temp), fd); fclose(temp); + } -static void close_stdout() { discard_output(STDOUT_FILENO); } +static void close_stdout() { + + discard_output(STDOUT_FILENO); + +} // Prevent the targeted code from writing to "stderr" but allow sanitizers and // this driver to do so. static void dup_and_close_stderr() { + int output_fileno = fileno(output_file); int output_fd = dup(output_fileno); - if (output_fd <= 0) - abort(); + if (output_fd <= 0) abort(); FILE *new_output_file = fdopen(output_fd, "w"); - if (!new_output_file) - abort(); - if (!__sanitizer_set_report_fd) - return; - __sanitizer_set_report_fd(reinterpret_cast<void *>(output_fd)); + if (!new_output_file) abort(); + if (!__sanitizer_set_report_fd) return; + __sanitizer_set_report_fd((void *)output_fd); discard_output(output_fileno); -} -static void Printf(const char *Fmt, ...) { - va_list ap; - va_start(ap, Fmt); - vfprintf(output_file, Fmt, ap); - va_end(ap); - fflush(output_file); } // Close stdout and/or stderr if user asks for it. static void maybe_close_fd_mask() { + char *fd_mask_str = getenv("AFL_DRIVER_CLOSE_FD_MASK"); - if (!fd_mask_str) - return; + if (!fd_mask_str) return; int fd_mask = atoi(fd_mask_str); - if (fd_mask & 2) - dup_and_close_stderr(); - if (fd_mask & 1) - close_stdout(); + if (fd_mask & 2) dup_and_close_stderr(); + if (fd_mask & 1) close_stdout(); + } // Define LLVMFuzzerMutate to avoid link failures for targets that use it // with libFuzzer's LLVMFuzzerCustomMutator. -extern "C" size_t LLVMFuzzerMutate(uint8_t *Data, size_t Size, size_t MaxSize) { - assert(false && "LLVMFuzzerMutate should not be called from afl_driver"); +size_t LLVMFuzzerMutate(uint8_t *Data, size_t Size, size_t MaxSize) { + + // assert(false && "LLVMFuzzerMutate should not be called from afl_driver"); return 0; + } // Execute any files provided as parameters. static int ExecuteFilesOnyByOne(int argc, char **argv) { + + unsigned char *buf = malloc(MAX_FILE); for (int i = 1; i < argc; i++) { - std::ifstream in(argv[i], std::ios::binary); - in.seekg(0, in.end); - size_t length = in.tellg(); - in.seekg (0, in.beg); - std::cout << "Reading " << length << " bytes from " << argv[i] << std::endl; - // Allocate exactly length bytes so that we reliably catch buffer overflows. - std::vector<char> bytes(length); - in.read(bytes.data(), bytes.size()); - assert(in); - LLVMFuzzerTestOneInput(reinterpret_cast<const uint8_t *>(bytes.data()), - bytes.size()); - std::cout << "Execution successful" << std::endl; + + int fd = open(argv[i], O_RDONLY); + if (fd == -1) continue; + ssize_t length = read(fd, buf, MAX_FILE); + if (length > 0) { + + printf("Reading %zu bytes from %s\n", length, argv[i]); + LLVMFuzzerTestOneInput(buf, length); + printf("Execution successful.\n"); + + } + } + + free(buf); return 0; + } int main(int argc, char **argv) { - Printf( + + printf( "======================= INFO =========================\n" "This binary is built for AFL-fuzz.\n" "To run the target function on individual input(s) execute this:\n" @@ -242,32 +252,39 @@ int main(int argc, char **argv) { "afl-fuzz will run N iterations before " "re-spawning the process (default: 1000)\n" "======================================================\n", - argv[0], argv[0], argv[0]); + argv[0], argv[0], argv[0]); + output_file = stderr; maybe_duplicate_stderr(); maybe_close_fd_mask(); - if (LLVMFuzzerInitialize) - LLVMFuzzerInitialize(&argc, &argv); + if (LLVMFuzzerInitialize) LLVMFuzzerInitialize(&argc, &argv); + // Do any other expensive one-time initialization here. - uint8_t dummy_input[1] = {0}; - int N = 100000; + uint8_t dummy_input[64] = {0}; + memcpy(dummy_input, (void *)AFL_PERSISTENT, sizeof(AFL_PERSISTENT)); + memcpy(dummy_input + 32, (void *)AFL_DEFER_FORKSVR, + sizeof(AFL_DEFER_FORKSVR)); + int N = INT_MAX; if (argc == 2 && argv[1][0] == '-') - N = atoi(argv[1] + 1); - else if(argc == 2 && (N = atoi(argv[1])) > 0) - Printf("WARNING: using the deprecated call style `%s %d`\n", argv[0], N); + N = atoi(argv[1] + 1); + else if (argc == 2 && (N = atoi(argv[1])) > 0) + printf("WARNING: using the deprecated call style `%s %d`\n", argv[0], N); else if (argc > 1) { -// if (!getenv("AFL_DRIVER_DONT_DEFER")) { - __afl_sharedmem_fuzzing = 0; - __afl_manual_init(); -// } + + // if (!getenv("AFL_DRIVER_DONT_DEFER")) { + + __afl_sharedmem_fuzzing = 0; + __afl_manual_init(); + // } return ExecuteFilesOnyByOne(argc, argv); exit(0); + } assert(N > 0); -// if (!getenv("AFL_DRIVER_DONT_DEFER")) + // if (!getenv("AFL_DRIVER_DONT_DEFER")) __afl_manual_init(); // Call LLVMFuzzerTestOneInput here so that coverage caused by initialization @@ -276,17 +293,26 @@ int main(int argc, char **argv) { int num_runs = 0; while (__afl_persistent_loop(N)) { + #ifdef _DEBUG - fprintf(stderr, "CLIENT crc: %016llx len: %u\n", hash64(__afl_fuzz_ptr, *__afl_fuzz_len, 0xa5b35705), *__afl_fuzz_len); + fprintf(stderr, "CLIENT crc: %016llx len: %u\n", + hash64(__afl_fuzz_ptr, *__afl_fuzz_len, 0xa5b35705), + *__afl_fuzz_len); fprintf(stderr, "RECV:"); for (int i = 0; i < *__afl_fuzz_len; i++) fprintf(stderr, "%02x", __afl_fuzz_ptr[i]); - fprintf(stderr,"\n"); + fprintf(stderr, "\n"); #endif if (*__afl_fuzz_len) { + num_runs++; LLVMFuzzerTestOneInput(__afl_fuzz_ptr, *__afl_fuzz_len); + } + } - Printf("%s: successfully executed %d input(s)\n", argv[0], num_runs); + + printf("%s: successfully executed %d input(s)\n", argv[0], num_runs); + } + diff --git a/examples/aflpp_driver/aflpp_driver_test.c b/examples/aflpp_driver/aflpp_driver_test.c new file mode 100644 index 00000000..e4567bbf --- /dev/null +++ b/examples/aflpp_driver/aflpp_driver_test.c @@ -0,0 +1,25 @@ +#include <stdio.h> +#include <stdlib.h> +#include <stdint.h> + +#include "hash.h" + +int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { + + fprintf(stderr, "FUNC crc: %016llx len: %lu\n", + hash64((u8 *)Data, (unsigned int)Size, + (unsigned long long int)0xa5b35705), + Size); + + if (Size < 5) return 0; + + if (Data[0] == 'F') + if (Data[1] == 'A') + if (Data[2] == '$') + if (Data[3] == '$') + if (Data[4] == '$') abort(); + + return 0; + +} + diff --git a/examples/aflpp_driver/aflpp_driver_test.cpp b/examples/aflpp_driver/aflpp_driver_test.cpp deleted file mode 100644 index 13dc09b9..00000000 --- a/examples/aflpp_driver/aflpp_driver_test.cpp +++ /dev/null @@ -1,22 +0,0 @@ -#include <stdio.h> -#include <stdlib.h> -#include <stdint.h> -#include "hash.h" - -extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) { - - fprintf(stderr, "FUNC crc: %016llx len: %lu\n", hash64((u8*)Data, (unsigned int) Size, (unsigned long long int) 0xa5b35705), Size); - - if (Size < 5) - return 0; - - if (Data[0] == 'F') - if (Data[1] == 'A') - if (Data[2] == '$') - if (Data[3] == '$') - if (Data[4] == '$') - abort(); - - return 0; - -} diff --git a/examples/custom_mutators/post_library_gif.so.c b/examples/custom_mutators/post_library_gif.so.c index 9b76ead5..2d72400c 100644 --- a/examples/custom_mutators/post_library_gif.so.c +++ b/examples/custom_mutators/post_library_gif.so.c @@ -83,7 +83,7 @@ typedef struct post_state { } post_state_t; -void *afl afl_custom_init(void *afl) { +void *afl_custom_init(void *afl) { post_state_t *state = malloc(sizeof(post_state_t)); if (!state) { diff --git a/examples/defork/Makefile b/examples/defork/Makefile new file mode 100644 index 00000000..e8240dba --- /dev/null +++ b/examples/defork/Makefile @@ -0,0 +1,64 @@ +# +# american fuzzy lop++ - defork +# ---------------------------------- +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at: +# +# http://www.apache.org/licenses/LICENSE-2.0 +# + +.PHONY: all install clean + +PREFIX ?= /usr/local +BIN_PATH = $(PREFIX)/bin +HELPER_PATH = $(PREFIX)/lib/afl + +CFLAGS = -fPIC -Wall -Wextra +LDFLAGS = -shared + +UNAME_SAYS_LINUX=$(shell uname | grep -E '^Linux|^GNU' >/dev/null; echo $$?) +UNAME_SAYS_LINUX:sh=uname | grep -E '^Linux|^GNU' >/dev/null; echo $$? + +_LDFLAGS_ADD=$(UNAME_SAYS_LINUX:1=) +LDFLAGS_ADD=$(_LDFLAGS_ADD:0=-ldl) +LDFLAGS += $(LDFLAGS_ADD) + +# on gcc for arm there is no -m32, but -mbe32 +M32FLAG = -m32 +M64FLAG = -m64 + +CC_IS_GCC=$(shell $(CC) --version 2>/dev/null | grep -q gcc; echo $$?) +CC_IS_GCC:sh=$(CC) --version 2>/dev/null | grep -q gcc; echo $$? +CC_IS_ARMCOMPILER=$(shell $(CC) -v 2>&1 >/dev/null | grep -q arm; echo $$?) +CC_IS_ARMCOMPILER:sh=$(CC) -v 2>&1 >/dev/null | grep -q arm; echo $$? + +_M32FLAG=$(CC_IS_GCC)$(CC_IS_ARMCOMPILER) +__M32FLAG=$(_M32FLAG:00=-mbe32) +___M32FLAG=$(__M32FLAG:$(CC_IS_GCC)$(CC_IS_ARMCOMPILER)=-m32) +M32FLAG=$(___M32FLAG) +#ifeq "$(findstring clang, $(shell $(CC) --version 2>/dev/null))" "" +# ifneq (,$(findstring arm, "$(shell $(CC) -v 2>&1 >/dev/null)")) +# M32FLAG = -mbe32 +# endif +#endif + +all: defork32.so defork64.so + +defork32.so: defork.c + -@$(CC) $(M32FLAG) $(CFLAGS) $^ $(LDFLAGS) -o $@ 2>/dev/null || echo "defork32 build failure (that's fine)" + +defork64.so: defork.c + -@$(CC) $(M64FLAG) $(CFLAGS) $^ $(LDFLAGS) -o $@ 2>/dev/null || echo "defork64 build failure (that's fine)" + +install: defork32.so defork64.so + install -d -m 755 $(DESTDIR)$(HELPER_PATH)/ + if [ -f defork32.so ]; then set -e; install -m 755 defork32.so $(DESTDIR)$(HELPER_PATH)/; fi + if [ -f defork64.so ]; then set -e; install -m 755 defork64.so $(DESTDIR)$(HELPER_PATH)/; fi + +target: + ../../afl-clang forking_target.c -o forking_target -Wall -Wextra -Werror + +clean: + rm -f defork32.so defork64.so forking_target diff --git a/examples/defork/README.md b/examples/defork/README.md new file mode 100644 index 00000000..7e950323 --- /dev/null +++ b/examples/defork/README.md @@ -0,0 +1,11 @@ +# defork + +when the target forks, this breaks all normal fuzzing runs. +Sometimes, though, it is enough to just run the child process. +If this is the case, then this LD_PRELOAD library will always return 0 on fork, +the target will belive it is running as the child, post-fork. + +This is defork.c from the amazing preeny project +https://github.com/zardus/preeny + +It is altered for afl++ to work with its fork-server: the initial fork will go through, the second fork will be blocked. diff --git a/examples/defork/defork.c b/examples/defork/defork.c new file mode 100644 index 00000000..f71d1124 --- /dev/null +++ b/examples/defork/defork.c @@ -0,0 +1,50 @@ +#define __GNU_SOURCE +#include <dlfcn.h> +#include <unistd.h> +#include <stdio.h> +#include <stdbool.h> + +#include "../../include/config.h" + +/* we want to fork once (for the afl++ forkserver), + then immediately return as child on subsequent forks. */ +static bool forked = 0; + +pid_t (*original_fork)(void); + +/* In case we are not running in afl, we use a dummy original_fork */ +static pid_t nop(void) { + + return 0; + +} + +__attribute__((constructor)) void preeny_fork_orig() { + + if (getenv(SHM_ENV_VAR)) { + + printf("defork: running in AFL++. Allowing forkserver.\n"); + original_fork = dlsym(RTLD_NEXT, "socket"); + + } else { + + printf("defork: no AFL++ detected. Disabling fork from the start.\n"); + original_fork = &nop; + + } + +} + +pid_t fork(void) { + + /* If we forked before, or if we're in the child (pid==0), + we don't want to fork anymore, else, we are still in the forkserver. + The forkserver parent needs to fork infinite times, each child should never + fork again. This can be written without branches and I hate myself for it. + */ + pid_t ret = !forked && original_fork(); + forked = !ret; + return ret; + +} + diff --git a/examples/defork/forking_target.c b/examples/defork/forking_target.c new file mode 100644 index 00000000..98f6365a --- /dev/null +++ b/examples/defork/forking_target.c @@ -0,0 +1,48 @@ +#include <stdio.h> +#include <unistd.h> +#include <stdint.h> +#include <sys/types.h> + +/* This is an example target for defork.c - fuzz using +``` +mkdir in; echo a > ./in/a +AFL_PRELOAD=./defork64.so ../../afl-fuzz -i in -o out -- ./forking_target @@ +``` +*/ + +int main(int argc, char **argv) { + + if (argc < 2) { + + printf("Example tool to test defork.\nUsage ./forking_target <input>\n"); + return -1; + + } + + pid_t pid = fork(); + if (pid == 0) { + + printf("We're in the child.\n"); + FILE *f = fopen(argv[1], "r"); + char buf[4096]; + fread(buf, 1, 4096, f); + uint32_t offset = buf[100] + (buf[101] << 8); + char test_val = buf[offset]; + return test_val < 100; + + } else if (pid < 0) { + + perror("fork"); + return -1; + + } else { + + printf("We are in the parent - defork didn't work! :( (pid=%d)\n", + (int)pid); + + } + + return 0; + +} + diff --git a/examples/persistent_demo/persistent_demo_new.c b/examples/persistent_demo/persistent_demo_new.c index e4e328b0..7f878c0c 100644 --- a/examples/persistent_demo/persistent_demo_new.c +++ b/examples/persistent_demo/persistent_demo_new.c @@ -28,6 +28,20 @@ #include <signal.h> #include <string.h> +/* this lets the source compile without afl-clang-fast/lto */ +#ifndef __AFL_FUZZ_TESTCASE_LEN + +ssize_t fuzz_len; +unsigned char fuzz_buf[1024000]; + + #define __AFL_FUZZ_TESTCASE_LEN fuzz_len + #define __AFL_FUZZ_TESTCASE_BUF fuzz_buf + #define __AFL_FUZZ_INIT() void sync(void); + #define __AFL_LOOP(x) ((fuzz_len = read(0, fuzz_buf, sizeof(fuzz_buf))) > 0 ? + #define __AFL_INIT() sync() + +#endif + __AFL_FUZZ_INIT(); /* Main entry point. */ diff --git a/gcc_plugin/GNUmakefile b/gcc_plugin/GNUmakefile index bf5c53e0..4a4f0dcd 100644 --- a/gcc_plugin/GNUmakefile +++ b/gcc_plugin/GNUmakefile @@ -70,9 +70,16 @@ ifeq "$(TEST_MMAP)" "1" endif ifneq "$(shell uname -s)" "Haiku" - LDFLAGS += -lrt + LDFLAGS += -lrt +else + CFLAGS_SAFE += -DUSEMMAP=1 +endif + +ifeq "$(shell uname -s)" "SunOS" + PLUGIN_FLAGS += -I/usr/include/gmp endif + PROGS = ../afl-gcc-fast ../afl-gcc-pass.so ../afl-gcc-rt.o diff --git a/gcc_plugin/afl-gcc-fast.c b/gcc_plugin/afl-gcc-fast.c index af0beca7..b1bacfbd 100644 --- a/gcc_plugin/afl-gcc-fast.c +++ b/gcc_plugin/afl-gcc-fast.c @@ -379,7 +379,7 @@ int main(int argc, char **argv, char **envp) { u32 map_size = atoi(ptr); if (map_size != MAP_SIZE) - FATAL("AFL_MAP_SIZE is not supported by afl-gcc-fast"); + WARNF("AFL_MAP_SIZE is not supported by afl-gcc-fast"); } diff --git a/include/afl-fuzz.h b/include/afl-fuzz.h index c9f84c61..b82ddb4a 100644 --- a/include/afl-fuzz.h +++ b/include/afl-fuzz.h @@ -82,6 +82,11 @@ #include <sys/sysctl.h> #endif /* __APPLE__ || __FreeBSD__ || __OpenBSD__ */ +#if defined(__HAIKU__) + #include <kernel/OS.h> + #include <kernel/scheduler.h> +#endif + /* For systems that have sched_setaffinity; right now just Linux, but one can hope... */ @@ -139,7 +144,8 @@ struct queue_entry { var_behavior, /* Variable behavior? */ favored, /* Currently favored? */ fs_redundant, /* Marked as redundant in the fs? */ - fully_colorized; /* Do not run redqueen stage again */ + fully_colorized, /* Do not run redqueen stage again */ + is_ascii; /* Is the input just ascii text? */ u32 bitmap_size, /* Number of bits set in bitmap */ fuzz_level; /* Number of fuzzing iterations */ @@ -333,7 +339,7 @@ typedef struct afl_env_vars { afl_dumb_forksrv, afl_import_first, afl_custom_mutator_only, afl_no_ui, afl_force_ui, afl_i_dont_care_about_missing_crashes, afl_bench_just_one, afl_bench_until_crash, afl_debug_child_output, afl_autoresume, - afl_cal_fast; + afl_cal_fast, afl_cycle_schedules, afl_expand_havoc; u8 *afl_tmpdir, *afl_custom_mutator_library, *afl_python_module, *afl_path, *afl_hang_tmout, *afl_skip_crashes, *afl_preload; @@ -347,6 +353,13 @@ struct afl_pass_stat { }; +struct foreign_sync { + + u8 * dir; + time_t ctime; + +}; + typedef struct afl_state { /* Position of this state in the global states list */ @@ -454,7 +467,9 @@ typedef struct afl_state { fixed_seed, /* do not reseed */ fast_cal, /* Try to calibrate faster? */ disable_trim, /* Never trim in fuzz_one */ - shmem_testcase_mode; /* If sharedmem testcases are used */ + shmem_testcase_mode, /* If sharedmem testcases are used */ + expand_havoc, /* perform expensive havoc after no find */ + cycle_schedules; /* cycle power schedules ? */ u8 *virgin_bits, /* Regions yet untouched by fuzzing */ *virgin_tmout, /* Bits we haven't seen in tmouts */ @@ -535,7 +550,8 @@ typedef struct afl_state { u64 total_bitmap_size, /* Total bit count for all bitmaps */ total_bitmap_entries; /* Number of bitmaps counted */ - s32 cpu_core_count; /* CPU core count */ + s32 cpu_core_count, /* CPU core count */ + cpu_to_bind; /* bind to specific CPU */ #ifdef HAVE_AFFINITY s32 cpu_aff; /* Selected CPU core */ @@ -546,6 +562,10 @@ typedef struct afl_state { *queue_top, /* Top of the list */ *q_prev100; /* Previous 100 marker */ + // growing buf + struct queue_entry **queue_buf; + size_t queue_size; + struct queue_entry **top_rated; /* Top entries for bitmap bytes */ struct extra_data *extras; /* Extra tokens to fuzz with */ @@ -574,6 +594,15 @@ typedef struct afl_state { u8 describe_op_buf_256[256]; /* describe_op will use this to return a string up to 256 */ + unsigned long long int last_avg_exec_update; + u32 last_avg_execs; + float last_avg_execs_saved; + +/* foreign sync */ +#define FOREIGN_SYNCS_MAX 32 + u8 foreign_sync_cnt; + struct foreign_sync foreign_syncs[FOREIGN_SYNCS_MAX]; + #ifdef _AFL_DOCUMENT_MUTATIONS u8 do_document; u32 document_counter; @@ -937,6 +966,7 @@ void fix_up_banner(afl_state_t *, u8 *); void check_if_tty(afl_state_t *); void setup_signal_handlers(void); void save_cmdline(afl_state_t *, u32, char **); +void read_foreign_testcases(afl_state_t *, int); /* CmpLog */ diff --git a/include/config.h b/include/config.h index e8f52f45..344a368f 100644 --- a/include/config.h +++ b/include/config.h @@ -28,7 +28,7 @@ /* Version string: */ // c = release, d = volatile github dev, e = experimental branch -#define VERSION "++2.66c" +#define VERSION "++2.66d" /****************************************************** * * @@ -380,6 +380,10 @@ #define CMPLOG_SHM_ENV_VAR "__AFL_CMPLOG_SHM_ID" +/* CPU Affinity lockfile env var */ + +#define CPU_AFFINITY_ENV_VAR "__AFL_LOCKFILE" + /* Uncomment this to use inferior block-coverage-based instrumentation. Note that you need to recompile the target binary for this to have any effect: */ @@ -397,5 +401,28 @@ // #define IGNORE_FINDS +/* Text mutations */ + +/* Minimum length of a queue input to be evaluated for "is_ascii"? */ + +#define AFL_TXT_MIN_LEN 12 + +/* What is the minimum percentage of ascii characters present to be classifed + as "is_ascii"? */ + +#define AFL_TXT_MIN_PERCENT 94 + +/* How often to perform ASCII mutations 0 = disable, 1-8 are good values */ + +#define AFL_TXT_BIAS 6 + +/* Maximum length of a string to tamper with */ + +#define AFL_TXT_STRING_MAX_LEN 1024 + +/* Maximum mutations on a string */ + +#define AFL_TXT_STRING_MAX_MUTATIONS 6 + #endif /* ! _HAVE_CONFIG_H */ diff --git a/include/envs.h b/include/envs.h index 86222418..7153ed47 100644 --- a/include/envs.h +++ b/include/envs.h @@ -34,6 +34,7 @@ static char *afl_environment_variables[] = { "AFL_CUSTOM_MUTATOR_LIBRARY", "AFL_CUSTOM_MUTATOR_ONLY", "AFL_CXX", + "AFL_CYCLE_SCHEDULES", "AFL_DEBUG", "AFL_DEBUG_CHILD_OUTPUT", "AFL_DEBUG_GDB", @@ -64,6 +65,7 @@ static char *afl_environment_variables[] = { "AFL_LLVM_CMPLOG", "AFL_LLVM_INSTRIM", "AFL_LLVM_CTX", + "AFL_LLVM_DOCUMENT_IDS", "AFL_LLVM_INSTRUMENT", "AFL_LLVM_INSTRIM_LOOPHEAD", "AFL_LLVM_LTO_AUTODICTIONARY", @@ -129,6 +131,7 @@ static char *afl_environment_variables[] = { "AFL_USE_CFISAN", "AFL_WINE_PATH", "AFL_NO_SNAPSHOT", + "AFL_EXPAND_HAVOC_NOW", NULL }; diff --git a/include/snapshot-inl.h b/include/snapshot-inl.h index b73a001e..a18187ef 100644 --- a/include/snapshot-inl.h +++ b/include/snapshot-inl.h @@ -25,8 +25,7 @@ // From AFL-Snapshot-LKM/include/afl_snapshot.h (must be kept synced) #include <sys/ioctl.h> -#include <sys/types.h> -#include <sys/stat.h> +#include <stdlib.h> #include <fcntl.h> #define AFL_SNAPSHOT_FILE_NAME "/dev/afl_snapshot" @@ -35,6 +34,35 @@ #define AFL_SNAPSHOT_IOCTL_DO _IO(AFL_SNAPSHOT_IOCTL_MAGIC, 1) #define AFL_SNAPSHOT_IOCTL_CLEAN _IO(AFL_SNAPSHOT_IOCTL_MAGIC, 2) +#define AFL_SNAPSHOT_EXCLUDE_VMRANGE \ + _IOR(AFL_SNAPSHOT_IOCTL_MAGIC, 3, struct afl_snapshot_vmrange_args *) +#define AFL_SNAPSHOT_INCLUDE_VMRANGE \ + _IOR(AFL_SNAPSHOT_IOCTL_MAGIC, 4, struct afl_snapshot_vmrange_args *) +#define AFL_SNAPSHOT_IOCTL_TAKE _IOR(AFL_SNAPSHOT_IOCTL_MAGIC, 5, int) +#define AFL_SNAPSHOT_IOCTL_RESTORE _IO(AFL_SNAPSHOT_IOCTL_MAGIC, 6) + +// Trace new mmaped ares and unmap them on restore. +#define AFL_SNAPSHOT_MMAP 1 +// Do not snapshot any page (by default all writeable not-shared pages +// are shanpshotted. +#define AFL_SNAPSHOT_BLOCK 2 +// Snapshot file descriptor state, close newly opened descriptors +#define AFL_SNAPSHOT_FDS 4 +// Snapshot registers state +#define AFL_SNAPSHOT_REGS 8 +// Perform a restore when exit_group is invoked +#define AFL_SNAPSHOT_EXIT 16 +// TODO(andrea) allow not COW snapshots (high perf on small processes) +// Disable COW, restore all the snapshotted pages +#define AFL_SNAPSHOT_NOCOW 32 +// Do not snapshot Stack pages +#define AFL_SNAPSHOT_NOSTACK 64 + +struct afl_snapshot_vmrange_args { + + unsigned long start, end; + +}; static int afl_snapshot_dev_fd; @@ -45,15 +73,43 @@ static int afl_snapshot_init(void) { } -static int afl_snapshot_do() { +static void afl_snapshot_exclude_vmrange(void *start, void *end) { + + struct afl_snapshot_vmrange_args args = {(unsigned long)start, + (unsigned long)end}; + ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_EXCLUDE_VMRANGE, &args); + +} + +static void afl_snapshot_include_vmrange(void *start, void *end) { + + struct afl_snapshot_vmrange_args args = {(unsigned long)start, + (unsigned long)end}; + ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_INCLUDE_VMRANGE, &args); + +} + +static int afl_snapshot_take(int config) { + + return ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_IOCTL_TAKE, config); + +} + +static int afl_snapshot_do(void) { return ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_IOCTL_DO); } -static int afl_snapshot_clean(void) { +static void afl_snapshot_restore(void) { + + ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_IOCTL_RESTORE); + +} + +static void afl_snapshot_clean(void) { - return ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_IOCTL_CLEAN); + ioctl(afl_snapshot_dev_fd, AFL_SNAPSHOT_IOCTL_CLEAN); } diff --git a/libdislocator/libdislocator.so.c b/libdislocator/libdislocator.so.c index 7a70fd15..2324e390 100644 --- a/libdislocator/libdislocator.so.c +++ b/libdislocator/libdislocator.so.c @@ -166,7 +166,7 @@ static u32 alloc_canary; static void *__dislocator_alloc(size_t len) { - u8 * ret; + u8 * ret, *base; size_t tlen; int flags, fd, sp; @@ -189,6 +189,7 @@ static void *__dislocator_alloc(size_t len) { /* We will also store buffer length and a canary below the actual buffer, so let's add 8 bytes for that. */ + base = NULL; tlen = (1 + PG_COUNT(rlen + 8)) * PAGE_SIZE; flags = MAP_PRIVATE | MAP_ANONYMOUS; fd = -1; @@ -201,12 +202,20 @@ static void *__dislocator_alloc(size_t len) { if (sp) flags |= MAP_HUGETLB; #elif defined(__FreeBSD__) if (sp) flags |= MAP_ALIGNED_SUPER; + #elif defined(__sun) + if (sp) { + + base = (void *)(caddr_t)(1 << 21); + flags |= MAP_ALIGN; + + } + #endif #else (void)sp; #endif - ret = (u8 *)mmap(NULL, tlen, PROT_READ | PROT_WRITE, flags, fd, 0); + ret = (u8 *)mmap(base, tlen, PROT_READ | PROT_WRITE, flags, fd, 0); #if defined(USEHUGEPAGE) /* We try one more time with regular call */ if (ret == MAP_FAILED) { @@ -217,6 +226,8 @@ static void *__dislocator_alloc(size_t len) { flags &= -MAP_HUGETLB; #elif defined(__FreeBSD__) flags &= -MAP_ALIGNED_SUPER; + #elif defined(__sun) + flags &= -MAP_ALIGN; #endif ret = (u8 *)mmap(NULL, tlen, PROT_READ | PROT_WRITE, flags, fd, 0); diff --git a/libtokencap/Makefile b/libtokencap/Makefile index 63b87bb0..8bdfa5ac 100644 --- a/libtokencap/Makefile +++ b/libtokencap/Makefile @@ -35,13 +35,14 @@ _UNIQ=_QINU_ _____OS_DL = $(____OS_DL:$(_UNIQ)$(UNAME_S)=) ______OS_DL = $(_____OS_DL:$(_UNIQ)="-ldl") - _OS_TARGET = $(____OS_DL:$(_UNIQ)FreeBSD=$(_UNIQ)) - __OS_TARGET = $(_OS_TARGET:$(_UNIQ)OpenBSD=$(_UNIQ)) - ___OS_TARGET = $(__OS_TARGET:$(_UNIQ)NetBSD=$(_UNIQ)) - ____OS_TARGET = $(___OS_TARGET:$(_UNIQ)Haiku=$(_UNIQ)) -_____OS_TARGET = $(___OS_TARGET:$(_UNIQ)$(UNAME_S)=) + _OS_TARGET = $(____OS_DL:$(_UNIQ)FreeBSD=$(_UNIQ)) + __OS_TARGET = $(_OS_TARGET:$(_UNIQ)OpenBSD=$(_UNIQ)) + ___OS_TARGET = $(__OS_TARGET:$(_UNIQ)NetBSD=$(_UNIQ)) + ____OS_TARGET = $(___OS_TARGET:$(_UNIQ)Haiku=$(_UNIQ)) + _____OS_TARGET = $(____OS_TARGET:$(_UNIQ)SunOS=$(_UNIQ)) +______OS_TARGET = $(____OS_TARGET:$(_UNIQ)$(UNAME_S)=) -TARGETS = $(____OS_TARGET:$(_UNIQ)=libtokencap.so) +TARGETS = $(_____OS_TARGET:$(_UNIQ)=libtokencap.so) LDFLAGS += $(______OS_DL) diff --git a/libtokencap/libtokencap.so.c b/libtokencap/libtokencap.so.c index 600d2a5d..21bac082 100644 --- a/libtokencap/libtokencap.so.c +++ b/libtokencap/libtokencap.so.c @@ -35,7 +35,7 @@ #if !defined __linux__ && !defined __APPLE__ && !defined __FreeBSD__ && \ !defined __OpenBSD__ && !defined __NetBSD__ && !defined __DragonFly__ && \ - !defined(__HAIKU__) + !defined(__HAIKU__) && !defined(__sun) #error "Sorry, this library is unsupported in this platform for now!" #endif /* !__linux__ && !__APPLE__ && ! __FreeBSD__ && ! __OpenBSD__ && \ !__NetBSD__*/ @@ -52,6 +52,10 @@ #include <sys/mman.h> #elif defined __HAIKU__ #include <kernel/image.h> +#elif defined __sun + /* For map addresses the old struct is enough */ + #include <sys/procfs.h> + #include <limits.h> #endif #include <dlfcn.h> @@ -237,6 +241,8 @@ static void __tokencap_load_mappings(void) { image_info ii; int32_t group = 0; + __tokencap_ro_loaded = 1; + while (get_next_image_info(0, &group, &ii) == B_OK) { __tokencap_ro[__tokencap_ro_cnt].st = ii.text; @@ -246,6 +252,38 @@ static void __tokencap_load_mappings(void) { } +#elif defined __sun + prmap_t *c, *map; + char path[PATH_MAX]; + ssize_t r; + size_t hint; + int fd; + + snprintf(path, sizeof(path), "/proc/%ld/map", getpid()); + fd = open(path, O_RDONLY); + hint = (1 << 20); + map = malloc(hint); + + __tokencap_ro_loaded = 1; + + for (; (r = pread(fd, map, hint, 0)) == hint;) { + + hint <<= 1; + map = realloc(map, hint); + + } + + for (c = map; r > 0; c++, r -= sizeof(prmap_t)) { + + __tokencap_ro[__tokencap_ro_cnt].st = c->pr_vaddr; + __tokencap_ro[__tokencap_ro_cnt].en = c->pr_vaddr + c->pr_size; + + if (++__tokencap_ro_cnt == MAX_MAPPINGS) break; + + } + + free(map); + close(fd); #endif } diff --git a/llvm_mode/GNUmakefile b/llvm_mode/GNUmakefile index b5d026ef..0813af34 100644 --- a/llvm_mode/GNUmakefile +++ b/llvm_mode/GNUmakefile @@ -32,15 +32,16 @@ ifeq "$(shell uname)" "OpenBSD" LLVM_CONFIG ?= $(BIN_PATH)/llvm-config HAS_OPT = $(shell test -x $(BIN_PATH)/opt && echo 0 || echo 1) ifeq "$(HAS_OPT)" "1" - $(error llvm_mode needs a complete llvm installation (versions 3.4 up to 11) -> e.g. "pkg_add llvm-7.0.1p9") + $(error llvm_mode needs a complete llvm installation (versions 3.4 up to 12) -> e.g. "pkg_add llvm-7.0.1p9") endif else LLVM_CONFIG ?= llvm-config endif LLVMVER = $(shell $(LLVM_CONFIG) --version 2>/dev/null | sed 's/git//' ) -LLVM_UNSUPPORTED = $(shell $(LLVM_CONFIG) --version 2>/dev/null | egrep -q '^3\.[0-3]|^1[2-9]' && echo 1 || echo 0 ) +LLVM_UNSUPPORTED = $(shell $(LLVM_CONFIG) --version 2>/dev/null | egrep -q '^3\.[0-3]|^19' && echo 1 || echo 0 ) LLVM_NEW_API = $(shell $(LLVM_CONFIG) --version 2>/dev/null | egrep -q '^1[0-9]' && echo 1 || echo 0 ) +LLVM_HAVE_LTO = $(shell $(LLVM_CONFIG) --version 2>/dev/null | egrep -q '^1[1-9]' && echo 1 || echo 0 ) LLVM_MAJOR = $(shell $(LLVM_CONFIG) --version 2>/dev/null | sed 's/\..*//') LLVM_BINDIR = $(shell $(LLVM_CONFIG) --bindir 2>/dev/null) LLVM_LIBDIR = $(shell $(LLVM_CONFIG) --libdir 2>/dev/null) @@ -53,20 +54,29 @@ ifeq "$(LLVMVER)" "" endif ifeq "$(LLVM_UNSUPPORTED)" "1" - $(warning llvm_mode only supports llvm versions 3.4 up to 11) + $(warning llvm_mode only supports llvm versions 3.4 up to 12) endif +LLVM_TOO_OLD=1 + ifeq "$(LLVM_MAJOR)" "9" $(info [+] llvm_mode detected llvm 9, enabling neverZero implementation) + LLVM_TOO_OLD=0 endif ifeq "$(LLVM_NEW_API)" "1" $(info [+] llvm_mode detected llvm 10+, enabling neverZero implementation and c++14) LLVM_STDCXX = c++14 + LLVM_TOO_OLD=0 +endif + +ifeq "$(LLVM_TOO_OLD)" "1" + $(info [!] llvm_mode detected an old version of llvm, upgrade to at least 9 or preferable 11!) + $(shell sleep 1) endif -ifeq "$(LLVM_MAJOR)" "11" - $(info [+] llvm_mode detected llvm 11, enabling afl-clang-lto LTO implementation) +ifeq "$(LLVM_HAVE_LTO)" "1" + $(info [+] llvm_mode detected llvm 11+, enabling afl-clang-lto LTO implementation) LLVM_LTO = 1 #TEST_MMAP = 1 endif @@ -182,18 +192,24 @@ ifeq "$(LLVM_LTO)" "1" ifneq "$(shell readlink $(LLVM_BINDIR)/ld.lld 2>&1)" "" AFL_REAL_LD = $(LLVM_BINDIR)/ld.lld else - $(warn ld.lld not found, can not enable LTO mode) + $(warn ld.lld not found, cannot enable LTO mode) LLVM_LTO = 0 endif endif + else + $(warn -flto is not working (LLVMgold.so not found?), cannot enable LTO mode) + LLVM_LTO = 0 endif endif AFL_CLANG_FUSELD= -ifneq "$(AFL_CLANG_FLTO)" "" -ifeq "$(shell echo 'int main() {return 0; }' | $(CLANG_BIN) -x c - -fuse-ld=`command -v ld` -o .test 2>/dev/null && echo 1 || echo 0 ; rm -f .test )" "1" - AFL_CLANG_FUSELD=1 -endif +ifeq "$(LLVM_LTO)" "1" + ifeq "$(shell echo 'int main() {return 0; }' | $(CLANG_BIN) -x c - -fuse-ld=`command -v ld` -o .test 2>/dev/null && echo 1 || echo 0 ; rm -f .test )" "1" + AFL_CLANG_FUSELD=1 + else + $(warn -fuse-ld is not working, cannot enable LTO mode) + LLVM_LTO = 0 + endif endif CFLAGS ?= -O3 -funroll-loops -fPIC -D_FORTIFY_SOURCE=2 diff --git a/llvm_mode/LLVMInsTrim.so.cc b/llvm_mode/LLVMInsTrim.so.cc index 75548266..4d8c4719 100644 --- a/llvm_mode/LLVMInsTrim.so.cc +++ b/llvm_mode/LLVMInsTrim.so.cc @@ -56,7 +56,6 @@ struct InsTrim : public ModulePass { protected: uint32_t function_minimum_size = 1; - uint32_t debug = 0; char * skip_nozero = NULL; private: @@ -102,7 +101,6 @@ struct InsTrim : public ModulePass { bool runOnModule(Module &M) override { - char be_quiet = 0; setvbuf(stdout, NULL, _IONBF, 0); if ((isatty(2) && !getenv("AFL_QUIET")) || getenv("AFL_DEBUG") != NULL) { diff --git a/llvm_mode/README.instrument_file.md b/llvm_mode/README.instrument_file.md index 29c40eec..46e45ba2 100644 --- a/llvm_mode/README.instrument_file.md +++ b/llvm_mode/README.instrument_file.md @@ -18,6 +18,8 @@ For this purpose, I have added a "partial instrumentation" support to the LLVM mode of AFLFuzz that allows you to specify on a source file level which files should be compiled with or without instrumentation. +Note: When using PCGUARD mode - and have llvm 12+ - you can use this instead: +https://clang.llvm.org/docs/SanitizerCoverage.html#partially-disabling-instrumentation ## 2) Building the LLVM module diff --git a/llvm_mode/README.laf-intel.md b/llvm_mode/README.laf-intel.md index 2fa4bc26..f63ab2bb 100644 --- a/llvm_mode/README.laf-intel.md +++ b/llvm_mode/README.laf-intel.md @@ -35,8 +35,8 @@ bit_width may be 64, 32 or 16. A new experimental feature is splitting floating point comparisons into a series of sign, exponent and mantissa comparisons followed by splitting each of them into 8 bit comparisons when necessary. -It is activated with the `AFL_LLVM_LAF_SPLIT_FLOATS` setting, available only -when `AFL_LLVM_LAF_SPLIT_COMPARES` is set. +It is activated with the `AFL_LLVM_LAF_SPLIT_FLOATS` setting. +Note that setting this automatically activates `AFL_LLVM_LAF_SPLIT_COMPARES` You can also set `AFL_LLVM_LAF_ALL` and have all of the above enabled :-) diff --git a/llvm_mode/README.lto.md b/llvm_mode/README.lto.md index 4641fa89..e521ac82 100644 --- a/llvm_mode/README.lto.md +++ b/llvm_mode/README.lto.md @@ -2,7 +2,7 @@ ## TLDR; -This version requires a current llvm 11 compiled from the github master. +This version requires a current llvm 11+ compiled from the github master. 1. Use afl-clang-lto/afl-clang-lto++ because it is faster and gives better coverage than anything else that is out there in the AFL world @@ -10,7 +10,7 @@ This version requires a current llvm 11 compiled from the github master. 2. You can use it together with llvm_mode: laf-intel and the instrument file listing features and can be combined with cmplog/Redqueen -3. It only works with llvm 11 (current github master state) +3. It only works with llvm 11+ 4. AUTODICTIONARY feature! see below @@ -61,9 +61,9 @@ AUTODICTIONARY: 11 strings found [+] Instrumented 12071 locations with no collisions (on average 1046 collisions would be in afl-gcc/afl-clang-fast) (non-hardened mode). ``` -## Getting llvm 11 +## Getting llvm 11+ -### Installing llvm 11 from the llvm repository +### Installing llvm from the llvm repository (version 11) Installing the llvm snapshot builds is easy and mostly painless: @@ -83,7 +83,7 @@ apt-get install -y clang-11 clang-tools-11 libc++1-11 libc++-11-dev \ libomp5-11 lld-11 lldb-11 llvm-11 llvm-11-dev llvm-11-runtime llvm-11-tools ``` -### Building llvm 11 yourself +### Building llvm yourself (version 12) Building llvm from github takes quite some long time and is not painless: ``` @@ -125,10 +125,9 @@ NOTE: some targets also need to set the linker, try both `afl-clang-lto` and ## AUTODICTIONARY feature -Setting `AFL_LLVM_LTO_AUTODICTIONARY` will generate a dictionary in the -target binary based on string compare and memory compare functions. -afl-fuzz will automatically get these transmitted when starting to fuzz. -This improves coverage on a lot of targets. +While compiling, automatically a dictionary based on string comparisons is +generated put into the target binary. This dictionary is transfered to afl-fuzz +on start. This improves coverage statistically by 5-10% :) ## Fixed memory map @@ -141,12 +140,20 @@ to be dynamic - the original afl way, which is slower). AFL_LLVM_MAP_DYNAMIC can be set so the shared memory address is dynamic (which is safer but also slower). +## Document edge IDs + +Setting `export AFL_LLVM_DOCUMENT_IDS=file` will document to a file which edge +ID was given to which function. This helps to identify functions with variable +bytes or which functions were touched by an input. + ## Solving difficult targets Some targets are difficult because the configure script does unusual stuff that is unexpected for afl. See the next chapter `Potential issues` how to solve these. +### Example: ffmpeg + An example of a hard to solve target is ffmpeg. Here is how to successfully instrument it: @@ -156,7 +163,7 @@ instrument it: when compiling, so we have to trick configure: ``` -./configure --enable-lto --disable-shared +./configure --enable-lto --disable-shared --disable-inline-asm ``` 3. Now the configuration is done - and we edit the settings in `./ffbuild/config.mak` @@ -186,6 +193,31 @@ instrument it: 4. Then type make, wait for a long time and you are done :) +### Example: WebKit jsc + +Building jsc is difficult as the build script has bugs. + +1. checkout Webkit: +``` +svn checkout https://svn.webkit.org/repository/webkit/trunk WebKit +cd WebKit +``` + +2. Fix the build environment: +``` +mkdir -p WebKitBuild/Release +cd WebKitBuild/Release +ln -s ../../../../../usr/bin/llvm-ar-12 llvm-ar-12 +ln -s ../../../../../usr/bin/llvm-ranlib-12 llvm-ranlib-12 +cd ../.. +``` + +3. Build :) + +``` +Tools/Scripts/build-jsc --jsc-only --cli --cmakeargs="-DCMAKE_AR='llvm-ar-12' -DCMAKE_RANLIB='llvm-ranlib-12' -DCMAKE_VERBOSE_MAKEFILE:BOOL=ON -DCMAKE_CC_FLAGS='-O3 -lrt' -DCMAKE_CXX_FLAGS='-O3 -lrt' -DIMPORTED_LOCATION='/lib/x86_64-linux-gnu/' -DCMAKE_CC=afl-clang-lto -DCMAKE_CXX=afl-clang-lto++ -DENABLE_STATIC_JSC=ON" +``` + ## Potential issues ### compiling libraries fails @@ -220,17 +252,17 @@ AS=llvm-as ... afl-clang-lto is still work in progress. Known issues: - * Anything that llvm 11 cannot compile, afl-clang-lto can not compile either - obviously + * Anything that llvm 11+ cannot compile, afl-clang-lto can not compile either - obviously * Anything that does not compile with LTO, afl-clang-lto can not compile either - obviously -Hence if building a target with afl-clang-lto fails try to build it with llvm11 -and LTO enabled (`CC=clang-11` `CXX=clang++-11` `CFLAGS=-flto=full` and +Hence if building a target with afl-clang-lto fails try to build it with llvm12 +and LTO enabled (`CC=clang-12` `CXX=clang++-12` `CFLAGS=-flto=full` and `CXXFLAGS=-flto=full`). If this succeeeds then there is an issue with afl-clang-lto. Please report at [https://github.com/AFLplusplus/AFLplusplus/issues/226](https://github.com/AFLplusplus/AFLplusplus/issues/226) -Even some targets where clang-11 fails can be build if the fail is just in +Even some targets where clang-12 fails can be build if the fail is just in `./configure`, see `Solving difficult targets` above. ### Target crashes immediately @@ -270,7 +302,7 @@ Still more problems came up though as this only works without bugs from llvm 9 onwards, and with high optimization the link optimization ruins the instrumented control flow graph. -This is all now fixed with llvm 11. The llvm's own linker is now able to +This is all now fixed with llvm 11+. The llvm's own linker is now able to load passes and this bypasses all problems we had. Happy end :) diff --git a/llvm_mode/README.md b/llvm_mode/README.md index e2e22751..22088dfd 100644 --- a/llvm_mode/README.md +++ b/llvm_mode/README.md @@ -6,7 +6,7 @@ ## 1) Introduction -! llvm_mode works with llvm versions 3.4 up to 11 ! +! llvm_mode works with llvm versions 3.4 up to 12 ! The code in this directory allows you to instrument programs for AFL using true compiler-level instrumentation, instead of the more crude @@ -183,4 +183,4 @@ AFL_LLVM_INSTRUMENT=PCGUARD make ``` Note that this us currently the default, as it is the best mode. -If you have llvm 11 and compiled afl-clang-lto - this is the only better mode. +If you have llvm 11+ and compiled afl-clang-lto - this is the only better mode. diff --git a/llvm_mode/README.persistent_mode.md b/llvm_mode/README.persistent_mode.md index 83cc7f4d..4f0bcb2e 100644 --- a/llvm_mode/README.persistent_mode.md +++ b/llvm_mode/README.persistent_mode.md @@ -52,6 +52,21 @@ afl-clang-fast -o fuzz_target fuzz_target.c -lwhat_you_need_for_your_target And that is it! The speed increase is usually x10 to x20. +If you want to be able to compile the target without afl-clang-fast/lto then +add this just after the includes: + +``` +#ifndef __AFL_FUZZ_TESTCASE_LEN + ssize_t fuzz_len; + #define __AFL_FUZZ_TESTCASE_LEN fuzz_len + unsigned char fuzz_buf[1024000]; + #define __AFL_FUZZ_TESTCASE_BUF fuzz_buf + #define __AFL_FUZZ_INIT() void sync(void); + #define __AFL_LOOP(x) ((fuzz_len = read(0, fuzz_buf, sizeof(fuzz_buf))) > 0 ? + #define __AFL_INIT() sync() +#endif +``` + ## 3) deferred initialization AFL tries to optimize performance by executing the targeted binary just once, diff --git a/llvm_mode/afl-clang-fast.c b/llvm_mode/afl-clang-fast.c index f1b03682..738433ac 100644 --- a/llvm_mode/afl-clang-fast.c +++ b/llvm_mode/afl-clang-fast.c @@ -162,6 +162,7 @@ static void find_obj(u8 *argv0) { static void edit_params(u32 argc, char **argv, char **envp) { u8 fortify_set = 0, asan_set = 0, x_set = 0, bit_mode = 0; + u8 have_pic = 0; u8 *name; cc_params = ck_alloc((argc + 128) * sizeof(u8 *)); @@ -268,7 +269,8 @@ static void edit_params(u32 argc, char **argv, char **envp) { } - if (getenv("LAF_SPLIT_COMPARES") || getenv("AFL_LLVM_LAF_SPLIT_COMPARES")) { + if (getenv("LAF_SPLIT_COMPARES") || getenv("AFL_LLVM_LAF_SPLIT_COMPARES") || + getenv("AFL_LLVM_LAF_SPLIT_FLOATS")) { cc_params[cc_par_cnt++] = "-Xclang"; cc_params[cc_par_cnt++] = "-load"; @@ -311,12 +313,15 @@ static void edit_params(u32 argc, char **argv, char **envp) { cc_params[cc_par_cnt++] = alloc_printf("-fuse-ld=%s", AFL_REAL_LD); cc_params[cc_par_cnt++] = "-Wl,--allow-multiple-definition"; - if (instrument_mode == INSTRUMENT_CFG) - cc_params[cc_par_cnt++] = - alloc_printf("-Wl,-mllvm=-load=%s/afl-llvm-lto-instrim.so", obj_path); - else - cc_params[cc_par_cnt++] = alloc_printf( - "-Wl,-mllvm=-load=%s/afl-llvm-lto-instrumentation.so", obj_path); + /* + The current LTO instrim mode is not good, so we disable it + if (instrument_mode == INSTRUMENT_CFG) + cc_params[cc_par_cnt++] = + alloc_printf("-Wl,-mllvm=-load=%s/afl-llvm-lto-instrim.so", + obj_path); else + */ + cc_params[cc_par_cnt++] = alloc_printf( + "-Wl,-mllvm=-load=%s/afl-llvm-lto-instrumentation.so", obj_path); cc_params[cc_par_cnt++] = lto_flag; } else { @@ -359,6 +364,19 @@ static void edit_params(u32 argc, char **argv, char **envp) { } + u32 idx; + if (lto_mode && argc > 1) { + + for (idx = 1; idx < argc; idx++) { + + if (!strncasecmp(argv[idx], "-fpic", 5)) have_pic = 1; + + } + + if (!have_pic) cc_params[cc_par_cnt++] = "-fPIC"; + + } + /* Detect stray -v calls from ./configure scripts. */ while (--argc) { @@ -379,6 +397,8 @@ static void edit_params(u32 argc, char **argv, char **envp) { if (!strcmp(cur, "-Wl,-z,defs") || !strcmp(cur, "-Wl,--no-undefined")) continue; + if (lto_mode && !strncmp(cur, "-fuse-ld=", 9)) continue; + cc_params[cc_par_cnt++] = cur; } @@ -500,13 +520,15 @@ static void edit_params(u32 argc, char **argv, char **envp) { "int __afl_sharedmem_fuzzing = 1;" "extern unsigned int *__afl_fuzz_len;" "extern unsigned char *__afl_fuzz_ptr;" - "unsigned char *__afl_fuzz_alt_ptr;"; + "unsigned char __afl_fuzz_alt[1024000];" + "unsigned char *__afl_fuzz_alt_ptr = __afl_fuzz_alt;"; cc_params[cc_par_cnt++] = "-D__AFL_FUZZ_TESTCASE_BUF=(__afl_fuzz_ptr ? __afl_fuzz_ptr : " - "(__afl_fuzz_alt_ptr = (unsigned char *) malloc(1 * 1024 * 1024)))"; + "__afl_fuzz_alt_ptr)"; cc_params[cc_par_cnt++] = - "-D__AFL_FUZZ_TESTCASE_LEN=(__afl_fuzz_ptr ? *__afl_fuzz_len : read(0, " - "__afl_fuzz_alt_ptr, 1 * 1024 * 1024))"; + "-D__AFL_FUZZ_TESTCASE_LEN=(__afl_fuzz_ptr ? *__afl_fuzz_len : " + "(*__afl_fuzz_len = read(0, __afl_fuzz_alt_ptr, 1024000)) == 0xffffffff " + "? 0 : *__afl_fuzz_len)"; cc_params[cc_par_cnt++] = "-D__AFL_LOOP(_A)=" @@ -617,6 +639,10 @@ int main(int argc, char **argv, char **envp) { } + if ((getenv("AFL_LLVM_INSTRUMENT_FILE") || getenv("AFL_LLVM_WHITELIST")) && + getenv("AFL_DONT_OPTIMIZE")) + FATAL("AFL_LLVM_INSTRUMENT_FILE and AFL_DONT_OPTIMIZE cannot be combined"); + if (getenv("AFL_LLVM_INSTRIM") || getenv("INSTRIM") || getenv("INSTRIM_LIB")) { @@ -660,7 +686,7 @@ int main(int argc, char **argv, char **envp) { } if (strncasecmp(ptr, "pc-guard", strlen("pc-guard")) == 0 || - strncasecmp(ptr, "pcguard", strlen("pcgard")) == 0) { + strncasecmp(ptr, "pcguard", strlen("pcguard")) == 0) { if (!instrument_mode || instrument_mode == INSTRUMENT_PCGUARD) instrument_mode = INSTRUMENT_PCGUARD; @@ -763,9 +789,19 @@ int main(int argc, char **argv, char **envp) { #if LLVM_VERSION_MAJOR <= 6 instrument_mode = INSTRUMENT_AFL; #else - if (getenv("AFL_LLVM_INSTRUMENT_FILE") || getenv("AFL_LLVM_WHITELIST")) + if (getenv("AFL_LLVM_INSTRUMENT_FILE") || getenv("AFL_LLVM_WHITELIST")) { + instrument_mode = INSTRUMENT_AFL; - else + WARNF( + "switching to classic instrumentation because " + "AFL_LLVM_INSTRUMENT_FILE does not work with PCGUARD. Use " + "-fsanitize-coverage-allowlist=allowlist.txt if you want to use " + "PCGUARD. Requires llvm 12+. See " + "https://clang.llvm.org/docs/" + "SanitizerCoverage.html#partially-disabling-instrumentation"); + + } else + instrument_mode = INSTRUMENT_PCGUARD; #endif @@ -813,9 +849,13 @@ int main(int argc, char **argv, char **envp) { if (instrument_mode == INSTRUMENT_PCGUARD && (getenv("AFL_LLVM_INSTRUMENT_FILE") || getenv("AFL_LLVM_WHITELIST"))) - WARNF( + FATAL( "Instrumentation type PCGUARD does not support " - "AFL_LLVM_INSTRUMENT_FILE!"); + "AFL_LLVM_INSTRUMENT_FILE! Use " + "-fsanitize-coverage-allowlist=allowlist.txt instead (requires llvm " + "12+), see " + "https://clang.llvm.org/docs/" + "SanitizerCoverage.html#partially-disabling-instrumentation"); if (argc < 2 || strcmp(argv[1], "-h") == 0) { @@ -871,6 +911,8 @@ int main(int argc, char **argv, char **envp) { "AFL_NO_BUILTIN: compile for use with libtokencap.so\n" "AFL_PATH: path to instrumenting pass and runtime " "(afl-llvm-rt.*o)\n" + "AFL_LLVM_DOCUMENT_IDS: document edge IDs given to which function (LTO " + "only)\n" "AFL_QUIET: suppress verbose output\n" "AFL_USE_ASAN: activate address sanitizer\n" "AFL_USE_CFISAN: activate control flow sanitizer\n" @@ -939,7 +981,7 @@ int main(int argc, char **argv, char **envp) { u32 map_size = atoi(ptr2); if (map_size != MAP_SIZE) - FATAL("AFL_MAP_SIZE is not supported by afl-clang-fast"); + WARNF("AFL_MAP_SIZE is not supported by afl-clang-fast"); } diff --git a/llvm_mode/afl-llvm-common.cc b/llvm_mode/afl-llvm-common.cc index 47b49358..9a884ded 100644 --- a/llvm_mode/afl-llvm-common.cc +++ b/llvm_mode/afl-llvm-common.cc @@ -14,6 +14,8 @@ #include <fstream> #include <llvm/Support/raw_ostream.h> + +#define IS_EXTERN extern #include "afl-llvm-common.h" using namespace llvm; @@ -88,6 +90,7 @@ void initInstrumentList() { char *instrumentListFilename = getenv("AFL_LLVM_INSTRUMENT_FILE"); if (!instrumentListFilename) instrumentListFilename = getenv("AFL_LLVM_WHITELIST"); + if (instrumentListFilename) { std::string line; @@ -105,6 +108,10 @@ void initInstrumentList() { } + if (debug) + SAYF(cMGN "[D] " cRST "loaded instrument list with %zu entries\n", + myInstrumentList.size()); + } bool isInInstrumentList(llvm::Function *F) { @@ -145,8 +152,6 @@ bool isInInstrumentList(llvm::Function *F) { } - (void)instLine; - /* Continue only if we know where we actually are */ if (!instFilename.str().empty()) { @@ -164,6 +169,10 @@ bool isInInstrumentList(llvm::Function *F) { if (fnmatch(("*" + *it).c_str(), instFilename.str().c_str(), 0) == 0) { + if (debug) + SAYF(cMGN "[D] " cRST + "Function %s is in the list (%s), instrumenting ... \n", + F->getName().str().c_str(), instFilename.str().c_str()); return true; } @@ -219,12 +228,15 @@ bool isInInstrumentList(llvm::Function *F) { // we could not find out the location. in this case we say it is not // in the the instrument file list - + if (!be_quiet) + WARNF( + "No debug information found for function %s, will not be " + "instrumented (recompile with -g -O[1-3]).", + F->getName().str().c_str()); return false; } - // return false; } diff --git a/llvm_mode/afl-llvm-common.h b/llvm_mode/afl-llvm-common.h index 38e0c830..5b96be43 100644 --- a/llvm_mode/afl-llvm-common.h +++ b/llvm_mode/afl-llvm-common.h @@ -38,5 +38,14 @@ void initInstrumentList(); bool isInInstrumentList(llvm::Function *F); unsigned long long int calculateCollisions(uint32_t edges); +#ifndef IS_EXTERN + #define IS_EXTERN +#endif + +IS_EXTERN int debug; +IS_EXTERN int be_quiet; + +#undef IS_EXTERN + #endif diff --git a/llvm_mode/afl-llvm-lto-instrim.so.cc b/llvm_mode/afl-llvm-lto-instrim.so.cc index ca2b5886..98e9ff9a 100644 --- a/llvm_mode/afl-llvm-lto-instrim.so.cc +++ b/llvm_mode/afl-llvm-lto-instrim.so.cc @@ -73,8 +73,8 @@ struct InsTrimLTO : public ModulePass { protected: uint32_t function_minimum_size = 1; char * skip_nozero = NULL; - int afl_global_id = 1, debug = 0, autodictionary = 0; - uint32_t be_quiet = 0, inst_blocks = 0, inst_funcs = 0; + int afl_global_id = 1, autodictionary = 1; + uint32_t inst_blocks = 0, inst_funcs = 0; uint64_t map_addr = 0x10000; public: @@ -127,10 +127,6 @@ struct InsTrimLTO : public ModulePass { /* Process environment variables */ - if (getenv("AFL_LLVM_AUTODICTIONARY") || - getenv("AFL_LLVM_LTO_AUTODICTIONARY")) - autodictionary = 1; - if (getenv("AFL_LLVM_MAP_DYNAMIC")) map_addr = 0; if ((ptr = getenv("AFL_LLVM_MAP_ADDR"))) { @@ -705,7 +701,7 @@ struct InsTrimLTO : public ModulePass { Value *Incr = IRB.CreateAdd(Counter, One); - if (skip_nozero) { + if (skip_nozero == NULL) { auto cf = IRB.CreateICmpEQ(Incr, Zero); auto carry = IRB.CreateZExt(cf, Int8Ty); diff --git a/llvm_mode/afl-llvm-lto-instrumentation.so.cc b/llvm_mode/afl-llvm-lto-instrumentation.so.cc index af2db3ff..5686eb56 100644 --- a/llvm_mode/afl-llvm-lto-instrumentation.so.cc +++ b/llvm_mode/afl-llvm-lto-instrumentation.so.cc @@ -86,9 +86,9 @@ class AFLLTOPass : public ModulePass { bool runOnModule(Module &M) override; protected: - int afl_global_id = 1, debug = 0, autodictionary = 0; + int afl_global_id = 1, autodictionary = 1; uint32_t function_minimum_size = 1; - uint32_t be_quiet = 0, inst_blocks = 0, inst_funcs = 0, total_instr = 0; + uint32_t inst_blocks = 0, inst_funcs = 0, total_instr = 0; uint64_t map_addr = 0x10000; char * skip_nozero = NULL; @@ -103,6 +103,7 @@ bool AFLLTOPass::runOnModule(Module &M) { std::vector<CallInst *> calls; DenseMap<Value *, std::string *> valueMap; char * ptr; + FILE * documentFile = NULL; IntegerType *Int8Ty = IntegerType::getInt8Ty(C); IntegerType *Int32Ty = IntegerType::getInt32Ty(C); @@ -120,15 +121,16 @@ bool AFLLTOPass::runOnModule(Module &M) { be_quiet = 1; - if (getenv("AFL_LLVM_AUTODICTIONARY") || - getenv("AFL_LLVM_LTO_AUTODICTIONARY")) - autodictionary = 1; + if ((ptr = getenv("AFL_LLVM_DOCUMENT_IDS")) != NULL) { + + if ((documentFile = fopen(ptr, "a")) == NULL) + WARNF("Cannot access document file %s", ptr); + + } if (getenv("AFL_LLVM_MAP_DYNAMIC")) map_addr = 0; - if (getenv("AFL_LLVM_INSTRIM_SKIPSINGLEBLOCK") || - getenv("AFL_LLVM_SKIPSINGLEBLOCK")) - function_minimum_size = 2; + if (getenv("AFL_LLVM_SKIPSINGLEBLOCK")) function_minimum_size = 2; if ((ptr = getenv("AFL_LLVM_MAP_ADDR"))) { @@ -204,7 +206,8 @@ bool AFLLTOPass::runOnModule(Module &M) { if (debug) fprintf(stderr, - "DEBUG: Function %s is not the instrument file listed\n", + "DEBUG: Function %s is not in a source file that was specified " + "in the instrument file list\n", F.getName().str().c_str()); continue; @@ -536,6 +539,8 @@ bool AFLLTOPass::runOnModule(Module &M) { uint32_t succ = 0; + if (F.size() == 1) InsBlocks.push_back(&BB); + for (succ_iterator SI = succ_begin(&BB), SE = succ_end(&BB); SI != SE; ++SI) if ((*SI)->size() > 0) succ++; @@ -554,9 +559,12 @@ bool AFLLTOPass::runOnModule(Module &M) { do { --i; + BasicBlock * newBB; BasicBlock * origBB = &(*InsBlocks[i]); std::vector<BasicBlock *> Successors; Instruction * TI = origBB->getTerminator(); + uint32_t fs = origBB->getParent()->size(); + uint32_t countto; for (succ_iterator SI = succ_begin(origBB), SE = succ_end(origBB); SI != SE; ++SI) { @@ -566,15 +574,25 @@ bool AFLLTOPass::runOnModule(Module &M) { } - if (TI == NULL || TI->getNumSuccessors() < 2) continue; + if (fs == 1) { + + newBB = origBB; + countto = 1; + + } else { + + if (TI == NULL || TI->getNumSuccessors() < 2) continue; + countto = Successors.size(); + + } // if (Successors.size() != TI->getNumSuccessors()) // FATAL("Different successor numbers %lu <-> %u\n", Successors.size(), // TI->getNumSuccessors()); - for (uint32_t j = 0; j < Successors.size(); j++) { + for (uint32_t j = 0; j < countto; j++) { - BasicBlock *newBB = llvm::SplitEdge(origBB, Successors[j]); + if (fs != 1) newBB = llvm::SplitEdge(origBB, Successors[j]); if (!newBB) { @@ -583,6 +601,13 @@ bool AFLLTOPass::runOnModule(Module &M) { } + if (documentFile) { + + fprintf(documentFile, "%s %u\n", F.getName().str().c_str(), + afl_global_id); + + } + BasicBlock::iterator IP = newBB->getFirstInsertionPt(); IRBuilder<> IRB(&(*IP)); @@ -615,7 +640,7 @@ bool AFLLTOPass::runOnModule(Module &M) { Value *Incr = IRB.CreateAdd(Counter, One); - if (skip_nozero) { + if (skip_nozero == NULL) { auto cf = IRB.CreateICmpEQ(Incr, Zero); auto carry = IRB.CreateZExt(cf, Int8Ty); @@ -638,6 +663,9 @@ bool AFLLTOPass::runOnModule(Module &M) { } + if (documentFile) fclose(documentFile); + documentFile = NULL; + // save highest location ID to global variable // do this after each function to fail faster if (!be_quiet && afl_global_id > MAP_SIZE && diff --git a/llvm_mode/afl-llvm-lto-instrumentlist.so.cc b/llvm_mode/afl-llvm-lto-instrumentlist.so.cc index 6e6199e9..ab7c0c58 100644 --- a/llvm_mode/afl-llvm-lto-instrumentlist.so.cc +++ b/llvm_mode/afl-llvm-lto-instrumentlist.so.cc @@ -104,7 +104,6 @@ class AFLcheckIfInstrument : public ModulePass { protected: std::list<std::string> myInstrumentList; - int debug = 0; }; @@ -116,7 +115,6 @@ bool AFLcheckIfInstrument::runOnModule(Module &M) { /* Show a banner */ - char be_quiet = 0; setvbuf(stdout, NULL, _IONBF, 0); if ((isatty(2) && !getenv("AFL_QUIET")) || getenv("AFL_DEBUG") != NULL) { @@ -164,32 +162,41 @@ bool AFLcheckIfInstrument::runOnModule(Module &M) { } + if (instFilename.str().empty()) { + + if (!be_quiet) + WARNF( + "Function %s has no source file name information and will " + "not be instrumented.", + F.getName().str().c_str()); + continue; + + } + } - (void)instLine; + //(void)instLine; + fprintf(stderr, "xxx %s %s\n", F.getName().str().c_str(), + instFilename.str().c_str()); if (debug) SAYF(cMGN "[D] " cRST "function %s is in file %s\n", F.getName().str().c_str(), instFilename.str().c_str()); - /* Continue only if we know where we actually are */ - if (!instFilename.str().empty()) { - for (std::list<std::string>::iterator it = myInstrumentList.begin(); - it != myInstrumentList.end(); ++it) { + for (std::list<std::string>::iterator it = myInstrumentList.begin(); + it != myInstrumentList.end(); ++it) { - /* We don't check for filename equality here because - * filenames might actually be full paths. Instead we - * check that the actual filename ends in the filename - * specified in the list. */ - if (instFilename.str().length() >= it->length()) { + /* We don't check for filename equality here because + * filenames might actually be full paths. Instead we + * check that the actual filename ends in the filename + * specified in the list. */ + if (instFilename.str().length() >= it->length()) { - if (fnmatch(("*" + *it).c_str(), instFilename.str().c_str(), 0) == - 0) { + if (fnmatch(("*" + *it).c_str(), instFilename.str().c_str(), 0) == + 0) { - instrumentFunction = true; - break; - - } + instrumentFunction = true; + break; } @@ -197,6 +204,15 @@ bool AFLcheckIfInstrument::runOnModule(Module &M) { } + } else { + + if (!be_quiet) + WARNF( + "No debug information found for function %s, recompile with -g " + "-O[1-3]", + F.getName().str().c_str()); + continue; + } /* Either we couldn't figure out our location or the location is diff --git a/llvm_mode/afl-llvm-pass.so.cc b/llvm_mode/afl-llvm-pass.so.cc index 90cf3eb4..618abe48 100644 --- a/llvm_mode/afl-llvm-pass.so.cc +++ b/llvm_mode/afl-llvm-pass.so.cc @@ -82,7 +82,6 @@ class AFLCoverage : public ModulePass { protected: uint32_t ngram_size = 0; - uint32_t debug = 0; uint32_t map_size = MAP_SIZE; uint32_t function_minimum_size = 1; char * ctx_str = NULL, *skip_nozero = NULL; @@ -139,7 +138,6 @@ bool AFLCoverage::runOnModule(Module &M) { /* Show a banner */ - char be_quiet = 0; setvbuf(stdout, NULL, _IONBF, 0); if (getenv("AFL_DEBUG")) debug = 1; diff --git a/llvm_mode/afl-llvm-rt.o.c b/llvm_mode/afl-llvm-rt.o.c index f81d13ee..88abcbe0 100644 --- a/llvm_mode/afl-llvm-rt.o.c +++ b/llvm_mode/afl-llvm-rt.o.c @@ -53,7 +53,11 @@ #define CONST_PRIO 5 #ifndef MAP_FIXED_NOREPLACE - #define MAP_FIXED_NOREPLACE MAP_FIXED + #ifdef MAP_EXCL + #define MAP_FIXED_NOREPLACE MAP_EXCL | MAP_FIXED + #else + #define MAP_FIXED_NOREPLACE MAP_FIXED + #endif #endif #include <sys/mman.h> @@ -510,12 +514,19 @@ static void __afl_start_snapshots(void) { if (!child_pid) { + //(void)nice(-20); // does not seem to improve + signal(SIGCHLD, old_sigchld_handler); close(FORKSRV_FD); close(FORKSRV_FD + 1); - if (!afl_snapshot_do()) { raise(SIGSTOP); } + if (!afl_snapshot_take(AFL_SNAPSHOT_MMAP | AFL_SNAPSHOT_FDS | + AFL_SNAPSHOT_REGS | AFL_SNAPSHOT_EXIT)) { + + raise(SIGSTOP); + + } __afl_area_ptr[0] = 1; memset(__afl_prev_loc, 0, NGRAM_SIZE_MAX * sizeof(PREV_LOC_T)); @@ -713,6 +724,8 @@ static void __afl_start_forkserver(void) { if (!child_pid) { + //(void)nice(-20); + signal(SIGCHLD, old_sigchld_handler); close(FORKSRV_FD); @@ -846,6 +859,35 @@ __attribute__((constructor(CONST_PRIO))) void __afl_auto_init(void) { void __sanitizer_cov_trace_pc_guard(uint32_t *guard) { + // For stability analysis, if you want to know to which function unstable + // edge IDs belong to - uncomment, recompile+install llvm_mode, recompile + // the target. libunwind and libbacktrace are better solutions. + // Set AFL_DEBUG_CHILD_OUTPUT=1 and run afl-fuzz with 2>file to capture + // the backtrace output + /* + uint32_t unstable[] = { ... unstable edge IDs }; + uint32_t idx; + char bt[1024]; + for (idx = 0; i < sizeof(unstable)/sizeof(uint32_t); i++) { + + if (unstable[idx] == __afl_area_ptr[*guard]) { + + int bt_size = backtrace(bt, 256); + if (bt_size > 0) { + + char **bt_syms = backtrace_symbols(bt, bt_size); + if (bt_syms) + fprintf(stderr, "DEBUG: edge=%u caller=%s\n", unstable[idx], + bt_syms[0]); + + } + + } + + } + + */ + __afl_area_ptr[*guard]++; } diff --git a/llvm_mode/cmplog-instructions-pass.cc b/llvm_mode/cmplog-instructions-pass.cc index f929361a..7c48d906 100644 --- a/llvm_mode/cmplog-instructions-pass.cc +++ b/llvm_mode/cmplog-instructions-pass.cc @@ -76,9 +76,6 @@ class CmpLogInstructions : public ModulePass { } - protected: - int be_quiet = 0; - private: bool hookInstrs(Module &M); diff --git a/llvm_mode/cmplog-routines-pass.cc b/llvm_mode/cmplog-routines-pass.cc index 318193a4..a0f8f64f 100644 --- a/llvm_mode/cmplog-routines-pass.cc +++ b/llvm_mode/cmplog-routines-pass.cc @@ -76,9 +76,6 @@ class CmpLogRoutines : public ModulePass { } - protected: - int be_quiet = 0; - private: bool hookRtns(Module &M); diff --git a/llvm_mode/compare-transform-pass.so.cc b/llvm_mode/compare-transform-pass.so.cc index 2d1ab1cc..bed3597a 100644 --- a/llvm_mode/compare-transform-pass.so.cc +++ b/llvm_mode/compare-transform-pass.so.cc @@ -75,9 +75,6 @@ class CompareTransform : public ModulePass { } - protected: - int be_quiet = 0; - private: bool transformCmps(Module &M, const bool processStrcmp, const bool processMemcmp, const bool processStrncmp, @@ -358,7 +355,7 @@ bool CompareTransform::transformCmps(Module &M, const bool processStrcmp, Value * VarStr; bool HasStr1 = getConstantStringInfo(Str1P, Str1); bool HasStr2 = getConstantStringInfo(Str2P, Str2); - uint64_t constStrLen, constSizedLen, unrollLen; + uint64_t constStrLen, unrollLen, constSizedLen = 0; bool isMemcmp = !callInst->getCalledFunction()->getName().compare(StringRef("memcmp")); bool isSizedcmp = isMemcmp || @@ -474,8 +471,8 @@ bool CompareTransform::transformCmps(Module &M, const bool processStrcmp, if (cur_lenchk_bb) { IRBuilder<> cur_lenchk_IRB(&*(cur_lenchk_bb->getFirstInsertionPt())); - Value * icmp = cur_lenchk_IRB.CreateICmpEQ(sizedValue, - ConstantInt::get(Int64Ty, i)); + Value * icmp = cur_lenchk_IRB.CreateICmpEQ( + sizedValue, ConstantInt::get(sizedValue->getType(), i)); cur_lenchk_IRB.CreateCondBr(icmp, end_bb, cur_cmp_bb); cur_lenchk_bb->getTerminator()->eraseFromParent(); diff --git a/llvm_mode/split-compares-pass.so.cc b/llvm_mode/split-compares-pass.so.cc index 651fa5b4..3630bd8c 100644 --- a/llvm_mode/split-compares-pass.so.cc +++ b/llvm_mode/split-compares-pass.so.cc @@ -71,15 +71,13 @@ class SplitComparesTransform : public ModulePass { } - protected: - int be_quiet = 0; - private: int enableFPSplit; size_t splitIntCompares(Module &M, unsigned bitw); size_t splitFPCompares(Module &M); bool simplifyCompares(Module &M); + bool simplifyFPCompares(Module &M); bool simplifyIntSignedness(Module &M); size_t nextPowerOfTwo(size_t in); @@ -89,12 +87,11 @@ class SplitComparesTransform : public ModulePass { char SplitComparesTransform::ID = 0; -/* This function splits ICMP instructions with xGE or xLE predicates into two - * ICMP instructions with predicate xGT or xLT and EQ */ -bool SplitComparesTransform::simplifyCompares(Module &M) { +/* This function splits FCMP instructions with xGE or xLE predicates into two + * FCMP instructions with predicate xGT or xLT and EQ */ +bool SplitComparesTransform::simplifyFPCompares(Module &M) { LLVMContext & C = M.getContext(); - std::vector<Instruction *> icomps; std::vector<Instruction *> fcomps; IntegerType * Int1Ty = IntegerType::getInt1Ty(C); @@ -112,24 +109,6 @@ bool SplitComparesTransform::simplifyCompares(Module &M) { if ((selectcmpInst = dyn_cast<CmpInst>(&IN))) { - if (selectcmpInst->getPredicate() == CmpInst::ICMP_UGE || - selectcmpInst->getPredicate() == CmpInst::ICMP_SGE || - selectcmpInst->getPredicate() == CmpInst::ICMP_ULE || - selectcmpInst->getPredicate() == CmpInst::ICMP_SLE) { - - auto op0 = selectcmpInst->getOperand(0); - auto op1 = selectcmpInst->getOperand(1); - - IntegerType *intTyOp0 = dyn_cast<IntegerType>(op0->getType()); - IntegerType *intTyOp1 = dyn_cast<IntegerType>(op1->getType()); - - /* this is probably not needed but we do it anyway */ - if (!intTyOp0 || !intTyOp1) { continue; } - - icomps.push_back(selectcmpInst); - - } - if (enableFPSplit && (selectcmpInst->getPredicate() == CmpInst::FCMP_OGE || selectcmpInst->getPredicate() == CmpInst::FCMP_UGE || @@ -159,105 +138,159 @@ bool SplitComparesTransform::simplifyCompares(Module &M) { } - if (!icomps.size() && !fcomps.size()) { return false; } + if (!fcomps.size()) { return false; } - for (auto &IcmpInst : icomps) { + /* transform for floating point */ + for (auto &FcmpInst : fcomps) { - BasicBlock *bb = IcmpInst->getParent(); + BasicBlock *bb = FcmpInst->getParent(); - auto op0 = IcmpInst->getOperand(0); - auto op1 = IcmpInst->getOperand(1); + auto op0 = FcmpInst->getOperand(0); + auto op1 = FcmpInst->getOperand(1); /* find out what the new predicate is going to be */ - auto pred = dyn_cast<CmpInst>(IcmpInst)->getPredicate(); + auto pred = dyn_cast<CmpInst>(FcmpInst)->getPredicate(); CmpInst::Predicate new_pred; switch (pred) { - case CmpInst::ICMP_UGE: - new_pred = CmpInst::ICMP_UGT; + case CmpInst::FCMP_UGE: + new_pred = CmpInst::FCMP_UGT; break; - case CmpInst::ICMP_SGE: - new_pred = CmpInst::ICMP_SGT; + case CmpInst::FCMP_OGE: + new_pred = CmpInst::FCMP_OGT; break; - case CmpInst::ICMP_ULE: - new_pred = CmpInst::ICMP_ULT; + case CmpInst::FCMP_ULE: + new_pred = CmpInst::FCMP_ULT; break; - case CmpInst::ICMP_SLE: - new_pred = CmpInst::ICMP_SLT; + case CmpInst::FCMP_OLE: + new_pred = CmpInst::FCMP_OLT; break; default: // keep the compiler happy continue; } - /* split before the icmp instruction */ - BasicBlock *end_bb = bb->splitBasicBlock(BasicBlock::iterator(IcmpInst)); + /* split before the fcmp instruction */ + BasicBlock *end_bb = bb->splitBasicBlock(BasicBlock::iterator(FcmpInst)); /* the old bb now contains a unconditional jump to the new one (end_bb) * we need to delete it later */ - /* create the ICMP instruction with new_pred and add it to the old basic - * block bb it is now at the position where the old IcmpInst was */ - Instruction *icmp_np; - icmp_np = CmpInst::Create(Instruction::ICmp, new_pred, op0, op1); + /* create the FCMP instruction with new_pred and add it to the old basic + * block bb it is now at the position where the old FcmpInst was */ + Instruction *fcmp_np; + fcmp_np = CmpInst::Create(Instruction::FCmp, new_pred, op0, op1); bb->getInstList().insert(BasicBlock::iterator(bb->getTerminator()), - icmp_np); + fcmp_np); - /* create a new basic block which holds the new EQ icmp */ - Instruction *icmp_eq; + /* create a new basic block which holds the new EQ fcmp */ + Instruction *fcmp_eq; /* insert middle_bb before end_bb */ BasicBlock *middle_bb = BasicBlock::Create(C, "injected", end_bb->getParent(), end_bb); - icmp_eq = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, op0, op1); - middle_bb->getInstList().push_back(icmp_eq); + fcmp_eq = CmpInst::Create(Instruction::FCmp, CmpInst::FCMP_OEQ, op0, op1); + middle_bb->getInstList().push_back(fcmp_eq); /* add an unconditional branch to the end of middle_bb with destination * end_bb */ BranchInst::Create(end_bb, middle_bb); /* replace the uncond branch with a conditional one, which depends on the - * new_pred icmp. True goes to end, false to the middle (injected) bb */ + * new_pred fcmp. True goes to end, false to the middle (injected) bb */ auto term = bb->getTerminator(); - BranchInst::Create(end_bb, middle_bb, icmp_np, bb); + BranchInst::Create(end_bb, middle_bb, fcmp_np, bb); term->eraseFromParent(); - /* replace the old IcmpInst (which is the first inst in end_bb) with a PHI + /* replace the old FcmpInst (which is the first inst in end_bb) with a PHI * inst to wire up the loose ends */ PHINode *PN = PHINode::Create(Int1Ty, 2, ""); - /* the first result depends on the outcome of icmp_eq */ - PN->addIncoming(icmp_eq, middle_bb); - /* if the source was the original bb we know that the icmp_np yielded true + /* the first result depends on the outcome of fcmp_eq */ + PN->addIncoming(fcmp_eq, middle_bb); + /* if the source was the original bb we know that the fcmp_np yielded true * hence we can hardcode this value */ PN->addIncoming(ConstantInt::get(Int1Ty, 1), bb); - /* replace the old IcmpInst with our new and shiny PHI inst */ - BasicBlock::iterator ii(IcmpInst); - ReplaceInstWithInst(IcmpInst->getParent()->getInstList(), ii, PN); + /* replace the old FcmpInst with our new and shiny PHI inst */ + BasicBlock::iterator ii(FcmpInst); + ReplaceInstWithInst(FcmpInst->getParent()->getInstList(), ii, PN); } - /* now for floating point */ - for (auto &FcmpInst : fcomps) { + return true; - BasicBlock *bb = FcmpInst->getParent(); +} - auto op0 = FcmpInst->getOperand(0); - auto op1 = FcmpInst->getOperand(1); +/* This function splits ICMP instructions with xGE or xLE predicates into two + * ICMP instructions with predicate xGT or xLT and EQ */ +bool SplitComparesTransform::simplifyCompares(Module &M) { + + LLVMContext & C = M.getContext(); + std::vector<Instruction *> icomps; + IntegerType * Int1Ty = IntegerType::getInt1Ty(C); + + /* iterate over all functions, bbs and instruction and add + * all integer comparisons with >= and <= predicates to the icomps vector */ + for (auto &F : M) { + + if (!isInInstrumentList(&F)) continue; + + for (auto &BB : F) { + + for (auto &IN : BB) { + + CmpInst *selectcmpInst = nullptr; + + if ((selectcmpInst = dyn_cast<CmpInst>(&IN))) { + + if (selectcmpInst->getPredicate() == CmpInst::ICMP_UGE || + selectcmpInst->getPredicate() == CmpInst::ICMP_SGE || + selectcmpInst->getPredicate() == CmpInst::ICMP_ULE || + selectcmpInst->getPredicate() == CmpInst::ICMP_SLE) { + + auto op0 = selectcmpInst->getOperand(0); + auto op1 = selectcmpInst->getOperand(1); + + IntegerType *intTyOp0 = dyn_cast<IntegerType>(op0->getType()); + IntegerType *intTyOp1 = dyn_cast<IntegerType>(op1->getType()); + + /* this is probably not needed but we do it anyway */ + if (!intTyOp0 || !intTyOp1) { continue; } + + icomps.push_back(selectcmpInst); + + } + + } + + } + + } + + } + + if (!icomps.size()) { return false; } + + for (auto &IcmpInst : icomps) { + + BasicBlock *bb = IcmpInst->getParent(); + + auto op0 = IcmpInst->getOperand(0); + auto op1 = IcmpInst->getOperand(1); /* find out what the new predicate is going to be */ - auto pred = dyn_cast<CmpInst>(FcmpInst)->getPredicate(); + auto pred = dyn_cast<CmpInst>(IcmpInst)->getPredicate(); CmpInst::Predicate new_pred; switch (pred) { - case CmpInst::FCMP_UGE: - new_pred = CmpInst::FCMP_UGT; + case CmpInst::ICMP_UGE: + new_pred = CmpInst::ICMP_UGT; break; - case CmpInst::FCMP_OGE: - new_pred = CmpInst::FCMP_OGT; + case CmpInst::ICMP_SGE: + new_pred = CmpInst::ICMP_SGT; break; - case CmpInst::FCMP_ULE: - new_pred = CmpInst::FCMP_ULT; + case CmpInst::ICMP_ULE: + new_pred = CmpInst::ICMP_ULT; break; - case CmpInst::FCMP_OLE: - new_pred = CmpInst::FCMP_OLT; + case CmpInst::ICMP_SLE: + new_pred = CmpInst::ICMP_SLT; break; default: // keep the compiler happy continue; @@ -265,25 +298,25 @@ bool SplitComparesTransform::simplifyCompares(Module &M) { } /* split before the icmp instruction */ - BasicBlock *end_bb = bb->splitBasicBlock(BasicBlock::iterator(FcmpInst)); + BasicBlock *end_bb = bb->splitBasicBlock(BasicBlock::iterator(IcmpInst)); /* the old bb now contains a unconditional jump to the new one (end_bb) * we need to delete it later */ /* create the ICMP instruction with new_pred and add it to the old basic * block bb it is now at the position where the old IcmpInst was */ - Instruction *fcmp_np; - fcmp_np = CmpInst::Create(Instruction::FCmp, new_pred, op0, op1); + Instruction *icmp_np; + icmp_np = CmpInst::Create(Instruction::ICmp, new_pred, op0, op1); bb->getInstList().insert(BasicBlock::iterator(bb->getTerminator()), - fcmp_np); + icmp_np); - /* create a new basic block which holds the new EQ fcmp */ - Instruction *fcmp_eq; + /* create a new basic block which holds the new EQ icmp */ + Instruction *icmp_eq; /* insert middle_bb before end_bb */ BasicBlock *middle_bb = BasicBlock::Create(C, "injected", end_bb->getParent(), end_bb); - fcmp_eq = CmpInst::Create(Instruction::FCmp, CmpInst::FCMP_OEQ, op0, op1); - middle_bb->getInstList().push_back(fcmp_eq); + icmp_eq = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, op0, op1); + middle_bb->getInstList().push_back(icmp_eq); /* add an unconditional branch to the end of middle_bb with destination * end_bb */ BranchInst::Create(end_bb, middle_bb); @@ -291,20 +324,20 @@ bool SplitComparesTransform::simplifyCompares(Module &M) { /* replace the uncond branch with a conditional one, which depends on the * new_pred icmp. True goes to end, false to the middle (injected) bb */ auto term = bb->getTerminator(); - BranchInst::Create(end_bb, middle_bb, fcmp_np, bb); + BranchInst::Create(end_bb, middle_bb, icmp_np, bb); term->eraseFromParent(); /* replace the old IcmpInst (which is the first inst in end_bb) with a PHI * inst to wire up the loose ends */ PHINode *PN = PHINode::Create(Int1Ty, 2, ""); /* the first result depends on the outcome of icmp_eq */ - PN->addIncoming(fcmp_eq, middle_bb); + PN->addIncoming(icmp_eq, middle_bb); /* if the source was the original bb we know that the icmp_np yielded true * hence we can hardcode this value */ PN->addIncoming(ConstantInt::get(Int1Ty, 1), bb); /* replace the old IcmpInst with our new and shiny PHI inst */ - BasicBlock::iterator ii(FcmpInst); - ReplaceInstWithInst(FcmpInst->getParent()->getInstList(), ii, PN); + BasicBlock::iterator ii(IcmpInst); + ReplaceInstWithInst(IcmpInst->getParent()->getInstList(), ii, PN); } @@ -640,7 +673,8 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { BranchInst::Create(end_bb, signequal_bb); - /* create a new bb which is executed if exponents are equal */ + /* create a new bb which is executed if exponents are satisfying the compare + */ BasicBlock *middle_bb = BasicBlock::Create(C, "injected", end_bb->getParent(), end_bb); @@ -695,7 +729,9 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { } /* compare the exponents of the operands */ + Instruction *icmp_exponents_equal; Instruction *icmp_exponent_result; + BasicBlock * signequal2_bb = signequal_bb; switch (FcmpInst->getPredicate()) { case CmpInst::FCMP_OEQ: @@ -707,22 +743,60 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { icmp_exponent_result = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_NE, m_e0, m_e1); break; + /* compare the exponents of the operands (signs are equal) + * if exponents are equal -> proceed to mantissa comparison + * else get result depending on sign + */ case CmpInst::FCMP_OGT: case CmpInst::FCMP_UGT: Instruction *icmp_exponent; + icmp_exponents_equal = + CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, m_e0, m_e1); + signequal_bb->getInstList().insert( + BasicBlock::iterator(signequal_bb->getTerminator()), + icmp_exponents_equal); + + // shortcut for unequal exponents + signequal2_bb = signequal_bb->splitBasicBlock( + BasicBlock::iterator(signequal_bb->getTerminator())); + + /* if the exponents are equal goto middle_bb else to signequal2_bb */ + term = signequal_bb->getTerminator(); + BranchInst::Create(middle_bb, signequal2_bb, icmp_exponents_equal, + signequal_bb); + term->eraseFromParent(); + icmp_exponent = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_UGT, m_e0, m_e1); - signequal_bb->getInstList().insert( - BasicBlock::iterator(signequal_bb->getTerminator()), icmp_exponent); + signequal2_bb->getInstList().insert( + BasicBlock::iterator(signequal2_bb->getTerminator()), + icmp_exponent); icmp_exponent_result = BinaryOperator::Create(Instruction::Xor, icmp_exponent, t_s0); break; case CmpInst::FCMP_OLT: case CmpInst::FCMP_ULT: + icmp_exponents_equal = + CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, m_e0, m_e1); + signequal_bb->getInstList().insert( + BasicBlock::iterator(signequal_bb->getTerminator()), + icmp_exponents_equal); + + // shortcut for unequal exponents + signequal2_bb = signequal_bb->splitBasicBlock( + BasicBlock::iterator(signequal_bb->getTerminator())); + + /* if the exponents are equal goto middle_bb else to signequal2_bb */ + term = signequal_bb->getTerminator(); + BranchInst::Create(middle_bb, signequal2_bb, icmp_exponents_equal, + signequal_bb); + term->eraseFromParent(); + icmp_exponent = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_ULT, m_e0, m_e1); - signequal_bb->getInstList().insert( - BasicBlock::iterator(signequal_bb->getTerminator()), icmp_exponent); + signequal2_bb->getInstList().insert( + BasicBlock::iterator(signequal2_bb->getTerminator()), + icmp_exponent); icmp_exponent_result = BinaryOperator::Create(Instruction::Xor, icmp_exponent, t_s0); break; @@ -731,15 +805,40 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { } - signequal_bb->getInstList().insert( - BasicBlock::iterator(signequal_bb->getTerminator()), + signequal2_bb->getInstList().insert( + BasicBlock::iterator(signequal2_bb->getTerminator()), icmp_exponent_result); { - auto term = signequal_bb->getTerminator(); - /* if the exponents are different do a fraction cmp */ - BranchInst::Create(middle_bb, end_bb, icmp_exponent_result, signequal_bb); + term = signequal2_bb->getTerminator(); + + switch (FcmpInst->getPredicate()) { + + case CmpInst::FCMP_OEQ: + /* if the exponents are satifying the compare do a fraction cmp in + * middle_bb */ + BranchInst::Create(middle_bb, end_bb, icmp_exponent_result, + signequal2_bb); + break; + case CmpInst::FCMP_ONE: + case CmpInst::FCMP_UNE: + /* if the exponents are satifying the compare do a fraction cmp in + * middle_bb */ + BranchInst::Create(end_bb, middle_bb, icmp_exponent_result, + signequal2_bb); + break; + case CmpInst::FCMP_OGT: + case CmpInst::FCMP_UGT: + case CmpInst::FCMP_OLT: + case CmpInst::FCMP_ULT: + BranchInst::Create(end_bb, signequal2_bb); + break; + default: + continue; + + } + term->eraseFromParent(); } @@ -800,44 +899,82 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { /* compare the fractions of the operands */ Instruction *icmp_fraction_result; + Instruction *icmp_fraction_result2; + BasicBlock * middle2_bb = middle_bb; + PHINode * PN2 = nullptr; switch (FcmpInst->getPredicate()) { case CmpInst::FCMP_OEQ: icmp_fraction_result = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, t_f0, t_f1); + middle2_bb->getInstList().insert( + BasicBlock::iterator(middle2_bb->getTerminator()), + icmp_fraction_result); + break; case CmpInst::FCMP_UNE: case CmpInst::FCMP_ONE: icmp_fraction_result = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_NE, t_f0, t_f1); + middle2_bb->getInstList().insert( + BasicBlock::iterator(middle2_bb->getTerminator()), + icmp_fraction_result); + break; case CmpInst::FCMP_OGT: case CmpInst::FCMP_UGT: - Instruction *icmp_fraction; - icmp_fraction = - CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_UGT, t_f0, t_f1); - middle_bb->getInstList().insert( - BasicBlock::iterator(middle_bb->getTerminator()), icmp_fraction); - icmp_fraction_result = - BinaryOperator::Create(Instruction::Xor, icmp_fraction, t_s0); - break; case CmpInst::FCMP_OLT: - case CmpInst::FCMP_ULT: - icmp_fraction = - CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_ULT, t_f0, t_f1); - middle_bb->getInstList().insert( - BasicBlock::iterator(middle_bb->getTerminator()), icmp_fraction); - icmp_fraction_result = - BinaryOperator::Create(Instruction::Xor, icmp_fraction, t_s0); - break; + case CmpInst::FCMP_ULT: { + + middle2_bb = middle_bb->splitBasicBlock( + BasicBlock::iterator(middle_bb->getTerminator())); + + BasicBlock *negative_bb = BasicBlock::Create( + C, "negative_value", middle2_bb->getParent(), middle2_bb); + BasicBlock *positive_bb = BasicBlock::Create( + C, "positive_value", negative_bb->getParent(), negative_bb); + + if (FcmpInst->getPredicate() == CmpInst::FCMP_OGT || + FcmpInst->getPredicate() == CmpInst::FCMP_UGT) { + + negative_bb->getInstList().push_back( + icmp_fraction_result = CmpInst::Create( + Instruction::ICmp, CmpInst::ICMP_ULT, t_f0, t_f1)); + positive_bb->getInstList().push_back( + icmp_fraction_result2 = CmpInst::Create( + Instruction::ICmp, CmpInst::ICMP_UGT, t_f0, t_f1)); + + } else { + + negative_bb->getInstList().push_back( + icmp_fraction_result = CmpInst::Create( + Instruction::ICmp, CmpInst::ICMP_UGT, t_f0, t_f1)); + positive_bb->getInstList().push_back( + icmp_fraction_result2 = CmpInst::Create( + Instruction::ICmp, CmpInst::ICMP_ULT, t_f0, t_f1)); + + } + + BranchInst::Create(middle2_bb, negative_bb); + BranchInst::Create(middle2_bb, positive_bb); + + term = middle_bb->getTerminator(); + BranchInst::Create(negative_bb, positive_bb, t_s0, middle_bb); + term->eraseFromParent(); + + PN2 = PHINode::Create(Int1Ty, 2, ""); + PN2->addIncoming(icmp_fraction_result, negative_bb); + PN2->addIncoming(icmp_fraction_result2, positive_bb); + middle2_bb->getInstList().insert( + BasicBlock::iterator(middle2_bb->getTerminator()), PN2); + + } break; + default: continue; } - middle_bb->getInstList().insert( - BasicBlock::iterator(middle_bb->getTerminator()), icmp_fraction_result); - PHINode *PN = PHINode::Create(Int1Ty, 3, ""); switch (FcmpInst->getPredicate()) { @@ -849,7 +986,7 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { /* unequal exponents cannot be equal values, too */ PN->addIncoming(ConstantInt::get(Int1Ty, 0), signequal_bb); /* fractions comparison */ - PN->addIncoming(icmp_fraction_result, middle_bb); + PN->addIncoming(icmp_fraction_result, middle2_bb); break; case CmpInst::FCMP_ONE: case CmpInst::FCMP_UNE: @@ -857,25 +994,25 @@ size_t SplitComparesTransform::splitFPCompares(Module &M) { /* goto true branch */ PN->addIncoming(ConstantInt::get(Int1Ty, 1), bb); /* unequal exponents are unequal values, too */ - PN->addIncoming(ConstantInt::get(Int1Ty, 1), signequal_bb); + PN->addIncoming(icmp_exponent_result, signequal_bb); /* fractions comparison */ - PN->addIncoming(icmp_fraction_result, middle_bb); + PN->addIncoming(icmp_fraction_result, middle2_bb); break; case CmpInst::FCMP_OGT: case CmpInst::FCMP_UGT: /* if op1 is negative goto true branch, else go on comparing */ PN->addIncoming(t_s1, bb); - PN->addIncoming(icmp_exponent_result, signequal_bb); - PN->addIncoming(icmp_fraction_result, middle_bb); + PN->addIncoming(icmp_exponent_result, signequal2_bb); + PN->addIncoming(PN2, middle2_bb); break; case CmpInst::FCMP_OLT: case CmpInst::FCMP_ULT: /* if op0 is negative goto true branch, else go on comparing */ PN->addIncoming(t_s0, bb); - PN->addIncoming(icmp_exponent_result, signequal_bb); - PN->addIncoming(icmp_fraction_result, middle_bb); + PN->addIncoming(icmp_exponent_result, signequal2_bb); + PN->addIncoming(PN2, middle2_bb); break; default: continue; @@ -1107,7 +1244,8 @@ size_t SplitComparesTransform::splitIntCompares(Module &M, unsigned bitw) { bool SplitComparesTransform::runOnModule(Module &M) { - int bitw = 64; + int bitw = 64; + size_t count; char *bitw_env = getenv("AFL_LLVM_LAF_SPLIT_COMPARES_BITW"); if (!bitw_env) bitw_env = getenv("LAF_SPLIT_COMPARES_BITW"); @@ -1115,30 +1253,44 @@ bool SplitComparesTransform::runOnModule(Module &M) { enableFPSplit = getenv("AFL_LLVM_LAF_SPLIT_FLOATS") != NULL; - simplifyCompares(M); - - simplifyIntSignedness(M); - if ((isatty(2) && getenv("AFL_QUIET") == NULL) || getenv("AFL_DEBUG") != NULL) { errs() << "Split-compare-pass by laf.intel@gmail.com, extended by " "heiko@hexco.de\n"; - if (enableFPSplit) - errs() << "Split-floatingpoint-compare-pass: " << splitFPCompares(M) + } else { + + be_quiet = 1; + + } + + if (enableFPSplit) { + + count = splitFPCompares(M); + + if (!be_quiet) { + + errs() << "Split-floatingpoint-compare-pass: " << count << " FP comparisons splitted\n"; - } else + } - be_quiet = 1; + simplifyFPCompares(M); + + } + + simplifyCompares(M); + + simplifyIntSignedness(M); switch (bitw) { case 64: + count = splitIntCompares(M, bitw); if (!be_quiet) - errs() << "Split-integer-compare-pass " << bitw - << "bit: " << splitIntCompares(M, bitw) << " splitted\n"; + errs() << "Split-integer-compare-pass " << bitw << "bit: " << count + << " splitted\n"; bitw >>= 1; #if LLVM_VERSION_MAJOR > 3 || \ @@ -1146,9 +1298,10 @@ bool SplitComparesTransform::runOnModule(Module &M) { [[clang::fallthrough]]; /*FALLTHRU*/ /* FALLTHROUGH */ #endif case 32: + count = splitIntCompares(M, bitw); if (!be_quiet) - errs() << "Split-integer-compare-pass " << bitw - << "bit: " << splitIntCompares(M, bitw) << " splitted\n"; + errs() << "Split-integer-compare-pass " << bitw << "bit: " << count + << " splitted\n"; bitw >>= 1; #if LLVM_VERSION_MAJOR > 3 || \ @@ -1156,9 +1309,10 @@ bool SplitComparesTransform::runOnModule(Module &M) { [[clang::fallthrough]]; /*FALLTHRU*/ /* FALLTHROUGH */ #endif case 16: + count = splitIntCompares(M, bitw); if (!be_quiet) - errs() << "Split-integer-compare-pass " << bitw - << "bit: " << splitIntCompares(M, bitw) << " splitted\n"; + errs() << "Split-integer-compare-pass " << bitw << "bit: " << count + << " splitted\n"; bitw >>= 1; break; diff --git a/llvm_mode/split-switches-pass.so.cc b/llvm_mode/split-switches-pass.so.cc index 44075c94..f025df77 100644 --- a/llvm_mode/split-switches-pass.so.cc +++ b/llvm_mode/split-switches-pass.so.cc @@ -91,9 +91,6 @@ class SplitSwitchesTransform : public ModulePass { typedef std::vector<CaseExpr> CaseVector; - protected: - int be_quiet = 0; - private: bool splitSwitches(Module &M); bool transformCmps(Module &M, const bool processStrcmp, diff --git a/qemu_mode/patches/afl-qemu-cpu-inl.h b/qemu_mode/patches/afl-qemu-cpu-inl.h index 5f579687..63b7581d 100644 --- a/qemu_mode/patches/afl-qemu-cpu-inl.h +++ b/qemu_mode/patches/afl-qemu-cpu-inl.h @@ -620,7 +620,13 @@ static void afl_wait_tsl(CPUState *cpu, int fd) { last_tb = tb_htable_lookup(cpu, c.last_tb.pc, c.last_tb.cs_base, c.last_tb.flags, c.cf_mask); - if (last_tb) { tb_add_jump(last_tb, c.tb_exit, tb); } +#define TB_JMP_RESET_OFFSET_INVALID 0xffff + if (last_tb && (last_tb->jmp_reset_offset[c.tb_exit] != + TB_JMP_RESET_OFFSET_INVALID)) { + + tb_add_jump(last_tb, c.tb_exit, tb); + + } } diff --git a/src/afl-analyze.c b/src/afl-analyze.c index 56284f6f..e6dd0fca 100644 --- a/src/afl-analyze.c +++ b/src/afl-analyze.c @@ -772,15 +772,38 @@ static void set_up_environment(void) { setenv("ASAN_OPTIONS", "abort_on_error=1:" "detect_leaks=0:" + "allocator_may_return_null=1:" "symbolize=0:" - "allocator_may_return_null=1", + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", + 0); + + setenv("UBSAN_OPTIONS", + "halt_on_error=1:" + "abort_on_error=1:" + "malloc_context_size=0:" + "allocator_may_return_null=1:" + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); setenv("MSAN_OPTIONS", "exit_code=" STRINGIFY(MSAN_ERROR) ":" - "symbolize=0:" "abort_on_error=1:" + "msan_track_origins=0" "allocator_may_return_null=1:" - "msan_track_origins=0", 0); + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); if (get_afl_env("AFL_PRELOAD")) { diff --git a/src/afl-common.c b/src/afl-common.c index c023789b..367dec72 100644 --- a/src/afl-common.c +++ b/src/afl-common.c @@ -145,7 +145,8 @@ char **get_qemu_argv(u8 *own_loc, u8 **target_path_p, int argc, char **argv) { char **new_argv = ck_alloc(sizeof(char *) * (argc + 4)); u8 * tmp, *cp = NULL, *rsl, *own_copy; - memcpy(new_argv + 3, argv + 1, (int)(sizeof(char *)) * argc); + memcpy(&new_argv[3], &argv[1], (int)(sizeof(char *)) * (argc - 1)); + new_argv[argc - 1] = NULL; new_argv[2] = *target_path_p; new_argv[1] = "--"; @@ -226,7 +227,8 @@ char **get_wine_argv(u8 *own_loc, u8 **target_path_p, int argc, char **argv) { char **new_argv = ck_alloc(sizeof(char *) * (argc + 3)); u8 * tmp, *cp = NULL, *rsl, *own_copy; - memcpy(new_argv + 2, argv + 1, (int)(sizeof(char *)) * argc); + memcpy(&new_argv[2], &argv[1], (int)(sizeof(char *)) * (argc - 1)); + new_argv[argc - 1] = NULL; new_argv[1] = *target_path_p; diff --git a/src/afl-forkserver.c b/src/afl-forkserver.c index 419ce28e..47493eba 100644 --- a/src/afl-forkserver.c +++ b/src/afl-forkserver.c @@ -434,7 +434,27 @@ void afl_fsrv_start(afl_forkserver_t *fsrv, char **argv, "detect_leaks=0:" "malloc_context_size=0:" "symbolize=0:" - "allocator_may_return_null=1", + "allocator_may_return_null=1:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", + 0); + + /* Set sane defaults for UBSAN if nothing else specified. */ + + setenv("UBSAN_OPTIONS", + "halt_on_error=1:" + "abort_on_error=1:" + "malloc_context_size=0:" + "allocator_may_return_null=1:" + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); /* MSAN is tricky, because it doesn't support abort_on_error=1 at this @@ -446,7 +466,12 @@ void afl_fsrv_start(afl_forkserver_t *fsrv, char **argv, "abort_on_error=1:" "malloc_context_size=0:" "allocator_may_return_null=1:" - "msan_track_origins=0", + "msan_track_origins=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); fsrv->init_child_func(fsrv, argv); diff --git a/src/afl-fuzz-init.c b/src/afl-fuzz-init.c index a2e849dc..396a20f0 100644 --- a/src/afl-fuzz-init.c +++ b/src/afl-fuzz-init.c @@ -28,38 +28,157 @@ #ifdef HAVE_AFFINITY -/* Build a list of processes bound to specific cores. Returns -1 if nothing - can be found. Assumes an upper bound of 4k CPUs. */ +/* bind process to a specific cpu. Returns 0 on failure. */ -void bind_to_free_cpu(afl_state_t *afl) { +static u8 bind_cpu(afl_state_t *afl, s32 cpuid) { #if defined(__linux__) || defined(__FreeBSD__) || defined(__DragonFly__) cpu_set_t c; #elif defined(__NetBSD__) - cpuset_t * c; + cpuset_t *c; + #elif defined(__sun) + psetid_t c; + #endif + + afl->cpu_aff = cpuid; + + #if defined(__linux__) || defined(__FreeBSD__) || defined(__DragonFly__) + + CPU_ZERO(&c); + CPU_SET(cpuid, &c); + + #elif defined(__NetBSD__) + + c = cpuset_create(); + if (c == NULL) { PFATAL("cpuset_create failed"); } + cpuset_set(cpuid, c); + + #elif defined(__sun) + + pset_create(&c); + if (pset_assign(c, cpuid, NULL)) { PFATAL("pset_assign failed"); } + + #endif + + #if defined(__linux__) + + return (sched_setaffinity(0, sizeof(c), &c) == 0); + + #elif defined(__FreeBSD__) || defined(__DragonFly__) + + return (pthread_setaffinity_np(pthread_self(), sizeof(c), &c) == 0); + + #elif defined(__NetBSD__) + + if (pthread_setaffinity_np(pthread_self(), cpuset_size(c), c)) { + + cpuset_destroy(c); + return 0; + + } + + cpuset_destroy(c); + return 1; + #elif defined(__sun) - psetid_t c; + + if (pset_bind(c, P_PID, getpid(), NULL)) { + + pset_destroy(c); + return 0; + + } + + pset_destroy(c); + return 1; + + #else + + // this will need something for other platforms + // TODO: Solaris/Illumos has processor_bind ... might worth a try + WARNF("Cannot bind to CPU yet on this platform."); + return 1; + #endif +} + +/* Build a list of processes bound to specific cores. Returns -1 if nothing + can be found. Assumes an upper bound of 4k CPUs. */ + +void bind_to_free_cpu(afl_state_t *afl) { + u8 cpu_used[4096] = {0}; + u8 lockfile[PATH_MAX] = ""; u32 i; - if (afl->cpu_core_count < 2) { return; } - if (afl->afl_env.afl_no_affinity) { + if (afl->cpu_to_bind != -1) { + + FATAL("-b and AFL_NO_AFFINITY are mututally exclusive."); + + } + WARNF("Not binding to a CPU core (AFL_NO_AFFINITY set)."); return; } + if (afl->cpu_to_bind != -1) { + + if (!bind_cpu(afl, afl->cpu_to_bind)) { + + FATAL( + "Could not bind to requested CPU %d! Make sure you passed a valid " + "-b.", + afl->cpu_to_bind); + + } + + return; + + } + + if (afl->cpu_core_count < 2) { return; } + + if (afl->sync_id) { + + s32 lockfd, first = 1; + + snprintf(lockfile, sizeof(lockfile), "%s/.affinity_lock", afl->sync_dir); + setenv(CPU_AFFINITY_ENV_VAR, lockfile, 1); + + do { + + if ((lockfd = open(lockfile, O_RDWR | O_CREAT | O_EXCL, 0600)) < 0) { + + if (first) { + + WARNF("CPU affinity lock file present, waiting ..."); + first = 0; + + } + + usleep(1000); + + } + + } while (lockfd < 0); + + close(lockfd); + + } + #if defined(__linux__) + DIR * d; struct dirent *de; d = opendir("/proc"); if (!d) { + if (lockfile[0]) unlink(lockfile); WARNF("Unable to access /proc - can't scan for free CPU cores."); return; @@ -67,11 +186,6 @@ void bind_to_free_cpu(afl_state_t *afl) { ACTF("Checking CPU core loadout..."); - /* Introduce some jitter, in case multiple AFL tasks are doing the same - thing at the same time... */ - - usleep(R(1000) * 250); - /* Scan all /proc/<pid>/status entries, checking for Cpus_allowed_list. Flag all processes bound to a specific CPU using cpu_used[]. This will fail for some exotic binding setups, but is likely good enough in almost @@ -114,20 +228,29 @@ void bind_to_free_cpu(afl_state_t *afl) { } closedir(d); + #elif defined(__FreeBSD__) || defined(__DragonFly__) + struct kinfo_proc *procs; size_t nprocs; size_t proccount; int s_name[] = {CTL_KERN, KERN_PROC, KERN_PROC_ALL}; size_t s_name_l = sizeof(s_name) / sizeof(s_name[0]); - if (sysctl(s_name, s_name_l, NULL, &nprocs, NULL, 0) != 0) return; + if (sysctl(s_name, s_name_l, NULL, &nprocs, NULL, 0) != 0) { + + if (lockfile[0]) unlink(lockfile); + return; + + } + proccount = nprocs / sizeof(*procs); nprocs = nprocs * 4 / 3; procs = ck_alloc(nprocs); if (sysctl(s_name, s_name_l, procs, &nprocs, NULL, 0) != 0) { + if (lockfile[0]) unlink(lockfile); ck_free(procs); return; @@ -136,6 +259,7 @@ void bind_to_free_cpu(afl_state_t *afl) { for (i = 0; i < proccount; i++) { #if defined(__FreeBSD__) + if (!strcmp(procs[i].ki_comm, "idle")) continue; // fix when ki_oncpu = -1 @@ -145,16 +269,21 @@ void bind_to_free_cpu(afl_state_t *afl) { if (oncpu != -1 && oncpu < sizeof(cpu_used) && procs[i].ki_pctcpu > 60) cpu_used[oncpu] = 1; + #elif defined(__DragonFly__) + if (procs[i].kp_lwp.kl_cpuid < sizeof(cpu_used) && procs[i].kp_lwp.kl_pctcpu > 10) cpu_used[procs[i].kp_lwp.kl_cpuid] = 1; + #endif } ck_free(procs); + #elif defined(__NetBSD__) + struct kinfo_proc2 *procs; size_t nprocs; size_t proccount; @@ -163,13 +292,20 @@ void bind_to_free_cpu(afl_state_t *afl) { CTL_KERN, KERN_PROC2, KERN_PROC_ALL, 0, sizeof(struct kinfo_proc2), 0}; size_t s_name_l = sizeof(s_name) / sizeof(s_name[0]); - if (sysctl(s_name, s_name_l, NULL, &nprocs, NULL, 0) != 0) return; + if (sysctl(s_name, s_name_l, NULL, &nprocs, NULL, 0) != 0) { + + if (lockfile[0]) unlink(lockfile); + return; + + } + proccount = nprocs / sizeof(struct kinfo_proc2); procs = ck_alloc(nprocs * sizeof(struct kinfo_proc2)); s_name[5] = proccount; if (sysctl(s_name, s_name_l, procs, &nprocs, NULL, 0) != 0) { + if (lockfile[0]) unlink(lockfile); ck_free(procs); return; @@ -183,7 +319,9 @@ void bind_to_free_cpu(afl_state_t *afl) { } ck_free(procs); + #elif defined(__sun) + kstat_named_t *n; kstat_ctl_t * m; kstat_t * k; @@ -198,6 +336,7 @@ void bind_to_free_cpu(afl_state_t *afl) { if (!k) { + if (lockfile[0]) unlink(lockfile); kstat_close(m); return; @@ -205,6 +344,7 @@ void bind_to_free_cpu(afl_state_t *afl) { if (kstat_read(m, k, NULL)) { + if (lockfile[0]) unlink(lockfile); kstat_close(m); return; @@ -220,6 +360,7 @@ void bind_to_free_cpu(afl_state_t *afl) { k = kstat_lookup(m, "cpu_stat", i, NULL); if (kstat_read(m, k, &cs)) { + if (lockfile[0]) unlink(lockfile); kstat_close(m); return; @@ -233,6 +374,7 @@ void bind_to_free_cpu(afl_state_t *afl) { } kstat_close(m); + #else #warning \ "For this platform we do not have free CPU binding code yet. If possible, please supply a PR to https://github.com/AFLplusplus/AFLplusplus" @@ -240,23 +382,38 @@ void bind_to_free_cpu(afl_state_t *afl) { size_t cpu_start = 0; - try: #if !defined(__ANDROID__) - for (i = cpu_start; i < afl->cpu_core_count; i++) { - if (!cpu_used[i]) { break; } + for (i = cpu_start; i < afl->cpu_core_count; i++) { - } + #else - if (i == afl->cpu_core_count) { + /* for some reason Android goes backwards */ - #else - for (i = afl->cpu_core_count - cpu_start - 1; i > -1; i--) - if (!cpu_used[i]) break; - if (i == -1) { + for (i = afl->cpu_core_count - 1; i > -1; i--) { #endif + if (cpu_used[i]) { continue; } + + OKF("Found a free CPU core, try binding to #%u.", i); + + if (bind_cpu(afl, i)) { + + /* Success :) */ + break; + + } + + WARNF("setaffinity failed to CPU %d, trying next CPU", i); + cpu_start++; + + } + + if (lockfile[0]) unlink(lockfile); + + if (i == afl->cpu_core_count || i == -1) { + SAYF("\n" cLRD "[-] " cRST "Uh-oh, looks like all %d CPU cores on your system are allocated to\n" " other instances of afl-fuzz (or similar CPU-locked tasks). " @@ -269,97 +426,175 @@ void bind_to_free_cpu(afl_state_t *afl) { } - OKF("Found a free CPU core, try binding to #%u.", i); - - afl->cpu_aff = i; +} - #if defined(__linux__) || defined(__FreeBSD__) || defined(__DragonFly__) - CPU_ZERO(&c); - CPU_SET(i, &c); - #elif defined(__NetBSD__) - c = cpuset_create(); - if (c == NULL) PFATAL("cpuset_create failed"); - cpuset_set(i, c); - #elif defined(__sun) -pset_create(&c); -if (pset_assign(c, i, NULL)) PFATAL("pset_assign failed"); - #endif +#endif /* HAVE_AFFINITY */ - #if defined(__linux__) - if (sched_setaffinity(0, sizeof(c), &c)) { +/* Shuffle an array of pointers. Might be slightly biased. */ - if (cpu_start == afl->cpu_core_count) { +static void shuffle_ptrs(afl_state_t *afl, void **ptrs, u32 cnt) { - PFATAL("sched_setaffinity failed for CPU %d, exit", i); + u32 i; - } + for (i = 0; i < cnt - 2; ++i) { - WARNF("sched_setaffinity failed to CPU %d, trying next CPU", i); - cpu_start++; - goto try - ; + u32 j = i + rand_below(afl, cnt - i); + void *s = ptrs[i]; + ptrs[i] = ptrs[j]; + ptrs[j] = s; } - #elif defined(__FreeBSD__) || defined(__DragonFly__) - if (pthread_setaffinity_np(pthread_self(), sizeof(c), &c)) { +} - if (cpu_start == afl->cpu_core_count) - PFATAL("pthread_setaffinity failed for cpu %d, exit", i); - WARNF("pthread_setaffinity failed to CPU %d, trying next CPU", i); - cpu_start++; - goto try - ; +/* Read all testcases from foreign input directories, then queue them for + testing. Called at startup and at sync intervals. + Does not descend into subdirectories! */ - } +void read_foreign_testcases(afl_state_t *afl, int first) { - #elif defined(__NetBSD__) -if (pthread_setaffinity_np(pthread_self(), cpuset_size(c), c)) { + if (!afl->foreign_sync_cnt) return; - if (cpu_start == afl->cpu_core_count) - PFATAL("pthread_setaffinity failed for cpu %d, exit", i); - WARNF("pthread_setaffinity failed to CPU %d, trying next CPU", i); - cpu_start++; - goto try - ; + struct dirent **nl; + s32 nl_cnt; + u32 i, iter; -} + u8 val_buf[2][STRINGIFY_VAL_SIZE_MAX]; -cpuset_destroy(c); - #elif defined(__sun) -if (pset_bind(c, P_PID, getpid(), NULL)) { + for (iter = 0; iter < afl->foreign_sync_cnt; iter++) { - if (cpu_start == afl->cpu_core_count) - PFATAL("pset_bind failed for cpu %d, exit", i); - WARNF("pthread_setaffinity failed to CPU %d, trying next CPU", i); - cpu_start++; - goto try - ; + if (afl->foreign_syncs[iter].dir != NULL && + afl->foreign_syncs[iter].dir[0] != 0) { -} + if (first) ACTF("Scanning '%s'...", afl->foreign_syncs[iter].dir); + time_t ctime_max = 0; -pset_destroy(c); - #else - // this will need something for other platforms - // TODO: Solaris/Illumos has processor_bind ... might worth a try - #endif + /* We use scandir() + alphasort() rather than readdir() because otherwise, + the ordering of test cases would vary somewhat randomly and would be + difficult to control. */ -} + nl_cnt = scandir(afl->foreign_syncs[iter].dir, &nl, NULL, NULL); -#endif /* HAVE_AFFINITY */ + if (nl_cnt < 0) { -/* Shuffle an array of pointers. Might be slightly biased. */ + if (first) { -static void shuffle_ptrs(afl_state_t *afl, void **ptrs, u32 cnt) { + WARNF("Unable to open directory '%s'", afl->foreign_syncs[iter].dir); + sleep(1); - u32 i; + } - for (i = 0; i < cnt - 2; ++i) { + continue; - u32 j = i + rand_below(afl, cnt - i); - void *s = ptrs[i]; - ptrs[i] = ptrs[j]; - ptrs[j] = s; + } + + if (nl_cnt == 0) { + + if (first) + WARNF("directory %s is currently empty", + afl->foreign_syncs[iter].dir); + continue; + + } + + /* Show stats */ + + snprintf(afl->stage_name_buf, STAGE_BUF_SIZE, "foreign sync %u", iter); + + afl->stage_name = afl->stage_name_buf; + afl->stage_cur = 0; + afl->stage_max = 0; + + for (i = 0; i < nl_cnt; ++i) { + + struct stat st; + + u8 *fn2 = + alloc_printf("%s/%s", afl->foreign_syncs[iter].dir, nl[i]->d_name); + + free(nl[i]); /* not tracked */ + + if (unlikely(lstat(fn2, &st) || access(fn2, R_OK))) { + + if (first) PFATAL("Unable to access '%s'", fn2); + continue; + + } + + /* we detect new files by their ctime */ + if (likely(st.st_ctime <= afl->foreign_syncs[iter].ctime)) { + + ck_free(fn2); + continue; + + } + + /* This also takes care of . and .. */ + + if (!S_ISREG(st.st_mode) || !st.st_size || strstr(fn2, "/README.txt")) { + + ck_free(fn2); + continue; + + } + + if (st.st_size > MAX_FILE) { + + if (first) + WARNF( + "Test case '%s' is too big (%s, limit is %s), skipping", fn2, + stringify_mem_size(val_buf[0], sizeof(val_buf[0]), st.st_size), + stringify_mem_size(val_buf[1], sizeof(val_buf[1]), MAX_FILE)); + ck_free(fn2); + continue; + + } + + // lets do not use add_to_queue(afl, fn2, st.st_size, 0); + // as this could add duplicates of the startup input corpus + + int fd = open(fn2, O_RDONLY); + if (fd < 0) { + + ck_free(fn2); + continue; + + } + + u8 fault; + u8 *mem = mmap(0, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0); + + if (mem == MAP_FAILED) { + + ck_free(fn2); + continue; + + } + + write_to_testcase(afl, mem, st.st_size); + fault = fuzz_run_target(afl, &afl->fsrv, afl->fsrv.exec_tmout); + afl->syncing_party = "foreign"; + afl->queued_imported += + save_if_interesting(afl, mem, st.st_size, fault); + afl->syncing_party = 0; + munmap(mem, st.st_size); + close(fd); + + if (st.st_ctime > ctime_max) ctime_max = st.st_ctime; + + } + + afl->foreign_syncs[iter].ctime = ctime_max; + free(nl); /* not tracked */ + + } + + } + + if (first) { + + afl->last_path_time = 0; + afl->queued_at_start = afl->queued_paths; } @@ -393,7 +628,7 @@ void read_testcases(afl_state_t *afl) { ACTF("Scanning '%s'...", afl->in_dir); /* We use scandir() + alphasort() rather than readdir() because otherwise, - the ordering of test cases would vary somewhat randomly and would be + the ordering of test cases would vary somewhat randomly and would be difficult to control. */ nl_cnt = scandir(afl->in_dir, &nl, NULL, alphasort); @@ -454,9 +689,11 @@ void read_testcases(afl_state_t *afl) { if (st.st_size > MAX_FILE) { - FATAL("Test case '%s' is too big (%s, limit is %s)", fn2, + WARNF("Test case '%s' is too big (%s, limit is %s), skipping", fn2, stringify_mem_size(val_buf[0], sizeof(val_buf[0]), st.st_size), stringify_mem_size(val_buf[1], sizeof(val_buf[1]), MAX_FILE)); + ck_free(fn2); + continue; } diff --git a/src/afl-fuzz-mutators.c b/src/afl-fuzz-mutators.c index 9fc77ffe..b30106a0 100644 --- a/src/afl-fuzz-mutators.c +++ b/src/afl-fuzz-mutators.c @@ -40,7 +40,7 @@ void setup_custom_mutators(afl_state_t *afl) { if (fn) { - if (afl->limit_time_sig) + if (afl->limit_time_sig && afl->limit_time_sig != -1) FATAL( "MOpt and custom mutator are mutually exclusive. We accept pull " "requests that integrates MOpt with the optional mutators " @@ -168,7 +168,8 @@ struct custom_mutator *load_custom_mutator(afl_state_t *afl, const char *fn) { /* "afl_custom_deinit", optional for backward compatibility */ mutator->afl_custom_deinit = dlsym(dh, "afl_custom_deinit"); - if (!mutator->afl_custom_deinit) FATAL("Symbol 'afl_custom_init' not found."); + if (!mutator->afl_custom_deinit) + FATAL("Symbol 'afl_custom_deinit' not found."); /* "afl_custom_post_process", optional */ mutator->afl_custom_post_process = dlsym(dh, "afl_custom_post_process"); @@ -282,22 +283,48 @@ u8 trim_case_custom(afl_state_t *afl, struct queue_entry *q, u8 *in_buf, } else if (unlikely(retlen > orig_len)) { - FATAL( - "Trimmed data returned by custom mutator is larger than original " - "data"); + /* Do not exit the fuzzer, even if the trimmed data returned by the custom + mutator is larger than the original data. For some use cases, like the + grammar mutator, the definition of "size" may have different meanings. + For example, the trimming function in a grammar mutator aims at + reducing the objects in a grammar structure, but does not guarantee to + generate a smaller binary buffer. + + Thus, we allow the custom mutator to generate the trimmed data that is + larger than the original data. */ + + if (afl->not_on_tty && afl->debug) { + + WARNF( + "Trimmed data returned by custom mutator is larger than original " + "data"); + + } + + } else if (unlikely(retlen == 0)) { + + /* Do not run the empty test case on the target. To keep the custom + trimming function running, we simply treat the empty test case as an + unsuccessful trimming and skip it, instead of aborting the trimming. */ + + ++afl->trim_execs; } - write_to_testcase(afl, retbuf, retlen); + if (likely(retlen)) { - fault = fuzz_run_target(afl, &afl->fsrv, afl->fsrv.exec_tmout); - ++afl->trim_execs; + write_to_testcase(afl, retbuf, retlen); - if (afl->stop_soon || fault == FSRV_RUN_ERROR) { goto abort_trimming; } + fault = fuzz_run_target(afl, &afl->fsrv, afl->fsrv.exec_tmout); + ++afl->trim_execs; - cksum = hash64(afl->fsrv.trace_bits, afl->fsrv.map_size, HASH_CONST); + if (afl->stop_soon || fault == FSRV_RUN_ERROR) { goto abort_trimming; } + + cksum = hash64(afl->fsrv.trace_bits, afl->fsrv.map_size, HASH_CONST); + + } - if (cksum == q->exec_cksum) { + if (likely(retlen && cksum == q->exec_cksum)) { q->len = retlen; memcpy(in_buf, retbuf, retlen); diff --git a/src/afl-fuzz-one.c b/src/afl-fuzz-one.c index 72383727..1f0bf30e 100644 --- a/src/afl-fuzz-one.c +++ b/src/afl-fuzz-one.c @@ -24,6 +24,8 @@ */ #include "afl-fuzz.h" +#include <string.h> +#include <limits.h> /* MOpt */ @@ -362,6 +364,8 @@ static void locate_diffs(u8 *ptr1, u8 *ptr2, u32 len, s32 *first, s32 *last) { #endif /* !IGNORE_FINDS */ +#define BUF_PARAMS(name) (void **)&afl->name##_buf, &afl->name##_size + /* Take the current entry from the queue, fuzz it for a while. This function is a tad too long... returns 0 if fuzzed successfully, 1 if skipped or bailed out. */ @@ -1854,6 +1858,21 @@ havoc_stage: /* We essentially just do several thousand runs (depending on perf_score) where we take the input file and make random stacked tweaks. */ + u32 r_max, r; + + if (unlikely(afl->expand_havoc)) { + + /* add expensive havoc cases here, they are activated after a full + cycle without finds happened */ + + r_max = 16 + ((afl->extras_cnt + afl->a_extras_cnt) ? 2 : 0); + + } else { + + r_max = 15 + ((afl->extras_cnt + afl->a_extras_cnt) ? 2 : 0); + + } + for (afl->stage_cur = 0; afl->stage_cur < afl->stage_max; ++afl->stage_cur) { u32 use_stacking = 1 << (1 + rand_below(afl, HAVOC_STACK_POW2)); @@ -1896,8 +1915,7 @@ havoc_stage: } - switch (rand_below( - afl, 15 + ((afl->extras_cnt + afl->a_extras_cnt) ? 2 : 0))) { + switch ((r = rand_below(afl, r_max))) { case 0: @@ -2192,85 +2210,198 @@ havoc_stage: } - /* Values 15 and 16 can be selected only if there are any extras - present in the dictionaries. */ + default: - case 15: { + if (likely(r <= 16 && (afl->extras_cnt || afl->a_extras_cnt))) { - /* Overwrite bytes with an extra. */ + /* Values 15 and 16 can be selected only if there are any extras + present in the dictionaries. */ - if (!afl->extras_cnt || (afl->a_extras_cnt && rand_below(afl, 2))) { + if (r == 15) { - /* No user-specified extras or odds in our favor. Let's use an - auto-detected one. */ + /* Overwrite bytes with an extra. */ - u32 use_extra = rand_below(afl, afl->a_extras_cnt); - u32 extra_len = afl->a_extras[use_extra].len; - u32 insert_at; + if (!afl->extras_cnt || + (afl->a_extras_cnt && rand_below(afl, 2))) { - if (extra_len > temp_len) { break; } + /* No user-specified extras or odds in our favor. Let's use an + auto-detected one. */ - insert_at = rand_below(afl, temp_len - extra_len + 1); - memcpy(out_buf + insert_at, afl->a_extras[use_extra].data, - extra_len); + u32 use_extra = rand_below(afl, afl->a_extras_cnt); + u32 extra_len = afl->a_extras[use_extra].len; + u32 insert_at; - } else { + if (extra_len > temp_len) { break; } - /* No auto extras or odds in our favor. Use the dictionary. */ + insert_at = rand_below(afl, temp_len - extra_len + 1); + memcpy(out_buf + insert_at, afl->a_extras[use_extra].data, + extra_len); - u32 use_extra = rand_below(afl, afl->extras_cnt); - u32 extra_len = afl->extras[use_extra].len; - u32 insert_at; + } else { - if (extra_len > temp_len) { break; } + /* No auto extras or odds in our favor. Use the dictionary. */ - insert_at = rand_below(afl, temp_len - extra_len + 1); - memcpy(out_buf + insert_at, afl->extras[use_extra].data, extra_len); + u32 use_extra = rand_below(afl, afl->extras_cnt); + u32 extra_len = afl->extras[use_extra].len; + u32 insert_at; - } + if (extra_len > temp_len) { break; } - break; + insert_at = rand_below(afl, temp_len - extra_len + 1); + memcpy(out_buf + insert_at, afl->extras[use_extra].data, + extra_len); - } + } - case 16: { + break; - u32 use_extra, extra_len, insert_at = rand_below(afl, temp_len + 1); - u8 *ptr; + } else { // case 16 + + u32 use_extra, extra_len, + insert_at = rand_below(afl, temp_len + 1); + u8 *ptr; + + /* Insert an extra. Do the same dice-rolling stuff as for the + previous case. */ + + if (!afl->extras_cnt || + (afl->a_extras_cnt && rand_below(afl, 2))) { + + use_extra = rand_below(afl, afl->a_extras_cnt); + extra_len = afl->a_extras[use_extra].len; + ptr = afl->a_extras[use_extra].data; + + } else { + + use_extra = rand_below(afl, afl->extras_cnt); + extra_len = afl->extras[use_extra].len; + ptr = afl->extras[use_extra].data; + + } + + if (temp_len + extra_len >= MAX_FILE) { break; } + + out_buf = ck_maybe_grow(BUF_PARAMS(out), temp_len + extra_len); + + /* Tail */ + memmove(out_buf + insert_at + extra_len, out_buf + insert_at, + temp_len - insert_at); + + /* Inserted part */ + memcpy(out_buf + insert_at, ptr, extra_len); - /* Insert an extra. Do the same dice-rolling stuff as for the - previous case. */ + temp_len += extra_len; - if (!afl->extras_cnt || (afl->a_extras_cnt && rand_below(afl, 2))) { + break; - use_extra = rand_below(afl, afl->a_extras_cnt); - extra_len = afl->a_extras[use_extra].len; - ptr = afl->a_extras[use_extra].data; + } } else { - use_extra = rand_below(afl, afl->extras_cnt); - extra_len = afl->extras[use_extra].len; - ptr = afl->extras[use_extra].data; + /* + switch (r) { - } + case 15: // fall through + case 16: + case 17: {*/ - if (temp_len + extra_len >= MAX_FILE) { break; } + /* Overwrite bytes with a randomly selected chunk from another + testcase or insert that chunk. */ - out_buf = ck_maybe_grow(BUF_PARAMS(out), temp_len + extra_len); + if (afl->queued_paths < 4) break; - /* Tail */ - memmove(out_buf + insert_at + extra_len, out_buf + insert_at, - temp_len - insert_at); + /* Pick a random queue entry and seek to it. */ - /* Inserted part */ - memcpy(out_buf + insert_at, ptr, extra_len); + u32 tid; + do + tid = rand_below(afl, afl->queued_paths); + while (tid == afl->current_entry); - temp_len += extra_len; + struct queue_entry *target = afl->queue_buf[tid]; - break; + /* Make sure that the target has a reasonable length. */ - } + while (target && (target->len < 2 || target == afl->queue_cur)) + target = target->next; + + if (!target) break; + + /* Read the testcase into a new buffer. */ + + fd = open(target->fname, O_RDONLY); + + if (unlikely(fd < 0)) { + + PFATAL("Unable to open '%s'", target->fname); + + } + + u32 new_len = target->len; + u8 *new_buf = ck_maybe_grow(BUF_PARAMS(in_scratch), new_len); + + ck_read(fd, new_buf, new_len, target->fname); + + close(fd); + + u8 overwrite = 0; + if (temp_len >= 2 && rand_below(afl, 2)) + overwrite = 1; + else if (temp_len + HAVOC_BLK_XL >= MAX_FILE) { + + if (temp_len >= 2) + overwrite = 1; + else + break; + + } + + if (overwrite) { + + u32 copy_from, copy_to, copy_len; + + copy_len = choose_block_len(afl, new_len - 1); + if (copy_len > temp_len) copy_len = temp_len; + + copy_from = rand_below(afl, new_len - copy_len + 1); + copy_to = rand_below(afl, temp_len - copy_len + 1); + + memmove(out_buf + copy_to, new_buf + copy_from, copy_len); + + } else { + + u32 clone_from, clone_to, clone_len; + + clone_len = choose_block_len(afl, new_len); + clone_from = rand_below(afl, new_len - clone_len + 1); + + clone_to = rand_below(afl, temp_len); + + u8 *temp_buf = + ck_maybe_grow(BUF_PARAMS(out_scratch), temp_len + clone_len); + + /* Head */ + + memcpy(temp_buf, out_buf, clone_to); + + /* Inserted part */ + + memcpy(temp_buf + clone_to, new_buf + clone_from, clone_len); + + /* Tail */ + memcpy(temp_buf + clone_to + clone_len, out_buf + clone_to, + temp_len - clone_to); + + swap_bufs(BUF_PARAMS(out), BUF_PARAMS(out_scratch)); + out_buf = temp_buf; + temp_len += clone_len; + + } + + break; + + } + + // end of default: } @@ -2357,20 +2488,7 @@ retry_splicing: } while (tid == afl->current_entry); afl->splicing_with = tid; - target = afl->queue; - - while (tid >= 100) { - - target = target->next_100; - tid -= 100; - - } - - while (tid--) { - - target = target->next; - - } + target = afl->queue_buf[tid]; /* Make sure that the target has a reasonable length. */ @@ -4750,7 +4868,7 @@ u8 fuzz_one(afl_state_t *afl) { return (key_val_lv_1 | key_val_lv_2); -#undef BUF_PARAMS - } +#undef BUF_PARAMS + diff --git a/src/afl-fuzz-queue.c b/src/afl-fuzz-queue.c index 7afdd9f1..71874283 100644 --- a/src/afl-fuzz-queue.c +++ b/src/afl-fuzz-queue.c @@ -24,6 +24,9 @@ #include "afl-fuzz.h" #include <limits.h> +#include <ctype.h> + +#define BUF_PARAMS(name) (void **)&afl->name##_buf, &afl->name##_size /* Mark deterministic checks as done for a particular queue entry. We use the .state file to avoid repeating deterministic fuzzing when resuming aborted @@ -100,6 +103,111 @@ void mark_as_redundant(afl_state_t *afl, struct queue_entry *q, u8 state) { } +/* check if ascii or UTF-8 */ + +static u8 check_if_text(struct queue_entry *q) { + + if (q->len < AFL_TXT_MIN_LEN) return 0; + + u8 buf[MAX_FILE]; + s32 fd, len = q->len, offset = 0, ascii = 0, utf8 = 0, comp; + + if ((fd = open(q->fname, O_RDONLY)) < 0) return 0; + if ((comp = read(fd, buf, len)) != len) return 0; + close(fd); + + while (offset < len) { + + // ASCII: <= 0x7F to allow ASCII control characters + if ((buf[offset + 0] == 0x09 || buf[offset + 0] == 0x0A || + buf[offset + 0] == 0x0D || + (0x20 <= buf[offset + 0] && buf[offset + 0] <= 0x7E))) { + + offset++; + utf8++; + ascii++; + continue; + + } + + if (isascii((int)buf[offset]) || isprint((int)buf[offset])) { + + ascii++; + // we continue though as it can also be a valid utf8 + + } + + // non-overlong 2-byte + if (((0xC2 <= buf[offset + 0] && buf[offset + 0] <= 0xDF) && + (0x80 <= buf[offset + 1] && buf[offset + 1] <= 0xBF)) && + len - offset > 1) { + + offset += 2; + utf8++; + comp--; + continue; + + } + + // excluding overlongs + if ((len - offset > 2) && + ((buf[offset + 0] == 0xE0 && + (0xA0 <= buf[offset + 1] && buf[offset + 1] <= 0xBF) && + (0x80 <= buf[offset + 2] && + buf[offset + 2] <= 0xBF)) || // straight 3-byte + (((0xE1 <= buf[offset + 0] && buf[offset + 0] <= 0xEC) || + buf[offset + 0] == 0xEE || buf[offset + 0] == 0xEF) && + (0x80 <= buf[offset + 1] && buf[offset + 1] <= 0xBF) && + (0x80 <= buf[offset + 2] && + buf[offset + 2] <= 0xBF)) || // excluding surrogates + (buf[offset + 0] == 0xED && + (0x80 <= buf[offset + 1] && buf[offset + 1] <= 0x9F) && + (0x80 <= buf[offset + 2] && buf[offset + 2] <= 0xBF)))) { + + offset += 3; + utf8++; + comp -= 2; + continue; + + } + + // planes 1-3 + if ((len - offset > 3) && + ((buf[offset + 0] == 0xF0 && + (0x90 <= buf[offset + 1] && buf[offset + 1] <= 0xBF) && + (0x80 <= buf[offset + 2] && buf[offset + 2] <= 0xBF) && + (0x80 <= buf[offset + 3] && + buf[offset + 3] <= 0xBF)) || // planes 4-15 + ((0xF1 <= buf[offset + 0] && buf[offset + 0] <= 0xF3) && + (0x80 <= buf[offset + 1] && buf[offset + 1] <= 0xBF) && + (0x80 <= buf[offset + 2] && buf[offset + 2] <= 0xBF) && + (0x80 <= buf[offset + 3] && buf[offset + 3] <= 0xBF)) || // plane 16 + (buf[offset + 0] == 0xF4 && + (0x80 <= buf[offset + 1] && buf[offset + 1] <= 0x8F) && + (0x80 <= buf[offset + 2] && buf[offset + 2] <= 0xBF) && + (0x80 <= buf[offset + 3] && buf[offset + 3] <= 0xBF)))) { + + offset += 4; + utf8++; + comp -= 3; + continue; + + } + + offset++; + + } + + u32 percent_utf8 = (utf8 * 100) / comp; + u32 percent_ascii = (ascii * 100) / len; + + if (percent_utf8 >= percent_ascii && percent_utf8 >= AFL_TXT_MIN_PERCENT) + return 2; + if (percent_ascii >= AFL_TXT_MIN_PERCENT) return 1; + return 0; + +} + /* Append new test case to the queue. */ void add_to_queue(afl_state_t *afl, u8 *fname, u32 len, u8 passed_det) { @@ -138,6 +246,10 @@ void add_to_queue(afl_state_t *afl, u8 *fname, u32 len, u8 passed_det) { } + struct queue_entry **queue_buf = ck_maybe_grow( + BUF_PARAMS(queue), afl->queued_paths * sizeof(struct queue_entry *)); + queue_buf[afl->queued_paths - 1] = q; + afl->last_path_time = get_cur_time(); if (afl->custom_mutators_count) { @@ -159,6 +271,9 @@ void add_to_queue(afl_state_t *afl, u8 *fname, u32 len, u8 passed_det) { } + /* only redqueen currently uses is_ascii */ + if (afl->shm.cmplog_mode) q->is_ascii = check_if_text(q); + } /* Destroy the entire queue. */ diff --git a/src/afl-fuzz-redqueen.c b/src/afl-fuzz-redqueen.c index 724da407..cb4c78df 100644 --- a/src/afl-fuzz-redqueen.c +++ b/src/afl-fuzz-redqueen.c @@ -24,6 +24,7 @@ */ +#include <limits.h> #include "afl-fuzz.h" #include "cmplog.h" @@ -177,6 +178,9 @@ static u8 colorization(afl_state_t *afl, u8 *buf, u32 len, u64 exec_cksum) { afl->stage_cycles[STAGE_COLORIZATION] += afl->stage_cur; ck_free(backup); + ck_free(rng); + rng = NULL; + while (ranges) { rng = ranges; @@ -186,9 +190,6 @@ static u8 colorization(afl_state_t *afl, u8 *buf, u32 len, u64 exec_cksum) { } - ck_free(rng); - rng = NULL; - // save the input with the high entropy if (needs_write) { @@ -262,11 +263,64 @@ static u8 its_fuzz(afl_state_t *afl, u8 *buf, u32 len, u8 *status) { } +static long long strntoll(const char *str, size_t sz, char **end, int base) { + + char buf[64]; + long long ret; + const char *beg = str; + + for (; beg && sz && *beg == ' '; beg++, sz--) {}; + + if (!sz || sz >= sizeof(buf)) { + + if (end) *end = (char *)str; + return 0; + + } + + memcpy(buf, beg, sz); + buf[sz] = '\0'; + ret = strtoll(buf, end, base); + if (ret == LLONG_MIN || ret == LLONG_MAX) return ret; + if (end) *end = (char *)beg + (*end - buf); + return ret; + +} + +static unsigned long long strntoull(const char *str, size_t sz, char **end, + int base) { + + char buf[64]; + unsigned long long ret; + const char * beg = str; + + for (; beg && sz && *beg == ' '; beg++, sz--) + ; + + if (!sz || sz >= sizeof(buf)) { + + if (end) *end = (char *)str; + return 0; + + } + + memcpy(buf, beg, sz); + buf[sz] = '\0'; + ret = strtoull(buf, end, base); + if (end) *end = (char *)beg + (*end - buf); + return ret; + +} + +#define BUF_PARAMS(name) (void **)&afl->name##_buf, &afl->name##_size + static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, u64 pattern, u64 repl, u64 o_pattern, u32 idx, u8 *orig_buf, u8 *buf, u32 len, u8 do_reverse, u8 *status) { + if (!buf) { FATAL("BUG: buf was NULL. Please report this.\n"); } + u64 *buf_64 = (u64 *)&buf[idx]; u32 *buf_32 = (u32 *)&buf[idx]; u16 *buf_16 = (u16 *)&buf[idx]; @@ -277,9 +331,56 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, u8 * o_buf_8 = &orig_buf[idx]; u32 its_len = len - idx; - *status = 0; + // *status = 0; + + u8 * endptr; + u8 use_num = 0, use_unum = 0; + unsigned long long unum; + long long num; + if (afl->queue_cur->is_ascii) { + + endptr = buf_8; + num = strntoll(buf_8, len - idx, (char **)&endptr, 0); + if (endptr == buf_8) { + + unum = strntoull(buf_8, len - idx, (char **)&endptr, 0); + if (endptr == buf_8) use_unum = 1; + + } else + + use_num = 1; + + } + + if (use_num && num == pattern) { - if (SHAPE_BYTES(h->shape) == 8) { + size_t old_len = endptr - buf_8; + size_t num_len = snprintf(NULL, 0, "%lld", num); + + u8 *new_buf = ck_maybe_grow(BUF_PARAMS(out_scratch), len + num_len); + memcpy(new_buf, buf, idx); + + snprintf(new_buf + idx, num_len, "%lld", num); + memcpy(new_buf + idx + num_len, buf_8 + old_len, len - idx - old_len); + + if (unlikely(its_fuzz(afl, new_buf, len, status))) { return 1; } + + } else if (use_unum && unum == pattern) { + + size_t old_len = endptr - buf_8; + size_t num_len = snprintf(NULL, 0, "%llu", unum); + + u8 *new_buf = ck_maybe_grow(BUF_PARAMS(out_scratch), len + num_len); + memcpy(new_buf, buf, idx); + + snprintf(new_buf + idx, num_len, "%llu", unum); + memcpy(new_buf + idx + num_len, buf_8 + old_len, len - idx - old_len); + + if (unlikely(its_fuzz(afl, new_buf, len, status))) { return 1; } + + } + + if (SHAPE_BYTES(h->shape) >= 8 && *status != 1) { if (its_len >= 8 && *buf_64 == pattern && *o_buf_64 == o_pattern) { @@ -290,7 +391,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } // reverse encoding - if (do_reverse) { + if (do_reverse && *status != 1) { if (unlikely(cmp_extend_encoding(afl, h, SWAP64(pattern), SWAP64(repl), SWAP64(o_pattern), idx, orig_buf, buf, @@ -304,7 +405,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } - if (SHAPE_BYTES(h->shape) == 4 || *status == 2) { + if (SHAPE_BYTES(h->shape) >= 4 && *status != 1) { if (its_len >= 4 && *buf_32 == (u32)pattern && *o_buf_32 == (u32)o_pattern) { @@ -316,7 +417,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } // reverse encoding - if (do_reverse) { + if (do_reverse && *status != 1) { if (unlikely(cmp_extend_encoding(afl, h, SWAP32(pattern), SWAP32(repl), SWAP32(o_pattern), idx, orig_buf, buf, @@ -330,7 +431,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } - if (SHAPE_BYTES(h->shape) == 2 || *status == 2) { + if (SHAPE_BYTES(h->shape) >= 2 && *status != 1) { if (its_len >= 2 && *buf_16 == (u16)pattern && *o_buf_16 == (u16)o_pattern) { @@ -342,7 +443,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } // reverse encoding - if (do_reverse) { + if (do_reverse && *status != 1) { if (unlikely(cmp_extend_encoding(afl, h, SWAP16(pattern), SWAP16(repl), SWAP16(o_pattern), idx, orig_buf, buf, @@ -356,7 +457,7 @@ static u8 cmp_extend_encoding(afl_state_t *afl, struct cmp_header *h, } - if (SHAPE_BYTES(h->shape) == 1 || *status == 2) { + if (SHAPE_BYTES(h->shape) >= 1 && *status != 1) { if (its_len >= 1 && *buf_8 == (u8)pattern && *o_buf_8 == (u8)o_pattern) { @@ -482,6 +583,7 @@ static u8 cmp_fuzz(afl_state_t *afl, u32 key, u8 *orig_buf, u8 *buf, u32 len) { for (idx = 0; idx < len && fails < 8; ++idx) { + status = 0; if (unlikely(cmp_extend_encoding(afl, h, o->v0, o->v1, orig_o->v0, idx, orig_buf, buf, len, 1, &status))) { @@ -499,6 +601,7 @@ static u8 cmp_fuzz(afl_state_t *afl, u32 key, u8 *orig_buf, u8 *buf, u32 len) { } + status = 0; if (unlikely(cmp_extend_encoding(afl, h, o->v1, o->v0, orig_o->v1, idx, orig_buf, buf, len, 1, &status))) { @@ -570,14 +673,15 @@ static u8 rtn_extend_encoding(afl_state_t *afl, struct cmp_header *h, for (i = 0; i < its_len; ++i) { - if (pattern[idx + i] != buf[idx + i] || - o_pattern[idx + i] != orig_buf[idx + i] || *status == 1) { + if (pattern[i] != buf[idx + i] || o_pattern[i] != orig_buf[idx + i] || + *status == 1) { break; } - buf[idx + i] = repl[idx + i]; + buf[idx + i] = repl[i]; + if (unlikely(its_fuzz(afl, buf, len, status))) { return 1; } } diff --git a/src/afl-fuzz-run.c b/src/afl-fuzz-run.c index 2a1664e2..44d3c522 100644 --- a/src/afl-fuzz-run.c +++ b/src/afl-fuzz-run.c @@ -134,6 +134,8 @@ void write_to_testcase(afl_state_t *afl, void *mem, u32 len) { } +#define BUF_PARAMS(name) (void **)&afl->name##_buf, &afl->name##_size + /* The same, but with an adjustable gap. Used for trimming. */ static void write_with_gap(afl_state_t *afl, void *mem, u32 len, u32 skip_at, @@ -142,18 +144,81 @@ static void write_with_gap(afl_state_t *afl, void *mem, u32 len, u32 skip_at, s32 fd = afl->fsrv.out_fd; u32 tail_len = len - skip_at - skip_len; + /* + This memory is used to carry out the post_processing(if present) after copying + the testcase by removing the gaps. This can break though + */ + u8 *mem_trimmed = ck_maybe_grow(BUF_PARAMS(out_scratch), len - skip_len + 1); + + ssize_t new_size = len - skip_len; + void * new_mem = mem; + u8 * new_buf = NULL; + + bool post_process_skipped = true; + + if (unlikely(afl->custom_mutators_count)) { + + new_mem = mem_trimmed; + + LIST_FOREACH(&afl->custom_mutator_list, struct custom_mutator, { + + if (el->afl_custom_post_process) { + + // We copy into the mem_trimmed only if we actually have custom mutators + // *with* post_processing installed + + if (post_process_skipped) { + + if (skip_at) { memcpy(mem_trimmed, (u8 *)mem, skip_at); } + + if (tail_len) { + + memcpy(mem_trimmed + skip_at, (u8 *)mem + skip_at + skip_len, + tail_len); + + } + + post_process_skipped = false; + + } + + new_size = + el->afl_custom_post_process(el->data, new_mem, new_size, &new_buf); + + if (unlikely(!new_buf || (new_size <= 0))) { + + FATAL("Custom_post_process failed (ret: %lu)", + (long unsigned)new_size); + + } + + } + + new_mem = new_buf; + + }); + + } + if (afl->fsrv.shmem_fuzz) { - if (skip_at) { memcpy(afl->fsrv.shmem_fuzz, mem, skip_at); } + if (!post_process_skipped) { + + // If we did post_processing, copy directly from the new_buf bufer + + memcpy(afl->fsrv.shmem_fuzz, new_buf, new_size); + + } + + else { - if (tail_len) { + memcpy(afl->fsrv.shmem_fuzz, mem, skip_at); - memcpy(afl->fsrv.shmem_fuzz + skip_at, (u8 *)mem + skip_at + skip_len, - tail_len); + memcpy(afl->fsrv.shmem_fuzz, mem + skip_at + skip_len, tail_len); } - *afl->fsrv.shmem_fuzz_len = len - skip_len; + *afl->fsrv.shmem_fuzz_len = new_size; #ifdef _DEBUG if (afl->debug) { @@ -197,18 +262,21 @@ static void write_with_gap(afl_state_t *afl, void *mem, u32 len, u32 skip_at, } - if (skip_at) { ck_write(fd, mem, skip_at, afl->fsrv.out_file); } + if (!post_process_skipped) { - u8 *memu8 = mem; - if (tail_len) { + ck_write(fd, new_buf, new_size, afl->fsrv.out_file); - ck_write(fd, memu8 + skip_at + skip_len, tail_len, afl->fsrv.out_file); + } else { + + ck_write(fd, mem, skip_at, afl->fsrv.out_file); + + ck_write(fd, mem + skip_at + skip_len, tail_len, afl->fsrv.out_file); } if (!afl->fsrv.out_file) { - if (ftruncate(fd, len - skip_len)) { PFATAL("ftruncate() failed"); } + if (ftruncate(fd, new_size)) { PFATAL("ftruncate() failed"); } lseek(fd, 0, SEEK_SET); } else { @@ -219,6 +287,8 @@ static void write_with_gap(afl_state_t *afl, void *mem, u32 len, u32 skip_at, } +#undef BUF_PARAMS + /* Calibrate a new test case. This is done when processing the input directory to warn about flaky or otherwise problematic test cases early on; and when new paths are discovered to detect variable behavior and so on. */ @@ -612,6 +682,8 @@ void sync_fuzzers(afl_state_t *afl) { } + if (afl->foreign_sync_cnt) read_foreign_testcases(afl, 0); + } /* Trim all new test cases to save cycles when doing deterministic checks. The diff --git a/src/afl-fuzz-state.c b/src/afl-fuzz-state.c index e0e43f54..e2d62bc6 100644 --- a/src/afl-fuzz-state.c +++ b/src/afl-fuzz-state.c @@ -94,6 +94,7 @@ void afl_state_init(afl_state_t *afl, uint32_t map_size) { afl->havoc_div = 1; /* Cycle count divisor for havoc */ afl->stage_name = "init"; /* Name of the current fuzz stage */ afl->splicing_with = -1; /* Splicing with which test case? */ + afl->cpu_to_bind = -1; #ifdef HAVE_AFFINITY afl->cpu_aff = -1; /* Selected CPU core */ @@ -293,6 +294,20 @@ void read_afl_environment(afl_state_t *afl, char **envp) { afl->afl_env.afl_autoresume = get_afl_env(afl_environment_variables[i]) ? 1 : 0; + } else if (!strncmp(env, "AFL_CYCLE_SCHEDULES", + + afl_environment_variable_len)) { + + afl->cycle_schedules = afl->afl_env.afl_cycle_schedules = + get_afl_env(afl_environment_variables[i]) ? 1 : 0; + + } else if (!strncmp(env, "AFL_EXPAND_HAVOC_NOW", + + afl_environment_variable_len)) { + + afl->expand_havoc = afl->afl_env.afl_expand_havoc = + get_afl_env(afl_environment_variables[i]) ? 1 : 0; + } else if (!strncmp(env, "AFL_CAL_FAST", afl_environment_variable_len)) { @@ -405,6 +420,7 @@ void afl_state_deinit(afl_state_t *afl) { if (afl->pass_stats) { ck_free(afl->pass_stats); } if (afl->orig_cmp_map) { ck_free(afl->orig_cmp_map); } + if (afl->queue_buf) { free(afl->queue_buf); } if (afl->out_buf) { free(afl->out_buf); } if (afl->out_scratch_buf) { free(afl->out_scratch_buf); } if (afl->eff_buf) { free(afl->eff_buf); } diff --git a/src/afl-fuzz-stats.c b/src/afl-fuzz-stats.c index fc93011b..7b30b5ea 100644 --- a/src/afl-fuzz-stats.c +++ b/src/afl-fuzz-stats.c @@ -39,7 +39,7 @@ void write_stats_file(afl_state_t *afl, double bitmap_cvg, double stability, u8 fn[PATH_MAX]; s32 fd; FILE * f; - uint32_t t_bytes = count_non_255_bytes(afl, afl->virgin_bits); + u32 t_bytes = count_non_255_bytes(afl, afl->virgin_bits); snprintf(fn, PATH_MAX, "%s/fuzzer_stats", afl->out_dir); @@ -67,89 +67,102 @@ void write_stats_file(afl_state_t *afl, double bitmap_cvg, double stability, } + if ((unlikely(!afl->last_avg_exec_update || + cur_time - afl->last_avg_exec_update >= 60000))) { + + afl->last_avg_execs_saved = + (float)(1000 * (afl->fsrv.total_execs - afl->last_avg_execs)) / + (float)(cur_time - afl->last_avg_exec_update); + afl->last_avg_execs = afl->fsrv.total_execs; + afl->last_avg_exec_update = cur_time; + + } + #ifndef __HAIKU__ if (getrusage(RUSAGE_CHILDREN, &rus)) { rus.ru_maxrss = 0; } #endif - fprintf( - f, - "start_time : %llu\n" - "last_update : %llu\n" - "run_time : %llu\n" - "fuzzer_pid : %u\n" - "cycles_done : %llu\n" - "cycles_wo_finds : %llu\n" - "execs_done : %llu\n" - "execs_per_sec : %0.02f\n" - // "real_execs_per_sec: %0.02f\n" // damn the name is too long - "paths_total : %u\n" - "paths_favored : %u\n" - "paths_found : %u\n" - "paths_imported : %u\n" - "max_depth : %u\n" - "cur_path : %u\n" /* Must match find_start_position() */ - "pending_favs : %u\n" - "pending_total : %u\n" - "variable_paths : %u\n" - "stability : %0.02f%%\n" - "bitmap_cvg : %0.02f%%\n" - "unique_crashes : %llu\n" - "unique_hangs : %llu\n" - "last_path : %llu\n" - "last_crash : %llu\n" - "last_hang : %llu\n" - "execs_since_crash : %llu\n" - "exec_timeout : %u\n" - "slowest_exec_ms : %u\n" - "peak_rss_mb : %lu\n" - "cpu_affinity : %d\n" - "edges_found : %u\n" - "var_byte_count : %u\n" - "afl_banner : %s\n" - "afl_version : " VERSION - "\n" - "target_mode : %s%s%s%s%s%s%s%s%s\n" - "command_line : %s\n", - afl->start_time / 1000, cur_time / 1000, - (cur_time - afl->start_time) / 1000, (u32)getpid(), - afl->queue_cycle ? (afl->queue_cycle - 1) : 0, afl->cycles_wo_finds, - afl->fsrv.total_execs, - afl->fsrv.total_execs / - ((double)(get_cur_time() - afl->start_time) / 1000), - afl->queued_paths, afl->queued_favored, afl->queued_discovered, - afl->queued_imported, afl->max_depth, afl->current_entry, - afl->pending_favored, afl->pending_not_fuzzed, afl->queued_variable, - stability, bitmap_cvg, afl->unique_crashes, afl->unique_hangs, - afl->last_path_time / 1000, afl->last_crash_time / 1000, - afl->last_hang_time / 1000, afl->fsrv.total_execs - afl->last_crash_execs, - afl->fsrv.exec_tmout, afl->slowest_exec_ms, + fprintf(f, + "start_time : %llu\n" + "last_update : %llu\n" + "run_time : %llu\n" + "fuzzer_pid : %u\n" + "cycles_done : %llu\n" + "cycles_wo_finds : %llu\n" + "execs_done : %llu\n" + "execs_per_sec : %0.02f\n" + "execs_ps_last_min : %0.02f\n" + "paths_total : %u\n" + "paths_favored : %u\n" + "paths_found : %u\n" + "paths_imported : %u\n" + "max_depth : %u\n" + "cur_path : %u\n" /* Must match find_start_position() */ + "pending_favs : %u\n" + "pending_total : %u\n" + "variable_paths : %u\n" + "stability : %0.02f%%\n" + "bitmap_cvg : %0.02f%%\n" + "unique_crashes : %llu\n" + "unique_hangs : %llu\n" + "last_path : %llu\n" + "last_crash : %llu\n" + "last_hang : %llu\n" + "execs_since_crash : %llu\n" + "exec_timeout : %u\n" + "slowest_exec_ms : %u\n" + "peak_rss_mb : %lu\n" + "cpu_affinity : %d\n" + "edges_found : %u\n" + "var_byte_count : %u\n" + "havoc_expansion : %u\n" + "afl_banner : %s\n" + "afl_version : " VERSION + "\n" + "target_mode : %s%s%s%s%s%s%s%s%s\n" + "command_line : %s\n", + afl->start_time / 1000, cur_time / 1000, + (cur_time - afl->start_time) / 1000, (u32)getpid(), + afl->queue_cycle ? (afl->queue_cycle - 1) : 0, afl->cycles_wo_finds, + afl->fsrv.total_execs, + afl->fsrv.total_execs / + ((double)(get_cur_time() - afl->start_time) / 1000), + afl->last_avg_execs_saved, afl->queued_paths, afl->queued_favored, + afl->queued_discovered, afl->queued_imported, afl->max_depth, + afl->current_entry, afl->pending_favored, afl->pending_not_fuzzed, + afl->queued_variable, stability, bitmap_cvg, afl->unique_crashes, + afl->unique_hangs, afl->last_path_time / 1000, + afl->last_crash_time / 1000, afl->last_hang_time / 1000, + afl->fsrv.total_execs - afl->last_crash_execs, afl->fsrv.exec_tmout, + afl->slowest_exec_ms, #ifndef __HAIKU__ #ifdef __APPLE__ - (unsigned long int)(rus.ru_maxrss >> 20), + (unsigned long int)(rus.ru_maxrss >> 20), #else - (unsigned long int)(rus.ru_maxrss >> 10), + (unsigned long int)(rus.ru_maxrss >> 10), #endif #else - -1UL, + -1UL, #endif #ifdef HAVE_AFFINITY - afl->cpu_aff, + afl->cpu_aff, #else - -1, + -1, #endif - t_bytes, afl->var_byte_count, afl->use_banner, - afl->unicorn_mode ? "unicorn" : "", afl->fsrv.qemu_mode ? "qemu " : "", - afl->non_instrumented_mode ? " non_instrumented " : "", - afl->no_forkserver ? "no_fsrv " : "", afl->crash_mode ? "crash " : "", - afl->persistent_mode ? "persistent " : "", - afl->shmem_testcase_mode ? "shmem_testcase " : "", - afl->deferred_mode ? "deferred " : "", - (afl->unicorn_mode || afl->fsrv.qemu_mode || afl->non_instrumented_mode || - afl->no_forkserver || afl->crash_mode || afl->persistent_mode || - afl->deferred_mode) - ? "" - : "default", - afl->orig_cmdline); + t_bytes, afl->var_byte_count, afl->expand_havoc, afl->use_banner, + afl->unicorn_mode ? "unicorn" : "", + afl->fsrv.qemu_mode ? "qemu " : "", + afl->non_instrumented_mode ? " non_instrumented " : "", + afl->no_forkserver ? "no_fsrv " : "", afl->crash_mode ? "crash " : "", + afl->persistent_mode ? "persistent " : "", + afl->shmem_testcase_mode ? "shmem_testcase " : "", + afl->deferred_mode ? "deferred " : "", + (afl->unicorn_mode || afl->fsrv.qemu_mode || + afl->non_instrumented_mode || afl->no_forkserver || + afl->crash_mode || afl->persistent_mode || afl->deferred_mode) + ? "" + : "default", + afl->orig_cmdline); /* ignore errors */ if (afl->debug) { diff --git a/src/afl-fuzz.c b/src/afl-fuzz.c index e4e2669c..da30797c 100644 --- a/src/afl-fuzz.c +++ b/src/afl-fuzz.c @@ -42,19 +42,21 @@ static void at_exit() { int i; char *list[4] = {SHM_ENV_VAR, SHM_FUZZ_ENV_VAR, CMPLOG_SHM_ENV_VAR, NULL}; - char *ptr = getenv("__AFL_TARGET_PID1"); + char *ptr; + ptr = getenv(CPU_AFFINITY_ENV_VAR); + if (ptr && *ptr) unlink(ptr); + + ptr = getenv("__AFL_TARGET_PID1"); if (ptr && *ptr && (i = atoi(ptr)) > 0) kill(i, SIGKILL); ptr = getenv("__AFL_TARGET_PID2"); - if (ptr && *ptr && (i = atoi(ptr)) > 0) kill(i, SIGKILL); i = 0; while (list[i] != NULL) { ptr = getenv(list[i]); - if (ptr && *ptr) { #ifdef USEMMAP @@ -129,18 +131,23 @@ static void usage(afl_state_t *afl, u8 *argv0, int more_help) { "executions.\n\n" "Other stuff:\n" - " -T text - text banner to show on the screen\n" " -M/-S id - distributed mode (see docs/parallel_fuzzing.md)\n" " use -D to force -S secondary to perform deterministic " "fuzzing\n" + " -F path - sync to a foreign fuzzer queue directory (requires " + "-M, can\n" + " be specified up to %u times)\n" + " -T text - text banner to show on the screen\n" " -I command - execute this command/script when a new crash is " "found\n" //" -B bitmap.txt - mutate a specific test case, use the out/fuzz_bitmap //" "file\n" " -C - crash exploration mode (the peruvian rabbit thing)\n" + " -b cpu_id - bind the fuzzing process to the specified CPU core " + "(0-...)\n" " -e ext - file extension for the fuzz test input file (if " "needed)\n\n", - argv0, EXEC_TIMEOUT, MEM_LIMIT); + argv0, EXEC_TIMEOUT, MEM_LIMIT, FOREIGN_SYNCS_MAX); if (more_help > 1) { @@ -156,11 +163,13 @@ static void usage(afl_state_t *afl, u8 *argv0, int more_help) { "AFL_BENCH_UNTIL_CRASH: exit soon when the first crashing input has been found\n" "AFL_CUSTOM_MUTATOR_LIBRARY: lib with afl_custom_fuzz() to mutate inputs\n" "AFL_CUSTOM_MUTATOR_ONLY: avoid AFL++'s internal mutators\n" + "AFL_CYCLE_SCHEDULES: after completing a cycle, switch to a different -p schedule\n" "AFL_DEBUG: extra debugging output for Python mode trimming\n" "AFL_DEBUG_CHILD_OUTPUT: do not suppress stdout/stderr from target\n" "AFL_DISABLE_TRIM: disable the trimming of test cases\n" "AFL_DUMB_FORKSRV: use fork server without feedback from target\n" "AFL_EXIT_WHEN_DONE: exit when all inputs are run and no new finds are found\n" + "AFL_EXPAND_HAVOC_NOW: immediately enable expand havoc mode (default: after 60 minutes and a cycle without finds)\n" "AFL_FAST_CAL: limit the calibration stage to three cycles for speedup\n" "AFL_FORCE_UI: force showing the status screen (for virtual consoles)\n" "AFL_HANG_TMOUT: override timeout value (in milliseconds)\n" @@ -264,9 +273,11 @@ int main(int argc, char **argv_orig, char **envp) { gettimeofday(&tv, &tz); rand_set_seed(afl, tv.tv_sec ^ tv.tv_usec ^ getpid()); - while ((opt = getopt(argc, argv, - "+c:i:I:o:f:m:t:T:dDnCB:S:M:x:QNUWe:p:s:V:E:L:hRP:")) > - 0) { + afl->shmem_testcase_mode = 1; // we always try to perform shmem fuzzing + + while ((opt = getopt( + argc, argv, + "+b:c:i:I:o:f:F:m:t:T:dDnCB:S:M:x:QNUWe:p:s:V:E:L:hRP:")) > 0) { switch (opt) { @@ -274,6 +285,17 @@ int main(int argc, char **argv_orig, char **envp) { afl->infoexec = optarg; break; + case 'b': { /* bind CPU core */ + + if (afl->cpu_to_bind != -1) FATAL("Multiple -b options not supported"); + + if (sscanf(optarg, "%u", &afl->cpu_to_bind) < 0 || optarg[0] == '-') + FATAL("Bad syntax used for -b"); + + break; + + } + case 'c': { afl->shm.cmplog_mode = 1; @@ -399,6 +421,19 @@ int main(int argc, char **argv_orig, char **envp) { afl->use_splicing = 1; break; + case 'F': /* foreign sync dir */ + + if (!afl->is_main_node) + FATAL( + "Option -F can only be specified after the -M option for the " + "main fuzzer of a fuzzing campaign"); + if (afl->foreign_sync_cnt >= FOREIGN_SYNCS_MAX) + FATAL("Maximum %u entried of -F option can be specified", + FOREIGN_SYNCS_MAX); + afl->foreign_syncs[afl->foreign_sync_cnt].dir = optarg; + afl->foreign_sync_cnt++; + break; + case 'f': /* target file */ if (afl->fsrv.out_file) { FATAL("Multiple -f options not supported"); } @@ -561,7 +596,6 @@ int main(int argc, char **argv_orig, char **envp) { if (afl->fsrv.qemu_mode) { FATAL("Multiple -Q options not supported"); } afl->fsrv.qemu_mode = 1; - afl->shmem_testcase_mode = 1; if (!mem_limit_given) { afl->fsrv.mem_limit = MEM_LIMIT_QEMU; } @@ -578,7 +612,6 @@ int main(int argc, char **argv_orig, char **envp) { if (afl->unicorn_mode) { FATAL("Multiple -U options not supported"); } afl->unicorn_mode = 1; - afl->shmem_testcase_mode = 1; if (!mem_limit_given) { afl->fsrv.mem_limit = MEM_LIMIT_UNICORN; } @@ -589,7 +622,6 @@ int main(int argc, char **argv_orig, char **envp) { if (afl->use_wine) { FATAL("Multiple -W options not supported"); } afl->fsrv.qemu_mode = 1; afl->use_wine = 1; - afl->shmem_testcase_mode = 1; if (!mem_limit_given) { afl->fsrv.mem_limit = 0; } @@ -899,6 +931,7 @@ int main(int argc, char **argv_orig, char **envp) { if (get_afl_env("AFL_NO_ARITH")) { afl->no_arith = 1; } if (get_afl_env("AFL_SHUFFLE_QUEUE")) { afl->shuffle_queue = 1; } if (get_afl_env("AFL_FAST_CAL")) { afl->fast_cal = 1; } + if (get_afl_env("AFL_EXPAND_HAVOC_NOW")) { afl->expand_havoc = 1; } if (afl->afl_env.afl_autoresume) { @@ -1011,16 +1044,23 @@ int main(int argc, char **argv_orig, char **envp) { } + check_crash_handling(); + check_cpu_governor(afl); + get_core_count(afl); + atexit(at_exit); + + setup_dirs_fds(afl); + #ifdef HAVE_AFFINITY bind_to_free_cpu(afl); #endif /* HAVE_AFFINITY */ - check_crash_handling(); - check_cpu_governor(afl); - - atexit(at_exit); + #ifdef __HAIKU__ + /* Prioritizes performance over power saving */ + set_scheduler_mode(SCHEDULER_MODE_LOW_LATENCY); + #endif afl->fsrv.trace_bits = afl_shm_init(&afl->shm, afl->fsrv.map_size, afl->non_instrumented_mode); @@ -1038,20 +1078,26 @@ int main(int argc, char **argv_orig, char **envp) { } - setup_dirs_fds(afl); - if (afl->is_secondary_node && check_main_node_exists(afl) == 0) { WARNF("no -M main node found. You need to run one main instance!"); - sleep(5); + sleep(3); } + #ifdef RAND_TEST_VALUES + u32 counter; + for (counter = 0; counter < 100000; counter++) + printf("DEBUG: rand %06d is %u\n", counter, rand_below(afl, 65536)); + #endif + setup_custom_mutators(afl); setup_cmdline_file(afl, argv + optind); read_testcases(afl); + // read_foreign_testcases(afl, 1); for the moment dont do this + load_auto(afl); pivot_inputs(afl); @@ -1209,6 +1255,7 @@ int main(int argc, char **argv_orig, char **envp) { } + // (void)nice(-20); // does not improve the speed // real start time, we reset, so this works correctly with -V afl->start_time = get_cur_time(); @@ -1245,11 +1292,43 @@ int main(int argc, char **argv_orig, char **envp) { /* If we had a full queue cycle with no new finds, try recombination strategies next. */ - if (afl->queued_paths == prev_queued) { + if (afl->queued_paths == prev_queued && + (get_cur_time() - afl->start_time) >= 3600) { if (afl->use_splicing) { ++afl->cycles_wo_finds; + switch (afl->expand_havoc) { + + case 0: + afl->expand_havoc = 1; + break; + case 1: + if (afl->limit_time_sig == 0 && !afl->custom_only && + !afl->python_only) { + + afl->limit_time_sig = -1; + afl->limit_time_puppet = 0; + + } + + afl->expand_havoc = 2; + break; + case 2: + // afl->cycle_schedules = 1; + afl->expand_havoc = 3; + break; + case 3: + // nothing else currently + break; + + } + + if (afl->expand_havoc) { + + } else + + afl->expand_havoc = 1; } else { @@ -1263,6 +1342,53 @@ int main(int argc, char **argv_orig, char **envp) { } + if (afl->cycle_schedules) { + + /* we cannot mix non-AFLfast schedules with others */ + + switch (afl->schedule) { + + case EXPLORE: + afl->schedule = EXPLOIT; + break; + case EXPLOIT: + afl->schedule = MMOPT; + break; + case MMOPT: + afl->schedule = SEEK; + break; + case SEEK: + afl->schedule = EXPLORE; + break; + case FAST: + afl->schedule = COE; + break; + case COE: + afl->schedule = LIN; + break; + case LIN: + afl->schedule = QUAD; + break; + case QUAD: + afl->schedule = RARE; + break; + case RARE: + afl->schedule = FAST; + break; + + } + + struct queue_entry *q = afl->queue; + // we must recalculate the scores of all queue entries + while (q) { + + update_bitmap_score(afl, q); + q = q->next; + + } + + } + prev_queued = afl->queued_paths; if (afl->sync_id && afl->queue_cycle == 1 && diff --git a/src/afl-gcc.c b/src/afl-gcc.c index b8ff7e77..22e6be8e 100644 --- a/src/afl-gcc.c +++ b/src/afl-gcc.c @@ -132,6 +132,9 @@ static void edit_params(u32 argc, char **argv) { name = argv[0]; + /* This should never happen but fixes a scan-build warning */ + if (!name) { FATAL("Empty argv set"); } + } else { ++name; @@ -465,7 +468,7 @@ int main(int argc, char **argv) { u32 map_size = atoi(ptr); if (map_size != MAP_SIZE) { - FATAL("AFL_MAP_SIZE is not supported by afl-gcc"); + WARNF("AFL_MAP_SIZE is not supported by afl-gcc"); } diff --git a/src/afl-showmap.c b/src/afl-showmap.c index 883398ff..71e975a1 100644 --- a/src/afl-showmap.c +++ b/src/afl-showmap.c @@ -456,15 +456,38 @@ static void set_up_environment(afl_forkserver_t *fsrv) { setenv("ASAN_OPTIONS", "abort_on_error=1:" "detect_leaks=0:" + "allocator_may_return_null=1:" "symbolize=0:" - "allocator_may_return_null=1", + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", + 0); + + setenv("UBSAN_OPTIONS", + "halt_on_error=1:" + "abort_on_error=1:" + "malloc_context_size=0:" + "allocator_may_return_null=1:" + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); setenv("MSAN_OPTIONS", "exit_code=" STRINGIFY(MSAN_ERROR) ":" - "symbolize=0:" "abort_on_error=1:" + "msan_track_origins=0" "allocator_may_return_null=1:" - "msan_track_origins=0", 0); + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); if (get_afl_env("AFL_PRELOAD")) { diff --git a/src/afl-tmin.c b/src/afl-tmin.c index 2db1eae7..68fcdd14 100644 --- a/src/afl-tmin.c +++ b/src/afl-tmin.c @@ -701,15 +701,38 @@ static void set_up_environment(afl_forkserver_t *fsrv) { setenv("ASAN_OPTIONS", "abort_on_error=1:" "detect_leaks=0:" + "allocator_may_return_null=1:" "symbolize=0:" - "allocator_may_return_null=1", + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", + 0); + + setenv("UBSAN_OPTIONS", + "halt_on_error=1:" + "abort_on_error=1:" + "malloc_context_size=0:" + "allocator_may_return_null=1:" + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); setenv("MSAN_OPTIONS", "exit_code=" STRINGIFY(MSAN_ERROR) ":" - "symbolize=0:" "abort_on_error=1:" + "msan_track_origins=0" "allocator_may_return_null=1:" - "msan_track_origins=0", 0); + "symbolize=0:" + "handle_segv=0:" + "handle_sigbus=0:" + "handle_abort=0:" + "handle_sigfpe=0:" + "handle_sigill=0", 0); if (get_afl_env("AFL_PRELOAD")) { diff --git a/test/test-cmplog.c b/test/test-cmplog.c new file mode 100644 index 00000000..b077e3ab --- /dev/null +++ b/test/test-cmplog.c @@ -0,0 +1,23 @@ +#include <stdio.h> +#include <string.h> +#include <stdarg.h> +#include <stdlib.h> +#include <stdint.h> +#include <unistd.h> +int main(int argc, char *argv[]) { + + char buf[1024]; + ssize_t i; + if ((i = read(0, buf, sizeof(buf) - 1)) < 24) return 0; + buf[i] = 0; + if (buf[0] != 'A') return 0; + if (buf[1] != 'B') return 0; + if (buf[2] != 'C') return 0; + if (buf[3] != 'D') return 0; + if (memcmp(buf + 4, "1234", 4) || memcmp(buf + 8, "EFGH", 4)) return 0; + if (strncmp(buf + 12, "IJKL", 4) == 0 && strcmp(buf + 16, "DEADBEEF") == 0) + abort(); + return 0; + +} + diff --git a/test/test-floatingpoint.c b/test/test-floatingpoint.c new file mode 100644 index 00000000..febfae05 --- /dev/null +++ b/test/test-floatingpoint.c @@ -0,0 +1,33 @@ +#include <stdlib.h> +#include <unistd.h> +#include <limits.h> +#include <stdint.h> + +__AFL_FUZZ_INIT(); + +int main(void) { + + ssize_t bytes_read; + + __AFL_INIT(); + float *magic = (float *)__AFL_FUZZ_TESTCASE_BUF; + + while (__AFL_LOOP(INT_MAX)) { + + int len = __AFL_FUZZ_TESTCASE_LEN; + if (len < sizeof(float)) return 1; + + /* 15 + 1/2 = 15.5 */ + /* 15 + 1/2 + 1/8 = 15.625 */ + /* 15 + 1/2 + 1/8 + 1/32 = 15.65625 */ + /* 15 + 1/2 + 1/8 + 1/32 + 1/128 = 15.6640625 */ + if ((*magic >= 15.0 + 0.5 + 0.125 + 0.03125) && + (*magic <= 15.0 + 0.5 + 0.125 + 0.03125 + 0.0078125)) + abort(); + + } + + return 0; + +} + diff --git a/test/test-fp_cases.c b/test/test-fp_cases.c new file mode 100644 index 00000000..b0f792bc --- /dev/null +++ b/test/test-fp_cases.c @@ -0,0 +1,213 @@ +/* test cases for floating point comparison transformations + * compile with -DFLOAT_TYPE=float + * or -DFLOAT_TYPE=double + * or -DFLOAT_TYPE="long double" + */ + +#include <assert.h> + +int main() { + + volatile FLOAT_TYPE a, b; + /* different values */ + a = -2.1; + b = -2; /* signs equal, exp equal, mantissa > */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = 1.8; + b = 2.1; /* signs equal, exp differ, mantissa > */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = 2; + b = 2.1; /* signs equal, exp equal, mantissa < */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -2; + b = -1.8; /* signs equal, exp differ, mantissa < */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -1; + b = 1; /* signs differ, exp equal, mantissa equal */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -1; + b = 0; /* signs differ, exp differ, mantissa equal */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -2; + b = 2.8; /* signs differ, exp equal, mantissa < */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -2; + b = 1.8; /* signs differ, exp differ, mantissa < */ + assert((a < b)); + assert((a <= b)); + assert(!(a > b)); + assert(!(a >= b)); + assert((a != b)); + assert(!(a == b)); + + a = -2; + b = -2.1; /* signs equal, exp equal, mantissa > */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 2.1; + b = 1.8; /* signs equal, exp differ, mantissa > */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 2.1; + b = 2; /* signs equal, exp equal, mantissa < */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = -1.8; + b = -2; /* signs equal, exp differ, mantissa < */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 1; + b = -1; /* signs differ, exp equal, mantissa equal */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 0; + b = -1; /* signs differ, exp differ, mantissa equal */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 2.8; + b = -2; /* signs differ, exp equal, mantissa < */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + a = 1.8; + b = -2; /* signs differ, exp differ, mantissa < */ + assert((a > b)); + assert((a >= b)); + assert(!(a < b)); + assert(!(a <= b)); + assert((a != b)); + assert(!(a == b)); + + /* equal values */ + a = 0; + b = 0; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + + a = -0; + b = 0; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + + a = 1; + b = 1; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + + a = 0.5; + b = 0.5; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + + a = -1; + b = -1; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + + a = -0.5; + b = -0.5; + assert(!(a < b)); + assert((a <= b)); + assert(!(a > b)); + assert((a >= b)); + assert(!(a != b)); + assert((a == b)); + +} + diff --git a/test/test.sh b/test/test.sh index 90920215..46843d4a 100755 --- a/test/test.sh +++ b/test/test.sh @@ -7,6 +7,10 @@ test -z "" 2>/dev/null || { echo Error: test command not found ; exit 1 ; } GREP=`type grep > /dev/null 2>&1 && echo OK` test "$GREP" = OK || { echo Error: grep command not found ; exit 1 ; } echo foobar | grep -qE 'asd|oob' 2>/dev/null || { echo Error: grep command does not support -q and/or -E option ; exit 1 ; } +test -e ./test.sh || cd $(dirname $0) || exit 1 +test -e ./test.sh || { echo Error: you must be in the test/ directory ; exit 1 ; } +export AFL_PATH=`pwd`/.. + echo 1 > test.1 echo 1 > test.2 OK=OK @@ -202,7 +206,7 @@ test "$SYS" = "i686" -o "$SYS" = "x86_64" -o "$SYS" = "amd64" -o "$SYS" = "i86pc rm -f in2/in* export AFL_QUIET=1 if command -v bash >/dev/null ; then { - AFL_PATH=`pwd`/.. ../afl-cmin.bash -m ${MEM_LIMIT} -i in -o in2 -- ./test-instr.plain >/dev/null + ../afl-cmin.bash -m ${MEM_LIMIT} -i in -o in2 -- ./test-instr.plain >/dev/null CNT=`ls in2/* 2>/dev/null | wc -l` case "$CNT" in *2) $ECHO "$GREEN[+] afl-cmin.bash correctly minimized the number of testcases" ;; @@ -326,7 +330,7 @@ test -e ../afl-clang-fast -a -e ../split-switches-pass.so && { rm -f in2/in* export AFL_QUIET=1 if type bash >/dev/null ; then { - AFL_PATH=`pwd`/.. ../afl-cmin.bash -m ${MEM_LIMIT} -i in -o in2 -- ./test-instr.plain >/dev/null + ../afl-cmin.bash -m ${MEM_LIMIT} -i in -o in2 -- ./test-instr.plain >/dev/null CNT=`ls in2/* 2>/dev/null | wc -l` case "$CNT" in *2) $ECHO "$GREEN[+] afl-cmin.bash correctly minimized the number of testcases" ;; @@ -372,8 +376,7 @@ test -e ../afl-clang-fast -a -e ../split-switches-pass.so && { $ECHO "$YELLOW[-] llvm_mode InsTrim not compiled, cannot test" INCOMPLETE=1 } - AFL_LLVM_INSTRUMENT=AFL - AFL_DEBUG=1 AFL_LLVM_LAF_SPLIT_SWITCHES=1 AFL_LLVM_LAF_TRANSFORM_COMPARES=1 AFL_LLVM_LAF_SPLIT_COMPARES=1 ../afl-clang-fast -o test-compcov.compcov test-compcov.c > test.out 2>&1 + AFL_LLVM_INSTRUMENT=AFL AFL_DEBUG=1 AFL_LLVM_LAF_SPLIT_SWITCHES=1 AFL_LLVM_LAF_TRANSFORM_COMPARES=1 AFL_LLVM_LAF_SPLIT_COMPARES=1 ../afl-clang-fast -o test-compcov.compcov test-compcov.c > test.out 2>&1 test -e test-compcov.compcov && test_compcov_binary_functionality ./test-compcov.compcov && { grep --binary-files=text -Eq " [ 123][0-9][0-9] location| [3-9][0-9] location" test.out && { $ECHO "$GREEN[+] llvm_mode laf-intel/compcov feature works correctly" @@ -386,6 +389,26 @@ test -e ../afl-clang-fast -a -e ../split-switches-pass.so && { CODE=1 } rm -f test-compcov.compcov test.out + AFL_LLVM_INSTRUMENT=AFL AFL_LLVM_LAF_SPLIT_FLOATS=1 ../afl-clang-fast -o test-floatingpoint test-floatingpoint.c >errors 2>&1 + test -e test-floatingpoint && { + mkdir -p in + echo ZZZZ > in/in + $ECHO "$GREY[*] running afl-fuzz with floating point splitting, this will take max. 30 seconds" + { + AFL_BENCH_UNTIL_CRASH=1 AFL_NO_UI=1 ../afl-fuzz -s 1 -V30 -m ${MEM_LIMIT} -i in -o out -- ./test-floatingpoint >>errors 2>&1 + } >>errors 2>&1 + test -n "$( ls out/crashes/id:* 2>/dev/null )" && { + $ECHO "$GREEN[+] llvm_mode laf-intel floatingpoint splitting feature works correctly" + } || { + cat errors + $ECHO "$RED[!] llvm_mode laf-intel floatingpoint splitting feature failed" + CODE=1 + } + } || { + $ECHO "$RED[!] llvm_mode laf-intel floatingpoint splitting feature compilation failed" + CODE=1 + } + rm -f test-floatingpoint test.out in/in echo foobar.c > instrumentlist.txt AFL_DEBUG=1 AFL_LLVM_INSTRUMENT_FILE=instrumentlist.txt ../afl-clang-fast -o test-compcov test-compcov.c > test.out 2>&1 test -e test-compcov && test_compcov_binary_functionality ./test-compcov && { @@ -400,6 +423,28 @@ test -e ../afl-clang-fast -a -e ../split-switches-pass.so && { CODE=1 } rm -f test-compcov test.out instrumentlist.txt + AFL_LLVM_CMPLOG=1 ../afl-clang-fast -o test-cmplog test-cmplog.c > /dev/null 2>&1 + test -e test-cmplog && { + $ECHO "$GREY[*] running afl-fuzz for llvm_mode cmplog, this will take approx 10 seconds" + { + mkdir -p in + echo 0000000000000000000000000 > in/in + ../afl-fuzz -m none -V10 -i in -o out -c./test-cmplog -- ./test-cmplog >>errors 2>&1 + } >>errors 2>&1 + test -n "$( ls out/crashes/id:000000* 2>/dev/null )" && { + $ECHO "$GREEN[+] afl-fuzz is working correctly with llvm_mode cmplog" + } || { + echo CUT------------------------------------------------------------------CUT + cat errors + echo CUT------------------------------------------------------------------CUT + $ECHO "$RED[!] afl-fuzz is not working correctly with llvm_mode cmplog" + CODE=1 + } + } || { + $ECHO "$YELLOW[-] we cannot test llvm_mode cmplog because it is not present" + INCOMPLETE=1 + } + rm -rf errors test-cmplog in ../afl-clang-fast -o test-persistent ../examples/persistent_demo/persistent_demo.c > /dev/null 2>&1 test -e test-persistent && { echo foo | ../afl-showmap -m ${MEM_LIMIT} -o /dev/null -q -r ./test-persistent && { |
