ARROW-2191: [C++] Only use specific version of jemalloc

ci/travis_install_toolchain.sh

@@ -24,7 +24,6 @@ source $TRAVIS_BUILD_DIR/ci/travis_install_conda.sh
 if [ ! -e $CPP_TOOLCHAIN ]; then
     # Set up C++ toolchain from conda-forge packages for faster builds
     conda create -y -q -p $CPP_TOOLCHAIN python=2.7 \
-        jemalloc=4.5.0.post \
         nomkl \
         boost-cpp \
         rapidjson \

cpp/CMakeLists.txt

@@ -127,10 +127,6 @@ if("${CMAKE_SOURCE_DIR}" STREQUAL "${CMAKE_CURRENT_SOURCE_DIR}")
     "Build the Arrow jemalloc-based allocator"
     OFF)
 
-  option(ARROW_JEMALLOC_USE_SHARED
-    "Rely on jemalloc shared libraries where relevant"
-    ON)
-
   option(ARROW_HDFS
     "Build the Arrow HDFS bridge"
     ON)
@@ -605,44 +601,28 @@ set(ARROW_TEST_LINK_LIBS ${ARROW_MIN_TEST_LIBS})
 
 if (ARROW_JEMALLOC)
   add_definitions(-DARROW_JEMALLOC)
-  # In the case that jemalloc is only available as a shared library also use it to
-  # link it in the static requirements. In contrast to other libraries we try in
-  # most cases to use the system provided version of jemalloc to better align with
-  # other potential users of jemalloc.
-  if (JEMALLOC_STATIC_LIB AND NOT ARROW_JEMALLOC_USE_SHARED)
-    set(ARROW_JEMALLOC_STATIC_LINK_LIBS jemalloc_static)
-  else()
-    set(ARROW_JEMALLOC_STATIC_LINK_LIBS jemalloc_shared)
-  endif()
-
-  if (NOT APPLE)
-      set(ARROW_JEMALLOC_STATIC_LINK_LIBS ${ARROW_JEMALLOC_STATIC_LINK_LIBS} ${PTHREAD_LIBRARY} rt)
-  endif()
-
-  if (ARROW_JEMALLOC_USE_SHARED)
-    set(ARROW_JEMALLOC_SHARED_LINK_LIBS jemalloc_shared)
+  add_definitions(-DARROW_JEMALLOC_INCLUDE_DIR=${JEMALLOC_INCLUDE_DIR})
+
+  if (CMAKE_COMPILER_IS_GNUCXX)
+    set(ARROW_JEMALLOC_LINK_LIBS
+      jemalloc_static
+      # For glibc <2.17 we need to link to librt.
+      # As we compile with --as-needed by default, the linker will omit this
+      # dependency if not required.
+      ${PTHREAD_LIBRARY}
+      rt
+      )
   else()
-    if (CMAKE_COMPILER_IS_GNUCXX)
-      set(ARROW_JEMALLOC_SHARED_LINK_LIBS
-        jemalloc_static
-        # For glibc <2.17 we need to link to librt.
-        # As we compile with --as-needed by default, the linker will omit this
-        # dependency if not required.
-        ${PTHREAD_LIBRARY}
-        rt
-        )
-    else()
-      set(ARROW_JEMALLOC_SHARED_LINK_LIBS
-        jemalloc_static
-        )
-    endif()
+    set(ARROW_JEMALLOC_LINK_LIBS
+      jemalloc_static
+      )
   endif()
   set(ARROW_SHARED_PRIVATE_LINK_LIBS
     ${ARROW_SHARED_PRIVATE_LINK_LIBS}
-    ${ARROW_JEMALLOC_SHARED_LINK_LIBS})
+    ${ARROW_JEMALLOC_LINK_LIBS})
   set(ARROW_STATIC_LINK_LIBS
     ${ARROW_STATIC_LINK_LIBS}
-    ${ARROW_JEMALLOC_STATIC_LINK_LIBS})
+    ${ARROW_JEMALLOC_LINK_LIBS})
 elseif (NOT MSVC)
   # We need to separate this as otherwise CMake would mess with the library
   # linking order.

cpp/cmake_modules/ThirdpartyToolchain.cmake

@@ -82,10 +82,6 @@ if (DEFINED ENV{RAPIDJSON_HOME})
   set(RAPIDJSON_HOME "$ENV{RAPIDJSON_HOME}")
 endif()
 
-if (DEFINED ENV{JEMALLOC_HOME})
-  set(JEMALLOC_HOME "$ENV{JEMALLOC_HOME}")
-endif()
-
 if (DEFINED ENV{GFLAGS_HOME})
   set(GFLAGS_HOME "$ENV{GFLAGS_HOME}")
 endif()
@@ -474,33 +470,37 @@ if (MSVC)
 endif()
 
 if (ARROW_JEMALLOC)
-  find_package(jemalloc)
-
-  if(NOT JEMALLOC_FOUND)
-    set(ARROW_JEMALLOC_USE_SHARED OFF)
-    set(JEMALLOC_PREFIX "${CMAKE_CURRENT_BINARY_DIR}/jemalloc_ep-prefix/src/jemalloc_ep/dist/")
-    set(JEMALLOC_HOME "${JEMALLOC_PREFIX}")
-    set(JEMALLOC_INCLUDE_DIR "${JEMALLOC_PREFIX}/include")
-    set(JEMALLOC_SHARED_LIB "${JEMALLOC_PREFIX}/lib/libjemalloc${CMAKE_SHARED_LIBRARY_SUFFIX}")
-    set(JEMALLOC_STATIC_LIB "${JEMALLOC_PREFIX}/lib/libjemalloc_pic${CMAKE_STATIC_LIBRARY_SUFFIX}")
-    set(JEMALLOC_VENDORED 1)
-    ExternalProject_Add(jemalloc_ep
-      URL ${CMAKE_CURRENT_SOURCE_DIR}/thirdparty/jemalloc/${JEMALLOC_VERSION}.tar.gz
-      CONFIGURE_COMMAND ./autogen.sh "--prefix=${JEMALLOC_PREFIX}" "--with-jemalloc-prefix=je_arrow_" "--with-private-namespace=je_arrow_private_" && touch doc/jemalloc.html && touch doc/jemalloc.3
-      ${EP_LOG_OPTIONS}
-      BUILD_IN_SOURCE 1
-      BUILD_COMMAND ${MAKE}
-      BUILD_BYPRODUCTS "${JEMALLOC_STATIC_LIB}" "${JEMALLOC_SHARED_LIB}"
-      INSTALL_COMMAND ${MAKE} -j1 install)
-  else()
-    set(JEMALLOC_VENDORED 0)
-  endif()
+  # We only use a vendored jemalloc as we want to control its version.
+  # Also our build of jemalloc is specially prefixed so that it will not 
+  # conflict with the default allocator as well as other jemalloc
+  # installations.
+  # find_package(jemalloc)
+
+  set(ARROW_JEMALLOC_USE_SHARED OFF)
+  set(JEMALLOC_PREFIX "${CMAKE_CURRENT_BINARY_DIR}/jemalloc_ep-prefix/src/jemalloc_ep/dist/")
+  set(JEMALLOC_HOME "${JEMALLOC_PREFIX}")
+  set(JEMALLOC_INCLUDE_DIR "${JEMALLOC_PREFIX}/include")
+  set(JEMALLOC_SHARED_LIB "${JEMALLOC_PREFIX}/lib/libjemalloc${CMAKE_SHARED_LIBRARY_SUFFIX}")
+  set(JEMALLOC_STATIC_LIB "${JEMALLOC_PREFIX}/lib/libjemalloc_pic${CMAKE_STATIC_LIBRARY_SUFFIX}")
+  set(JEMALLOC_VENDORED 1)
+  ExternalProject_Add(jemalloc_ep
+    URL ${CMAKE_CURRENT_SOURCE_DIR}/thirdparty/jemalloc/${JEMALLOC_VERSION}.tar.gz
+    PATCH_COMMAND touch doc/jemalloc.3 doc/jemalloc.html
+    CONFIGURE_COMMAND ./autogen.sh "--prefix=${JEMALLOC_PREFIX}" "--with-jemalloc-prefix=je_arrow_" "--with-private-namespace=je_arrow_private_"
+    ${EP_LOG_OPTIONS}
+    BUILD_IN_SOURCE 1
+    BUILD_COMMAND ${MAKE}
+    BUILD_BYPRODUCTS "${JEMALLOC_STATIC_LIB}" "${JEMALLOC_SHARED_LIB}"
+    INSTALL_COMMAND ${MAKE} -j1 install)
 
-  include_directories(SYSTEM ${JEMALLOC_INCLUDE_DIR})
+  # Don't use the include directory directly so that we can point to a path
+  # that is unique to our codebase.
+  include_directories(SYSTEM "${CMAKE_CURRENT_BINARY_DIR}/jemalloc_ep-prefix/src/")
   ADD_THIRDPARTY_LIB(jemalloc
     STATIC_LIB ${JEMALLOC_STATIC_LIB}
     SHARED_LIB ${JEMALLOC_SHARED_LIB}
     DEPS ${PTHREAD_LIBRARY})
+  add_dependencies(jemalloc_static jemalloc_ep)
 endif()
 
 ## Google PerfTools

cpp/src/arrow/CMakeLists.txt

@@ -65,6 +65,10 @@ if (ARROW_GPU)
   add_subdirectory(gpu)
 endif()
 
+if (ARROW_JEMALLOC AND JEMALLOC_VENDORED)
+  add_dependencies(arrow_dependencies jemalloc_static)
+endif()
+
 if (ARROW_WITH_BROTLI)
   add_definitions(-DARROW_WITH_BROTLI)
   SET(ARROW_SRCS util/compression_brotli.cc ${ARROW_SRCS})

cpp/src/arrow/memory_pool.cc

@@ -31,7 +31,8 @@
 #ifdef ARROW_JEMALLOC
 // Needed to support jemalloc 3 and 4
 #define JEMALLOC_MANGLE
-#include <jemalloc/jemalloc.h>
+// Explicitly link to our version of jemalloc
+#include "jemalloc_ep/dist/include/jemalloc/jemalloc.h"
 #endif
 
 namespace arrow {