Properly generate lists of exported symbols for sanitizer runtimes

This change adds a Python script that is invoked for
the just-built sanitizer runtime to generate the list of exported symbols
passed to the linker. By default, it contains interceptors and sanitizer
interface functions, but can be extended with tool-specific lists.


git-svn-id: https://llvm.org/svn/llvm-project/compiler-rt/trunk@189356 91177308-0d34-0410-b5e6-96231b3b80d8

18 files changed, 137 insertions, 100 deletions

diff --git a/CMakeLists.txt b/CMakeLists.txt
index 67cc51ed9..70178488d 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -40,6 +40,7 @@ set(COMPILER_RT_BINARY_DIR ${CMAKE_CURRENT_BINARY_DIR})
 # Setup custom SDK sysroots.
 set(COMPILER_RT_DARWIN_SDK_SYSROOT ${COMPILER_RT_SOURCE_DIR}/SDKs/darwin)
 set(COMPILER_RT_LINUX_SDK_SYSROOT ${COMPILER_RT_SOURCE_DIR}/SDKs/linux)
+include(SanitizerUtils)
 
 # Detect whether the current target platform is 32-bit or 64-bit, and setup
 # the correct commandline flags needed to attempt to target 32-bit and 64-bit.
diff --git a/cmake/Modules/AddCompilerRT.cmake b/cmake/Modules/AddCompilerRT.cmake
index bf114a401..1ece3eb90 100644
--- a/cmake/Modules/AddCompilerRT.cmake
+++ b/cmake/Modules/AddCompilerRT.cmake
@@ -36,11 +36,10 @@ endmacro()
 # add_compiler_rt_static_runtime(<name> <arch>
 #                                SOURCES <source files>
 #                                CFLAGS <compile flags>
-#                                DEFS <compile definitions>
-#                                SYMS <symbols file>)
+#                                DEFS <compile definitions>)
 macro(add_compiler_rt_static_runtime name arch)
   if(CAN_TARGET_${arch})
-    parse_arguments(LIB "SOURCES;CFLAGS;DEFS;SYMS" "" ${ARGN})
+    parse_arguments(LIB "SOURCES;CFLAGS;DEFS" "" ${ARGN})
     add_library(${name} STATIC ${LIB_SOURCES})
     # Setup compile flags and definitions.
     set_target_compile_flags(${name}
@@ -53,13 +52,6 @@ macro(add_compiler_rt_static_runtime name arch)
     # Add installation command.
     install(TARGETS ${name}
       ARCHIVE DESTINATION ${COMPILER_RT_LIBRARY_INSTALL_DIR})
-    # Generate the .syms file if possible.
-    if(LIB_SYMS)
-      get_target_property(libfile ${name} LOCATION)
-      configure_file(${LIB_SYMS} ${libfile}.syms)
-      install(FILES ${libfile}.syms
-        DESTINATION ${COMPILER_RT_LIBRARY_INSTALL_DIR})
-    endif(LIB_SYMS)
   else()
     message(FATAL_ERROR "Archtecture ${arch} can't be targeted")
   endif()
diff --git a/cmake/Modules/SanitizerUtils.cmake b/cmake/Modules/SanitizerUtils.cmake
new file mode 100644
index 000000000..7a4f8670c
--- /dev/null
+++ b/cmake/Modules/SanitizerUtils.cmake
@@ -0,0 +1,22 @@
+include(LLVMParseArguments)
+
+set(SANITIZER_GEN_DYNAMIC_LIST
+  ${COMPILER_RT_SOURCE_DIR}/lib/sanitizer_common/scripts/gen_dynamic_list.py)
+
+# Create a target "<name>-symbols" that would generate the list of symbols
+# that need to be exported from sanitizer runtime "<name>". Function
+# interceptors are exported automatically, user can also provide files with
+# symbol names that should be exported as well.
+#   add_sanitizer_rt_symbols(<name> <files with extra symbols to export>)
+macro(add_sanitizer_rt_symbols name)
+  get_target_property(libfile ${name} LOCATION)
+  set(symsfile "${libfile}.syms")
+  add_custom_target(${name}-symbols ALL
+    COMMAND ${SANITIZER_GEN_DYNAMIC_LIST} ${libfile} ${ARGN}
+      > ${symsfile}
+    DEPENDS ${name} ${SANITIZER_GEN_DYNAMIC_LIST} ${ARGN}
+    WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
+    VERBATIM
+    SOURCES ${SANITIZER_GEN_DYNAMIC_LIST} ${ARGN})
+  install(FILES ${symsfile} DESTINATION ${COMPILER_RT_LIBRARY_INSTALL_DIR})
+endmacro()
diff --git a/lib/asan/CMakeLists.txt b/lib/asan/CMakeLists.txt
index fe14cc3df..5a2ce5466 100644
--- a/lib/asan/CMakeLists.txt
+++ b/lib/asan/CMakeLists.txt
@@ -102,16 +102,16 @@ else()
     add_compiler_rt_static_runtime(clang_rt.asan-${arch} ${arch}
       SOURCES ${ASAN_SOURCES} ${ASAN_SOURCE_LIBS}
       CFLAGS ${ASAN_CFLAGS}
-      DEFS ${ASAN_COMMON_DEFINITIONS}
-      SYMS asan.syms)
-    list(APPEND ASAN_RUNTIME_LIBRARIES clang_rt.asan-${arch})
+      DEFS ${ASAN_COMMON_DEFINITIONS})
+    add_sanitizer_rt_symbols(clang_rt.asan-${arch} asan.syms.extra)
+    list(APPEND ASAN_RUNTIME_LIBRARIES clang_rt.asan-${arch}
+      clang_rt.asan-${arch}-symbols)
 
     if (WIN32)
       add_compiler_rt_static_runtime(clang_rt.asan_dll_thunk-${arch} ${arch}
       SOURCES asan_dll_thunk.cc
       CFLAGS ${ASAN_CFLAGS} -DASAN_DLL_THUNK
-      DEFS ${ASAN_COMMON_DEFINITIONS}
-      SYMS asan.syms)
+      DEFS ${ASAN_COMMON_DEFINITIONS})
       list(APPEND ASAN_RUNTIME_LIBRARIES clang_rt.asan_dll_thunk-${arch})
     endif()
   endforeach()
diff --git a/lib/asan/asan.syms b/lib/asan/asan.syms
deleted file mode 100644
index fce367314..000000000
--- a/lib/asan/asan.syms
+++ /dev/null
@@ -1,5 +0,0 @@
-{
-  __asan_*;
-  __sanitizer_syscall_pre_*;
-  __sanitizer_syscall_post_*;
-};
diff --git a/lib/asan/asan.syms.extra b/lib/asan/asan.syms.extra
new file mode 100644
index 000000000..9e5ba4682
--- /dev/null
+++ b/lib/asan/asan.syms.extra
@@ -0,0 +1 @@
+__asan_*
diff --git a/lib/msan/CMakeLists.txt b/lib/msan/CMakeLists.txt
index b3e88ea29..98c296148 100644
--- a/lib/msan/CMakeLists.txt
+++ b/lib/msan/CMakeLists.txt
@@ -25,9 +25,10 @@ if(CAN_TARGET_${arch})
             $<TARGET_OBJECTS:RTInterception.${arch}>
             $<TARGET_OBJECTS:RTSanitizerCommon.${arch}>
             $<TARGET_OBJECTS:RTSanitizerCommonLibc.${arch}>
-    CFLAGS ${MSAN_RTL_CFLAGS}
-    SYMS msan.syms)
-  list(APPEND MSAN_RUNTIME_LIBRARIES clang_rt.msan-${arch})
+    CFLAGS ${MSAN_RTL_CFLAGS})
+  add_sanitizer_rt_symbols(clang_rt.msan-${arch} msan.syms.extra)
+  list(APPEND MSAN_RUNTIME_LIBRARIES clang_rt.msan-${arch}
+    clang_rt.msan-${arch}-symbols)
 endif()
 
 add_compiler_rt_resource_file(msan_blacklist msan_blacklist.txt)
diff --git a/lib/msan/msan.syms b/lib/msan/msan.syms
deleted file mode 100644
index 24bbaba47..000000000
--- a/lib/msan/msan.syms
+++ /dev/null
@@ -1,5 +0,0 @@
-{
-  __msan_*;
-  __sanitizer_syscall_pre_*;
-  __sanitizer_syscall_post_*;
-};
diff --git a/lib/msan/msan.syms.extra b/lib/msan/msan.syms.extra
new file mode 100644
index 000000000..aad41cf11
--- /dev/null
+++ b/lib/msan/msan.syms.extra
@@ -0,0 +1 @@
+__msan_*
diff --git a/lib/sanitizer_common/scripts/gen_dynamic_list.py b/lib/sanitizer_common/scripts/gen_dynamic_list.py
new file mode 100755
index 000000000..23b861eed
--- /dev/null
+++ b/lib/sanitizer_common/scripts/gen_dynamic_list.py
@@ -0,0 +1,76 @@
+#!/usr/bin/env python
+#===- lib/sanitizer_common/scripts/gen_dynamic_list.py ---------------------===#
+#
+#                     The LLVM Compiler Infrastructure
+#
+# This file is distributed under the University of Illinois Open Source
+# License. See LICENSE.TXT for details.
+#
+#===------------------------------------------------------------------------===#
+#
+# Generates the list of functions that should be exported from sanitizer
+# runtimes. The output format is recognized by --dynamic-list linker option.
+# Usage:
+#   gen_dynamic_list.py libclang_rt.*san*.a [ files ... ]
+#
+#===------------------------------------------------------------------------===#
+import re
+import os
+import subprocess
+import sys
+
+def get_global_functions(library):
+  functions = []
+  nm_proc = subprocess.Popen(['nm', library], stdout=subprocess.PIPE)
+  nm_out = nm_proc.communicate()[0].split('\n')
+  if nm_proc.returncode != 0:
+    raise subprocess.CalledProcessError(nm_proc.returncode, 'nm')
+  for line in nm_out:
+    cols = line.split(' ')
+    if (len(cols) == 3 and cols[1] in ('T', 'W')) :
+      functions.append(cols[2])
+  return functions
+
+result = []
+
+library = sys.argv[1]
+all_functions = get_global_functions(library)
+function_set = set(all_functions)
+new_delete = set(['_ZdaPv', '_ZdaPvRKSt9nothrow_t',
+                  '_ZdlPv', '_ZdlPvRKSt9nothrow_t',
+                  '_Znam', '_ZnamRKSt9nothrow_t',
+                  '_Znwm', '_ZnwmRKSt9nothrow_t'])
+versioned_functions = set(['memcpy', 'pthread_cond_broadcast',
+                           'pthread_cond_destroy', 'pthread_cond_signal',
+                           'pthread_cond_timedwait', 'pthread_cond_wait',
+                           'realpath', 'sched_getaffinity'])
+for func in all_functions:
+  # Export new/delete operators.
+  if func in new_delete:
+    result.append(func)
+    continue
+  # Export interceptors.
+  match = re.match('__interceptor_(.*)', func)
+  if match:
+    result.append(func)
+    # We have to avoid exporting the interceptors for versioned library
+    # functions due to gold internal error.
+    orig_name = match.group(1)
+    if orig_name in function_set and orig_name not in versioned_functions:
+      result.append(orig_name)
+    continue
+  # Export sanitizer interface functions.
+  if re.match('__sanitizer_(.*)', func):
+    result.append(func)
+
+# Additional exported functions from files.
+for fname in sys.argv[2:]:
+  f = open(fname, 'r')
+  for line in f:
+    result.append(line.rstrip())
+
+print '{'
+result.sort()
+for f in result:
+  print '  ' + f + ';'
+print '};'
diff --git a/lib/tsan/Makefile.old b/lib/tsan/Makefile.old
index 64dff8370..3443a0753 100644
--- a/lib/tsan/Makefile.old
+++ b/lib/tsan/Makefile.old
@@ -64,8 +64,6 @@ run: all
 
 presubmit:
 	../sanitizer_common/scripts/check_lint.sh
-	#./gen_dynamic_list.sh > rtl/tsan.syms.new
-	#diff rtl/tsan.syms rtl/tsan.syms.new
 	# Debug build with clang.
 	$(MAKE) -f Makefile.old clean
 	$(MAKE) -f Makefile.old run DEBUG=1 -j 16 CC=$(CLANG) CXX=$(CLANG)++
diff --git a/lib/tsan/gen_dynamic_list.sh b/lib/tsan/gen_dynamic_list.sh
deleted file mode 100755
index 4868b42be..000000000
--- a/lib/tsan/gen_dynamic_list.sh
+++ /dev/null
@@ -1,56 +0,0 @@
-#!/bin/bash
-set -e
-
-# Collect interceptor names from a source file.
-function collect() {
-  while read line ; do
-    if [[ $line =~ ^(.*)((TSAN_INTERCEPT|INTERCEPT_FUNCTION)\()([a-z,A-Z,0-9,_]+)(.*)$ ]] ; then
-      results+=" ${BASH_REMATCH[4]}"
-      results+=" __interceptor_${BASH_REMATCH[4]}"
-    fi
-  done < "$1"
-}
-
-# Interface functions.
-results+=" __tsan_init"
-results+=" __tsan_read*"
-results+=" __tsan_write*"
-results+=" __tsan_vptr*"
-results+=" __tsan_func*"
-results+=" __tsan_atomic*"
-results+=" __tsan_java*"
-results+=" __tsan_unaligned*"
-results+=" __tsan_release"
-results+=" __tsan_acquire"
-results+=" __sanitizer_unaligned*"
-results+=" __sanitizer_syscall*"
-results+=" _Znwm"
-results+=" _Znam"
-results+=" _ZnwmRKSt9nothrow_t"
-results+=" _ZnamRKSt9nothrow_t"
-results+=" _ZdlPv"
-results+=" _ZdlPvRKSt9nothrow_t"
-results+=" _ZdaPv"
-results+=" _ZdaPvRKSt9nothrow_t"
-results+=" Annotate*"
-results+=" WTFAnnotate*"
-results+=" RunningOnValgrind"
-
-collect rtl/tsan_interceptors.cc
-collect ../sanitizer_common/sanitizer_common_interceptors.inc
-
-results=`for i in $results; do echo $i; done | sort -f`
-echo "# AUTO GENERATED by compiler-rt/lib/tsan/gen_dynamic_list.sh; EDITING IS FUTILE."
-echo "{"
-NM=`nm rtl/libtsan.a`
-for i in $results; do
-  # Remove symbols that are not present in the library.
-  if [[ $NM =~ " $i" ]]; then
-    echo "  $i;"
-  else if [[ $i == *"*" ]]; then
-    echo "  $i;"
-  fi
-  fi
-done
-echo "};"
-
diff --git a/lib/tsan/rtl/CMakeLists.txt b/lib/tsan/rtl/CMakeLists.txt
index f1a8ff4d6..461c3b8c1 100644
--- a/lib/tsan/rtl/CMakeLists.txt
+++ b/lib/tsan/rtl/CMakeLists.txt
@@ -45,7 +45,8 @@ if(CAN_TARGET_x86_64 AND UNIX AND NOT APPLE)
             $<TARGET_OBJECTS:RTSanitizerCommon.${arch}>
             $<TARGET_OBJECTS:RTSanitizerCommonLibc.${arch}>
     CFLAGS ${TSAN_CFLAGS}
-    DEFS ${TSAN_COMMON_DEFINITIONS}
-    SYMS tsan.syms)
-  list(APPEND TSAN_RUNTIME_LIBRARIES clang_rt.tsan-${arch})
+    DEFS ${TSAN_COMMON_DEFINITIONS})
+  add_sanitizer_rt_symbols(clang_rt.tsan-${arch} tsan.syms.extra)
+  list(APPEND TSAN_RUNTIME_LIBRARIES clang_rt.tsan-${arch}
+    clang_rt.tsan-${arch}-symbols)
 endif()
diff --git a/lib/tsan/rtl/tsan.syms b/lib/tsan/rtl/tsan.syms
deleted file mode 100644
index 4464a0a23..000000000
--- a/lib/tsan/rtl/tsan.syms
+++ /dev/null
@@ -1,5 +0,0 @@
-{
-  __tsan_*;
-  __sanitizer_syscall_pre_*;
-  __sanitizer_syscall_post_*;
-};
diff --git a/lib/tsan/rtl/tsan.syms.extra b/lib/tsan/rtl/tsan.syms.extra
new file mode 100644
index 000000000..1558f66e4
--- /dev/null
+++ b/lib/tsan/rtl/tsan.syms.extra
@@ -0,0 +1,13 @@
+__tsan_init
+__tsan_read*
+__tsan_write*
+__tsan_vptr*
+__tsan_func*
+__tsan_atomic*
+__tsan_java*
+__tsan_unaligned*
+__tsan_release
+__tsan_acquire
+Annotate*
+WTFAnnotate*
+RunningOnValgrind
diff --git a/lib/ubsan/CMakeLists.txt b/lib/ubsan/CMakeLists.txt
index c8470bc6d..6a9c8c37c 100644
--- a/lib/ubsan/CMakeLists.txt
+++ b/lib/ubsan/CMakeLists.txt
@@ -34,17 +34,19 @@ else()
     # Main UBSan runtime.
     add_compiler_rt_static_runtime(clang_rt.ubsan-${arch} ${arch}
       SOURCES ${UBSAN_SOURCES}
-      CFLAGS ${UBSAN_CFLAGS}
-      SYMS ubsan.syms)
+      CFLAGS ${UBSAN_CFLAGS})
+    add_sanitizer_rt_symbols(clang_rt.ubsan-${arch} ubsan.syms.extra)
     # C++-specific parts of UBSan runtime. Requires a C++ ABI library.
     add_compiler_rt_static_runtime(clang_rt.ubsan_cxx-${arch} ${arch}
       SOURCES ${UBSAN_CXX_SOURCES}
-      CFLAGS ${UBSAN_CFLAGS}
-      SYMS ubsan.syms)
+      CFLAGS ${UBSAN_CFLAGS})
+    add_sanitizer_rt_symbols(clang_rt.ubsan_cxx-${arch} ubsan.syms.extra)
     list(APPEND UBSAN_RUNTIME_LIBRARIES
            clang_rt.san-${arch}
            clang_rt.ubsan-${arch}
-           clang_rt.ubsan_cxx-${arch})
+           clang_rt.ubsan-${arch}-symbols
+           clang_rt.ubsan_cxx-${arch}
+           clang_rt.ubsan_cxx-${arch}-symbols)
   endforeach()
 endif()
 
diff --git a/lib/ubsan/ubsan.syms b/lib/ubsan/ubsan.syms
deleted file mode 100644
index e74de33f0..000000000
--- a/lib/ubsan/ubsan.syms
+++ /dev/null
@@ -1 +0,0 @@
-{ __ubsan_*; };
diff --git a/lib/ubsan/ubsan.syms.extra b/lib/ubsan/ubsan.syms.extra
new file mode 100644
index 000000000..7f8be6944
--- /dev/null
+++ b/lib/ubsan/ubsan.syms.extra
@@ -0,0 +1 @@
+__ubsan_*