From 0ca6df6f09243109a14ab080d75be4fff8a5d836 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Tue, 2 Jul 2019 11:51:09 +0200 Subject: typo fix --- docs/README | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'docs/README') diff --git a/docs/README b/docs/README index ca38223d..f9734528 100644 --- a/docs/README +++ b/docs/README @@ -21,7 +21,7 @@ american fuzzy lop plus plus https://github.com/andreafioraldi/afl and got the community patches applied to it. - C. Hoellers afl-fuzz Python mutator module and llvm_mode whitelist support + C. Hollers afl-fuzz Python mutator module and llvm_mode whitelist support was added too (https://github.com/choller/afl) So all in all this is the best-of AFL that is currently out there :-) -- cgit 1.4.1 From f7d9019b8c6d4587b4874791636a9933f4b1a63a Mon Sep 17 00:00:00 2001 From: van Hauser Date: Wed, 10 Jul 2019 16:14:30 +0200 Subject: Readme updates --- docs/README | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) (limited to 'docs/README') diff --git a/docs/README b/docs/README index f9734528..3f19d328 100644 --- a/docs/README +++ b/docs/README @@ -2,28 +2,28 @@ american fuzzy lop plus plus ============================ - Written by Michal Zalewski + Originally written by Michal Zalewski Repository: https://github.com/vanhauser-thc/AFLplusplus - afl++ is maintained by Marc Heuse and Heiko Eissfeldt - as there have been no updates to afl since - November 2017. + afl++ is maintained by Marc Heuse , Heiko Eissfeldt + and Andrea Fioraldi as there have been no + updates to afl since November 2017. - This version has several bug fixes, new features and speed enhancements - based on community patches from https://github.com/vanhauser-thc/afl-patches - To see the list of which patches have been applied, see the PATCHES file. + Many improvements were made, e.g. more performant llvm_mode, supporting + llvm up to version 8, Qemu 3.1, more speed for Qemu, etc. Additionally AFLfast's power schedules by Marcel Boehme from - github.com/mboehme/aflfast have been incorporated. - - Plus it was upgraded to qemu 3.1 from 2.1 with the work of - https://github.com/andreafioraldi/afl and got the community patches applied - to it. + https://github.com/mboehme/aflfast have been incorporated. C. Hollers afl-fuzz Python mutator module and llvm_mode whitelist support was added too (https://github.com/choller/afl) + The newest additions is the excellent MOpt mutator from + https://github.com/puppet-meteor/MOpt-AFL + + A more thorough list is available in the PATCHES file. + So all in all this is the best-of AFL that is currently out there :-) -- cgit 1.4.1 From e66402485342088e6fcaecfe2abbba291a48bda5 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Sun, 14 Jul 2019 10:50:13 +0200 Subject: whitelist features works now --- docs/PATCHES | 1 + docs/README | 6 +++++- llvm_mode/LLVMInsTrim.so.cc | 19 ++++++++++++++----- llvm_mode/README.llvm | 10 ++++++---- 4 files changed, 26 insertions(+), 10 deletions(-) (limited to 'docs/README') diff --git a/docs/PATCHES b/docs/PATCHES index 06da053e..f6ca9284 100644 --- a/docs/PATCHES +++ b/docs/PATCHES @@ -17,6 +17,7 @@ afl-qemu-optimize-entrypoint.diff by mh(at)mh-sec(dot)de afl-qemu-speed.diff by abiondo on github afl-qemu-optimize-map.diff by mh(at)mh-sec(dot)de ++ instrim (https://github.com/csienslab/instrim) was integrated + MOpt (github.com/puppet-meteor/MOpt-AFL) was imported + AFLfast additions (github.com/mboehme/aflfast) were incorporated. + Qemu 3.1 upgrade with enhancement patches (github.com/andreafioraldi/afl) diff --git a/docs/README b/docs/README index 3f19d328..54e3e4a4 100644 --- a/docs/README +++ b/docs/README @@ -19,9 +19,13 @@ american fuzzy lop plus plus C. Hollers afl-fuzz Python mutator module and llvm_mode whitelist support was added too (https://github.com/choller/afl) - The newest additions is the excellent MOpt mutator from + New is the excellent MOpt mutator from https://github.com/puppet-meteor/MOpt-AFL + Also newly integrated is instrim, a very effective CFG llvm_mode + instrumentation implementation which replaced the original afl one and is + from https://github.com/csienslab/instrim + A more thorough list is available in the PATCHES file. So all in all this is the best-of AFL that is currently out there :-) diff --git a/llvm_mode/LLVMInsTrim.so.cc b/llvm_mode/LLVMInsTrim.so.cc index 62977e97..51640870 100644 --- a/llvm_mode/LLVMInsTrim.so.cc +++ b/llvm_mode/LLVMInsTrim.so.cc @@ -1,3 +1,6 @@ +#include +#include +#include #include #include "llvm/ADT/DenseMap.h" @@ -131,12 +134,16 @@ namespace { if (!myWhitelist.empty()) { bool instrumentBlock = false; - BasicBlock &BB = F.getEntryBlock(); - BasicBlock::iterator IP = BB.getFirstInsertionPt(); - IRBuilder<> IRB(&(*IP)); - DebugLoc Loc = IP->getDebugLoc(); + DebugLoc Loc; StringRef instFilename; + for (auto &BB : F) { + BasicBlock::iterator IP = BB.getFirstInsertionPt(); + IRBuilder<> IRB(&(*IP)); + if (!Loc) + Loc = IP->getDebugLoc(); + } + if ( Loc ) { DILocation *cDILoc = dyn_cast(Loc.getAsMDNode()); @@ -169,7 +176,9 @@ namespace { * not whitelisted, so we skip instrumentation. */ if (!instrumentBlock) { if (!instFilename.str().empty()) - SAYF( "Not in whitelist, skipping %s ...\n", instFilename.str().c_str()); + SAYF(cYEL "[!] " cBRI "Not in whitelist, skipping %s ...\n", instFilename.str().c_str()); + else + SAYF(cYEL "[!] " cBRI "No filename information found, skipping it"); continue; } } diff --git a/llvm_mode/README.llvm b/llvm_mode/README.llvm index dc860e97..b4e05a7a 100644 --- a/llvm_mode/README.llvm +++ b/llvm_mode/README.llvm @@ -38,8 +38,8 @@ co-exists with the original code. The idea and much of the implementation comes from Laszlo Szekeres. -2) How to use -------------- +2) How to use this +------------------ In order to leverage this mechanism, you need to have clang installed on your system. You should also make sure that the llvm-config tool is in your path @@ -69,8 +69,10 @@ operating mode of AFL, e.g.: Be sure to also include CXX set to afl-clang-fast++ for C++ code. The tool honors roughly the same environmental variables as afl-gcc (see -../docs/env_variables.txt). This includes AFL_INST_RATIO, AFL_USE_ASAN, -AFL_HARDEN, and AFL_DONT_OPTIMIZE. +../docs/env_variables.txt). This includes AFL_USE_ASAN, +AFL_HARDEN, and AFL_DONT_OPTIMIZE. However AFL_INST_RATIO is not honored +as it does not serve a good purpose with the more effective instrim CFG +analysis. Note: if you want the LLVM helper to be installed on your system for all users, you need to build it before issuing 'make install' in the parent -- cgit 1.4.1 From 32525238238e96ec0ce64a36f70558f76bc90ff5 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Mon, 15 Jul 2019 11:22:54 +0200 Subject: fixing commit fuckup --- Makefile | 2 +- docs/ChangeLog | 8 +++----- docs/README | 3 +-- docs/env_variables.txt | 22 +++++++++++++++++----- llvm_mode/LLVMInsTrim.so.cc | 2 +- llvm_mode/Makefile | 13 ++++++++----- llvm_mode/README.llvm | 19 ++++++++++++------- llvm_mode/afl-clang-fast.c | 16 +++++++++------- 8 files changed, 52 insertions(+), 33 deletions(-) (limited to 'docs/README') diff --git a/Makefile b/Makefile index 60dfde18..6b580381 100644 --- a/Makefile +++ b/Makefile @@ -194,7 +194,7 @@ install: all rm -f $${DESTDIR}$(BIN_PATH)/afl-as if [ -f afl-qemu-trace ]; then install -m 755 afl-qemu-trace $${DESTDIR}$(BIN_PATH); fi ifndef AFL_TRACE_PC - if [ -f afl-clang-fast -a -f libLLVMInsTrim.so -a -f afl-llvm-rt.o ]; then set -e; install -m 755 afl-clang-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang-fast++; install -m 755 libLLVMInsTrim.so afl-llvm-rt.o $${DESTDIR}$(HELPER_PATH); fi + if [ -f afl-clang-fast -a -f libLLVMInsTrim.so -a -f afl-llvm-rt.o ]; then set -e; install -m 755 afl-clang-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang-fast++; install -m 755 libLLVMInsTrim.so afl-llvm-pass.so afl-llvm-rt.o $${DESTDIR}$(HELPER_PATH); fi else if [ -f afl-clang-fast -a -f afl-llvm-rt.o ]; then set -e; install -m 755 afl-clang-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang-fast++; install -m 755 afl-llvm-rt.o $${DESTDIR}$(HELPER_PATH); fi endif diff --git a/docs/ChangeLog b/docs/ChangeLog index 9cdca49b..116029ea 100644 --- a/docs/ChangeLog +++ b/docs/ChangeLog @@ -17,9 +17,9 @@ sending a mail to . Version ++2.52d (tbd): ----------------------------- - - added instrim a much better llvm_mode instrumentation - (https://github.com/csienslab/instrim) - - added MOpt (github.com/puppet-meteor/MOpt-AFL) mode + - added instrim, a much faster llvm_mode instrumentation at the cost of + path discovery. See llvm_mode/README.instrim (https://github.com/csienslab/instrim) + - added MOpt (github.com/puppet-meteor/MOpt-AFL) mode, see docs/README.MOpt - added code to make it more portable to other platforms than Intel Linux - added never zero counters for afl-gcc and optional (because of an optimization issue in llvm < 9) for llvm_mode (AFL_LLVM_NEVER_ZERO=1) @@ -41,8 +41,6 @@ Version ++2.52d (tbd): tests as the random numbers are deterministic then - llvm_mode LAF_... env variables can now be specified as AFL_LLVM_LAF_... that is longer but in line with other llvm specific env vars - - ... your idea or patch? - ----------------------------- diff --git a/docs/README b/docs/README index 54e3e4a4..3a6c2921 100644 --- a/docs/README +++ b/docs/README @@ -23,8 +23,7 @@ american fuzzy lop plus plus https://github.com/puppet-meteor/MOpt-AFL Also newly integrated is instrim, a very effective CFG llvm_mode - instrumentation implementation which replaced the original afl one and is - from https://github.com/csienslab/instrim + instrumentation implementation from https://github.com/csienslab/instrim A more thorough list is available in the PATCHES file. diff --git a/docs/env_variables.txt b/docs/env_variables.txt index 8e2723d7..e58327b4 100644 --- a/docs/env_variables.txt +++ b/docs/env_variables.txt @@ -109,11 +109,21 @@ Then there are a few specific features that are only available in llvm_mode: See llvm_mode/README.whitelist for more information. - OTHER - ===== - - Setting LOOPHEAD=1 optimized loops. afl-fuzz will only be able to - see the path the loop took, but not how many times it was called - (unless its a complex loop). + INSTRIM + ======= + This feature increases the speed by whopping 20% but at the cost of a + lower path discovery and thefore coverage. + + - Setting AFL_LLVM_INSTRIM activates this mode + + - Setting AFL_LLVM_INSTRIM LOOPHEAD=1 expands on INSTRIM to optimize loops. + afl-fuzz will only be able to see the path the loop took, but not how + many times it was called (unless its a complex loop). + + See llvm_mode/README.instrim + + NOT_ZERO + ======== - Setting AFL_LLVM_NOT_ZERO=1 during compilation will use counters that skip zero on overflow. This is the default for llvm >= 9, @@ -121,6 +131,8 @@ Then there are a few specific features that are only available in llvm_mode: slowdown due a performance issue that is only fixed in llvm 9+. This feature increases path discovery by a little bit. + See llvm_mode/README.neverzero + 3) Settings for afl-fuzz ------------------------ diff --git a/llvm_mode/LLVMInsTrim.so.cc b/llvm_mode/LLVMInsTrim.so.cc index 8e9f7667..81cf98c4 100644 --- a/llvm_mode/LLVMInsTrim.so.cc +++ b/llvm_mode/LLVMInsTrim.so.cc @@ -96,7 +96,7 @@ namespace { OKF("LLVM neverZero activated (by hexcoder)\n"); #endif - if (getenv("LOOPHEAD")) { + if (getenv("AFL_LLVM_INSTRIM_LOOPHEAD") != NULL || getenv("LOOPHEAD") != NULL) { LoopHeadOpt = true; } diff --git a/llvm_mode/Makefile b/llvm_mode/Makefile index d0d4b690..2b685ddc 100644 --- a/llvm_mode/Makefile +++ b/llvm_mode/Makefile @@ -94,7 +94,7 @@ endif ifndef AFL_TRACE_PC - PROGS = ../afl-clang-fast ../libLLVMInsTrim.so ../afl-llvm-rt.o ../afl-llvm-rt-32.o ../afl-llvm-rt-64.o ../compare-transform-pass.so ../split-compares-pass.so ../split-switches-pass.so + PROGS = ../afl-clang-fast ../afl-llvm-pass.so ../libLLVMInsTrim.so ../afl-llvm-rt.o ../afl-llvm-rt-32.o ../afl-llvm-rt-64.o ../compare-transform-pass.so ../split-compares-pass.so ../split-switches-pass.so else PROGS = ../afl-clang-fast ../afl-llvm-rt.o ../afl-llvm-rt-32.o ../afl-llvm-rt-64.o ../compare-transform-pass.so ../split-compares-pass.so ../split-switches-pass.so endif @@ -104,7 +104,7 @@ ifneq "$(CLANGVER)" "$(LLVMVER)" CXX = $(shell llvm-config --bindir)/clang++ endif -all: test_deps test_shm $(PROGS) test_build all_done +all: test_shm test_deps $(PROGS) test_build all_done ifeq "$(SHMAT_OK)" "1" @@ -132,10 +132,10 @@ endif @which $(CC) >/dev/null 2>&1 || ( echo "[-] Oops, can't find '$(CC)'. Make sure that it's in your \$$PATH (or set \$$CC and \$$CXX)."; exit 1 ) @echo "[*] Checking for matching versions of '$(CC)' and '$(LLVM_CONFIG)'" ifneq "$(CLANGVER)" "$(LLVMVER)" - @echo "WARNING: we have llvm-config version $(LLVMVER) and a clang version $(CLANGVER)" - @echo "Retrying with the clang compiler from llvm: CC=`llvm-config --bindir`/clang" + @echo "[!] WARNING: we have llvm-config version $(LLVMVER) and a clang version $(CLANGVER)" + @echo "[!] Retrying with the clang compiler from llvm: CC=`llvm-config --bindir`/clang" else - @echo "we have llvm-config version $(LLVMVER) with a clang version $(CLANGVER), good." + @echo "[*] We have llvm-config version $(LLVMVER) with a clang version $(CLANGVER), good." endif @echo "[*] Checking for '../afl-showmap'..." @test -f ../afl-showmap || ( echo "[-] Oops, can't find '../afl-showmap'. Be sure to compile AFL first."; exit 1 ) @@ -148,6 +148,9 @@ endif ../libLLVMInsTrim.so: LLVMInsTrim.so.cc MarkNodes.cc | test_deps $(CXX) $(CLANG_CFL) -DLLVMInsTrim_EXPORTS -fno-rtti -fPIC -std=gnu++11 -shared $< MarkNodes.cc -o $@ $(CLANG_LFL) +../afl-llvm-pass.so: afl-llvm-pass.so.cc | test_deps + $(CXX) $(CLANG_CFL) -DLLVMInsTrim_EXPORTS -fno-rtti -fPIC -std=gnu++11 -shared $< -o $@ $(CLANG_LFL) + # laf ../split-switches-pass.so: split-switches-pass.so.cc | test_deps $(CXX) $(CLANG_CFL) -shared $< -o $@ $(CLANG_LFL) diff --git a/llvm_mode/README.llvm b/llvm_mode/README.llvm index 77c406f8..779ff47c 100644 --- a/llvm_mode/README.llvm +++ b/llvm_mode/README.llvm @@ -88,13 +88,18 @@ which C/C++ files to actually intrument. See README.whitelist For splitting memcmp, strncmp, etc. please see README.laf-intel -As the original afl llvm_mode implementation has been replaced with -then much more effective instrim (https://github.com/csienslab/instrim/) -there is an option for optimizing loops. This optimization shows which -part of the loop has been selected, but not how many time a loop has been -called in a row (unless its a complex loop and a block inside was -instrumented). If you want to enable this set the environment variable -LOOPHEAD=1 +Then there is an optimized instrumentation strategy that uses CFGs and +markers to just instrument what is needed. This increases speed by 20-25% +however has a lower path discovery. +If you want to use this, set AFL_LLVM_INSTRIM=1 +See README.instrim + +Finally if your llvm version is 8 or lower, you can activate a mode that +prevents that a counter overflow result in a 0 value. This is good for +path discovery, but the llvm implementation for intel for this functionality +is not optimal and was only fixed in llvm 9. +You can set this with AFL_LLVM_NOT_ZERO=1 +See README.neverzero 4) Gotchas, feedback, bugs diff --git a/llvm_mode/afl-clang-fast.c b/llvm_mode/afl-clang-fast.c index 249eea7d..19bad86c 100644 --- a/llvm_mode/afl-clang-fast.c +++ b/llvm_mode/afl-clang-fast.c @@ -88,7 +88,7 @@ static void find_obj(u8* argv0) { return; } - FATAL("Unable to find 'afl-llvm-rt.o' or 'libLLVMInsTrim.so'. Please set AFL_PATH"); + FATAL("Unable to find 'afl-llvm-rt.o' or 'afl-llvm-pass.so.cc'. Please set AFL_PATH"); } @@ -113,11 +113,11 @@ static void edit_params(u32 argc, char** argv) { cc_params[0] = alt_cc ? alt_cc : (u8*)"clang"; } - /* There are two ways to compile afl-clang-fast. In the traditional mode, we - use libLLVMInsTrim.so to inject instrumentation. In the experimental + /* There are three ways to compile with afl-clang-fast. In the traditional + mode, we use afl-llvm-pass.so, then there is libLLVMInsTrim.so which is + much faster but has less coverage. Finally tere is the experimental 'trace-pc-guard' mode, we use native LLVM instrumentation callbacks - instead. The latter is a very recent addition - see: - + instead. For trace-pc-guard see: http://clang.llvm.org/docs/SanitizerCoverage.html#tracing-pcs-with-guards */ // laf @@ -151,8 +151,10 @@ static void edit_params(u32 argc, char** argv) { cc_params[cc_par_cnt++] = "-Xclang"; cc_params[cc_par_cnt++] = "-load"; cc_params[cc_par_cnt++] = "-Xclang"; - cc_params[cc_par_cnt++] = alloc_printf("%s/libLLVMInsTrim.so", obj_path); -// cc_params[cc_par_cnt++] = alloc_printf("%s/afl-llvm-pass.so", obj_path); + if (getenv("AFL_LLVM_INSTRIM") != NULL || getenv("INSTRIM_LIB") != NULL) + cc_params[cc_par_cnt++] = alloc_printf("%s/libLLVMInsTrim.so", obj_path); + else + cc_params[cc_par_cnt++] = alloc_printf("%s/afl-llvm-pass.so", obj_path); #endif /* ^USE_TRACE_PC */ cc_params[cc_par_cnt++] = "-Qunused-arguments"; -- cgit 1.4.1 From 8a4cdd56d4d7557a54b0f28af7edceaddfb30663 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Tue, 16 Jul 2019 08:34:17 +0200 Subject: added gcc_plugin --- Makefile | 7 +- afl-fuzz.c | 62 ++----- docs/ChangeLog | 4 + docs/PATCHES | 1 + docs/README | 3 + gcc_plugin/Makefile | 75 +++++++++ gcc_plugin/README.gcc | 157 ++++++++++++++++++ gcc_plugin/afl-gcc-fast.c | 283 ++++++++++++++++++++++++++++++++ gcc_plugin/afl-gcc-pass.so.cc | 372 ++++++++++++++++++++++++++++++++++++++++++ gcc_plugin/afl-gcc-rt.o.c | 224 +++++++++++++++++++++++++ 10 files changed, 1137 insertions(+), 51 deletions(-) create mode 100644 gcc_plugin/Makefile create mode 100644 gcc_plugin/README.gcc create mode 100644 gcc_plugin/afl-gcc-fast.c create mode 100644 gcc_plugin/afl-gcc-pass.so.cc create mode 100644 gcc_plugin/afl-gcc-rt.o.c (limited to 'docs/README') diff --git a/Makefile b/Makefile index 6b580381..69bd7535 100644 --- a/Makefile +++ b/Makefile @@ -181,7 +181,7 @@ all_done: test_build .NOTPARALLEL: clean clean: - rm -f $(PROGS) afl-as as afl-g++ afl-clang afl-clang++ *.o *~ a.out core core.[1-9][0-9]* *.stackdump test .test .test1 .test2 test-instr .test-instr0 .test-instr1 qemu_mode/qemu-3.1.0.tar.xz afl-qemu-trace + rm -f $(PROGS) afl-as as afl-g++ afl-clang afl-clang++ *.o *~ a.out core core.[1-9][0-9]* *.stackdump test .test .test1 .test2 test-instr .test-instr0 .test-instr1 qemu_mode/qemu-3.1.0.tar.xz afl-qemu-trace afl-gcc-fast afl-gcc-pass.so afl-gcc-rt.o afl-g++-fast rm -rf out_dir qemu_mode/qemu-3.1.0 $(MAKE) -C llvm_mode clean $(MAKE) -C libdislocator clean @@ -193,6 +193,7 @@ install: all install -m 755 $(PROGS) $(SH_PROGS) $${DESTDIR}$(BIN_PATH) rm -f $${DESTDIR}$(BIN_PATH)/afl-as if [ -f afl-qemu-trace ]; then install -m 755 afl-qemu-trace $${DESTDIR}$(BIN_PATH); fi + if [ -f afl-gcc-fast ]; then set e; install -m 755 afl-gcc-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-gcc-fast $${DESTDIR}$(BIN_PATH)/afl-g++-fast; install -m 755 afl-gcc-pass.so afl-gcc-rt.o $${DESTDIR}$(HELPER_PATH); fi ifndef AFL_TRACE_PC if [ -f afl-clang-fast -a -f libLLVMInsTrim.so -a -f afl-llvm-rt.o ]; then set -e; install -m 755 afl-clang-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang-fast++; install -m 755 libLLVMInsTrim.so afl-llvm-pass.so afl-llvm-rt.o $${DESTDIR}$(HELPER_PATH); fi else @@ -204,7 +205,9 @@ endif if [ -f split-compares-pass.so ]; then set -e; install -m 755 split-compares-pass.so $${DESTDIR}$(HELPER_PATH); fi if [ -f split-switches-pass.so ]; then set -e; install -m 755 split-switches-pass.so $${DESTDIR}$(HELPER_PATH); fi - set -e; for i in afl-g++ afl-clang afl-clang++; do ln -sf afl-gcc $${DESTDIR}$(BIN_PATH)/$$i; done + set -e; ln -sf afl-gcc $${DESTDIR}$(BIN_PATH)/afl-g++ + set -e; if [ -f afl-clang-fast ] ; then ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang ; ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang++ ; else ln -sf afl-gcc $${DESTDIR}$(BIN_PATH)/afl-clang ; ln -sf afl-gcc $${DESTDIR}$(BIN_PATH)/afl-clang++; fi + install -m 755 afl-as $${DESTDIR}$(HELPER_PATH) ln -sf afl-as $${DESTDIR}$(HELPER_PATH)/as install -m 644 docs/README docs/ChangeLog docs/*.txt $${DESTDIR}$(DOC_PATH) diff --git a/afl-fuzz.c b/afl-fuzz.c index 3a8e5750..81dff98c 100644 --- a/afl-fuzz.c +++ b/afl-fuzz.c @@ -5234,6 +5234,7 @@ static u32 calculate_score(struct queue_entry* q) { // MOpt mode if (limit_time_sig != 0 && max_depth - q->depth < 3) perf_score *= 2; + else if (perf_score < 1) perf_score = 1; // Add a lower bound to AFLFast's energy assignment strategies /* Make sure that we don't go over limit. */ @@ -10224,8 +10225,7 @@ static u8 core_fuzzing(char** argv) { stage_max = (doing_det ? HAVOC_CYCLES_INIT : HAVOC_CYCLES) * perf_score / havoc_div / 100; - } - else { + } else { static u8 tmp[32]; @@ -10243,20 +10243,13 @@ static u8 core_fuzzing(char** argv) { //for (; swarm_now < swarm_num; swarm_now++) { - - - if (key_puppet == 1) - { - if (unlikely(orig_hit_cnt_puppet == 0)) - { + if (key_puppet == 1) { + if (unlikely(orig_hit_cnt_puppet == 0)) { orig_hit_cnt_puppet = queued_paths + unique_crashes; last_limit_time_start = get_cur_time(); - SPLICE_CYCLES_puppet = (UR(SPLICE_CYCLES_puppet_up - SPLICE_CYCLES_puppet_low + 1) + SPLICE_CYCLES_puppet_low); } } - - { havoc_stage_puppet: @@ -10266,14 +10259,11 @@ static u8 core_fuzzing(char** argv) { splice_cycle variable is set, generate different descriptions and such. */ if (!splice_cycle) { - stage_name = "MOpt core avoc"; stage_short = "MOpt core havoc"; stage_max = (doing_det ? HAVOC_CYCLES_INIT : HAVOC_CYCLES) * perf_score / havoc_div / 100; - - } - else { + } else { static u8 tmp[32]; perf_score = orig_perf; sprintf(tmp, "MOpt core splice %u", splice_cycle); @@ -10282,31 +10272,20 @@ static u8 core_fuzzing(char** argv) { stage_max = SPLICE_HAVOC * perf_score / havoc_div / 100; } - - if (stage_max < HAVOC_MIN) stage_max = HAVOC_MIN; - temp_len = len; - orig_hit_cnt = queued_paths + unique_crashes; - havoc_queued = queued_paths; - - for (stage_cur = 0; stage_cur < stage_max; stage_cur++) { u32 use_stacking = 1 << (1 + UR(HAVOC_STACK_POW2)); - stage_cur_val = use_stacking; - - for (i = 0; i < operator_num; i++) - { + for (i = 0; i < operator_num; i++) { core_operator_cycles_puppet_v3[i] = core_operator_cycles_puppet_v2[i]; } - for (i = 0; i < use_stacking; i++) { switch (select_algorithm()) { @@ -10366,8 +10345,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { u32 pos = UR(temp_len - 1); *(u16*)(out_buf + pos) -= 1 + UR(ARITH_MAX); - } - else { + } else { u32 pos = UR(temp_len - 1); u16 num = 1 + UR(ARITH_MAX); *(u16*)(out_buf + pos) = @@ -10377,8 +10355,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { u32 pos = UR(temp_len - 1); *(u16*)(out_buf + pos) += 1 + UR(ARITH_MAX); - } - else { + } else { u32 pos = UR(temp_len - 1); u16 num = 1 + UR(ARITH_MAX); *(u16*)(out_buf + pos) = @@ -10394,8 +10371,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { u32 pos = UR(temp_len - 3); *(u32*)(out_buf + pos) -= 1 + UR(ARITH_MAX); - } - else { + } else { u32 pos = UR(temp_len - 3); u32 num = 1 + UR(ARITH_MAX); *(u32*)(out_buf + pos) = @@ -10405,8 +10381,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { u32 pos = UR(temp_len - 3); *(u32*)(out_buf + pos) += 1 + UR(ARITH_MAX); - } - else { + } else { u32 pos = UR(temp_len - 3); u32 num = 1 + UR(ARITH_MAX); *(u32*)(out_buf + pos) = @@ -10429,8 +10404,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { *(u16*)(out_buf + UR(temp_len - 1)) = interesting_16[UR(sizeof(interesting_16) >> 1)]; - } - else { + } else { *(u16*)(out_buf + UR(temp_len - 1)) = SWAP16( interesting_16[UR(sizeof(interesting_16) >> 1)]); } @@ -10446,8 +10420,7 @@ static u8 core_fuzzing(char** argv) { if (UR(2)) { *(u32*)(out_buf + UR(temp_len - 3)) = interesting_32[UR(sizeof(interesting_32) >> 2)]; - } - else { + } else { *(u32*)(out_buf + UR(temp_len - 3)) = SWAP32( interesting_32[UR(sizeof(interesting_32) >> 2)]); } @@ -10466,7 +10439,6 @@ static u8 core_fuzzing(char** argv) { break; - case 13: { /* Delete bytes. We're making this a bit more likely @@ -10507,8 +10479,7 @@ static u8 core_fuzzing(char** argv) { clone_len = choose_block_len(temp_len); clone_from = UR(temp_len - clone_len + 1); - } - else { + } else { clone_len = choose_block_len(HAVOC_BLK_XL); clone_from = 0; @@ -10575,17 +10546,10 @@ static u8 core_fuzzing(char** argv) { } - tmp_core_time += 1; - - - u64 temp_total_found = queued_paths + unique_crashes; - - - if (common_fuzz_stuff(argv, out_buf, temp_len)) goto abandon_entry_puppet; diff --git a/docs/ChangeLog b/docs/ChangeLog index 116029ea..8c1aa994 100644 --- a/docs/ChangeLog +++ b/docs/ChangeLog @@ -17,6 +17,10 @@ sending a mail to . Version ++2.52d (tbd): ----------------------------- + - if llvm_mode was compiled, afl-clang/afl-clang++ will point to these + instead of afl-gcc + - added gcc_plugin which is like llvm_mode but for gcc. This version + supports gcc version 5 to 8. See gcc_plugin/README (https://github.com/T12z/afl) - added instrim, a much faster llvm_mode instrumentation at the cost of path discovery. See llvm_mode/README.instrim (https://github.com/csienslab/instrim) - added MOpt (github.com/puppet-meteor/MOpt-AFL) mode, see docs/README.MOpt diff --git a/docs/PATCHES b/docs/PATCHES index f6ca9284..78050208 100644 --- a/docs/PATCHES +++ b/docs/PATCHES @@ -17,6 +17,7 @@ afl-qemu-optimize-entrypoint.diff by mh(at)mh-sec(dot)de afl-qemu-speed.diff by abiondo on github afl-qemu-optimize-map.diff by mh(at)mh-sec(dot)de ++ gcc_plugin (https://github.com/T12z/afl) + instrim (https://github.com/csienslab/instrim) was integrated + MOpt (github.com/puppet-meteor/MOpt-AFL) was imported + AFLfast additions (github.com/mboehme/aflfast) were incorporated. diff --git a/docs/README b/docs/README index 3a6c2921..a8ec715b 100644 --- a/docs/README +++ b/docs/README @@ -25,6 +25,9 @@ american fuzzy lop plus plus Also newly integrated is instrim, a very effective CFG llvm_mode instrumentation implementation from https://github.com/csienslab/instrim + Similar to llvm_mode a gcc_plugin that supports versions 5 to 8 is + now available (from https://github.com/T12z/afl) + A more thorough list is available in the PATCHES file. So all in all this is the best-of AFL that is currently out there :-) diff --git a/gcc_plugin/Makefile b/gcc_plugin/Makefile new file mode 100644 index 00000000..721ce0ad --- /dev/null +++ b/gcc_plugin/Makefile @@ -0,0 +1,75 @@ +# +# american fuzzy lop - GCC plugin instrumentation +# ----------------------------------------------- +# +# Written by Austin Seipp and +# Laszlo Szekeres and +# Michal Zalewski +# +# GCC integration design is based on the LLVM design, which comes +# from Laszlo Szekeres. +# +# Copyright 2015 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 +# + +PREFIX ?= /usr/local +HELPER_PATH = $(PREFIX)/lib/afl +BIN_PATH = $(PREFIX)/bin + +CFLAGS ?= -O3 -g -funroll-loops +CFLAGS += -Wall -D_FORTIFY_SOURCE=2 -Wno-pointer-sign \ + -DAFL_PATH=\"$(HELPER_PATH)\" -DBIN_PATH=\"$(BIN_PATH)\" \ + +CXXFLAGS ?= -O3 -g -funroll-loops +CXXEFLAGS := $(CXXFLAGS) -Wall -D_FORTIFY_SOURCE=2 + +CC ?= gcc +CXX ?= g++ + +PLUGIN_FLAGS = -fPIC -fno-rtti -I"$(shell $(CC) -print-file-name=plugin)/include" + +PROGS = ../afl-gcc-fast ../afl-gcc-pass.so ../afl-gcc-rt.o + +all: test_deps $(PROGS) test_build all_done + +test_deps: + @echo "[*] Checking for working '$(CC)'..." + @which $(CC) >/dev/null 2>&1 || ( echo "[-] Oops, can't find '$(CC)'. Make sure that it's in your \$$PATH (or set \$$CC and \$$CXX)."; exit 1 ) + @echo "[*] Checking for '../afl-showmap'..." + @test -f ../afl-showmap || ( echo "[-] Oops, can't find '../afl-showmap'. Be sure to compile AFL first."; exit 1 ) + @echo "[+] All set and ready to build." + +../afl-gcc-fast: afl-gcc-fast.c | test_deps + $(CC) $(CFLAGS) $< -o $@ $(LDFLAGS) + ln -sf afl-gcc-fast ../afl-g++-fast + +../afl-gcc-pass.so: afl-gcc-pass.so.cc | test_deps + $(CXX) $(CXXEFLAGS) $(PLUGIN_FLAGS) -shared $< -o $@ + +../afl-gcc-rt.o: afl-gcc-rt.o.c | test_deps + $(CC) $(CFLAGS) -fPIC -c $< -o $@ + +test_build: $(PROGS) + @echo "[*] Testing the CC wrapper and instrumentation output..." +# unset AFL_USE_ASAN AFL_USE_MSAN; AFL_QUIET=1 AFL_INST_RATIO=100 AFL_PATH=. AFL_CC=$(CC) ../afl-gcc-fast $(CFLAGS) ../test-instr.c -o test-instr $(LDFLAGS) + unset AFL_USE_ASAN AFL_USE_MSAN; AFL_INST_RATIO=100 AFL_PATH=. AFL_CC=$(CC) ../afl-gcc-fast $(CFLAGS) ../test-instr.c -o test-instr $(LDFLAGS) + echo 0 | ../afl-showmap -m none -q -o .test-instr0 ./test-instr + echo 1 | ../afl-showmap -m none -q -o .test-instr1 ./test-instr + @rm -f test-instr + @cmp -s .test-instr0 .test-instr1; DR="$$?"; rm -f .test-instr0 .test-instr1; if [ "$$DR" = "0" ]; then echo; echo "Oops, the instrumentation does not seem to be behaving correctly!"; echo; echo "Please ping to troubleshoot the issue."; echo; exit 1; fi + @echo "[+] All right, the instrumentation seems to be working!" + +all_done: test_build + @echo "[+] All done! You can now use '../afl-gcc-fast' to compile programs." + +.NOTPARALLEL: clean + +clean: + rm -f *.o *.so *~ a.out core core.[1-9][0-9]* test-instr .test-instr0 .test-instr1 + rm -f $(PROGS) ../afl-g++-fast diff --git a/gcc_plugin/README.gcc b/gcc_plugin/README.gcc new file mode 100644 index 00000000..b3e9c853 --- /dev/null +++ b/gcc_plugin/README.gcc @@ -0,0 +1,157 @@ +=========================================== +Fast GCC-based instrumentation for afl-fuzz +=========================================== + + (See ../docs/README for the general instruction manual.) + (See ../llvm_mode/README.llvm for the LLVM-based instrumentation.) + +1) Introduction +--------------- + +The code in this directory allows you to instrument programs for AFL using +true compiler-level instrumentation, instead of the more crude +assembly-level rewriting approach taken by afl-gcc and afl-clang. This has +several interesting properties: + + - The compiler can make many optimizations that are hard to pull off when + manually inserting assembly. As a result, some slow, CPU-bound programs will + run up to around 2x faster. + + The gains are less pronounced for fast binaries, where the speed is limited + chiefly by the cost of creating new processes. In such cases, the gain will + probably stay within 10%. + + - The instrumentation is CPU-independent. At least in principle, you should + be able to rely on it to fuzz programs on non-x86 architectures (after + building afl-fuzz with AFL_NOX86=1). + + - Because the feature relies on the internals of GCC, it is gcc-specific + and will *not* work with LLVM (see ../llvm_plugin for an alternative). + +Once this implementation is shown to be sufficiently robust and portable, it +will probably replace afl-gcc. For now, it can be built separately and +co-exists with the original code. + +The idea and much of the implementation comes from Laszlo Szekeres. + +2) How to use +------------- + +In order to leverage this mechanism, you need to have GCC and the plugin headers +installed on your system. That should be all you need. On Debian machines, these +headers can be acquired by installing the `gcc--plugin-dev` packages. + +To build the instrumentation itself, type 'make'. This will generate binaries +called afl-gcc-fast and afl-g++-fast in the parent directory. Once this +is done, you can instrument third-party code in a way similar to the standard +operating mode of AFL, e.g.: + + CC=/path/to/afl/afl-gcc-fast ./configure [...options...] + make + +Be sure to also include CXX set to afl-g++-fast for C++ code. + +The tool honors roughly the same environmental variables as afl-gcc (see +../docs/env_variables.txt). This includes AFL_INST_RATIO, AFL_USE_ASAN, +AFL_HARDEN, and AFL_DONT_OPTIMIZE. + +Note: if you want the GCC plugin to be installed on your system for all +users, you need to build it before issuing 'make install' in the parent +directory. + +3) Gotchas, feedback, bugs +-------------------------- + +This is an early-stage mechanism, so field reports are welcome. You can send bug +reports to . + +4) Bonus feature #1: deferred instrumentation +--------------------------------------------- + +AFL tries to optimize performance by executing the targeted binary just once, +stopping it just before main(), and then cloning this "master" process to get +a steady supply of targets to fuzz. + +Although this approach eliminates much of the OS-, linker- and libc-level +costs of executing the program, it does not always help with binaries that +perform other time-consuming initialization steps - say, parsing a large config +file before getting to the fuzzed data. + +In such cases, it's beneficial to initialize the forkserver a bit later, once +most of the initialization work is already done, but before the binary attempts +to read the fuzzed input and parse it; in some cases, this can offer a 10x+ +performance gain. You can implement delayed initialization in LLVM mode in a +fairly simple way. + +First, locate a suitable location in the code where the delayed cloning can +take place. This needs to be done with *extreme* care to avoid breaking the +binary. In particular, the program will probably malfunction if you select +a location after: + + - The creation of any vital threads or child processes - since the forkserver + can't clone them easily. + + - The initialization of timers via setitimer() or equivalent calls. + + - The creation of temporary files, network sockets, offset-sensitive file + descriptors, and similar shared-state resources - but only provided that + their state meaningfully influences the behavior of the program later on. + + - Any access to the fuzzed input, including reading the metadata about its + size. + +With the location selected, add this code in the appropriate spot: + +#ifdef __AFL_HAVE_MANUAL_CONTROL + __AFL_INIT(); +#endif + +You don't need the #ifdef guards, but they will make the program still work as +usual when compiled with a tool other than afl-gcc-fast/afl-clang-fast. + +Finally, recompile the program with afl-gcc-fast (afl-gcc or afl-clang will +*not* generate a deferred-initialization binary) - and you should be all set! + +5) Bonus feature #2: persistent mode +------------------------------------ + +Some libraries provide APIs that are stateless, or whose state can be reset in +between processing different input files. When such a reset is performed, a +single long-lived process can be reused to try out multiple test cases, +eliminating the need for repeated fork() calls and the associated OS overhead. + +The basic structure of the program that does this would be: + + while (__AFL_LOOP(1000)) { + + /* Read input data. */ + /* Call library code to be fuzzed. */ + /* Reset state. */ + + } + + /* Exit normally */ + +The numerical value specified within the loop controls the maximum number +of iterations before AFL will restart the process from scratch. This minimizes +the impact of memory leaks and similar glitches; 1000 is a good starting point. + +A more detailed template is shown in ../experimental/persistent_demo/. +Similarly to the previous mode, the feature works only with afl-gcc-fast or +afl-clang-fast; #ifdef guards can be used to suppress it when using other +compilers. + +Note that as with the previous mode, the feature is easy to misuse; if you +do not reset the critical state fully, you may end up with false positives or +waste a whole lot of CPU power doing nothing useful at all. Be particularly +wary of memory leaks and the state of file descriptors. + +When running in this mode, the execution paths will inherently vary a bit +depending on whether the input loop is being entered for the first time or +executed again. To avoid spurious warnings, the feature implies +AFL_NO_VAR_CHECK and hides the "variable path" warnings in the UI. + +PS. Because there are task switches still involved, the mode isn't as fast as +"pure" in-process fuzzing offered, say, by LLVM's LibFuzzer; but it is a lot +faster than the normal fork() model, and compared to in-process fuzzing, +should be a lot more robust. diff --git a/gcc_plugin/afl-gcc-fast.c b/gcc_plugin/afl-gcc-fast.c new file mode 100644 index 00000000..018b3f5f --- /dev/null +++ b/gcc_plugin/afl-gcc-fast.c @@ -0,0 +1,283 @@ +/* + american fuzzy lop - GCC wrapper for GCC plugin + ------------------------------------------------ + + Written by Austin Seipp and + Laszlo Szekeres and + Michal Zalewski + + GCC integration design is based on the LLVM design, which comes + from Laszlo Szekeres. + + Copyright 2015 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 + + This program is a drop-in replacement for gcc, similar in most + respects to ../afl-gcc, but with compiler instrumentation through a + plugin. It tries to figure out compilation mode, adds a bunch of + flags, and then calls the real compiler. + + */ + +#define AFL_MAIN + +#include "../config.h" +#include "../types.h" +#include "../debug.h" +#include "../alloc-inl.h" + +#include +#include +#include +#include + +static u8* obj_path; /* Path to runtime libraries */ +static u8** cc_params; /* Parameters passed to the real CC */ +static u32 cc_par_cnt = 1; /* Param count, including argv0 */ + + +/* Try to find the runtime libraries. If that fails, abort. */ + +static void find_obj(u8* argv0) { + + u8 *afl_path = getenv("AFL_PATH"); + u8 *slash, *tmp; + + if (afl_path) { + + tmp = alloc_printf("%s/afl-gcc-rt.o", afl_path); + + if (!access(tmp, R_OK)) { + obj_path = afl_path; + ck_free(tmp); + return; + } + + ck_free(tmp); + + } + + slash = strrchr(argv0, '/'); + + if (slash) { + + u8 *dir; + + *slash = 0; + dir = ck_strdup(argv0); + *slash = '/'; + + tmp = alloc_printf("%s/afl-gcc-rt.o", dir); + + if (!access(tmp, R_OK)) { + obj_path = dir; + ck_free(tmp); + return; + } + + ck_free(tmp); + ck_free(dir); + + } + + if (!access(AFL_PATH "/afl-gcc-rt.o", R_OK)) { + obj_path = AFL_PATH; + return; + } + + FATAL("Unable to find 'afl-gcc-rt.o' or 'afl-gcc-pass.so'. Please set AFL_PATH"); +} + + +/* Copy argv to cc_params, making the necessary edits. */ + +static void edit_params(u32 argc, char** argv) { + + u8 fortify_set = 0, asan_set = 0, x_set = 0, maybe_linking = 1; + u8 *name; + + cc_params = ck_alloc((argc + 64) * sizeof(u8*)); + + name = strrchr(argv[0], '/'); + if (!name) name = argv[0]; else name++; + + if (!strcmp(name, "afl-g++-fast")) { + u8* alt_cxx = getenv("AFL_CXX"); + cc_params[0] = alt_cxx ? alt_cxx : (u8*)"g++"; + } else { + u8* alt_cc = getenv("AFL_CC"); + cc_params[0] = alt_cc ? alt_cc : (u8*)"gcc"; + } + + char* fplugin_arg = alloc_printf("-fplugin=%s/afl-gcc-pass.so", obj_path); + cc_params[cc_par_cnt++] = fplugin_arg; + + while (--argc) { + u8* cur = *(++argv); + +#if defined(__x86_64__) + if (!strcmp(cur, "-m32")) FATAL("-m32 is not supported"); +#endif + + if (!strcmp(cur, "-x")) x_set = 1; + + if (!strcmp(cur, "-c") || !strcmp(cur, "-S") || !strcmp(cur, "-E") || + !strcmp(cur, "-v")) maybe_linking = 0; + + if (!strcmp(cur, "-fsanitize=address") || + !strcmp(cur, "-fsanitize=memory")) asan_set = 1; + + if (strstr(cur, "FORTIFY_SOURCE")) fortify_set = 1; + + cc_params[cc_par_cnt++] = cur; + + } + + if (getenv("AFL_HARDEN")) { + + cc_params[cc_par_cnt++] = "-fstack-protector-all"; + + if (!fortify_set) + cc_params[cc_par_cnt++] = "-D_FORTIFY_SOURCE=2"; + + } + + if (!asan_set) { + + if (getenv("AFL_USE_ASAN")) { + + cc_params[cc_par_cnt++] = "-fsanitize=address"; + + if (getenv("AFL_USE_MSAN")) + FATAL("ASAN and MSAN are mutually exclusive"); + + } else if (getenv("AFL_USE_MSAN")) { + + cc_params[cc_par_cnt++] = "-fsanitize=memory"; + + if (getenv("AFL_USE_ASAN")) + FATAL("ASAN and MSAN are mutually exclusive"); + + } + + } + + if (!getenv("AFL_DONT_OPTIMIZE")) { + + cc_params[cc_par_cnt++] = "-g"; + cc_params[cc_par_cnt++] = "-O3"; + cc_params[cc_par_cnt++] = "-funroll-loops"; + + } + + cc_params[cc_par_cnt++] = "-D__AFL_HAVE_MANUAL_CONTROL=1"; + + /* When the user tries to use persistent or deferred forkserver modes by + appending a single line to the program, we want to reliably inject a + signature into the binary (to be picked up by afl-fuzz) and we want + to call a function from the runtime .o file. This is unnecessarily + painful for three reasons: + + 1) We need to convince the compiler not to optimize out the signature. + This is done with __attribute__((used)). + + 2) We need to convince the linker, when called with -Wl,--gc-sections, + not to do the same. This is done by forcing an assignment to a + 'volatile' pointer. + + 3) We need to declare __afl_persistent_loop() in the global namespace, + but doing this within a method in a class is hard - :: and extern "C" + are forbidden and __attribute__((alias(...))) doesn't work. Hence the + __asm__ aliasing trick. + + */ + + cc_params[cc_par_cnt++] = "-D__AFL_LOOP(_A)=" + "({ static volatile char *_B __attribute__((used)); " + " _B = (char*)\"" PERSIST_SIG "\"; " +#ifdef __APPLE__ + "int _L(unsigned int) __asm__(\"___afl_persistent_loop\"); " +#else + "int _L(unsigned int) __asm__(\"__afl_persistent_loop\"); " +#endif /* ^__APPLE__ */ + "_L(_A); })"; + + cc_params[cc_par_cnt++] = "-D__AFL_INIT()=" + "do { static volatile char *_A __attribute__((used)); " + " _A = (char*)\"" DEFER_SIG "\"; " +#ifdef __APPLE__ + "void _I(void) __asm__(\"___afl_manual_init\"); " +#else + "void _I(void) __asm__(\"__afl_manual_init\"); " +#endif /* ^__APPLE__ */ + "_I(); } while (0)"; + + if (maybe_linking) { + + if (x_set) { + cc_params[cc_par_cnt++] = "-x"; + cc_params[cc_par_cnt++] = "none"; + } + + cc_params[cc_par_cnt++] = alloc_printf("%s/afl-gcc-rt.o", obj_path); + + } + + cc_params[cc_par_cnt] = NULL; + +} + + +/* Main entry point */ + +int main(int argc, char** argv) { + + if (isatty(2) && !getenv("AFL_QUIET")) { + + SAYF(cCYA "afl-gcc-fast" cRST " initial version 1.94 by , updated to " cBRI VERSION cRST " by \n"); + + } + + if (argc < 2) { + + SAYF("\n" + "This is a helper application for afl-fuzz. It serves as a drop-in replacement\n" + "for gcc, letting you recompile third-party code with the required runtime\n" + "instrumentation. A common use pattern would be one of the following:\n\n" + + " CC=%s/afl-gcc-fast ./configure\n" + " CXX=%s/afl-g++-fast ./configure\n\n" + + "In contrast to the traditional afl-gcc tool, this version is implemented as\n" + "a GCC plugin and tends to offer improved performance with slow programs\n" + "(similarly to the LLVM plugin used by afl-clang-fast).\n\n" + + "You can specify custom next-stage toolchain via AFL_CC and AFL_CXX. Setting\n" + "AFL_HARDEN enables hardening optimizations in the compiled code.\n\n", + BIN_PATH, BIN_PATH); + + exit(1); + + } + + + find_obj(argv[0]); + + edit_params(argc, argv); +/*if (isatty(2) && !getenv("AFL_QUIET")) { + printf("Calling \"%s\" with:\n", cc_params[0]); + for(int i=1; i with bits from + Emese Revfy + + Updated by Thorsten Schulz + + GCC integration design is based on the LLVM design, which comes + from Laszlo Szekeres. Some of the boilerplate code below for + afl_pass to adapt to different GCC versions was taken from Emese + Revfy's Size Overflow plugin for GCC, licensed under the GPLv2/v3. + + (NOTE: this plugin code is under GPLv3, in order to comply with the + GCC runtime library exception, which states that you may distribute + "Target Code" from the compiler under a license of your choice, as + long as the "Compilation Process" is "Eligible", and contains no + GPL-incompatible software in GCC "during the process of + transforming high level code to target code". In this case, the + plugin will be used to generate "Target Code" during the + "Compilation Process", and thus it must be GPLv3 to be "eligible".) + + Copyright (C) 2015 Austin Seipp + + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 3 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see . + + */ + +//#define BUILD_INLINE_INST + +#include "../config.h" +#include "../debug.h" + +/* clear helper AFL types pulls in, which intervene with gcc-plugin geaders from GCC-8 */ +#ifdef likely +#undef likely +#endif +#ifdef unlikely +#undef unlikely +#endif + + +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* -------------------------------------------------------------------------- */ +/* -- AFL instrumentation pass ---------------------------------------------- */ + +static int be_quiet = 0; +static unsigned int inst_ratio = 100; +static bool inst_ext = true; /* I reckon inline is broken / unfunctional */ + + +static unsigned int ext_call_instrument(function *fun) { + /* Instrument all the things! */ + basic_block bb; + unsigned finst_blocks = 0; + unsigned fcnt_blocks = 0; + + FOR_ALL_BB_FN(bb, fun) { + gimple_seq fcall; + gimple_seq seq = NULL; + gimple_stmt_iterator bentry; + + if (!fcnt_blocks++) continue; /* skip block 0 */ + + // TODO: if the predecessor does not have to destinations + // then skip this block :TODO + + /* Bail on this block if we trip the specified ratio */ + if (R(100) >= inst_ratio) continue; + + /* Make up cur_loc */ + + unsigned int rand_loc = R(MAP_SIZE); + tree cur_loc = build_int_cst(uint64_type_node, rand_loc); + + /* Update bitmap via external call */ + /* to quote: + * /+ Trace a basic block with some ID +/ + * void __afl_trace(u16 x); + */ + + tree fntype = build_function_type_list( + void_type_node, /* return */ + uint16_type_node, /* args */ + NULL_TREE); /* done */ + tree fndecl = build_fn_decl("__afl_trace", fntype); + TREE_STATIC(fndecl) = 1; /* Defined elsewhere */ + TREE_PUBLIC(fndecl) = 1; /* Public */ + DECL_EXTERNAL(fndecl) = 1; /* External linkage */ + DECL_ARTIFICIAL(fndecl) = 1; /* Injected by compiler */ + + fcall = gimple_build_call(fndecl, 1, cur_loc); /* generate the function _call_ to above built reference, with *1* parameter -> the random const for the location */ + gimple_seq_add_stmt(&seq, fcall); /* and insert into a sequence */ + + /* Done - grab the entry to the block and insert sequence */ + bentry = gsi_start_bb(bb); + gsi_insert_seq_before(&bentry, seq, GSI_SAME_STMT); + + finst_blocks++; + } + fcnt_blocks--; /* discard the first in the count */ + + /* Say something nice. */ + if (!be_quiet) { + if (!finst_blocks) + WARNF(G_("No instrumentation targets found in " cBRI "%s" cRST ), + function_name(fun)); + else if (finst_blocks < fcnt_blocks) + OKF(G_("Instrumented %2u /%2u locations in " cBRI "%s" cRST ), + finst_blocks, fcnt_blocks, + function_name(fun)); + else + OKF(G_("Instrumented %2u locations in " cBRI "%s" cRST ), + finst_blocks, + function_name(fun)); + } + + return 0; +} + +static unsigned int inline_instrument(function *fun) { +#ifdef BUILD_INLINE_INST /* ifdef inline away, so I don't have to refactor it */ + /* Instrument all the things! */ + basic_block bb; + unsigned finst_blocks = 0; + unsigned fcnt_blocks = 0; + + /* Set up global type declarations */ + tree map_type = build_pointer_type(unsigned_char_type_node); + tree map_ptr_g = build_decl(UNKNOWN_LOCATION, VAR_DECL, get_identifier_with_length("__afl_area_ptr",14), map_type); + TREE_USED(map_ptr_g) = 1; + TREE_STATIC(map_ptr_g) = 1; /* Defined elsewhere */ + DECL_EXTERNAL(map_ptr_g) = 1; /* External linkage */ + DECL_PRESERVE_P(map_ptr_g) = 1; + DECL_ARTIFICIAL(map_ptr_g) = 1; + rest_of_decl_compilation(map_ptr_g, 1, 0); + + tree prev_loc_g = build_decl(UNKNOWN_LOCATION, VAR_DECL, get_identifier_with_length("__afl_prev_loc",14), uint16_type_node); + TREE_USED(prev_loc_g) = 1; + TREE_STATIC(prev_loc_g) = 1; /* Defined elsewhere */ + DECL_EXTERNAL(prev_loc_g) = 1; /* External linkage */ + DECL_PRESERVE_P(prev_loc_g) = 1; + DECL_ARTIFICIAL(prev_loc_g) = 1; + rest_of_decl_compilation(prev_loc_g, 1, 0); + + FOR_ALL_BB_FN(bb, fun) { + gimple *g; + gimple_seq seq = NULL; + gimple_stmt_iterator bentry; + + if (!fcnt_blocks++) continue; /* skip block 0 */ + + /* Bail on this block if we trip the specified ratio */ + if (R(100) >= inst_ratio) continue; + + /* Make up cur_loc */ + + unsigned int rand_loc = R(MAP_SIZE); + tree cur_loc = build_int_cst(uint64_type_node, rand_loc); + + /* Load prev_loc, xor with cur_loc */ + + tree area_off = create_tmp_var(uint64_type_node, "area_off"); + g = gimple_build_assign(area_off, BIT_XOR_EXPR, prev_loc_g, cur_loc); + gimple_seq_add_stmt(&seq, g); // area_off = prev_loc ^ cur_loc + + /* Update bitmap */ + +// tree zero = build_int_cst(unsigned_char_type_node, 0); + tree one = build_int_cst(unsigned_char_type_node, 1); + + tree tmp1 = create_tmp_var(map_type, "tmp1"); + g = gimple_build_assign(tmp1, PLUS_EXPR, map_ptr_g, area_off); + gimple_seq_add_stmt(&seq, g); // tmp1 = __afl_area_ptr + area_off + SAYF(G_("%d,"), fcnt_blocks); + + tree tmp2 = create_tmp_var(unsigned_char_type_node, "tmp2"); + //tree tmp1_ptr = build_simple_mem_ref_loc(UNKNOWN_LOCATION, tmp1); + g = gimple_build_assign(tmp2, INDIRECT_REF, tmp1); + gimple_seq_add_stmt(&seq, g); // tmp2 = *tmp1 + + tree tmp3 = create_tmp_var(unsigned_char_type_node, "tmp3"); + g = gimple_build_assign(tmp3, PLUS_EXPR, tmp2, one); + gimple_seq_add_stmt(&seq, g); // tmp3 = tmp2 + 1 + +// tree tmp4 = create_tmp_var(map_type, "tmp4"); +// g = gimple_build_assign(tmp4, PLUS_EXPR, map_ptr_g, area_off); +// gimple_seq_add_stmt(&seq, g); // tmp4 = __afl_area_ptr + area_off + +// tree deref2 = build2(MEM_REF, map_type, tmp4, zero); + tree deref2 = build4(ARRAY_REF, map_type, map_ptr_g, area_off, NULL, NULL); + g = gimple_build_assign(deref2, MODIFY_EXPR, tmp3); + gimple_seq_add_stmt(&seq, g); // *tmp4 = tmp3 + SAYF(G_("+%d,"), fcnt_blocks); + + /* Set prev_loc to cur_loc >> 1 */ + + tree shifted_loc = build_int_cst(TREE_TYPE(prev_loc_g), rand_loc >> 1); + g = gimple_build_assign(prev_loc_g, MODIFY_EXPR, shifted_loc); + gimple_seq_add_stmt(&seq, g); // __afl_pred_loc = cur_loc >> 1 + + /* Done - grab the entry to the block and insert sequence */ + + bentry = gsi_start_bb(bb); + gsi_insert_seq_before(&bentry, seq, GSI_SAME_STMT); + + inst_blocks++; + finst_blocks++; + } + + /* Say something nice. */ + if (!be_quiet) { + if (!finst_blocks) + WARNF(G_("No instrumentation targets found in " cBRI "%s" cRST ), + function_name(fun)); + else if (finst_blocks < fcnt_blocks) + OKF(G_("Instrumented %2u /%2u locations in " cBRI "%s" cRST ), + finst_blocks, fcnt_blocks, + function_name(fun)); + else + OKF(G_("Instrumented %2u locations in " cBRI "%s" cRST ), + finst_blocks, + function_name(fun)); + } + +#endif + return 0; +} + +/* -------------------------------------------------------------------------- */ +/* -- Boilerplate and initialization ---------------------------------------- */ + +static const struct pass_data afl_pass_data = { + + .type = GIMPLE_PASS, + .name = "afl-inst", + .optinfo_flags = OPTGROUP_NONE, + + .tv_id = TV_NONE, + .properties_required = 0, + .properties_provided = 0, + .properties_destroyed = 0, + .todo_flags_start = 0, + // NOTE(aseipp): it's very, very important to include + // at least 'TODO_update_ssa' here so that GCC will + // properly update the resulting SSA form, e.g., to + // include new PHI nodes for newly added symbols or + // names. Do not remove this. Do not taunt Happy Fun + // Ball. + .todo_flags_finish = TODO_update_ssa | TODO_verify_il | TODO_cleanup_cfg, +}; + +namespace { + +class afl_pass : public gimple_opt_pass { +private: + bool do_ext_call; + +public: + afl_pass(bool ext_call, gcc::context *g) : gimple_opt_pass(afl_pass_data, g), do_ext_call(ext_call) {} + + virtual unsigned int execute(function *fun) { + // TODO: implement whitelist feature here :TODO + return do_ext_call ? ext_call_instrument(fun) : inline_instrument(fun); + } +}; /* class afl_pass */ + +} /* anon namespace */ + +static struct opt_pass *make_afl_pass(bool ext_call, gcc::context *ctxt) { + return new afl_pass(ext_call, ctxt); +} + +/* -------------------------------------------------------------------------- */ +/* -- Initialization -------------------------------------------------------- */ + +int plugin_is_GPL_compatible = 1; + +static struct plugin_info afl_plugin_info = { + .version = "20181200", + .help = "AFL gcc plugin\n", +}; + +int plugin_init(struct plugin_name_args *plugin_info, + struct plugin_gcc_version *version) { + + struct register_pass_info afl_pass_info; + struct timeval tv; + struct timezone tz; + u32 rand_seed; + + /* Setup random() so we get Actually Random(TM) outputs from R() */ + gettimeofday(&tv, &tz); + rand_seed = tv.tv_sec ^ tv.tv_usec ^ getpid(); + srandom(rand_seed); + + /* Pass information */ + afl_pass_info.pass = make_afl_pass(inst_ext, g); + afl_pass_info.reference_pass_name = "ssa"; + afl_pass_info.ref_pass_instance_number = 1; + afl_pass_info.pos_op = PASS_POS_INSERT_AFTER; + + if (!plugin_default_version_check(version, &gcc_version)) { + FATAL(G_("Incompatible gcc/plugin versions!")); + } + + /* Show a banner */ + if (isatty(2) && !getenv("AFL_QUIET")) { + SAYF(G_(cCYA "afl-gcc-pass" cRST " initial version 1.94 by , updated to " cBRI VERSION cRST " by \n")); + } else + be_quiet = 1; + + /* Decide instrumentation ratio */ + char* inst_ratio_str = getenv("AFL_INST_RATIO"); + + if (inst_ratio_str) { + if (sscanf(inst_ratio_str, "%u", &inst_ratio) != 1 || !inst_ratio || inst_ratio > 100) + FATAL(G_("Bad value of AFL_INST_RATIO (must be between 1 and 100)")); + else { + if (!be_quiet) + ACTF(G_("%s instrumentation at ratio of %u%% in %s mode."), + inst_ext ? G_("Call-based") : G_("Inline"), + inst_ratio, + getenv("AFL_HARDEN") ? G_("hardened") : G_("non-hardened")); + } + } + + /* Go go gadget */ + register_callback(plugin_info->base_name, PLUGIN_INFO, NULL, &afl_plugin_info); + register_callback(plugin_info->base_name, PLUGIN_PASS_MANAGER_SETUP, NULL, &afl_pass_info); + return 0; +} diff --git a/gcc_plugin/afl-gcc-rt.o.c b/gcc_plugin/afl-gcc-rt.o.c new file mode 100644 index 00000000..4e238fa8 --- /dev/null +++ b/gcc_plugin/afl-gcc-rt.o.c @@ -0,0 +1,224 @@ +/* + american fuzzy lop - GCC plugin instrumentation bootstrap + --------------------------------------------------------- + + Written by Austin Seipp and + Laszlo Szekeres and + Michal Zalewski + + GCC integration design is based on the LLVM design, which comes + from Laszlo Szekeres. + + Copyright 2015 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 + + This code is the rewrite of afl-as.h's main_payload. + +*/ + +#include "../config.h" +#include "../types.h" + +#include +#include +#include +#include + +#include +#include +#include +#include + + +/* Globals needed by the injected instrumentation. The __afl_area_initial region + is used for instrumentation output before __afl_map_shm() has a chance to run. + It will end up as .comm, so it shouldn't be too wasteful. */ + +u8 __afl_area_initial[MAP_SIZE]; +u8* __afl_area_ptr = __afl_area_initial; +u16 __afl_prev_loc; + + +/* Running in persistent mode? */ + +static u8 is_persistent; + +/* Trace a basic block with some ID */ +void __afl_trace(u16 x) { + u16 l = __afl_prev_loc; + u16 n = l ^ x; + *(__afl_area_ptr+n) += 1; + __afl_prev_loc = (x >> 1); + return; +} + +/* SHM setup. */ + +static void __afl_map_shm(void) { + + u8 *id_str = getenv(SHM_ENV_VAR); + + /* If we're running under AFL, attach to the appropriate region, replacing the + early-stage __afl_area_initial region that is needed to allow some really + hacky .init code to work correctly in projects such as OpenSSL. */ + + if (id_str) { + + u32 shm_id = atoi(id_str); + + __afl_area_ptr = shmat(shm_id, NULL, 0); + + /* Whooooops. */ + + if (__afl_area_ptr == (void *)-1) exit(1); + + /* Write something into the bitmap so that even with low AFL_INST_RATIO, + our parent doesn't give up on us. */ + + __afl_area_ptr[0] = 1; + + } + +} + + +/* Fork server logic. */ + +static void __afl_start_forkserver(void) { + + static u8 tmp[4]; + s32 child_pid; + + u8 child_stopped = 0; + + /* Phone home and tell the parent that we're OK. If parent isn't there, + assume we're not running in forkserver mode and just execute program. */ + + if (write(FORKSRV_FD + 1, tmp, 4) != 4) return; + + while (1) { + + u32 was_killed; + int status; + + /* Wait for parent by reading from the pipe. Abort if read fails. */ + + if (read(FORKSRV_FD, &was_killed, 4) != 4) exit(1); + + /* If we stopped the child in persistent mode, but there was a race + condition and afl-fuzz already issued SIGKILL, write off the old + process. */ + + if (child_stopped && was_killed) { + child_stopped = 0; + if (waitpid(child_pid, &status, 0) < 0) exit(1); + } + + if (!child_stopped) { + + /* Once woken up, create a clone of our process. */ + + child_pid = fork(); + if (child_pid < 0) exit(1); + + /* In child process: close fds, resume execution. */ + + if (!child_pid) { + + close(FORKSRV_FD); + close(FORKSRV_FD + 1); + return; + + } + + } else { + + /* Special handling for persistent mode: if the child is alive but + currently stopped, simply restart it with SIGCONT. */ + + kill(child_pid, SIGCONT); + child_stopped = 0; + + } + + /* In parent process: write PID to pipe, then wait for child. */ + + if (write(FORKSRV_FD + 1, &child_pid, 4) != 4) exit(1); + + if (waitpid(child_pid, &status, is_persistent ? WUNTRACED : 0) < 0) + exit(1); + + /* In persistent mode, the child stops itself with SIGSTOP to indicate + a successful run. In this case, we want to wake it up without forking + again. */ + + if (WIFSTOPPED(status)) child_stopped = 1; + + /* Relay wait status to pipe, then loop back. */ + + if (write(FORKSRV_FD + 1, &status, 4) != 4) exit(1); + + } + +} + + +/* A simplified persistent mode handler, used as explained in README.llvm. */ + +int __afl_persistent_loop(unsigned int max_cnt) { + + static u8 first_pass = 1; + static u32 cycle_cnt; + + if (first_pass) { + + cycle_cnt = max_cnt; + first_pass = 0; + return 1; + + } + + if (is_persistent && --cycle_cnt) { + + raise(SIGSTOP); + return 1; + + } else return 0; + +} + + +/* This one can be called from user code when deferred forkserver mode + is enabled. */ + +void __afl_manual_init(void) { + + static u8 init_done; + + if (!init_done) { + + __afl_map_shm(); + __afl_start_forkserver(); + init_done = 1; + + } + +} + + +/* Proper initialization routine. */ + +__attribute__((constructor(101))) void __afl_auto_init(void) { + + is_persistent = !!getenv(PERSIST_ENV_VAR); + + if (getenv(DEFER_ENV_VAR)) return; + + __afl_manual_init(); + +} -- cgit 1.4.1 From f697752b52e619bdda2a5b629568e56c8f28fe47 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Sun, 21 Jul 2019 20:24:40 +0200 Subject: moved gcc_plugin to a branch, it is nowhere near "ok" --- Makefile | 2 +- afl-gcc.c | 2 +- docs/ChangeLog | 3 --- docs/PATCHES | 1 - docs/README | 3 +-- 5 files changed, 3 insertions(+), 8 deletions(-) (limited to 'docs/README') diff --git a/Makefile b/Makefile index 69bd7535..d2840848 100644 --- a/Makefile +++ b/Makefile @@ -193,7 +193,7 @@ install: all install -m 755 $(PROGS) $(SH_PROGS) $${DESTDIR}$(BIN_PATH) rm -f $${DESTDIR}$(BIN_PATH)/afl-as if [ -f afl-qemu-trace ]; then install -m 755 afl-qemu-trace $${DESTDIR}$(BIN_PATH); fi - if [ -f afl-gcc-fast ]; then set e; install -m 755 afl-gcc-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-gcc-fast $${DESTDIR}$(BIN_PATH)/afl-g++-fast; install -m 755 afl-gcc-pass.so afl-gcc-rt.o $${DESTDIR}$(HELPER_PATH); fi + #if [ -f afl-gcc-fast ]; then set e; install -m 755 afl-gcc-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-gcc-fast $${DESTDIR}$(BIN_PATH)/afl-g++-fast; install -m 755 afl-gcc-pass.so afl-gcc-rt.o $${DESTDIR}$(HELPER_PATH); fi ifndef AFL_TRACE_PC if [ -f afl-clang-fast -a -f libLLVMInsTrim.so -a -f afl-llvm-rt.o ]; then set -e; install -m 755 afl-clang-fast $${DESTDIR}$(BIN_PATH); ln -sf afl-clang-fast $${DESTDIR}$(BIN_PATH)/afl-clang-fast++; install -m 755 libLLVMInsTrim.so afl-llvm-pass.so afl-llvm-rt.o $${DESTDIR}$(HELPER_PATH); fi else diff --git a/afl-gcc.c b/afl-gcc.c index d109f6e2..f6ededeb 100644 --- a/afl-gcc.c +++ b/afl-gcc.c @@ -308,7 +308,7 @@ int main(int argc, char** argv) { if (isatty(2) && !getenv("AFL_QUIET")) { SAYF(cCYA "afl-cc" VERSION cRST " by \n"); - SAYF(cYEL "[!] " cBRI "WARNING: " cRST "afl-gcc is deprecated, gcc_plugin is faster, llvm_mode even faster\n"); + SAYF(cYEL "[!] " cBRI "NOTE: " cRST "afl-gcc is deprecated, llvm_mode is much faster and has more options\n"); } else be_quiet = 1; diff --git a/docs/ChangeLog b/docs/ChangeLog index 44eb4012..120893d9 100644 --- a/docs/ChangeLog +++ b/docs/ChangeLog @@ -24,9 +24,6 @@ Version ++2.52d (tbd): - all queue, hang and crash files now have their discovery time in their name - if llvm_mode was compiled, afl-clang/afl-clang++ will point to these instead of afl-gcc - - added gcc_plugin which is like llvm_mode but for gcc. This version - supports gcc version 5 to 8. See gcc_plugin/README.gcc - (https://github.com/T12z/afl) - added instrim, a much faster llvm_mode instrumentation at the cost of path discovery. See llvm_mode/README.instrim (https://github.com/csienslab/instrim) - added MOpt (github.com/puppet-meteor/MOpt-AFL) mode, see docs/README.MOpt diff --git a/docs/PATCHES b/docs/PATCHES index 78050208..f6ca9284 100644 --- a/docs/PATCHES +++ b/docs/PATCHES @@ -17,7 +17,6 @@ afl-qemu-optimize-entrypoint.diff by mh(at)mh-sec(dot)de afl-qemu-speed.diff by abiondo on github afl-qemu-optimize-map.diff by mh(at)mh-sec(dot)de -+ gcc_plugin (https://github.com/T12z/afl) + instrim (https://github.com/csienslab/instrim) was integrated + MOpt (github.com/puppet-meteor/MOpt-AFL) was imported + AFLfast additions (github.com/mboehme/aflfast) were incorporated. diff --git a/docs/README b/docs/README index a8ec715b..b606f6cf 100644 --- a/docs/README +++ b/docs/README @@ -10,6 +10,7 @@ american fuzzy lop plus plus and Andrea Fioraldi as there have been no updates to afl since November 2017. + Many improvements were made, e.g. more performant llvm_mode, supporting llvm up to version 8, Qemu 3.1, more speed for Qemu, etc. @@ -25,8 +26,6 @@ american fuzzy lop plus plus Also newly integrated is instrim, a very effective CFG llvm_mode instrumentation implementation from https://github.com/csienslab/instrim - Similar to llvm_mode a gcc_plugin that supports versions 5 to 8 is - now available (from https://github.com/T12z/afl) A more thorough list is available in the PATCHES file. -- cgit 1.4.1 From 3789a56225354b1ac13586dc4ef29c1c50fd4af1 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Tue, 23 Jul 2019 17:04:04 +0200 Subject: updated changelog and readme --- docs/ChangeLog | 3 +++ docs/README | 3 ++- 2 files changed, 5 insertions(+), 1 deletion(-) (limited to 'docs/README') diff --git a/docs/ChangeLog b/docs/ChangeLog index 120893d9..0b4927bc 100644 --- a/docs/ChangeLog +++ b/docs/ChangeLog @@ -17,6 +17,9 @@ sending a mail to . Version ++2.52d (tbd): ----------------------------- + - fix a crash in qemu_mode which also exists in stock afl + - added libcompcov, a laf-intel implementation for qemu! :) + see qemu_mode/libcompcov/README.libcompcov - afl-fuzz now displays the selected core in the status screen (blue {#}) - updated afl-fuzz and afl-system-config for new scaling governor location in modern kernels diff --git a/docs/README b/docs/README index b606f6cf..d0ffe369 100644 --- a/docs/README +++ b/docs/README @@ -12,7 +12,8 @@ american fuzzy lop plus plus Many improvements were made, e.g. more performant llvm_mode, supporting - llvm up to version 8, Qemu 3.1, more speed for Qemu, etc. + llvm up to version 8, Qemu 3.1, more speed and crashfixes for Qemu, + laf-intel feature for Qemu (with libcompcov) etc. Additionally AFLfast's power schedules by Marcel Boehme from https://github.com/mboehme/aflfast have been incorporated. -- cgit 1.4.1 From 9246f21f2a75fbe4113dd7340f870679a7953b24 Mon Sep 17 00:00:00 2001 From: Andrea Fioraldi Date: Wed, 24 Jul 2019 15:54:05 +0200 Subject: remove the unreadable AFLFast schedules tabel in markdown from README --- docs/README | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-) (limited to 'docs/README') diff --git a/docs/README b/docs/README index d0ffe369..5e0c4419 100644 --- a/docs/README +++ b/docs/README @@ -7,8 +7,8 @@ american fuzzy lop plus plus Repository: https://github.com/vanhauser-thc/AFLplusplus afl++ is maintained by Marc Heuse , Heiko Eissfeldt - and Andrea Fioraldi as there have been no - updates to afl since November 2017. + and Andrea Fioraldi as + there have been no updates to afl since November 2017. Many improvements were made, e.g. more performant llvm_mode, supporting @@ -194,15 +194,14 @@ The power schedules were copied from Marcel Böhme's excellent AFLfast implementation and expands on the ability to discover new paths and therefore the coverage. -| AFL flag | Power Schedule | -| ------------- | -------------------------- | -| `-p explore` (default)| ![EXPLORE](http://latex.codecogs.com/gif.latex?p%28i%29%3D%5Cfrac%7B%5Calpha%28i%29%7D%7B%5Cbeta%7D) | -| `-p fast` | ![FAST](http://latex.codecogs.com/gif.latex?p(i)=\\min\\left(\\frac{\\alpha(i)}{\\beta}\\cdot\\frac{2^{s(i)}}{f(i)},M\\right)) | -| `-p coe` | ![COE](http://latex.codecogs.com/gif.latex?p%28i%29%3D%5Cbegin%7Bcases%7D%200%20%26%20%5Ctext%7B%20if%20%7D%20f%28i%29%20%3E%20%5Cmu%5C%5C%20%5Cmin%5Cleft%28%5Cfrac%7B%5Calpha%28i%29%7D%7B%5Cbeta%7D%5Ccdot%202%5E%7Bs%28i%29%7D%2C%20M%5Cright%29%20%26%20%5Ctext%7B%20otherwise.%7D%20%5Cend%7Bcases%7D) | -| `-p quad` | ![QUAD](http://latex.codecogs.com/gif.latex?p%28i%29%20%3D%20%5Cmin%5Cleft%28%5Cfrac%7B%5Calpha%28i%29%7D%7B%5Cbeta%7D%5Ccdot%5Cfrac%7Bs%28i%29%5E2%7D%7Bf%28i%29%7D%2CM%5Cright%29) | -| `-p lin` | ![LIN](http://latex.codecogs.com/gif.latex?p%28i%29%20%3D%20%5Cmin%5Cleft%28%5Cfrac%7B%5Calpha%28i%29%7D%7B%5Cbeta%7D%5Ccdot%5Cfrac%7Bs%28i%29%7D%7Bf%28i%29%7D%2CM%5Cright%29) | -| `-p exploit` (AFL) | ![LIN](http://latex.codecogs.com/gif.latex?p%28i%29%20%3D%20%5Calpha%28i%29) | -where *α(i)* is the performance score that AFL uses to compute for the seed input *i*, *β(i)>1* is a constant, *s(i)* is the number of times that seed *i* has been chosen from the queue, *f(i)* is the number of generated inputs that exercise the same path as seed *i*, and *μ* is the average number of generated inputs exercising a path. +The available schedules are: + + - explore (default) + - fast + - coe + - quad + - lin + - exploit In parallel mode (-M/-S, several instances with shared queue), we suggest to run the master using the exploit schedule (-p exploit) and the slaves with a @@ -212,11 +211,12 @@ and explore (-p explore) schedules. In single mode, using -p fast is usually more beneficial than the default explore mode. (We don't want to change the default behaviour of afl, so "fast" has not been -made the default mode) - -More details can be found in the paper: -[23rd ACM Conference on Computer and Communications Security (CCS'16)](https://www.sigsac.org/ccs/CCS2016/accepted-papers/). +made the default mode). +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/ 6) Choosing initial test cases ------------------------------ -- cgit 1.4.1 From dfb3bd8e33b4f38188ebc67fd522225b111044c9 Mon Sep 17 00:00:00 2001 From: van Hauser Date: Thu, 25 Jul 2019 09:00:22 +0200 Subject: documentation update --- docs/ChangeLog | 3 ++- docs/PATCHES | 1 + docs/README | 4 ++++ docs/binaryonly_fuzzing.txt | 16 +++++++++++++++- unicorn_mode/README.md | 13 +++++++++---- 5 files changed, 31 insertions(+), 6 deletions(-) (limited to 'docs/README') diff --git a/docs/ChangeLog b/docs/ChangeLog index 0b4927bc..d393f92c 100644 --- a/docs/ChangeLog +++ b/docs/ChangeLog @@ -14,9 +14,10 @@ sending a mail to . ----------------------------- -Version ++2.52d (tbd): +Version ++2.52d (dev): ----------------------------- + - unicorn_mode got added - thank to domenukk for the patch! - fix a crash in qemu_mode which also exists in stock afl - added libcompcov, a laf-intel implementation for qemu! :) see qemu_mode/libcompcov/README.libcompcov diff --git a/docs/PATCHES b/docs/PATCHES index f6ca9284..8b188814 100644 --- a/docs/PATCHES +++ b/docs/PATCHES @@ -17,6 +17,7 @@ afl-qemu-optimize-entrypoint.diff by mh(at)mh-sec(dot)de afl-qemu-speed.diff by abiondo on github afl-qemu-optimize-map.diff by mh(at)mh-sec(dot)de ++ unicorn_mode (modernized and updated by domenukk) + instrim (https://github.com/csienslab/instrim) was integrated + MOpt (github.com/puppet-meteor/MOpt-AFL) was imported + AFLfast additions (github.com/mboehme/aflfast) were incorporated. diff --git a/docs/README b/docs/README index 5e0c4419..c2c93f38 100644 --- a/docs/README +++ b/docs/README @@ -27,6 +27,10 @@ american fuzzy lop plus plus Also newly integrated is instrim, a very effective CFG llvm_mode instrumentation implementation from https://github.com/csienslab/instrim + And finally the newest addition is the unicorn_mode which allows fuzzing + of binaries from completely different platforms - provided by domenukk! + The unicorn afl mode is not the stock version but like afl++ contains + various patches from forks that make it better :) A more thorough list is available in the PATCHES file. diff --git a/docs/binaryonly_fuzzing.txt b/docs/binaryonly_fuzzing.txt index 04e449c0..fd260450 100644 --- a/docs/binaryonly_fuzzing.txt +++ b/docs/binaryonly_fuzzing.txt @@ -12,7 +12,7 @@ The following is a description of how these can be fuzzed with afl++ !!!!! TL;DR: try DYNINST with afl-dyninst. If it produces too many crashes then - use afl -Q qemu_mode. + use afl -Q qemu_mode, or better: use both in parallel. !!!!! @@ -121,6 +121,20 @@ Pintool solutions: https://github.com/spinpx/afl_pin_mode <= only old Pintool version supported +Non-AFL solutions +----------------- + +There are many binary-only fuzzing frameworks. Some are great for CTFs but don't +work with large binaries, other are very slow but have good path discovery, +some are very hard to set-up ... + +QSYM: https://github.com/sslab-gatech/qsym +Manticore: https://github.com/trailofbits/manticore +S2E: https://github.com/S2E + + + + That's it! News, corrections, updates? Email vh@thc.org diff --git a/unicorn_mode/README.md b/unicorn_mode/README.md index 5d8b3ce0..24151387 100644 --- a/unicorn_mode/README.md +++ b/unicorn_mode/README.md @@ -7,10 +7,15 @@ ``` -afl-unicorn lets you fuzz any piece of binary that can be emulated by [Unicorn Engine](http://www.unicorn-engine.org/). +afl-unicorn lets you fuzz any piece of binary that can be emulated by +[Unicorn Engine](http://www.unicorn-engine.org/). -For an in-depth description of what this is, how to install it, and how to use it check out this [blog post](https://medium.com/@njvoss299/afl-unicorn-fuzzing-arbitrary-binary-code-563ca28936bf). +For the full readme please see docs/unicorn_mode.txt -For general help with AFL, please refer to both the official [AFL website](http://lcamtuf.coredump.cx/afl/) and the documents in the /doc/ directory. +For an in-depth description of what this is, how to install it, and how to use +it check out this [blog post](https://medium.com/@njvoss299/afl-unicorn-fuzzing-arbitrary-binary-code-563ca28936bf). -Created by Nathan Voss, originally funded by [Battelle](https://www.battelle.org/cyber). \ No newline at end of file +For general help with AFL, please refer to the documents in the /doc/ directory. + +Created by Nathan Voss, originally funded by +[Battelle](https://www.battelle.org/cyber). -- cgit 1.4.1 From 8f4f45c524d217236a2e64be0d95d0a6de11df9c Mon Sep 17 00:00:00 2001 From: van Hauser Date: Fri, 26 Jul 2019 10:35:58 +0200 Subject: incorporated most of the 2.53b changes --- README | 1 - README.md | 607 ++++++++++++++++++++++++++++++++++++++++++++++++++ afl-fuzz.c | 2 + docs/README | 592 ------------------------------------------------ llvm_mode/README.llvm | 6 +- types.h | 2 +- 6 files changed, 612 insertions(+), 598 deletions(-) delete mode 120000 README create mode 100644 README.md delete mode 100644 docs/README (limited to 'docs/README') diff --git a/README b/README deleted file mode 120000 index a90f4af9..00000000 --- a/README +++ /dev/null @@ -1 +0,0 @@ -docs/README \ No newline at end of file diff --git a/README.md b/README.md new file mode 100644 index 00000000..e1371175 --- /dev/null +++ b/README.md @@ -0,0 +1,607 @@ +# american fuzzy lop plus plus (afl++) + + Originally developed by Michal "lcamtuf" Zalewski. + + Repository: [https://github.com/vanhauser-thc/AFLplusplus](https://github.com/vanhauser-thc/AFLplusplus) + + afl++ is maintained by Marc Heuse , Heiko Eissfeldt + and Andrea Fioraldi . + +## The enhancements compared to the original stock afl + + Many improvements were made over the official afl release - which did not + get any improvements since November 2017. + + Among others afl++ has, e.g. more performant llvm_mode, supporting + llvm up to version 8, Qemu 3.1, more speed and crashfixes for Qemu, + laf-intel feature for Qemu (with libcompcov) and more. + + Additionally the following patches have been integrated: + + * AFLfast's power schedules by Marcel Boehme: [https://github.com/mboehme/aflfast](https://github.com/mboehme/aflfast) + + * C. Hollers afl-fuzz Python mutator module and llvm_mode whitelist support: [https://github.com/choller/afl](https://github.com/choller/afl) + + * 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) + + * unicorn_mode which allows fuzzing of binaries from completely different platforms (integration provided by domenukk) + + A more thorough list is available in the PATCHES file. + + So all in all this is the best-of AFL that is currently out there :-) + + For new versions and additional information, check out: + [https://github.com/vanhauser-thc/AFLplusplus](https://github.com/vanhauser-thc/AFLplusplus) + + To compare notes with other users or get notified about major new features, + send a mail to . + + See [docs/QuickStartGuide.txt](docs/QuickStartGuide.txt) if you don't have time to + read this file. + + +## 1) Challenges of guided fuzzing +------------------------------- + +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. + +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. + +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 program state, and are less +useful for guiding the fuzzing effort in the long haul. + +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. + + +## 2) The afl-fuzz approach + +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. + +Simplifying a bit, the overall algorithm can be summed up as: + + 1) Load user-supplied initial test cases into the queue, + + 2) Take next input file from the queue, + + 3) Attempt to trim the test case to the smallest size that doesn't alter + the measured behavior of the program, + + 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. + + +## 3) Instrumenting programs for use with AFL + +PLEASE NOTE: llvm_mode compilation with afl-clang-fast/afl-clang-fast++ +instead of afl-gcc/afl-g++ is much faster and has a few cool features. +See llvm_mode/ - however few code does not compile with llvm. +We support llvm versions 4.0 to 8. + +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. + +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. + +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: + +```shell +$ CC=/path/to/afl/afl-gcc ./configure +$ make clean all +``` + +For C++ programs, you'd would also want to set `CXX=/path/to/afl/afl-g++`. + +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.llvm](llvm_mode/README.llvm). +Clang/LLVM has a much better performance and works with LLVM version 4.0 to 8. + +Using the LAF Intel performance enhancements are also recommended, see +[llvm_mode/README.laf-intel](llvm_mode/README.laf-intel) + +Using partial instrumentation is also recommended, see +[llvm_mode/README.whitelist](llvm_mode/README.whitelist) + +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: + +```shell +$ CC=/path/to/afl/afl-gcc ./configure --disable-shared +``` + +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.dislocator](libdislocator/README.dislocator)) can help uncover heap corruption issues, too. + +PS. ASAN users are advised to review [docs/notes_for_asan.txt](docs/notes_for_asan.txt) +file for important caveats. + + +## 4) Instrumenting binary-only apps +--------------------------------- + +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. + +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.qemu](qemu_mode/README.qemu). + +The mode is approximately 2-5x slower than compile-time instrumentation, is +less conductive to parallelization, and may have some other quirks. + +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. + +A more comprehensive description of these and other options can be found in +[docs/binaryonly_fuzzing.txt](docs/binaryonly_fuzzing.txt) + + +## 5) Power schedules +------------------ + +The power schedules were copied from Marcel Böhme's excellent AFLfast +implementation and expands on the ability to discover new paths and +therefore the coverage. + +The available schedules are: + + - explore (default) + - fast + - coe + - quad + - lin + - exploit + +In parallel mode (-M/-S, several instances with shared queue), we suggest to +run the master using the exploit schedule (-p exploit) and the slaves with a +combination of cut-off-exponential (-p coe), exponential (-p fast; default), +and explore (-p explore) schedules. + +In single mode, using -p fast is usually more beneficial than the default +explore mode. +(We don't want to change the default behaviour of afl, so "fast" has not been +made the default mode). + +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/)[https://www.sigsac.org/ccs/CCS2016/accepted-papers/] + + +## 6) 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 +application. There are two basic rules: + + - Keep the files small. Under 1 kB is ideal, although not strictly necessary. + For a discussion of why size matters, see [perf_tips.txt](docs/perf_tips.txt). + + - Use multiple test cases only if they are functionally different from + each other. There is no point in using fifty different vacation photos + to fuzz an image library. + +You can find many good examples of starting files in the testcases/ subdirectory +that comes with this tool. + +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. + + +## 7) 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.txt](docs/perf_tips.txt). + +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. + + +## 8) Interpreting output +---------------------- + +See the [docs/status_screen.txt](docs/status_screen.txt) file for information on +how to interpret the displayed stats and monitor the health of the process. Be +sure to consult this file especially if any UI elements are highlighted in red. + +The fuzzing process will continue until you press Ctrl-C. At minimum, you want +to allow the fuzzer to complete one queue cycle, which may take anywhere from a +couple of hours to a week or so. + +There are three subdirectories created within the output directory and updated +in real time: + + - queue/ - test cases for every distinctive execution path, plus all the + starting files given by the user. This is the synthesized corpus + mentioned in section 2. + + Before using this corpus for any other purposes, you can shrink + it to a smaller size using the afl-cmin tool. The tool will find + a smaller subset of files offering equivalent edge coverage. + + - crashes/ - unique test cases that cause the tested program to receive a + fatal signal (e.g., SIGSEGV, SIGILL, SIGABRT). The entries are + grouped by the received signal. + + - hangs/ - unique test cases that cause the tested program to time out. The + default time limit before something is classified as a hang is + the larger of 1 second and the value of the -t parameter. + The value can be fine-tuned by setting AFL_HANG_TMOUT, but this + is rarely necessary. + +Crashes and hangs are considered "unique" if the associated execution paths +involve any state transitions not seen in previously-recorded faults. If a +single bug can be reached in multiple ways, there will be some count inflation +early in the process, but this should quickly taper off. + +The file names for crashes and hangs are correlated with parent, non-faulting +queue entries. This should help with debugging. + +When you can't reproduce a crash found by afl-fuzz, the most likely cause is +that you are not setting the same memory limit as used by the tool. Try: + +```shell +$ LIMIT_MB=50 +$ ( ulimit -Sv $[LIMIT_MB << 10]; /path/to/tested_binary ... ) +``` + +Change LIMIT_MB to match the -m parameter passed to afl-fuzz. On OpenBSD, +also change -Sv to -Sd. + +Any existing output directory can be also used to resume aborted jobs; try: + +```shell +$ ./afl-fuzz -i- -o existing_output_dir [...etc...] +``` + +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/). + + +## 9) 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 [parallel_fuzzing.txt](docs/parallel_fuzzing.txt). + +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 [parallel_fuzzing.txt](docs/parallel_fuzzing.txt) for tips. + + +## 10) 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.dictionaries](ictionaries/README.dictionaries); +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.tokencap](libtokencap/README.tokencap). + + +## 11) 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. +Every crash is also traceable to its parent non-crashing test case in the +queue, making it easier to diagnose faults. + +Having said that, it's important to acknowledge that some fuzzing crashes can be +difficult to quickly evaluate for exploitability without a lot of debugging and +code analysis work. To assist with this task, afl-fuzz supports a very unique +"crash exploration" mode enabled with the -C flag. + +In this mode, the fuzzer takes one or more crashing test cases as the input, +and uses its feedback-driven fuzzing strategies to very quickly enumerate all +code paths that can be reached in the program while keeping it in the +crashing state. + +Mutations that do not result in a crash are rejected; so are any changes that +do not affect the execution path. + +The output is a small corpus of files that can be very rapidly examined to see +what degree of control the attacker has over the faulting address, or whether +it is possible to get past an initial out-of-bounds read - and see what lies +beneath. + +Oh, one more thing: for test case minimization, give afl-tmin a try. The tool +can be operated in a very simple way: + +```shell +$ ./afl-tmin -i test_case -o minimized_result -- /path/to/program [...] +``` + +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 +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 +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 +insights into complex file formats. More info about its operation can be found +near the end of [docs/technical_details.txt](docs/technical_details.txt). + + +## 12) Going beyond crashes +------------------------ + +Fuzzing is a wonderful and underutilized technique for discovering non-crashing +design and implementation errors, too. Quite a few interesting bugs have been +found by modifying the target programs to call abort() when, say: + + - Two bignum libraries produce different outputs when given the same + fuzzer-generated input, + + - An image library produces different outputs when asked to decode the same + input image several times in a row, + + - A serialization / deserialization library fails to produce stable outputs + when iteratively serializing and deserializing fuzzer-supplied data, + + - A compression library produces an output inconsistent with the input file + when asked to compress and then decompress a particular blob. + +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). + + +## 13) Common-sense risks +---------------------- + +Please keep in mind that, similarly to many other computationally-intensive +tasks, fuzzing may put strain on your hardware and on the OS. In particular: + + - Your CPU will run hot and will need adequate cooling. In most cases, if + cooling is insufficient or stops working properly, CPU speeds will be + automatically throttled. That said, especially when fuzzing on less + suitable hardware (laptops, smartphones, etc), it's not entirely impossible + 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 + 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. + + - Fuzzing involves billions of reads and writes to the filesystem. On modern + systems, this will be usually heavily cached, resulting in fairly modest + "physical" I/O - but there are many factors that may alter this equation. + It is your responsibility to monitor for potential trouble; with very heavy + I/O, the lifespan of many HDDs and SSDs may be reduced. + + A good way to monitor disk I/O on Linux is the 'iostat' command: + +```shell + $ iostat -d 3 -x -k [...optional disk ID...] +``` + + +## 14) Known limitations & areas for improvement +--------------------------------------------- + +Here are some of the most important caveats for AFL: + + - 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 processed spawned by the fuzzed target may evade + detection unless you manually add some code to catch that. + + - As with any other brute-force tool, the fuzzer offers limited coverage if + encryption, checksums, cryptographic signatures, or compression are used to + wholly wrap the actual data format to be tested. + + To work around this, you can comment out the relevant checks (see + experimental/libpng_no_checksum/ for inspiration); if this is not possible, + you can also write a postprocessor, as explained in + experimental/post_library/ (with AFL_POST_LIBRARY) + + - There are some unfortunate trade-offs with ASAN and 64-bit binaries. This + isn't due to any specific fault of afl-fuzz; see [docs/notes_for_asan.txt](docs/notes_for_asan.txt) + for tips. + + - There is no direct support for fuzzing network services, background + daemons, or interactive apps that require UI interaction to work. You may + need to make simple code changes to make them behave in a more traditional + way. Preeny may offer a relatively simple option, too - see: + [https://github.com/zardus/preeny](https://github.com/zardus/preeny) + + 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/). + +Beyond this, see INSTALL for platform-specific tips. + + +## 15) Special thanks +------------------ + +Many of the improvements to the original afl wouldn't be possible without +feedback, bug reports, or patches from: + +``` + Jann Horn Hanno Boeck + Felix Groebert Jakub Wilk + Richard W. M. Jones Alexander Cherepanov + Tom Ritter Hovik Manucharyan + Sebastian Roschke Eberhard Mattes + Padraig Brady Ben Laurie + @dronesec Luca Barbato + Tobias Ospelt Thomas Jarosch + Martin Carpenter Mudge Zatko + Joe Zbiciak Ryan Govostes + Michael Rash William Robinet + Jonathan Gray Filipe Cabecinhas + Nico Weber Jodie Cunningham + Andrew Griffiths Parker Thompson + Jonathan Neuschfer Tyler Nighswander + Ben Nagy Samir Aguiar + Aidan Thornton Aleksandar Nikolich + Sam Hakim Laszlo Szekeres + David A. Wheeler Turo Lamminen + Andreas Stieger Richard Godbee + Louis Dassy teor2345 + Alex Moneger Dmitry Vyukov + Keegan McAllister Kostya Serebryany + Richo Healey Martijn Bogaard + rc0r Jonathan Foote + Christian Holler Dominique Pelle + Jacek Wielemborek Leo Barnes + Jeremy Barnes Jeff Trull + Guillaume Endignoux ilovezfs + Daniel Godas-Lopez Franjo Ivancic + Austin Seipp Daniel Komaromy + Daniel Binderman Jonathan Metzman + Vegard Nossum Jan Kneschke + Kurt Roeckx Marcel Bohme + Van-Thuan Pham Abhik Roychoudhury + Joshua J. Drake Toby Hutton + Rene Freingruber Sergey Davidoff + Sami Liedes Craig Young + Andrzej Jackowski Daniel Hodson +``` + +Thank you! + + +## 16) Contact +----------- + +Questions? Concerns? Bug reports? The contributors can be reached via +[https://github.com/vanhauser-thc/AFLplusplus](https://github.com/vanhauser-thc/AFLplusplus) + +There is also a mailing list for the afl project; to join, send a mail to +. 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/afl-fuzz.c b/afl-fuzz.c index 8f4e1344..f974268f 100644 --- a/afl-fuzz.c +++ b/afl-fuzz.c @@ -23,7 +23,9 @@ #define AFL_MAIN #define MESSAGES_TO_STDOUT +#ifndef _GNU_SOURCE #define _GNU_SOURCE +#endif #define _FILE_OFFSET_BITS 64 #include "config.h" diff --git a/docs/README b/docs/README deleted file mode 100644 index c2c93f38..00000000 --- a/docs/README +++ /dev/null @@ -1,592 +0,0 @@ -============================ -american fuzzy lop plus plus -============================ - - Originally written by Michal Zalewski - - Repository: https://github.com/vanhauser-thc/AFLplusplus - - afl++ is maintained by Marc Heuse , Heiko Eissfeldt - and Andrea Fioraldi as - there have been no updates to afl since November 2017. - - - Many improvements were made, e.g. more performant llvm_mode, supporting - llvm up to version 8, Qemu 3.1, more speed and crashfixes for Qemu, - laf-intel feature for Qemu (with libcompcov) etc. - - Additionally AFLfast's power schedules by Marcel Boehme from - https://github.com/mboehme/aflfast have been incorporated. - - C. Hollers afl-fuzz Python mutator module and llvm_mode whitelist support - was added too (https://github.com/choller/afl) - - New is the excellent MOpt mutator from - https://github.com/puppet-meteor/MOpt-AFL - - Also newly integrated is instrim, a very effective CFG llvm_mode - instrumentation implementation from https://github.com/csienslab/instrim - - And finally the newest addition is the unicorn_mode which allows fuzzing - of binaries from completely different platforms - provided by domenukk! - The unicorn afl mode is not the stock version but like afl++ contains - various patches from forks that make it better :) - - A more thorough list is available in the PATCHES file. - - So all in all this is the best-of AFL that is currently out there :-) - - - Copyright 2013, 2014, 2015, 2016 Google Inc. All rights reserved. - Released under terms and conditions of Apache License, Version 2.0. - - For new versions and additional information, check out: - https://github.com/vanhauser-thc/AFLplusplus - - To compare notes with other users or get notified about major new features, - send a mail to . - - ** See QuickStartGuide.txt if you don't have time to read this file. ** - - -1) Challenges of guided fuzzing -------------------------------- - -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. - -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. - -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 program state, and are less -useful for guiding the fuzzing effort in the long haul. - -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. - - -2) The afl-fuzz approach ------------------------- - -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. - -Simplifying a bit, the overall algorithm can be summed up as: - - 1) Load user-supplied initial test cases into the queue, - - 2) Take next input file from the queue, - - 3) Attempt to trim the test case to the smallest size that doesn't alter - the measured behavior of the program, - - 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. - - -3) Instrumenting programs for use with AFL ------------------------------------------- - -PLEASE NOTE: llvm_mode compilation with afl-clang-fast/afl-clang-fast++ -instead of afl-gcc/afl-g++ is much faster and has a few cool features. -See llvm_mode/ - however few code does not compile with llvm. -We support llvm versions 4.0 to 8. - -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. - -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. - -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: - -$ CC=/path/to/afl/afl-gcc ./configure -$ make clean all - -For C++ programs, you'd would also want to set CXX=/path/to/afl/afl-g++. - -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.llvm. -Clang/LLVM has a much better performance and works from LLVM version 4.0 to 8. -Using the LAF Intel performance enhancements are also recommended, see -llvm_mode/README.laf-intel -Using partial instrumentation is also recommended, see -llvm_mode/README.whitelist - -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: - -$ CC=/path/to/afl/afl-gcc ./configure --disable-shared - -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.dislocator) can help uncover heap corruption issues, too. - -PS. ASAN users are advised to docs/review notes_for_asan.txt file for -important caveats. - - -4) Instrumenting binary-only apps ---------------------------------- - -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. - -QEMU is a project separate from AFL, but you can conveniently build the -feature by doing: - -$ cd qemu_mode -$ ./build_qemu_support.sh - -For additional instructions and caveats, see qemu_mode/README.qemu. - -The mode is approximately 2-5x slower than compile-time instrumentation, is -less conductive to parallelization, and may have some other quirks. - -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. - - -5) Power schedules ------------------- - -The power schedules were copied from Marcel Böhme's excellent AFLfast -implementation and expands on the ability to discover new paths and -therefore the coverage. - -The available schedules are: - - - explore (default) - - fast - - coe - - quad - - lin - - exploit - -In parallel mode (-M/-S, several instances with shared queue), we suggest to -run the master using the exploit schedule (-p exploit) and the slaves with a -combination of cut-off-exponential (-p coe), exponential (-p fast; default), -and explore (-p explore) schedules. - -In single mode, using -p fast is usually more beneficial than the default -explore mode. -(We don't want to change the default behaviour of afl, so "fast" has not been -made the default mode). - -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/ - -6) 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 -application. There are two basic rules: - - - Keep the files small. Under 1 kB is ideal, although not strictly necessary. - For a discussion of why size matters, see perf_tips.txt. - - - Use multiple test cases only if they are functionally different from - each other. There is no point in using fifty different vacation photos - to fuzz an image library. - -You can find many good examples of starting files in the testcases/ subdirectory -that comes with this tool. - -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. - - -7) 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: - -$ ./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: - -$ ./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.txt. - -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. - - -8) Interpreting output ----------------------- - -See the status_screen.txt file for information on how to interpret the -displayed stats and monitor the health of the process. Be sure to consult this -file especially if any UI elements are highlighted in red. - -The fuzzing process will continue until you press Ctrl-C. At minimum, you want -to allow the fuzzer to complete one queue cycle, which may take anywhere from a -couple of hours to a week or so. - -There are three subdirectories created within the output directory and updated -in real time: - - - queue/ - test cases for every distinctive execution path, plus all the - starting files given by the user. This is the synthesized corpus - mentioned in section 2. - - Before using this corpus for any other purposes, you can shrink - it to a smaller size using the afl-cmin tool. The tool will find - a smaller subset of files offering equivalent edge coverage. - - - crashes/ - unique test cases that cause the tested program to receive a - fatal signal (e.g., SIGSEGV, SIGILL, SIGABRT). The entries are - grouped by the received signal. - - - hangs/ - unique test cases that cause the tested program to time out. The - default time limit before something is classified as a hang is - the larger of 1 second and the value of the -t parameter. - The value can be fine-tuned by setting AFL_HANG_TMOUT, but this - is rarely necessary. - -Crashes and hangs are considered "unique" if the associated execution paths -involve any state transitions not seen in previously-recorded faults. If a -single bug can be reached in multiple ways, there will be some count inflation -early in the process, but this should quickly taper off. - -The file names for crashes and hangs are correlated with parent, non-faulting -queue entries. This should help with debugging. - -When you can't reproduce a crash found by afl-fuzz, the most likely cause is -that you are not setting the same memory limit as used by the tool. Try: - -$ LIMIT_MB=50 -$ ( ulimit -Sv $[LIMIT_MB << 10]; /path/to/tested_binary ... ) - -Change LIMIT_MB to match the -m parameter passed to afl-fuzz. On OpenBSD, -also change -Sv to -Sd. - -Any existing output directory can be also used to resume aborted jobs; try: - -$ ./afl-fuzz -i- -o existing_output_dir [...etc...] - -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/. - - -9) 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 parallel_fuzzing.txt. - -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 parallel_fuzzing.txt for tips. - - -10) 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 - -To use this feature, you first need to create a dictionary in one of the two -formats discussed in dictionaries/README.dictionaries; 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 - -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.tokencap. - - -11) 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. -Every crash is also traceable to its parent non-crashing test case in the -queue, making it easier to diagnose faults. - -Having said that, it's important to acknowledge that some fuzzing crashes can be -difficult to quickly evaluate for exploitability without a lot of debugging and -code analysis work. To assist with this task, afl-fuzz supports a very unique -"crash exploration" mode enabled with the -C flag. - -In this mode, the fuzzer takes one or more crashing test cases as the input, -and uses its feedback-driven fuzzing strategies to very quickly enumerate all -code paths that can be reached in the program while keeping it in the -crashing state. - -Mutations that do not result in a crash are rejected; so are any changes that -do not affect the execution path. - -The output is a small corpus of files that can be very rapidly examined to see -what degree of control the attacker has over the faulting address, or whether -it is possible to get past an initial out-of-bounds read - and see what lies -beneath. - -Oh, one more thing: for test case minimization, give afl-tmin a try. The tool -can be operated in a very simple way: - -$ ./afl-tmin -i test_case -o minimized_result -- /path/to/program [...] - -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 -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 -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 -insights into complex file formats. More info about its operation can be found -near the end of technical_details.txt. - - -12) Going beyond crashes ------------------------- - -Fuzzing is a wonderful and underutilized technique for discovering non-crashing -design and implementation errors, too. Quite a few interesting bugs have been -found by modifying the target programs to call abort() when, say: - - - Two bignum libraries produce different outputs when given the same - fuzzer-generated input, - - - An image library produces different outputs when asked to decode the same - input image several times in a row, - - - A serialization / deserialization library fails to produce stable outputs - when iteratively serializing and deserializing fuzzer-supplied data, - - - A compression library produces an output inconsistent with the input file - when asked to compress and then decompress a particular blob. - -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). - - -13) Common-sense risks ----------------------- - -Please keep in mind that, similarly to many other computationally-intensive -tasks, fuzzing may put strain on your hardware and on the OS. In particular: - - - Your CPU will run hot and will need adequate cooling. In most cases, if - cooling is insufficient or stops working properly, CPU speeds will be - automatically throttled. That said, especially when fuzzing on less - suitable hardware (laptops, smartphones, etc), it's not entirely impossible - 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 - 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. - - - Fuzzing involves billions of reads and writes to the filesystem. On modern - systems, this will be usually heavily cached, resulting in fairly modest - "physical" I/O - but there are many factors that may alter this equation. - It is your responsibility to monitor for potential trouble; with very heavy - I/O, the lifespan of many HDDs and SSDs may be reduced. - - A good way to monitor disk I/O on Linux is the 'iostat' command: - - $ iostat -d 3 -x -k [...optional disk ID...] - - -14) Known limitations & areas for improvement ---------------------------------------------- - -Here are some of the most important caveats for AFL: - - - 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 processed spawned by the fuzzed target may evade - detection unless you manually add some code to catch that. - - - As with any other brute-force tool, the fuzzer offers limited coverage if - encryption, checksums, cryptographic signatures, or compression are used to - wholly wrap the actual data format to be tested. - - To work around this, you can comment out the relevant checks (see - experimental/libpng_no_checksum/ for inspiration); if this is not possible, - you can also write a postprocessor, as explained in - experimental/post_library/ (with AFL_POST_LIBRARY) - - - There are some unfortunate trade-offs with ASAN and 64-bit binaries. This - isn't due to any specific fault of afl-fuzz; see notes_for_asan.txt for - tips. - - - There is no direct support for fuzzing network services, background - daemons, or interactive apps that require UI interaction to work. You may - need to make simple code changes to make them behave in a more traditional - way. Preeny may offer a relatively simple option, too - see: - https://github.com/zardus/preeny - - 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 - - - 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 - - - Occasionally, sentient machines rise against their creators. If this - happens to you, please consult http://lcamtuf.coredump.cx/prep/. - -Beyond this, see INSTALL for platform-specific tips. - - -15) Special thanks ------------------- - -Many of the improvements to afl-fuzz wouldn't be possible without feedback, -bug reports, or patches from: - - Jann Horn Hanno Boeck - Felix Groebert Jakub Wilk - Richard W. M. Jones Alexander Cherepanov - Tom Ritter Hovik Manucharyan - Sebastian Roschke Eberhard Mattes - Padraig Brady Ben Laurie - @dronesec Luca Barbato - Tobias Ospelt Thomas Jarosch - Martin Carpenter Mudge Zatko - Joe Zbiciak Ryan Govostes - Michael Rash William Robinet - Jonathan Gray Filipe Cabecinhas - Nico Weber Jodie Cunningham - Andrew Griffiths Parker Thompson - Jonathan Neuschfer Tyler Nighswander - Ben Nagy Samir Aguiar - Aidan Thornton Aleksandar Nikolich - Sam Hakim Laszlo Szekeres - David A. Wheeler Turo Lamminen - Andreas Stieger Richard Godbee - Louis Dassy teor2345 - Alex Moneger Dmitry Vyukov - Keegan McAllister Kostya Serebryany - Richo Healey Martijn Bogaard - rc0r Jonathan Foote - Christian Holler Dominique Pelle - Jacek Wielemborek Leo Barnes - Jeremy Barnes Jeff Trull - Guillaume Endignoux ilovezfs - Daniel Godas-Lopez Franjo Ivancic - Austin Seipp Daniel Komaromy - Daniel Binderman Jonathan Metzman - Vegard Nossum Jan Kneschke - Kurt Roeckx Marcel Bohme - Van-Thuan Pham Abhik Roychoudhury - Joshua J. Drake Toby Hutton - Rene Freingruber Sergey Davidoff - Sami Liedes Craig Young - Andrzej Jackowski Daniel Hodson - -Thank you! - - -16) Contact ------------ - -Questions? Concerns? Bug reports? The contributors can be reached via -https://github.com/vanhauser-thc/AFLplusplus - -There is also a mailing list for the afl project; to join, send a mail to -. Or, if you prefer to browse -archives first, try: - - https://groups.google.com/group/afl-users diff --git a/llvm_mode/README.llvm b/llvm_mode/README.llvm index 00528a46..a0c40211 100644 --- a/llvm_mode/README.llvm +++ b/llvm_mode/README.llvm @@ -205,10 +205,8 @@ post-process the assembly or install any compiler plugins. See: http://clang.llvm.org/docs/SanitizerCoverage.html#tracing-pcs-with-guards -As of this writing, the feature is only available on SVN trunk, and is yet to -make it to an official release of LLVM. Nevertheless, if you have a -sufficiently recent compiler and want to give it a try, build afl-clang-fast -this way: +If you have a sufficiently recent compiler and want to give it a try, build +afl-clang-fast this way: AFL_TRACE_PC=1 make clean all diff --git a/types.h b/types.h index 3497bb2b..7606d4ed 100644 --- a/types.h +++ b/types.h @@ -78,7 +78,7 @@ typedef int64_t s64; #define STRINGIFY(x) STRINGIFY_INTERNAL(x) #define MEM_BARRIER() \ - asm volatile("" ::: "memory") + __asm__ volatile("" ::: "memory") #if __GNUC__ < 6 #define likely(_x) (_x) -- cgit 1.4.1