1#!/usr/bin/env bash
2#
3# Run as: CLANG=bin/clang build_symbolizer.sh out.o
4# If you want to use a local copy of zlib, set ZLIB_SRC.
5# zlib can be downloaded from http://www.zlib.net.
6#
7# Script compiles self-contained object file with symbolization code.
8#
9# Symbols exported by the object file will be used by Sanitizer runtime
10# libraries to symbolize code/data in-process.
11#
12# FIXME: We should really be using a simpler approach to building this object
13# file, and it should be available as a regular cmake rule. Conceptually, we
14# want to be doing "ld -r" followed by "objcopy -G" to create a relocatable
15# object file with only our entry points exposed. However, this does not work at
16# present, see https://github.com/llvm/llvm-project/issues/30098.
17
18set -x
19set -e
20set -u
21
22SCRIPT_DIR=$(cd "$(dirname "$0")" && pwd)
23SRC_DIR=$(readlink -f $SCRIPT_DIR/..)
24
25if [[ $# -ne 1 ]]; then
26  echo "Missing output file"
27  exit 1
28fi
29
30OUTPUT=$(readlink -f $1)
31COMPILER_RT_SRC=$(readlink -f ${SCRIPT_DIR}/../../../..)
32LLVM_SRC=${LLVM_SRC:-${COMPILER_RT_SRC}/../llvm}
33LLVM_SRC=$(readlink -f $LLVM_SRC)
34
35CLANG="${CLANG:-`which clang`}"
36CLANG_DIR=$(readlink -f $(dirname "$CLANG"))
37
38CC=$CLANG_DIR/clang
39CXX=$CLANG_DIR/clang++
40TBLGEN=$CLANG_DIR/llvm-tblgen
41OPT=$CLANG_DIR/opt
42AR=$CLANG_DIR/llvm-ar
43LINK=$CLANG_DIR/llvm-link
44
45for F in $CC $CXX $TBLGEN $LINK $OPT $AR; do
46  if [[ ! -x "$F" ]]; then
47    echo "Missing $F"
48     exit 1
49  fi
50done
51
52BUILD_DIR=${PWD}/symbolizer
53mkdir -p $BUILD_DIR
54cd $BUILD_DIR
55
56ZLIB_BUILD=${BUILD_DIR}/zlib
57LIBCXX_BUILD=${BUILD_DIR}/libcxx
58LLVM_BUILD=${BUILD_DIR}/llvm
59SYMBOLIZER_BUILD=${BUILD_DIR}/symbolizer
60
61FLAGS=${FLAGS:-}
62ZLIB_SRC=${ZLIB_SRC:-}
63TARGET_TRIPLE=$($CC -print-target-triple $FLAGS)
64if [[ "$FLAGS" =~ "-m32" ]] ; then
65  # Avoid new wrappers.
66  FLAGS+=" -U_FILE_OFFSET_BITS"
67fi
68FLAGS+=" -fPIC -flto -Oz -g0 -DNDEBUG -target $TARGET_TRIPLE -Wno-unused-command-line-argument"
69FLAGS+=" -include ${SRC_DIR}/../sanitizer_redefine_builtins.h -DSANITIZER_COMMON_REDEFINE_BUILTINS_IN_STD -Wno-language-extension-token"
70
71LINKFLAGS="-fuse-ld=lld -target $TARGET_TRIPLE"
72
73# Build zlib.
74if [[ ! -d ${ZLIB_BUILD} ]]; then
75  if [[ -z "${ZLIB_SRC}" ]]; then
76    git clone https://github.com/madler/zlib ${ZLIB_BUILD}
77  else
78    ZLIB_SRC=$(readlink -f $ZLIB_SRC)
79    mkdir -p ${ZLIB_BUILD}
80    cp -r ${ZLIB_SRC}/* ${ZLIB_BUILD}/
81  fi
82fi
83
84cd ${ZLIB_BUILD}
85AR="${AR}" CC="${CC}" CFLAGS="$FLAGS -Wno-deprecated-non-prototype" RANLIB=/bin/true ./configure --static
86make -j libz.a
87
88# Build and install libcxxabi and libcxx.
89if [[ ! -f ${LLVM_BUILD}/build.ninja ]]; then
90  rm -rf ${LIBCXX_BUILD}
91  mkdir -p ${LIBCXX_BUILD}
92  cd ${LIBCXX_BUILD}
93  LIBCXX_FLAGS="${FLAGS} -Wno-macro-redefined"
94  cmake -GNinja \
95    -DLLVM_ENABLE_RUNTIMES="libcxx;libcxxabi" \
96    -DCMAKE_BUILD_TYPE=Release \
97    -DCMAKE_C_COMPILER_WORKS=ON \
98    -DCMAKE_CXX_COMPILER_WORKS=ON \
99    -DCMAKE_C_COMPILER=$CC \
100    -DCMAKE_CXX_COMPILER=$CXX \
101    -DLIBCXX_ABI_NAMESPACE=__InternalSymbolizer \
102    '-DLIBCXX_EXTRA_SITE_DEFINES=_LIBCPP_DISABLE_VISIBILITY_ANNOTATIONS;_LIBCXXABI_DISABLE_VISIBILITY_ANNOTATIONS' \
103    -DCMAKE_C_FLAGS_RELEASE="${LIBCXX_FLAGS}" \
104    -DCMAKE_CXX_FLAGS_RELEASE="${LIBCXX_FLAGS}" \
105    -DLIBCXXABI_ENABLE_ASSERTIONS=OFF \
106    -DLIBCXXABI_ENABLE_EXCEPTIONS=OFF \
107    -DLIBCXXABI_USE_LLVM_UNWINDER=OFF \
108    -DLIBCXX_ENABLE_ASSERTIONS=OFF \
109    -DLIBCXX_ENABLE_EXCEPTIONS=OFF \
110    -DLIBCXX_ENABLE_RTTI=OFF \
111    -DCMAKE_SHARED_LINKER_FLAGS="$LINKFLAGS" \
112    -DLIBCXX_ENABLE_SHARED=OFF \
113    -DLIBCXXABI_ENABLE_SHARED=OFF \
114  $LLVM_SRC/../runtimes
115fi
116cd ${LIBCXX_BUILD}
117ninja cxx cxxabi
118
119FLAGS="${FLAGS} -fno-rtti -fno-exceptions"
120LLVM_CFLAGS="${FLAGS} -Wno-global-constructors"
121LLVM_CXXFLAGS="${LLVM_CFLAGS} -nostdinc++ -I${ZLIB_BUILD} -isystem ${LIBCXX_BUILD}/include -isystem ${LIBCXX_BUILD}/include/c++/v1"
122
123# Build LLVM.
124if [[ ! -f ${LLVM_BUILD}/build.ninja ]]; then
125  rm -rf ${LLVM_BUILD}
126  mkdir -p ${LLVM_BUILD}
127  cd ${LLVM_BUILD}
128  cmake -GNinja \
129    -DCMAKE_BUILD_TYPE=Release \
130    -DCMAKE_C_COMPILER_WORKS=ON \
131    -DCMAKE_CXX_COMPILER_WORKS=ON \
132    -DCMAKE_C_COMPILER=$CC \
133    -DCMAKE_CXX_COMPILER=$CXX \
134    -DLLVM_ENABLE_LIBCXX=ON \
135    -DCMAKE_C_FLAGS_RELEASE="${LLVM_CFLAGS}" \
136    -DCMAKE_CXX_FLAGS_RELEASE="${LLVM_CXXFLAGS}" \
137    -DCMAKE_EXE_LINKER_FLAGS="$LINKFLAGS -stdlib=libc++ -L${LIBCXX_BUILD}/lib" \
138    -DLLVM_TABLEGEN=$TBLGEN \
139    -DLLVM_INCLUDE_TESTS=OFF \
140    -DLLVM_ENABLE_ZLIB=ON \
141    -DLLVM_ENABLE_ZSTD=OFF \
142    -DLLVM_ENABLE_TERMINFO=OFF \
143    -DLLVM_ENABLE_THREADS=OFF \
144  $LLVM_SRC
145fi
146cd ${LLVM_BUILD}
147ninja LLVMSymbolize LLVMObject LLVMBinaryFormat LLVMDebugInfoDWARF LLVMSupport LLVMDebugInfoPDB LLVMDebuginfod LLVMMC LLVMDemangle LLVMTextAPI LLVMTargetParser
148
149cd ${BUILD_DIR}
150rm -rf ${SYMBOLIZER_BUILD}
151mkdir ${SYMBOLIZER_BUILD}
152cd ${SYMBOLIZER_BUILD}
153
154echo "Compiling..."
155SYMBOLIZER_FLAGS="$LLVM_CXXFLAGS -I${LLVM_SRC}/include -I${LLVM_BUILD}/include -std=c++17"
156$CXX $SYMBOLIZER_FLAGS ${SRC_DIR}/sanitizer_symbolize.cpp ${SRC_DIR}/sanitizer_wrappers.cpp -c
157$AR rc symbolizer.a sanitizer_symbolize.o sanitizer_wrappers.o
158
159SYMBOLIZER_API_LIST=__sanitizer_symbolize_code
160SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_data
161SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_frame
162SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_flush
163SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_demangle
164SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_set_demangle
165SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_set_inline_frames
166
167LIBCXX_ARCHIVE_DIR=$(dirname $(find $LIBCXX_BUILD -name libc++.a | head -n1))
168
169# Merge all the object files together and copy the resulting library back.
170$LINK $LIBCXX_ARCHIVE_DIR/libc++.a \
171      $LIBCXX_ARCHIVE_DIR/libc++abi.a \
172      $LLVM_BUILD/lib/libLLVMSymbolize.a \
173      $LLVM_BUILD/lib/libLLVMObject.a \
174      $LLVM_BUILD/lib/libLLVMBinaryFormat.a \
175      $LLVM_BUILD/lib/libLLVMDebugInfoDWARF.a \
176      $LLVM_BUILD/lib/libLLVMSupport.a \
177      $LLVM_BUILD/lib/libLLVMDebugInfoPDB.a \
178      $LLVM_BUILD/lib/libLLVMDebugInfoMSF.a \
179      $LLVM_BUILD/lib/libLLVMDebugInfoCodeView.a \
180      $LLVM_BUILD/lib/libLLVMDebuginfod.a \
181      $LLVM_BUILD/lib/libLLVMDemangle.a \
182      $LLVM_BUILD/lib/libLLVMMC.a \
183      $LLVM_BUILD/lib/libLLVMTextAPI.a \
184      $LLVM_BUILD/lib/libLLVMTargetParser.a \
185      $ZLIB_BUILD/libz.a \
186      symbolizer.a \
187      -ignore-non-bitcode -o all.bc
188
189echo "Optimizing..."
190$OPT -passes=internalize -internalize-public-api-list=${SYMBOLIZER_API_LIST} all.bc -o opt.bc
191$CC $FLAGS -fno-lto -c opt.bc -o symbolizer.o
192
193echo "Checking undefined symbols..."
194nm -f posix -g symbolizer.o | cut -f 1,2 -d \  | LC_COLLATE=C sort -u > undefined.new
195(diff -u $SCRIPT_DIR/global_symbols.txt undefined.new | grep -E "^\+[^+]") && \
196  (echo "Failed: unexpected symbols"; exit 1)
197
198cp -f symbolizer.o $OUTPUT
199
200echo "Success!"
201