compiler-rt/lib/sanitizer_common/symbolizer/scripts/build_symbolizer.sh

   1 #!/usr/bin/env bash
   2 #
   3 # Run as: CLANG=bin/clang ZLIB_SRC=src/zlib \
   4 #             build_symbolizer.sh runtime_build/lib/clang/4.0.0/lib/linux/
   5 # zlib can be downloaded from http://www.zlib.net.
   6 #
   7 # Script compiles self-contained object file with symbolization code and injects
   8 # it into the given set of runtime libraries. Script updates only libraries
   9 # which has unresolved __sanitizer_symbolize_* symbols and matches architecture.
  10 # Object file is be compiled from LLVM sources with dependencies like libc++ and
  11 # zlib. Then it internalizes symbols in the file, so that it can be linked
  12 # into arbitrary programs, avoiding conflicts with the program own symbols and
  13 # avoiding dependencies on any program symbols. The only acceptable dependencies
  14 # are libc and __sanitizer::internal_* from sanitizer runtime.
  15 #
  16 # Symbols exported by the object file will be used by Sanitizer runtime
  17 # libraries to symbolize code/data in-process.
  18 #
  19 # The script will modify the output directory which is given as the first
  20 # argument to the script.
  21 #
  22 # FIXME: We should really be using a simpler approach to building this object
  23 # file, and it should be available as a regular cmake rule. Conceptually, we
  24 # want to be doing "ld -r" followed by "objcopy -G" to create a relocatable
  25 # object file with only our entry points exposed. However, this does not work at
  26 # present, see PR30750.
  27
  28 set -x
  29 set -e
  30 set -u
  31
  32 SCRIPT_DIR=$(cd "$(dirname "$0")" && pwd)
  33 SRC_DIR=$(readlink -f $SCRIPT_DIR/..)
  34 TARGE_DIR=$(readlink -f $1)
  35 COMPILER_RT_SRC=$(readlink -f ${SCRIPT_DIR}/../../../..)
  36 LLVM_SRC=${LLVM_SRC:-${COMPILER_RT_SRC}/../llvm}
  37 LLVM_SRC=$(readlink -f $LLVM_SRC)
  38 if [[ ! -d "${LLVM_SRC}/../llvm" ]] ; then
  39   LLVM_SRC=$(readlink -f ${COMPILER_RT_SRC}/../../../llvm)
  40 fi
  41 LIBCXX_SRC=$(readlink -f ${COMPILER_RT_SRC}/../libcxx)
  42 LIBCXXABI_SRC=$(readlink -f ${COMPILER_RT_SRC}/../libcxxabi)
  43
  44 if [[ ! -d "${LLVM_SRC}/../llvm" ||
  45       ! -d "${LIBCXX_SRC}" ||
  46       ! -d "${LIBCXXABI_SRC}" ]]; then
  47   echo "Missing or incomplete LLVM_SRC"
  48   exit 1
  49 fi
  50
  51 if [[ "$ZLIB_SRC" == ""  ||
  52       ! -x "${ZLIB_SRC}/configure" ||
  53       ! -f "${ZLIB_SRC}/zlib.h" ]]; then
  54   echo "Missing or incomplete ZLIB_SRC"
  55   exit 1
  56 fi
  57 ZLIB_SRC=$(readlink -f $ZLIB_SRC)
  58
  59 J="${J:-50}"
  60
  61 CLANG="${CLANG:-`which clang`}"
  62 CLANG_DIR=$(readlink -f $(dirname "$CLANG"))
  63
  64 BUILD_DIR=$(readlink -f ./symbolizer)
  65 mkdir -p $BUILD_DIR
  66 cd $BUILD_DIR
  67
  68 CC=$CLANG_DIR/clang
  69 CXX=$CLANG_DIR/clang++
  70 TBLGEN=$CLANG_DIR/llvm-tblgen
  71 OPT=$CLANG_DIR/opt
  72 export AR=$CLANG_DIR/llvm-ar
  73 export LINK=$CLANG_DIR/llvm-link
  74 TARGET_TRIPLE=$($CC -print-target-triple)
  75
  76 for F in $CC $CXX $TBLGEN $LINK $OPT $AR; do
  77   if [[ ! -x "$F" ]]; then
  78     echo "Missing $F"
  79      exit 1
  80   fi
  81 done
  82
  83 ZLIB_BUILD=${BUILD_DIR}/zlib
  84 LIBCXX_BUILD=${BUILD_DIR}/libcxx
  85 LLVM_BUILD=${BUILD_DIR}/llvm
  86 SYMBOLIZER_BUILD=${BUILD_DIR}/symbolizer
  87
  88 FLAGS=${FLAGS:-}
  89 FLAGS="$FLAGS -fPIC -flto -Os -g0 -DNDEBUG"
  90
  91 # Build zlib.
  92 mkdir -p ${ZLIB_BUILD}
  93 cd ${ZLIB_BUILD}
  94 cp -r ${ZLIB_SRC}/* .
  95 CC=$CC CFLAGS="$FLAGS" RANLIB=/bin/true ./configure --static
  96 make -j${J} libz.a
  97
  98 # Build and install libcxxabi and libcxx.
  99 if [[ ! -d ${LIBCXX_BUILD} ]]; then
 100   mkdir -p ${LIBCXX_BUILD}
 101   cd ${LIBCXX_BUILD}
 102   LIBCXX_FLAGS="${FLAGS} -Wno-macro-redefined"
 103   PROJECTS=
 104   if [[ ! -d $LLVM_SRC/projects/libcxxabi ]] ; then
 105     PROJECTS="-DLLVM_ENABLE_PROJECTS='libcxx;libcxxabi'"
 106   fi
 107   cmake -GNinja \
 108     ${PROJECTS} \
 109     -DCMAKE_BUILD_TYPE=Release \
 110     -DCMAKE_C_COMPILER=$CC \
 111     -DCMAKE_CXX_COMPILER=$CXX \
 112     -DCMAKE_C_FLAGS_RELEASE="${LIBCXX_FLAGS}" \
 113     -DCMAKE_CXX_FLAGS_RELEASE="${LIBCXX_FLAGS}" \
 114     -DLIBCXXABI_ENABLE_ASSERTIONS=OFF \
 115     -DLIBCXXABI_ENABLE_EXCEPTIONS=OFF \
 116     -DLIBCXXABI_ENABLE_SHARED=OFF \
 117     -DLIBCXX_ENABLE_ASSERTIONS=OFF \
 118     -DLIBCXX_ENABLE_EXCEPTIONS=OFF \
 119     -DLIBCXX_ENABLE_RTTI=OFF \
 120     -DLIBCXX_ENABLE_SHARED=OFF \
 121   $LLVM_SRC
 122 fi
 123 cd ${LIBCXX_BUILD}
 124 ninja cxx cxxabi
 125
 126 FLAGS="${FLAGS} -fno-rtti -fno-exceptions"
 127 LLVM_FLAGS="${FLAGS} -nostdinc++ -I${ZLIB_BUILD} -isystem ${LIBCXX_BUILD}/include/${TARGET_TRIPLE}/c++/v1 -isystem ${LIBCXX_BUILD}/include/c++/v1 -Wno-error=global-constructors"
 128
 129 # Build LLVM.
 130 if [[ ! -d ${LLVM_BUILD} ]]; then
 131   mkdir -p ${LLVM_BUILD}
 132   cd ${LLVM_BUILD}
 133   cmake -GNinja \
 134     -DCMAKE_BUILD_TYPE=Release \
 135     -DCMAKE_C_COMPILER=$CC \
 136     -DCMAKE_CXX_COMPILER=$CXX \
 137     -DCMAKE_C_FLAGS_RELEASE="${LLVM_FLAGS}" \
 138     -DCMAKE_CXX_FLAGS_RELEASE="${LLVM_FLAGS}" \
 139     -DLLVM_TABLEGEN=$TBLGEN \
 140     -DLLVM_ENABLE_ZLIB=ON \
 141     -DLLVM_ENABLE_TERMINFO=OFF \
 142     -DLLVM_ENABLE_THREADS=OFF \
 143     -DLLVM_DISABLE_ASSEMBLY_FILES=ON \
 144   $LLVM_SRC
 145 fi
 146 cd ${LLVM_BUILD}
 147 ninja LLVMSymbolize LLVMObject LLVMBinaryFormat LLVMDebugInfoDWARF LLVMSupport LLVMDebugInfoPDB LLVMDebuginfod LLVMMC LLVMDemangle LLVMTextAPI
 148
 149 cd ${BUILD_DIR}
 150 rm -rf ${SYMBOLIZER_BUILD}
 151 mkdir ${SYMBOLIZER_BUILD}
 152 cd ${SYMBOLIZER_BUILD}
 153
 154 echo "Compiling..."
 155 SYMBOLIZER_FLAGS="$LLVM_FLAGS -I${LLVM_SRC}/include -I${LLVM_BUILD}/include -std=c++14"
 156 $CXX $SYMBOLIZER_FLAGS ${SRC_DIR}/sanitizer_symbolize.cpp ${SRC_DIR}/sanitizer_wrappers.cpp -c
 157 $AR rc symbolizer.a sanitizer_symbolize.o sanitizer_wrappers.o
 158
 159 SYMBOLIZER_API_LIST=__sanitizer_symbolize_code
 160 SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_data
 161 SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_flush
 162 SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_demangle
 163 SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_set_demangle
 164 SYMBOLIZER_API_LIST+=,__sanitizer_symbolize_set_inline_frames
 165
 166 LIBCXX_ARCHIVE_DIR=$(dirname $(find $LIBCXX_BUILD -name libc++.a | head -n1))
 167
 168 # Merge all the object files together and copy the resulting library back.
 169 $SCRIPT_DIR/ar_to_bc.sh $LIBCXX_ARCHIVE_DIR/libc++.a \
 170                         $LIBCXX_ARCHIVE_DIR/libc++abi.a \
 171                         $LLVM_BUILD/lib/libLLVMSymbolize.a \
 172                         $LLVM_BUILD/lib/libLLVMObject.a \
 173                         $LLVM_BUILD/lib/libLLVMBinaryFormat.a \
 174                         $LLVM_BUILD/lib/libLLVMDebugInfoDWARF.a \
 175                         $LLVM_BUILD/lib/libLLVMSupport.a \
 176                         $LLVM_BUILD/lib/libLLVMDebugInfoPDB.a \
 177                         $LLVM_BUILD/lib/libLLVMDebugInfoMSF.a \
 178                         $LLVM_BUILD/lib/libLLVMDebugInfoCodeView.a \
 179                         $LLVM_BUILD/lib/libLLVMDebuginfod.a \
 180                         $LLVM_BUILD/lib/libLLVMDemangle.a \
 181                         $LLVM_BUILD/lib/libLLVMMC.a \
 182                         $LLVM_BUILD/lib/libLLVMTextAPI.a \
 183                         $ZLIB_BUILD/libz.a \
 184                         symbolizer.a \
 185                         all.bc
 186
 187 echo "Optimizing..."
 188 $OPT -internalize -internalize-public-api-list=${SYMBOLIZER_API_LIST} all.bc -o opt.bc
 189 $CC $FLAGS -fno-lto -c opt.bc -o symbolizer.o
 190
 191 echo "Checking undefined symbols..."
 192 nm -f posix -g symbolizer.o | cut -f 1,2 -d \  | LC_COLLATE=C sort -u > undefined.new
 193 (diff -u $SCRIPT_DIR/global_symbols.txt undefined.new | grep -E "^\+[^+]") && \
 194   (echo "Failed: unexpected symbols"; exit 1)
 195
 196 arch() {
 197   objdump -f $1 | grep -m1 -Po "(?<=file format ).*$"
 198 }
 199
 200 SYMBOLIZER_FORMAT=$(arch symbolizer.o)
 201 echo "Injecting $SYMBOLIZER_FORMAT symbolizer..."
 202 for A in $TARGE_DIR/libclang_rt.*san*.a; do
 203   A_FORMAT=$(arch $A)
 204   if [[ "$A_FORMAT" != "$SYMBOLIZER_FORMAT" ]] ; then
 205     continue
 206   fi
 207   (nm -u $A 2>/dev/null | grep -E "__sanitizer_symbolize_code" >/dev/null) || continue
 208   echo "$A"
 209   $AR rcs $A symbolizer.o
 210 done
 211
 212 echo "Success!"