add llvm instrumentation

Dockerfile

@@ -12,7 +12,7 @@ ENV HOPPER_BIN=/hopper/hopper \
 RUN apt-get update \
     && apt-get -y upgrade \
     && apt-get -y install build-essential wget curl cmake git unzip xxd protobuf-compiler libprotobuf-dev \
-    && apt-get -y install llvm-dev libclang-dev clang \
+    && apt-get -y install llvm llvm-dev libclang-dev clang \
     && apt-get clean
 
 # ENV RUSTUP_DIST_SERVER="https://mirrors.ustc.edu.cn/rust-static"

README.md

@@ -21,7 +21,7 @@ To learn more about Hopper, check out our [paper](https://arxiv.org/pdf/2309.034
 ```
 
 The script will create a `install` directory in hopper's root directory, then you can use `hopper`.
-To use the command anywhere, you can set your project directory in your PATH variable.
+To use the command anywhere, you can set the project directory in your PATH variable.
 
 ### Using Docker
 You can choose to use the Dockerfile, which build the requirements and Hopper.
@@ -38,7 +38,7 @@ hopper compile --header ./cJSON.h --library ./libcjson.so --output output
 
 Use `hopper compile --help` to see detailed usage. If the compiling reports errors about header file, refer to the usage of [rust-bindgen](https://rust-lang.github.io/rust-bindgen/), which we used for parsing header file.
 You may wrap the header file with the missing definitions.
-Hopper uses [E9Patch](https://github.com/GJDuck/e9patch) to instrument binaries by default.
+Hopper uses [E9Patch](https://github.com/GJDuck/e9patch) to instrument binaries by default. Optionally, you can use [LLVM](./hopper-instrument/llvm-mode/) for source code instrumentation.
 
 After running `compile`, you will find that it generates the following files in the output directory:
 - `bin/hopper-fuzzer`:  generates inputs, maintatins states, and use `harness` to excuted the inputs.
@@ -97,7 +97,7 @@ echo core | sudo tee /proc/sys/kernel/core_pattern
 ```
 
 ### Function pattern 
-Hopper generates inputs for all functions in libiries by default. However, there are two ways to filter functions in Hopper: exlucding functions or including functions. This way, it can be focus on intersting functions.
+Hopper generates inputs for all functions that appear in both headers and libiries by default. However, there are two ways to filter functions in Hopper: exlucding functions or including functions. This way, it can be focus on intersting functions.
 
 #### `--func-pattern`
 ```

build.sh

@@ -19,24 +19,28 @@ if [[ "$OSTYPE" == "linux-gnu"* ]]; then
 
     if [ ! -x $INSTALL_DIR/patchelf ]; then
         info "download patchelf ..."
-        cd install
+        cd $INSTALL_DIR
         mkdir -p tmp
         cd tmp
         wget https://github.com/NixOS/patchelf/releases/download/${PATCHELF_VERSION}/patchelf-${PATCHELF_VERSION}-x86_64.tar.gz
         tar -xvf patchelf-${PATCHELF_VERSION}-x86_64.tar.gz
         cp bin/patchelf ../.
-        cd ../../
     fi
 fi
 
-# info "start install hopper's llvm plugins ..."
-# cd hopper-instrument/llvm-mode
-# make PREFIX=$INSTALL_DIR
+info "start install hopper's llvm plugins ..."
+cd $INSTALL_DIR
+rm -rf llvm_build
+mkdir llvm_build && cd llvm_build 
+cmake -DHOPPER_BIN_DIR=$INSTALL_DIR $ROOT_DIR/hopper-instrument/llvm-mode
+make
+make install
 
-BUILD_TYPE=${BUILD_TYPE:-debug}
-# BUILD_TYPE=${BUILD_TYPE:-release}
+# BUILD_TYPE=${BUILD_TYPE:-debug}
+BUILD_TYPE=${BUILD_TYPE:-release}
 
 info "start build and install hopper fuzzer ..."
+cd $ROOT_DIR
 if [[ "$BUILD_TYPE" == "debug" ]]; then
     cargo build
 else

examples/re2/hopper.config

@@ -5,7 +5,7 @@ TEST_HEADER=/data/workspace/fuzzing_bench/re2/cre2.h
 
 # Full path to shared library
 TEST_LIB_DIR=/data/workspace/fuzzing_bench/re2/build/hopper_build/
-TEST_LIBRARY=${TEST_LIB_DIR}/cre2/lib/libcre2.so ${TEST_LIB_DIR}/lib/libre2.so.10.0.0
+TEST_LIBRARY=${TEST_LIB_DIR}/cre2/lib/libcre2.so ${TEST_LIB_DIR}/lib/libre2.so
 
 # Output directory
 OUT_DIR=output
@@ -19,8 +19,7 @@ HOPPER_MAP_SIZE_POW2=20
 # set seeds for hopper
 # HOPPER_SEED_DIR=seeds
 
-
 # target library
 HOPPER_TEST_LIBRARY=cre2/lib/libre2.so
 
-HOPPER_CUSTOM_RULES=$SRC_DIR/../custom_rule
+HOPPER_CUSTOM_RULES=$SRC_DIR/../custom_rule

hopper

@@ -35,6 +35,9 @@ source ${ROOT_DIR}/tools/core_affinity.sh
 if [[ -e "hopper.config" ]]; then
     while IFS='=' read -r key value; do
     if [[ $key && $value ]]; then
+        if [[ $key = \#* ]] ; then
+            continue
+        fi
         declare -x "$key=$value"
         echo "config file set: $key = $value"
     fi
@@ -69,6 +72,14 @@ check_output() {
         warn "$USAGE"
         exit 1
     fi
+    if [ "$OUTPUT_DIR" == "" ];then
+        error "output directory can't be empty!"
+        exit 1
+    fi
+    if [ "$OUTPUT_DIR" == "/" ];then
+        error "output directory can't be root directory '/'!"
+        exit 1
+    fi
 }
 
 [ ! -d "${ROOT_DIR}/install" ] && warn "Please run ./build.sh to build hopper's code"
@@ -183,6 +194,10 @@ clean)
     find . -maxdepth 1 -type f ! -executable ! -name "test*" ! -name "*.log" ! -name "func_list" ! -name "custom_rule" ! -name "hopper.config" -delete
     info "clean files in '$OUTPUT_DIR' directory"
     ;;
+clang)
+    echo "CC=${ROOT_DIR}/install/hopper-clang"
+    echo "CXX=${ROOT_DIR}/install/hopper-clang++"
+    ;;
 help)
     warn "$USAGE"
     exit 0

hopper-compiler/Cargo.toml

@@ -17,6 +17,8 @@ gimli = { version = "0.26", default-features = false, features = ["read"] }
 object = { version = "0.28", default-features = false, features = ["read"] }
 eyre = "0.6"
 twoway = "0.2"
+# NO GPL
+# patchelf = "0.2.1"
 
 [features]
 default = ["elf", "dwarf"]

hopper-compiler/src/binary_info.rs

@@ -11,6 +11,8 @@ pub struct BinaryInfo {
     pub str_list: Vec<String>,
     // exported API in the symbol table
     pub func_list: Vec<FuncInfo>,
+    // needed libraries
+    pub needed: Vec<String>
 }
 
 #[derive(Debug)]
@@ -29,6 +31,7 @@ impl BinaryInfo {
         let lib_type;
         let mut str_list = vec![];
         let mut func_list = vec![];
+        let mut needed = vec![];
         match result {
             Object::Elf(elf) => {
                 lib_type = "elf";
@@ -67,6 +70,12 @@ impl BinaryInfo {
                         })
                     }
                 }
+                if let Some(dy) = elf.dynamic.as_ref() {
+                    for name in dy.get_libraries(&elf.dynstrtab) {
+                        needed.push(name.to_string());
+                    }
+                }
+
             }
             Object::PE(pe) => {
                 lib_type = "pe";
@@ -112,6 +121,7 @@ impl BinaryInfo {
             lib_type,
             str_list,
             func_list,
+            needed,
         })
     }
 

hopper-compiler/src/check.rs

@@ -52,7 +52,7 @@ pub fn output_lib_name(file: &str) -> String {
 pub fn check_llvm_runtime(libraries: &[String]) -> bool {
     libraries
         .iter()
-        .any(|l| check_file_contains(l, "HOOPER_LLVM_MARK"))
+        .any(|l| check_file_contains(l, "__hopper_area_ptr"))
 }
 
 pub fn check_file_contains(target: &str, s: &str) -> bool {

hopper-compiler/src/main.rs

@@ -75,7 +75,11 @@ fn instrument(library: &Path, output: &Path, config: &Config, lib_info: &BinaryI
         }
     }
     #[cfg(target_os = "linux")]
-    patch::patchelf_set_so_name(&lib_name, output_lib.to_str().context("fail to be str")?)?;
+    {
+        let output_lib_path = output_lib.to_str().context("fail to be str")?;
+        patch::patchelf_set_so_name(&lib_name, output_lib_path)?;
+        patch::remove_prev_needed(&config.library, output_lib_path, lib_info)?;
+    }
     Ok(output_lib)
 }