Fix WriteBatchBase::DeleteRange API comment

The `DeleteRange` end key is exclusive, not inclusive. Updated API comment accordingly.

.deprecated_arcconfig

@@ -1,17 +0,0 @@
-{
-  "project_id" : "rocksdb",
-  "conduit_uri" : "https://phabricator.fb.com/api/",
-  "copyright_holder" : "Facebook",
-  "load" : [
-    "arcanist_util"
-  ],
-  "lint.engine" : "FacebookFbcodeLintEngine",
-  "lint.engine.single.linter" : "FbcodeCppLinter",
-  "unit.engine" : "FacebookFbcodeUnitTestEngine",
-  "arcanist_configuration" : "FacebookArcanistConfiguration",
-  "base" : "git:HEAD^, hg:.^",
-  "git.default-relative-commit" : "HEAD^",
-  "git:arc.feature.start.default" : "origin/master",
-  "arc.feature.start.default" : "master",
-  "history.immutable" : false
-}

.gitignore

@@ -39,13 +39,14 @@ coverage/COVERAGE_REPORT
 .gdbhistory
 .gdb_history
 package/
-.phutil_module_cache
 unity.a
 tags
 etags
 rocksdb_dump
 rocksdb_undump
 db_test2
+trace_analyzer
+trace_analyzer_test
 
 java/out
 java/target

.lgtm.yml

@@ -0,0 +1,4 @@
+extraction:
+  cpp:
+    index:
+      build_command: make static_lib

.travis.yml

@@ -15,12 +15,23 @@ cache:
   - apt
 
 addons:
-   apt:
-      packages: ['zlib1g-dev', 'libbz2-dev', 'libsnappy-dev', 'curl', 'libgflags-dev', 'mingw-w64']
+  apt:
+    sources:
+      - ubuntu-toolchain-r-test
+    packages:
+      - curl
+      - g++-8
+      - libbz2-dev
+      - libgflags-dev
+      - libsnappy-dev
+      - mingw-w64
+      - zlib1g-dev
 env:
   - TEST_GROUP=platform_dependent # 16-18 minutes
   - TEST_GROUP=1 # 33-35 minutes
-  - TEST_GROUP=2 # 30-32 minutes
+  - TEST_GROUP=2 # 18-20 minutes
+  - TEST_GROUP=3 # 20-22 minutes
+  - TEST_GROUP=4 # 12-14 minutes
   # Run java tests
   - JOB_NAME=java_test # 4-11 minutes
   # Build ROCKSDB_LITE
@@ -28,6 +39,7 @@ env:
   # Build examples
   - JOB_NAME=examples # 5-7 minutes
   - JOB_NAME=cmake # 3-5 minutes
+  - JOB_NAME=cmake-gcc8 # 3-5 minutes
   - JOB_NAME=cmake-mingw # 3 minutes
 
 matrix:
@@ -36,6 +48,12 @@ matrix:
     env: TEST_GROUP=1
   - os: osx
     env: TEST_GROUP=2
+  - os: osx
+    env: TEST_GROUP=3
+  - os: osx
+    env: TEST_GROUP=4
+  - os: osx
+    env: JOB_NAME=cmake-gcc8
   - os : osx
     env: JOB_NAME=cmake-mingw
   - os : linux
@@ -46,9 +64,15 @@ matrix:
 # https://docs.travis-ci.com/user/caching/#ccache-cache
 install:
   - if [ "${TRAVIS_OS_NAME}" == osx ]; then
-      brew install ccache;
+      brew install ccache zstd lz4 snappy xz;
       PATH=$PATH:/usr/local/opt/ccache/libexec;
     fi
+  - if [ "${JOB_NAME}" == cmake-gcc8 ]; then
+      CC=gcc-8 && CXX=g++-8;
+    fi
+  - if [[ "${JOB_NAME}" == cmake* ]] && [ "${TRAVIS_OS_NAME}" == linux ]; then
+      mkdir cmake-dist && curl -sfSL https://cmake.org/files/v3.8/cmake-3.8.1-Linux-x86_64.tar.gz | tar --strip-components=1 -C cmake-dist -xz && export PATH=$PWD/cmake-dist/bin:$PATH;
+    fi
 
 before_script:
   # Increase the maximum number of open file descriptors, since some tests use
@@ -57,14 +81,41 @@ before_script:
 
 script:
   - ${CXX} --version
-  - if [ "${TEST_GROUP}" == 'platform_dependent' ]; then OPT=-DTRAVIS V=1 make -j4 all; OPT=-DTRAVIS V=1 ROCKSDBTESTS_END=db_block_cache_test make -j4 check_some; fi
-  - if [ "${TEST_GROUP}" == '1' ]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=db_block_cache_test ROCKSDBTESTS_END=comparator_db_test make -j4 check_some; fi
-  - if [ "${TEST_GROUP}" == '2' ]; then OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=comparator_db_test make -j4 check_some; fi
-  - if [ "${JOB_NAME}" == 'java_test' ]; then OPT=-DTRAVIS V=1 make clean jclean && make rocksdbjava jtest; fi
-  - if [ "${JOB_NAME}" == 'lite_build' ]; then OPT="-DTRAVIS -DROCKSDB_LITE" V=1 make -j4 static_lib; fi
-  - if [ "${JOB_NAME}" == 'examples' ]; then OPT=-DTRAVIS V=1 make -j4 static_lib; cd examples; make -j4; fi
-  - if [ "${JOB_NAME}" == 'cmake' ]; then mkdir build && cd build && cmake .. && make -j4 rocksdb; fi
-  - if [ "${JOB_NAME}" == 'cmake-mingw' ]; then mkdir build && cd build && cmake .. -DCMAKE_C_COMPILER=x86_64-w64-mingw32-gcc -DCMAKE_CXX_COMPILER=x86_64-w64-mingw32-g++ -DCMAKE_SYSTEM_NAME=Windows && make -j4 rocksdb; fi
+  - if [ `command -v ccache` ]; then ccache -C; fi
+  - case $TEST_GROUP in
+    platform_dependent)
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_END=db_block_cache_test make -j4 all_but_some_tests check_some
+      ;;
+    1)
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=db_block_cache_test ROCKSDBTESTS_END=full_filter_block_test make -j4 check_some
+      ;;
+    2)
+      OPT=-DTRAVIS V=1 make -j4 tools && OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=full_filter_block_test ROCKSDBTESTS_END=write_batch_with_index_test make -j4 check_some
+      ;;
+    3)
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=write_batch_with_index_test ROCKSDBTESTS_END=write_prepared_transaction_test make -j4 check_some
+      ;;
+    4)
+      OPT=-DTRAVIS V=1 ROCKSDBTESTS_START=write_prepared_transaction_test make -j4 check_some
+      ;;
+    esac
+  - case $JOB_NAME in
+    java_test)
+      OPT=-DTRAVIS V=1 make clean jclean && make rocksdbjava jtest
+      ;;
+    lite_build)
+      OPT='-DTRAVIS -DROCKSDB_LITE' V=1 make -j4 static_lib tools
+      ;;
+    examples)
+      OPT=-DTRAVIS V=1 make -j4 static_lib && cd examples && make -j4
+      ;;
+    cmake-mingw)
+      mkdir build && cd build && cmake -DJNI=1 .. -DCMAKE_C_COMPILER=x86_64-w64-mingw32-gcc -DCMAKE_CXX_COMPILER=x86_64-w64-mingw32-g++ -DCMAKE_SYSTEM_NAME=Windows && make -j4 rocksdb rocksdbjni
+      ;;
+    cmake*)
+      mkdir build && cd build && cmake -DJNI=1 .. -DCMAKE_BUILD_TYPE=Release && make -j4 rocksdb rocksdbjni
+      ;;
+    esac
 notifications:
     email:
       - leveldb@fb.com

AUTHORS

@@ -9,3 +9,4 @@ Sanjay Ghemawat <sanjay@google.com>
 # Partial list of contributors:
 Kevin Regan <kevin.d.regan@gmail.com>
 Johan Bilien <jobi@litl.com>
+Matthew Von-Maszewski <https://github.com/matthewvon> (Basho Technologies)

CMakeLists.txt

@@ -14,9 +14,9 @@
 #        cd build
 # 3. Run cmake to generate project files for Windows, add more options to enable required third-party libraries.
 #    See thirdparty.inc for more information.
-#        sample command: cmake -G "Visual Studio 14 Win64" -DGFLAGS=1 -DSNAPPY=1 -DJEMALLOC=1 -DJNI=1 ..
+#        sample command: cmake -G "Visual Studio 15 Win64" -DWITH_GFLAGS=1 -DWITH_SNAPPY=1 -DWITH_JEMALLOC=1 -DWITH_JNI=1 ..
 # 4. Then build the project in debug mode (you may want to add /m[:<N>] flag to run msbuild in <N> parallel threads
-#                                          or simply /m ot use all avail cores)
+#                                          or simply /m to use all avail cores)
 #        msbuild rocksdb.sln
 #
 #        rocksdb.sln build features exclusions of test only code in Release. If you build ALL_BUILD then everything
@@ -32,21 +32,36 @@
 # 3. cmake ..
 # 4. make -j
 
-cmake_minimum_required(VERSION 2.6)
+cmake_minimum_required(VERSION 2.8.12)
 project(rocksdb)
+enable_language(CXX)
+enable_language(C)
+enable_language(ASM)
 
 if(POLICY CMP0042)
   cmake_policy(SET CMP0042 NEW)
 endif()
 
-list(APPEND CMAKE_MODULE_PATH "${CMAKE_SOURCE_DIR}/cmake/modules/")
+list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_LIST_DIR}/cmake/modules/")
 
+option(WITH_JEMALLOC "build with JeMalloc" OFF)
+option(WITH_SNAPPY "build with SNAPPY" OFF)
+option(WITH_LZ4 "build with lz4" OFF)
+option(WITH_ZLIB "build with zlib" OFF)
+option(WITH_ZSTD "build with zstd" OFF)
+option(WITH_WINDOWS_UTF8_FILENAMES "use UTF8 as characterset for opening files, regardles of the system code page" OFF)
+if (WITH_WINDOWS_UTF8_FILENAMES)
+  add_definitions(-DROCKSDB_WINDOWS_UTF8_FILENAMES)
+endif()
 if(MSVC)
+  # Defaults currently different for GFLAGS.
+  #  We will address find_package work a little later
+  option(WITH_GFLAGS "build with GFlags" OFF)
+  option(WITH_XPRESS "build with windows built in compression" OFF)
   include(${CMAKE_CURRENT_SOURCE_DIR}/thirdparty.inc)
 else()
-  option(WITH_JEMALLOC "build with JeMalloc" OFF)
   if(CMAKE_SYSTEM_NAME MATCHES "FreeBSD")
-    # FreeBSD has jemaloc as default malloc
+    # FreeBSD has jemalloc as default malloc
     # but it does not have all the jemalloc files in include/...
     set(WITH_JEMALLOC ON)
   else()
@@ -54,28 +69,65 @@ else()
       find_package(JeMalloc REQUIRED)
       add_definitions(-DROCKSDB_JEMALLOC -DJEMALLOC_NO_DEMANGLE)
       include_directories(${JEMALLOC_INCLUDE_DIR})
+      list(APPEND THIRDPARTY_LIBS ${JEMALLOC_LIBRARIES})
     endif()
   endif()
-  option(WITH_SNAPPY "build with SNAPPY" OFF)
+
+  # No config file for this
+  option(WITH_GFLAGS "build with GFlags" ON)
+  if(WITH_GFLAGS)
+    find_package(gflags)
+    if(gflags_FOUND)
+      add_definitions(-DGFLAGS=1)
+      include_directories(${gflags_INCLUDE_DIR})
+      list(APPEND THIRDPARTY_LIBS ${gflags_LIBRARIES})
+    endif()
+  endif()
+
   if(WITH_SNAPPY)
     find_package(snappy REQUIRED)
     add_definitions(-DSNAPPY)
     include_directories(${SNAPPY_INCLUDE_DIR})
     list(APPEND THIRDPARTY_LIBS ${SNAPPY_LIBRARIES})
   endif()
+
+  if(WITH_ZLIB)
+    find_package(zlib REQUIRED)
+    add_definitions(-DZLIB)
+    if(ZLIB_INCLUDE_DIRS)
+      # CMake 3
+      include_directories(${ZLIB_INCLUDE_DIRS})
+    else()
+      # CMake 2
+      include_directories(${ZLIB_INCLUDE_DIR})
+    endif()
+    list(APPEND THIRDPARTY_LIBS ${ZLIB_LIBRARIES})
+  endif()
+
+  option(WITH_BZ2 "build with bzip2" OFF)
+  if(WITH_BZ2)
+    find_package(bzip2 REQUIRED)
+    add_definitions(-DBZIP2)
+    include_directories(${BZIP2_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${BZIP2_LIBRARIES})
+  endif()
+
+  if(WITH_LZ4)
+    find_package(lz4 REQUIRED)
+    add_definitions(-DLZ4)
+    include_directories(${LZ4_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${LZ4_LIBRARIES})
+  endif()
+
+  if(WITH_ZSTD)
+    find_package(zstd REQUIRED)
+    add_definitions(-DZSTD)
+    include_directories(${ZSTD_INCLUDE_DIR})
+    list(APPEND THIRDPARTY_LIBS ${ZSTD_LIBRARIES})
+  endif()
 endif()
 
-if(WIN32)
-  execute_process(COMMAND powershell -noprofile -Command "Get-Date -format MM_dd_yyyy" OUTPUT_VARIABLE DATE)
-  execute_process(COMMAND powershell -noprofile -Command "Get-Date -format HH:mm:ss" OUTPUT_VARIABLE TIME)
-  string(REGEX REPLACE "(..)_(..)_..(..).*" "\\1/\\2/\\3" DATE "${DATE}")
-  string(REGEX REPLACE "(..):(.....).*" " \\1:\\2" TIME "${TIME}")
-  set(GIT_DATE_TIME "${DATE} ${TIME}")
-else()
-  execute_process(COMMAND date "+%Y/%m/%d %H:%M:%S" OUTPUT_VARIABLE DATETIME)
-  string(REGEX REPLACE "\n" "" DATETIME ${DATETIME})
-  set(GIT_DATE_TIME "${DATETIME}")
-endif()
+string(TIMESTAMP GIT_DATE_TIME "%Y/%m/%d %H:%M:%S" UTC)
 
 find_package(Git)
 
@@ -91,20 +143,17 @@ endif()
 
 string(REGEX REPLACE "[^0-9a-f]+" "" GIT_SHA "${GIT_SHA}")
 
-if(NOT WIN32)
-  execute_process(COMMAND
-      "./build_tools/version.sh" "full"
-      WORKING_DIRECTORY ${PROJECT_SOURCE_DIR}
-      OUTPUT_VARIABLE ROCKSDB_VERSION
-  )
-  string(STRIP "${ROCKSDB_VERSION}" ROCKSDB_VERSION)
-  execute_process(COMMAND
-      "./build_tools/version.sh" "major"
-      WORKING_DIRECTORY ${PROJECT_SOURCE_DIR}
-      OUTPUT_VARIABLE ROCKSDB_VERSION_MAJOR
-  )
-  string(STRIP "${ROCKSDB_VERSION_MAJOR}" ROCKSDB_VERSION_MAJOR)
-endif()
+
+# Read rocksdb version from version.h header file.
+file(READ include/rocksdb/version.h version_header_file)
+string(REGEX MATCH "#define ROCKSDB_MAJOR ([0-9]+)" _ ${version_header_file})
+set(ROCKSDB_VERSION_MAJOR ${CMAKE_MATCH_1})
+string(REGEX MATCH "#define ROCKSDB_MINOR ([0-9]+)" _ ${version_header_file})
+set(ROCKSDB_VERSION_MINOR ${CMAKE_MATCH_1})
+string(REGEX MATCH "#define ROCKSDB_PATCH ([0-9]+)" _ ${version_header_file})
+set(ROCKSDB_VERSION_PATCH ${CMAKE_MATCH_1})
+set(ROCKSDB_VERSION ${ROCKSDB_VERSION_MAJOR}.${ROCKSDB_VERSION_MINOR}.${ROCKSDB_VERSION_PATCH})
+
 
 option(WITH_MD_LIBRARY "build with MD" ON)
 if(WIN32 AND MSVC)
@@ -115,33 +164,81 @@ if(WIN32 AND MSVC)
   endif()
 endif()
 
+set(BUILD_VERSION_CC ${CMAKE_BINARY_DIR}/build_version.cc)
+configure_file(util/build_version.cc.in ${BUILD_VERSION_CC} @ONLY)
+add_library(build_version OBJECT ${BUILD_VERSION_CC})
+target_include_directories(build_version PRIVATE
+  ${CMAKE_CURRENT_SOURCE_DIR}/util)
+if(MSVC)
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W4 /wd4127 /wd4800 /wd4996 /wd4351 /wd4100 /wd4204 /wd4324")
+else()
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -W -Wextra -Wall")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wsign-compare -Wshadow -Wno-unused-parameter -Wno-unused-variable -Woverloaded-virtual -Wnon-virtual-dtor -Wno-missing-field-initializers -Wno-strict-aliasing")
+  if(MINGW)
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-format")
+  endif()
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
+  if(NOT CMAKE_BUILD_TYPE STREQUAL "Debug")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-omit-frame-pointer")
+    include(CheckCXXCompilerFlag)
+    CHECK_CXX_COMPILER_FLAG("-momit-leaf-frame-pointer" HAVE_OMIT_LEAF_FRAME_POINTER)
+    if(HAVE_OMIT_LEAF_FRAME_POINTER)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -momit-leaf-frame-pointer")
+    endif()
+  endif()
+endif()
+
+include(CheckCCompilerFlag)
+if(CMAKE_SYSTEM_PROCESSOR MATCHES "ppc64le")
+  CHECK_C_COMPILER_FLAG("-maltivec" HAS_ALTIVEC)
+  if(HAS_ALTIVEC)
+    message(STATUS " HAS_ALTIVEC yes")
+    set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -maltivec")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -maltivec")
+    set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mcpu=power8")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mcpu=power8")
+  endif(HAS_ALTIVEC)
+endif(CMAKE_SYSTEM_PROCESSOR MATCHES "ppc64le")
+
 option(PORTABLE "build a portable binary" OFF)
 option(FORCE_SSE42 "force building with SSE4.2, even when PORTABLE=ON" OFF)
 if(PORTABLE)
   # MSVC does not need a separate compiler flag to enable SSE4.2; if nmmintrin.h
   # is available, it is available by default.
   if(FORCE_SSE42 AND NOT MSVC)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse4.2")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -msse4.2 -mpclmul")
   endif()
 else()
   if(MSVC)
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /arch:AVX2")
   else()
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native")
+    if(NOT HAVE_POWER8)
+      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native")
+    endif()
   endif()
 endif()
 
-set(CMAKE_REQUIRED_FLAGS ${CMAKE_CXX_FLAGS})
 include(CheckCXXSourceCompiles)
+if(NOT MSVC)
+  set(CMAKE_REQUIRED_FLAGS "-msse4.2 -mpclmul")
+endif()
 CHECK_CXX_SOURCE_COMPILES("
 #include <cstdint>
 #include <nmmintrin.h>
+#include <wmmintrin.h>
 int main() {
   volatile uint32_t x = _mm_crc32_u32(0, 0);
+  const auto a = _mm_set_epi64x(0, 0);
+  const auto b = _mm_set_epi64x(0, 0);
+  const auto c = _mm_clmulepi64_si128(a, b, 0x00);
+  auto d = _mm_cvtsi128_si64(c);
 }
 " HAVE_SSE42)
+unset(CMAKE_REQUIRED_FLAGS)
 if(HAVE_SSE42)
   add_definitions(-DHAVE_SSE42)
+  add_definitions(-DHAVE_PCLMUL)
 elseif(FORCE_SSE42)
   message(FATAL_ERROR "FORCE_SSE42=ON but unable to compile with SSE4.2 enabled")
 endif()
@@ -158,32 +255,6 @@ if(HAVE_THREAD_LOCAL)
   add_definitions(-DROCKSDB_SUPPORT_THREAD_LOCAL)
 endif()
 
-set(BUILD_VERSION_CC ${CMAKE_BINARY_DIR}/build_version.cc)
-configure_file(util/build_version.cc.in ${BUILD_VERSION_CC} @ONLY)
-add_library(build_version OBJECT ${BUILD_VERSION_CC})
-target_include_directories(build_version PRIVATE
-  ${CMAKE_CURRENT_SOURCE_DIR}/util)
-if(MSVC)
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /Zi /nologo /EHsc /GS /Gd /GR /GF /fp:precise /Zc:wchar_t /Zc:forScope /errorReport:queue")
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /FC /d2Zi+ /W3 /wd4127 /wd4800 /wd4996 /wd4351")
-else()
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -W -Wextra -Wall")
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wsign-compare -Wshadow -Wno-unused-parameter -Wno-unused-variable -Woverloaded-virtual -Wnon-virtual-dtor -Wno-missing-field-initializers")
-  if(MINGW)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-format")
-  endif()
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")
-  if(NOT CMAKE_BUILD_TYPE STREQUAL "Debug")
-    add_definitions(-DNDEBUG)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -O2 -fno-omit-frame-pointer")
-    include(CheckCXXCompilerFlag)
-    CHECK_CXX_COMPILER_FLAG("-momit-leaf-frame-pointer" HAVE_OMIT_LEAF_FRAME_POINTER)
-    if(HAVE_OMIT_LEAF_FRAME_POINTER)
-      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -momit-leaf-frame-pointer")
-    endif()
-  endif()
-endif()
-
 option(FAIL_ON_WARNINGS "Treat compile warnings as errors" ON)
 if(FAIL_ON_WARNINGS)
   if(MSVC)
@@ -224,21 +295,69 @@ if(WITH_UBSAN)
   endif()
 endif()
 
-# Used to run CI build and tests so we can run faster
-set(OPTIMIZE_DEBUG_DEFAULT 0)        # Debug build is unoptimized by default use -DOPTDBG=1 to optimize
-
-if(DEFINED OPTDBG)
-   set(OPTIMIZE_DEBUG ${OPTDBG})
-else()
-   set(OPTIMIZE_DEBUG ${OPTIMIZE_DEBUG_DEFAULT})
+option(WITH_NUMA "build with NUMA policy support" OFF)
+if(WITH_NUMA)
+  find_package(NUMA REQUIRED)
+  add_definitions(-DNUMA)
+  include_directories(${NUMA_INCLUDE_DIR})
+  list(APPEND THIRDPARTY_LIBS ${NUMA_LIBRARIES})
 endif()
 
-if(MSVC)
-  if((${OPTIMIZE_DEBUG} EQUAL 1))
-    message(STATUS "Debug optimization is enabled")
-    set(CMAKE_CXX_FLAGS_DEBUG "/Oxt /${RUNTIME_LIBRARY}d")
+option(WITH_TBB "build with Threading Building Blocks (TBB)" OFF)
+if(WITH_TBB)
+  find_package(TBB REQUIRED)
+  add_definitions(-DTBB)
+  include_directories(${TBB_INCLUDE_DIR})
+  list(APPEND THIRDPARTY_LIBS ${TBB_LIBRARIES})
+endif()
+
+# Stall notifications eat some performance from inserts
+option(DISABLE_STALL_NOTIF "Build with stall notifications" OFF)
+if(DISABLE_STALL_NOTIF)
+  add_definitions(-DROCKSDB_DISABLE_STALL_NOTIFICATION)
+endif()
+
+
+if(DEFINED USE_RTTI)
+  if(USE_RTTI)
+    message(STATUS "Enabling RTTI")
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DROCKSDB_USE_RTTI")
+    set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DROCKSDB_USE_RTTI")
   else()
-    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /RTC1 /Gm /${RUNTIME_LIBRARY}d")
+    if(MSVC)
+      message(STATUS "Disabling RTTI in Release builds. Always on in Debug.")
+      set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DROCKSDB_USE_RTTI")
+      set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /GR-")
+    else()
+      message(STATUS "Disabling RTTI in Release builds")
+      set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fno-rtti")
+      set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -fno-rtti")
+    endif()
+  endif()
+else()
+  message(STATUS "Enabling RTTI in Debug builds only (default)")
+  set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -DROCKSDB_USE_RTTI")
+  if(MSVC)
+     set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /GR-")
+  else()
+    set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -fno-rtti")
+  endif()
+endif()
+
+# Used to run CI build and tests so we can run faster
+option(OPTDBG "Build optimized debug build with MSVC" OFF)
+option(WITH_RUNTIME_DEBUG "build with debug version of runtime library" ON)
+if(MSVC)
+  if(OPTDBG)
+    message(STATUS "Debug optimization is enabled")
+    set(CMAKE_CXX_FLAGS_DEBUG "/Oxt")
+  else()
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /Od /RTC1 /Gm")
+  endif()
+  if(WITH_RUNTIME_DEBUG)
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /${RUNTIME_LIBRARY}d")
+  else()
+    set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /${RUNTIME_LIBRARY}")
   endif()
   set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Oxt /Zp8 /Gm- /Gy /${RUNTIME_LIBRARY}")
 
@@ -253,7 +372,7 @@ endif()
 option(ROCKSDB_LITE "Build RocksDBLite version" OFF)
 if(ROCKSDB_LITE)
   add_definitions(-DROCKSDB_LITE)
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions -Os")
 endif()
 
 if(CMAKE_SYSTEM_NAME MATCHES "Cygwin")
@@ -291,10 +410,8 @@ if(NOT WIN32)
 endif()
 
 option(WITH_FALLOCATE "build with fallocate" ON)
-
 if(WITH_FALLOCATE)
-  include(CheckCSourceCompiles)
-  CHECK_C_SOURCE_COMPILES("
+  CHECK_CXX_SOURCE_COMPILES("
 #include <fcntl.h>
 #include <linux/falloc.h>
 int main() {
@@ -307,12 +424,38 @@ int main() {
   endif()
 endif()
 
-include(CheckFunctionExists)
-CHECK_FUNCTION_EXISTS(malloc_usable_size HAVE_MALLOC_USABLE_SIZE)
+CHECK_CXX_SOURCE_COMPILES("
+#include <fcntl.h>
+int main() {
+  int fd = open(\"/dev/null\", 0);
+  sync_file_range(fd, 0, 1024, SYNC_FILE_RANGE_WRITE);
+}
+" HAVE_SYNC_FILE_RANGE_WRITE)
+if(HAVE_SYNC_FILE_RANGE_WRITE)
+  add_definitions(-DROCKSDB_RANGESYNC_PRESENT)
+endif()
+
+CHECK_CXX_SOURCE_COMPILES("
+#include <pthread.h>
+int main() {
+  (void) PTHREAD_MUTEX_ADAPTIVE_NP;
+}
+" HAVE_PTHREAD_MUTEX_ADAPTIVE_NP)
+if(HAVE_PTHREAD_MUTEX_ADAPTIVE_NP)
+  add_definitions(-DROCKSDB_PTHREAD_ADAPTIVE_MUTEX)
+endif()
+
+include(CheckCXXSymbolExists)
+check_cxx_symbol_exists(malloc_usable_size malloc.h HAVE_MALLOC_USABLE_SIZE)
 if(HAVE_MALLOC_USABLE_SIZE)
   add_definitions(-DROCKSDB_MALLOC_USABLE_SIZE)
 endif()
 
+check_cxx_symbol_exists(sched_getcpu sched.h HAVE_SCHED_GETCPU)
+if(HAVE_SCHED_GETCPU)
+  add_definitions(-DROCKSDB_SCHED_GETCPU_PRESENT)
+endif()
+
 include_directories(${PROJECT_SOURCE_DIR})
 include_directories(${PROJECT_SOURCE_DIR}/include)
 include_directories(SYSTEM ${PROJECT_SOURCE_DIR}/third-party/gtest-1.7.0/fused-src)
@@ -334,6 +477,7 @@ set(SOURCES
         db/compaction_iterator.cc
         db/compaction_job.cc
         db/compaction_picker.cc
+        db/compaction_picker_fifo.cc
         db/compaction_picker_universal.cc
         db/convenience.cc
         db/db_filesnapshot.cc
@@ -348,6 +492,7 @@ set(SOURCES
         db/db_info_dumper.cc
         db/db_iter.cc
         db/dbformat.cc
+        db/error_handler.cc
         db/event_helpers.cc
         db/experimental.cc
         db/external_sst_file_ingestion_job.cc
@@ -356,15 +501,16 @@ set(SOURCES
         db/flush_scheduler.cc
         db/forward_iterator.cc
         db/internal_stats.cc
+        db/logs_with_prep_tracker.cc
         db/log_reader.cc
         db/log_writer.cc
         db/malloc_stats.cc
-        db/managed_iterator.cc
         db/memtable.cc
         db/memtable_list.cc
         db/merge_helper.cc
         db/merge_operator.cc
         db/range_del_aggregator.cc
+        db/range_tombstone_fragmenter.cc
         db/repair.cc
         db/snapshot_impl.cc
         db/table_cache.cc
@@ -415,11 +561,14 @@ set(SOURCES
         table/block_based_table_factory.cc
         table/block_based_table_reader.cc
         table/block_builder.cc
+        table/block_fetcher.cc
         table/block_prefix_index.cc
         table/bloom_block.cc
         table/cuckoo_table_builder.cc
         table/cuckoo_table_factory.cc
         table/cuckoo_table_reader.cc
+        table/data_block_hash_index.cc
+        table/data_block_footer.cc
         table/flush_block_policy.cc
         table/format.cc
         table/full_filter_block.cc
@@ -435,6 +584,7 @@ set(SOURCES
         table/plain_table_index.cc
         table/plain_table_key_coding.cc
         table/plain_table_reader.cc
+        table/sst_file_reader.cc
         table/sst_file_writer.cc
         table/table_properties.cc
         table/two_level_iterator.cc
@@ -443,13 +593,16 @@ set(SOURCES
         tools/ldb_cmd.cc
         tools/ldb_tool.cc
         tools/sst_dump_tool.cc
+        tools/trace_analyzer_tool.cc
         util/arena.cc
         util/auto_roll_logger.cc
         util/bloom.cc
         util/coding.cc
         util/compaction_job_stats_impl.cc
         util/comparator.cc
+        util/compression_context_cache.cc
         util/concurrent_arena.cc
+        util/concurrent_task_limiter_impl.cc
         util/crc32c.cc
         util/delete_scheduler.cc
         util/dynamic_bloom.cc
@@ -459,6 +612,7 @@ set(SOURCES
         util/filename.cc
         util/filter_policy.cc
         util/hash.cc
+        util/jemalloc_nodump_allocator.cc
         util/log_buffer.cc
         util/murmurhash.cc
         util/random.cc
@@ -466,43 +620,39 @@ set(SOURCES
         util/slice.cc
         util/sst_file_manager_impl.cc
         util/status.cc
-        util/status_message.cc
         util/string_util.cc
         util/sync_point.cc
+        util/sync_point_impl.cc
         util/testutil.cc
         util/thread_local.cc
         util/threadpool_imp.cc
+        util/trace_replay.cc
         util/transaction_test_util.cc
         util/xxhash.cc
         utilities/backupable/backupable_db.cc
+        utilities/blob_db/blob_compaction_filter.cc
         utilities/blob_db/blob_db.cc
         utilities/blob_db/blob_db_impl.cc
+        utilities/blob_db/blob_db_impl_filesnapshot.cc
         utilities/blob_db/blob_dump_tool.cc
         utilities/blob_db/blob_file.cc
         utilities/blob_db/blob_log_reader.cc
         utilities/blob_db/blob_log_writer.cc
         utilities/blob_db/blob_log_format.cc
-        utilities/blob_db/ttl_extractor.cc
+        utilities/cassandra/cassandra_compaction_filter.cc
+        utilities/cassandra/format.cc
+        utilities/cassandra/merge_operator.cc
         utilities/checkpoint/checkpoint_impl.cc
-        utilities/col_buf_decoder.cc
-        utilities/col_buf_encoder.cc
-        utilities/column_aware_encoding_util.cc
         utilities/compaction_filters/remove_emptyvalue_compactionfilter.cc
-        utilities/date_tiered/date_tiered_db_impl.cc
         utilities/debug.cc
-        utilities/document/document_db.cc
-        utilities/document/json_document.cc
-        utilities/document/json_document_builder.cc
         utilities/env_mirror.cc
         utilities/env_timed.cc
-        utilities/geodb/geodb_impl.cc
         utilities/leveldb_options/leveldb_options.cc
         utilities/lua/rocks_lua_compaction_filter.cc
         utilities/memory/memory_util.cc
+        utilities/merge_operators/bytesxor.cc
         utilities/merge_operators/max.cc
         utilities/merge_operators/put.cc
-        utilities/merge_operators/cassandra/format.cc
-        utilities/merge_operators/cassandra/merge_operator.cc
         utilities/merge_operators/string_append/stringappend.cc
         utilities/merge_operators/string_append/stringappend2.cc
         utilities/merge_operators/uint64add.cc
@@ -513,23 +663,39 @@ set(SOURCES
         utilities/persistent_cache/block_cache_tier_metadata.cc
         utilities/persistent_cache/persistent_cache_tier.cc
         utilities/persistent_cache/volatile_tier_impl.cc
-        utilities/redis/redis_lists.cc
         utilities/simulator_cache/sim_cache.cc
-        utilities/spatialdb/spatial_db.cc
         utilities/table_properties_collectors/compact_on_deletion_collector.cc
+        utilities/trace/file_trace_reader_writer.cc
         utilities/transactions/optimistic_transaction_db_impl.cc
-        utilities/transactions/optimistic_transaction_impl.cc
+        utilities/transactions/optimistic_transaction.cc
+        utilities/transactions/pessimistic_transaction.cc
+        utilities/transactions/pessimistic_transaction_db.cc
+        utilities/transactions/snapshot_checker.cc
         utilities/transactions/transaction_base.cc
-        utilities/transactions/transaction_db_impl.cc
         utilities/transactions/transaction_db_mutex_impl.cc
-        utilities/transactions/transaction_impl.cc
         utilities/transactions/transaction_lock_mgr.cc
         utilities/transactions/transaction_util.cc
+        utilities/transactions/write_prepared_txn.cc
+        utilities/transactions/write_prepared_txn_db.cc
+        utilities/transactions/write_unprepared_txn.cc
+        utilities/transactions/write_unprepared_txn_db.cc
         utilities/ttl/db_ttl_impl.cc
         utilities/write_batch_with_index/write_batch_with_index.cc
         utilities/write_batch_with_index/write_batch_with_index_internal.cc
         $<TARGET_OBJECTS:build_version>)
 
+if(HAVE_SSE42 AND NOT MSVC)
+  set_source_files_properties(
+    util/crc32c.cc
+    PROPERTIES COMPILE_FLAGS "-msse4.2 -mpclmul")
+endif()
+
+if(HAVE_POWER8)
+  list(APPEND SOURCES
+    util/crc32c_ppc.c
+    util/crc32c_ppc_asm.S)
+endif(HAVE_POWER8)
+
 if(WIN32)
   list(APPEND SOURCES
     port/win/io_win.cc
@@ -537,8 +703,18 @@ if(WIN32)
     port/win/env_default.cc
     port/win/port_win.cc
     port/win/win_logger.cc
-    port/win/win_thread.cc
+    port/win/win_thread.cc)
+
+if(WITH_XPRESS)
+  list(APPEND SOURCES
     port/win/xpress_win.cc)
+endif()
+
+if(WITH_JEMALLOC)
+  list(APPEND SOURCES
+    port/win/win_jemalloc.cc)
+endif()
+
 else()
   list(APPEND SOURCES
     port/port_posix.cc
@@ -549,6 +725,14 @@ endif()
 set(ROCKSDB_STATIC_LIB rocksdb${ARTIFACT_SUFFIX})
 set(ROCKSDB_SHARED_LIB rocksdb-shared${ARTIFACT_SUFFIX})
 set(ROCKSDB_IMPORT_LIB ${ROCKSDB_SHARED_LIB})
+
+option(WITH_LIBRADOS "Build with librados" OFF)
+if(WITH_LIBRADOS)
+  list(APPEND SOURCES
+    utilities/env_librados.cc)
+  list(APPEND THIRDPARTY_LIBS rados)
+endif()
+
 if(WIN32)
   set(SYSTEM_LIBS ${SYSTEM_LIBS} Shlwapi.lib Rpcrt4.lib)
   set(LIBS ${ROCKSDB_STATIC_LIB} ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
@@ -567,13 +751,6 @@ else()
                         OUTPUT_NAME "rocksdb")
 endif()
 
-option(WITH_LIBRADOS "Build with librados" OFF)
-if(WITH_LIBRADOS)
-  list(APPEND SOURCES
-    utilities/env_librados.cc)
-  list(APPEND THIRDPARTY_LIBS rados)
-endif()
-
 add_library(${ROCKSDB_STATIC_LIB} STATIC ${SOURCES})
 target_link_libraries(${ROCKSDB_STATIC_LIB}
   ${THIRDPARTY_LIBS} ${SYSTEM_LIBS})
@@ -600,7 +777,73 @@ else()
   message(STATUS "JNI library is disabled")
 endif()
 
-set(TESTS
+# Installation and packaging
+if(WIN32)
+  option(ROCKSDB_INSTALL_ON_WINDOWS "Enable install target on Windows" OFF)
+endif()
+if(NOT WIN32 OR ROCKSDB_INSTALL_ON_WINDOWS)
+  if(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT)
+    if(${CMAKE_SYSTEM_NAME} STREQUAL "Linux")
+      # Change default installation prefix on Linux to /usr
+      set(CMAKE_INSTALL_PREFIX /usr CACHE PATH "Install path prefix, prepended onto install directories." FORCE)
+    endif()
+  endif()
+
+  include(GNUInstallDirs)
+  include(CMakePackageConfigHelpers)
+
+  set(package_config_destination ${CMAKE_INSTALL_LIBDIR}/cmake/rocksdb)
+
+  configure_package_config_file(
+    ${CMAKE_CURRENT_LIST_DIR}/cmake/RocksDBConfig.cmake.in RocksDBConfig.cmake
+    INSTALL_DESTINATION ${package_config_destination}
+  )
+
+  write_basic_package_version_file(
+    RocksDBConfigVersion.cmake
+    VERSION ${ROCKSDB_VERSION}
+    COMPATIBILITY SameMajorVersion
+  )
+
+  install(DIRECTORY include/rocksdb COMPONENT devel DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}")
+
+  install(
+    TARGETS ${ROCKSDB_STATIC_LIB}
+    EXPORT RocksDBTargets
+    COMPONENT devel
+    ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
+  )
+
+  install(
+    TARGETS ${ROCKSDB_SHARED_LIB}
+    EXPORT RocksDBTargets
+    COMPONENT runtime
+    ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}"
+    LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}"
+    INCLUDES DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
+  )
+
+  install(
+    EXPORT RocksDBTargets
+    COMPONENT devel
+    DESTINATION ${package_config_destination}
+    NAMESPACE RocksDB::
+  )
+
+  install(
+    FILES
+    ${CMAKE_CURRENT_BINARY_DIR}/RocksDBConfig.cmake
+    ${CMAKE_CURRENT_BINARY_DIR}/RocksDBConfigVersion.cmake
+    COMPONENT devel
+    DESTINATION ${package_config_destination}
+  )
+endif()
+
+option(WITH_TESTS "build with tests" ON)
+if(WITH_TESTS)
+  set(TESTS
         cache/cache_test.cc
         cache/lru_cache_test.cc
         db/column_family_test.cc
@@ -613,6 +856,7 @@ set(TESTS
         db/corruption_test.cc
         db/cuckoo_table_db_test.cc
         db/db_basic_test.cc
+        db/db_blob_index_test.cc
         db/db_block_cache_test.cc
         db/db_bloom_filter_test.cc
         db/db_compaction_filter_test.cc
@@ -622,6 +866,7 @@ set(TESTS
         db/db_inplace_update_test.cc
         db/db_io_failure_test.cc
         db/db_iter_test.cc
+        db/db_iter_stress_test.cc
         db/db_iterator_test.cc
         db/db_log_iter_test.cc
         db/db_memtable_test.cc
@@ -640,6 +885,8 @@ set(TESTS
         db/db_write_test.cc
         db/dbformat_test.cc
         db/deletefile_test.cc
+        db/error_handler_test.cc
+        db/obsolete_files_test.cc
         db/external_sst_file_basic_test.cc
         db/external_sst_file_test.cc
         db/fault_injection_test.cc
@@ -656,6 +903,8 @@ set(TESTS
         db/perf_context_test.cc
         db/plain_table_db_test.cc
         db/prefix_test.cc
+        db/range_del_aggregator_test.cc
+        db/range_tombstone_fragmenter_test.cc
         db/repair_test.cc
         db/table_properties_collector_test.cc
         db/version_builder_test.cc
@@ -678,14 +927,18 @@ set(TESTS
         options/options_test.cc
         table/block_based_filter_block_test.cc
         table/block_test.cc
+        table/cleanable_test.cc
         table/cuckoo_table_builder_test.cc
         table/cuckoo_table_reader_test.cc
+        table/data_block_hash_index_test.cc
         table/full_filter_block_test.cc
         table/merger_test.cc
+        table/sst_file_reader_test.cc
         table/table_test.cc
         tools/ldb_cmd_test.cc
         tools/reduce_levels_test.cc
         tools/sst_dump_test.cc
+        tools/trace_analyzer_test.cc
         util/arena_test.cc
         util/auto_roll_logger_test.cc
         util/autovector_test.cc
@@ -700,124 +953,117 @@ set(TESTS
         util/hash_test.cc
         util/heap_test.cc
         util/rate_limiter_test.cc
+        util/repeatable_thread_test.cc
         util/slice_transform_test.cc
         util/timer_queue_test.cc
         util/thread_list_test.cc
         util/thread_local_test.cc
         utilities/backupable/backupable_db_test.cc
         utilities/blob_db/blob_db_test.cc
+        utilities/cassandra/cassandra_functional_test.cc
+        utilities/cassandra/cassandra_format_test.cc
+        utilities/cassandra/cassandra_row_merge_test.cc
+        utilities/cassandra/cassandra_serialize_test.cc
         utilities/checkpoint/checkpoint_test.cc
-        utilities/column_aware_encoding_test.cc
-        utilities/date_tiered/date_tiered_test.cc
-        utilities/document/document_db_test.cc
-        utilities/document/json_document_test.cc
-        utilities/geodb/geodb_test.cc
         utilities/lua/rocks_lua_test.cc
         utilities/memory/memory_test.cc
-        utilities/merge_operators/cassandra/cassandra_merge_test.cc
-        utilities/merge_operators/cassandra/cassandra_format_test.cc
-        utilities/merge_operators/cassandra/cassandra_row_merge_test.cc
-        utilities/merge_operators/cassandra/cassandra_serialize_test.cc
         utilities/merge_operators/string_append/stringappend_test.cc
         utilities/object_registry_test.cc
         utilities/option_change_migration/option_change_migration_test.cc
         utilities/options/options_util_test.cc
         utilities/persistent_cache/hash_table_test.cc
         utilities/persistent_cache/persistent_cache_test.cc
-        utilities/redis/redis_lists_test.cc
-        utilities/spatialdb/spatial_db_test.cc
+        utilities/simulator_cache/sim_cache_test.cc
         utilities/table_properties_collectors/compact_on_deletion_collector_test.cc
         utilities/transactions/optimistic_transaction_test.cc
         utilities/transactions/transaction_test.cc
+        utilities/transactions/write_prepared_transaction_test.cc
+        utilities/transactions/write_unprepared_transaction_test.cc
         utilities/ttl/ttl_test.cc
         utilities/write_batch_with_index/write_batch_with_index_test.cc
-)
-if(WITH_LIBRADOS)
-  list(APPEND TESTS utilities/env_librados_test.cc)
-endif()
+  )
+  if(WITH_LIBRADOS)
+    list(APPEND TESTS utilities/env_librados_test.cc)
+  endif()
 
-set(BENCHMARKS
-  cache/cache_bench.cc
-  memtable/memtablerep_bench.cc
-  tools/db_bench.cc
-  table/table_reader_bench.cc
-  utilities/column_aware_encoding_exp.cc
-  utilities/persistent_cache/hash_table_bench.cc)
-add_library(testharness OBJECT util/testharness.cc)
-foreach(sourcefile ${BENCHMARKS})
-  get_filename_component(exename ${sourcefile} NAME_WE)
-  add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile}
-    $<TARGET_OBJECTS:testharness>)
-  target_link_libraries(${exename}${ARTIFACT_SUFFIX} gtest ${LIBS})
-endforeach(sourcefile ${BENCHMARKS})
-
-# For test util library that is build only in DEBUG mode
-# and linked to tests. Add test only code that is not #ifdefed for Release here.
-set(TESTUTIL_SOURCE
-    db/db_test_util.cc
-    monitoring/thread_status_updater_debug.cc
-    table/mock_table.cc
-    util/fault_injection_test_env.cc
-    utilities/merge_operators/cassandra/test_utils.cc
-)
-# test utilities are only build in debug
-enable_testing()
-add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND})
-set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
-add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
-if(MSVC)
-  set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
-endif()
-set_target_properties(${TESTUTILLIB}
-      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
-      )
-
-# Tests are excluded from Release builds
-set(TEST_EXES ${TESTS})
-
-foreach(sourcefile ${TEST_EXES})
+  set(BENCHMARKS
+    cache/cache_bench.cc
+    memtable/memtablerep_bench.cc
+    db/range_del_aggregator_bench.cc
+    tools/db_bench.cc
+    table/table_reader_bench.cc
+    utilities/persistent_cache/hash_table_bench.cc)
+  add_library(testharness OBJECT util/testharness.cc)
+  foreach(sourcefile ${BENCHMARKS})
     get_filename_component(exename ${sourcefile} NAME_WE)
     add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile}
       $<TARGET_OBJECTS:testharness>)
-    set_target_properties(${exename}${ARTIFACT_SUFFIX}
-      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
-      )
-    target_link_libraries(${exename}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} gtest ${LIBS})
-    if(NOT "${exename}" MATCHES "db_sanity_test")
+    target_link_libraries(${exename}${ARTIFACT_SUFFIX} gtest ${LIBS})
+  endforeach(sourcefile ${BENCHMARKS})
+
+  # For test util library that is build only in DEBUG mode
+  # and linked to tests. Add test only code that is not #ifdefed for Release here.
+  set(TESTUTIL_SOURCE
+      db/db_test_util.cc
+      monitoring/thread_status_updater_debug.cc
+      table/mock_table.cc
+      util/fault_injection_test_env.cc
+      utilities/cassandra/test_utils.cc
+  )
+  # test utilities are only build in debug
+  enable_testing()
+  add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND})
+  set(TESTUTILLIB testutillib${ARTIFACT_SUFFIX})
+  add_library(${TESTUTILLIB} STATIC ${TESTUTIL_SOURCE})
+  if(MSVC)
+    set_target_properties(${TESTUTILLIB} PROPERTIES COMPILE_FLAGS "/Fd${CMAKE_CFG_INTDIR}/testutillib${ARTIFACT_SUFFIX}.pdb")
+  endif()
+  set_target_properties(${TESTUTILLIB}
+        PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+        )
+
+  # Tests are excluded from Release builds
+  set(TEST_EXES ${TESTS})
+
+  foreach(sourcefile ${TEST_EXES})
+      get_filename_component(exename ${sourcefile} NAME_WE)
+      add_executable(${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX} ${sourcefile}
+        $<TARGET_OBJECTS:testharness>)
+      set_target_properties(${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX}
+        PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+        OUTPUT_NAME ${exename}${ARTIFACT_SUFFIX}
+        )
+      target_link_libraries(${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX} testutillib${ARTIFACT_SUFFIX} gtest ${LIBS})
+      if(NOT "${exename}" MATCHES "db_sanity_test")
+        add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
+        add_dependencies(check ${CMAKE_PROJECT_NAME}_${exename}${ARTIFACT_SUFFIX})
+      endif()
+  endforeach(sourcefile ${TEST_EXES})
+
+  # C executables must link to a shared object
+  set(C_TESTS db/c_test.c)
+  set(C_TEST_EXES ${C_TESTS})
+
+  foreach(sourcefile ${C_TEST_EXES})
+      string(REPLACE ".c" "" exename ${sourcefile})
+      string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
+      add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
+      set_target_properties(${exename}${ARTIFACT_SUFFIX}
+        PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
+        EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
+        )
+      target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${ROCKSDB_IMPORT_LIB} testutillib${ARTIFACT_SUFFIX})
       add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
       add_dependencies(check ${exename}${ARTIFACT_SUFFIX})
-    endif()
-endforeach(sourcefile ${TEST_EXES})
-
-# C executables must link to a shared object
-set(C_TESTS db/c_test.c)
-set(C_TEST_EXES ${C_TESTS})
-
-foreach(sourcefile ${C_TEST_EXES})
-    string(REPLACE ".c" "" exename ${sourcefile})
-    string(REGEX REPLACE "^((.+)/)+" "" exename ${exename})
-    add_executable(${exename}${ARTIFACT_SUFFIX} ${sourcefile})
-    set_target_properties(${exename}${ARTIFACT_SUFFIX}
-      PROPERTIES EXCLUDE_FROM_DEFAULT_BUILD_RELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_MINRELEASE 1
-      EXCLUDE_FROM_DEFAULT_BUILD_RELWITHDEBINFO 1
-      )
-    target_link_libraries(${exename}${ARTIFACT_SUFFIX} ${ROCKSDB_IMPORT_LIB} testutillib${ARTIFACT_SUFFIX})
-    add_test(NAME ${exename} COMMAND ${exename}${ARTIFACT_SUFFIX})
-    add_dependencies(check ${exename}${ARTIFACT_SUFFIX})
-endforeach(sourcefile ${C_TEST_EXES})
-add_subdirectory(tools)
-
-# Installation and packaging for Linux
-if(NOT WIN32)
-install(TARGETS ${ROCKSDB_STATIC_LIB} COMPONENT devel ARCHIVE DESTINATION lib64)
-install(TARGETS ${ROCKSDB_SHARED_LIB} COMPONENT runtime DESTINATION lib64)
-install(DIRECTORY "${PROJECT_SOURCE_DIR}/include/rocksdb/"
-        COMPONENT devel
-        DESTINATION include/rocksdb)
-set(CMAKE_INSTALL_PREFIX /usr)
+  endforeach(sourcefile ${C_TEST_EXES})
+endif()
+
+option(WITH_TOOLS "build with tools" ON)
+if(WITH_TOOLS)
+  add_subdirectory(tools)
 endif()

CODE_OF_CONDUCT.md

@@ -0,0 +1,3 @@
+# Code of Conduct
+
+Facebook has adopted a Code of Conduct that we expect project participants to adhere to. Please [read the full text](https://code.facebook.com/codeofconduct) so that you can understand what actions will and will not be tolerated.

CONTRIBUTING.md

@@ -1,5 +1,8 @@
 # Contributing to RocksDB
 
+## Code of Conduct
+The code of conduct is described in [`CODE_OF_CONDUCT.md`](CODE_OF_CONDUCT.md)
+
 ## Contributor License Agreement ("CLA")
 
 In order to accept your pull request, we need you to submit a CLA. You

HISTORY.md

@@ -1,25 +1,280 @@
 # Rocksdb Change Log
-## 5.7.5 (11/28/2017)
+## Unreleased
+### New Features
+* Enabled checkpoint on readonly db (DBImplReadOnly).
+* Make DB ignore dropped column families while committing results of atomic flush.
+* RocksDB may choose to preopen some files even if options.max_open_files != -1. This may make DB open slightly longer.
+* For users of dictionary compression with ZSTD v0.7.0+, we now reuse the same digested dictionary when compressing each of an SST file's data blocks for faster compression speeds.
+* For all users of dictionary compression who set `cache_index_and_filter_blocks == true`, we now store dictionary data used for decompression in the block cache for better control over memory usage. For users of ZSTD v1.1.4+ who compile with -DZSTD_STATIC_LINKING_ONLY, this includes a digested dictionary, which is used to increase decompression speed.
+* Add support for block checksums verification for external SST files before ingestion.
+
+### Public API Change
+* CompactionPri = kMinOverlappingRatio also uses compensated file size, which boosts file with lots of tombstones to be compacted first.
+* Transaction::GetForUpdate is extended with a do_validate parameter with default value of true. If false it skips validating the snapshot before doing the read. Similarly ::Merge, ::Put, ::Delete, and ::SingleDelete are extended with assume_tracked with default value of false. If true it indicates that call is assumed to be after a ::GetForUpdate.
+* `TableProperties::num_entries` and `TableProperties::num_deletions` now also account for number of range tombstones.
+* Remove geodb, spatial_db, document_db, json_document, date_tiered_db, and redis_lists.
+* With "ldb ----try_load_options", when wal_dir specified by the option file doesn't exist, ignore it.
+* Change time resolution in FileOperationInfo.
+* Deleting Blob files also go through SStFileManager.
+* Remove PlainTable's store_index_in_file feature. When opening an existing DB with index in SST files, the index and bloom filter will still be rebuild while SST files are opened, in the same way as there is no index in the file.
+
+### Bug Fixes
+* Fix a deadlock caused by compaction and file ingestion waiting for each other in the event of write stalls.
+* Fix a memory leak when files with range tombstones are read in mmap mode and block cache is enabled
+* Fix handling of corrupt range tombstone blocks such that corruptions cannot cause deleted keys to reappear
+* Lock free MultiGet
+* Fix incorrect `NotFound` point lookup result when querying the endpoint of a file that has been extended by a range tombstone.
+* Fix with pipelined write, write leaders's callback failure lead to the whole write group fail.
+
+### Change Default Options
+* Change options.compaction_pri's default to kMinOverlappingRatio
+
+## 5.18.0 (11/30/2018)
+### New Features
+* Introduced `JemallocNodumpAllocator` memory allocator. When being use, block cache will be excluded from core dump.
+* Introduced `PerfContextByLevel` as part of `PerfContext` which allows storing perf context at each level. Also replaced `__thread` with `thread_local` keyword for perf_context. Added per-level perf context for bloom filter and `Get` query.
+* With level_compaction_dynamic_level_bytes = true, level multiplier may be adjusted automatically when Level 0 to 1 compaction is lagged behind.
+* Introduced DB option `atomic_flush`. If true, RocksDB supports flushing multiple column families and atomically committing the result to MANIFEST. Useful when WAL is disabled.
+* Added `num_deletions` and `num_merge_operands` members to `TableProperties`.
+* Added "rocksdb.min-obsolete-sst-number-to-keep" DB property that reports the lower bound on SST file numbers that are being kept from deletion, even if the SSTs are obsolete.
+* Add xxhash64 checksum support
+* Introduced `MemoryAllocator`, which lets the user specify custom memory allocator for block based table.
+* Improved `DeleteRange` to prevent read performance degradation. The feature is no longer marked as experimental.
+
+### Public API Change
+* `DBOptions::use_direct_reads` now affects reads issued by `BackupEngine` on the database's SSTs.
+* `NO_ITERATORS` is divided into two counters `NO_ITERATOR_CREATED` and `NO_ITERATOR_DELETE`. Both of them are only increasing now, just as other counters.
+
+### Bug Fixes
+* Fix corner case where a write group leader blocked due to write stall blocks other writers in queue with WriteOptions::no_slowdown set.
+* Fix in-memory range tombstone truncation to avoid erroneously covering newer keys at a lower level, and include range tombstones in compacted files whose largest key is the range tombstone's start key.
+* Properly set the stop key for a truncated manual CompactRange
+* Fix slow flush/compaction when DB contains many snapshots. The problem became noticeable to us in DBs with 100,000+ snapshots, though it will affect others at different thresholds.
+* Fix the bug that WriteBatchWithIndex's SeekForPrev() doesn't see the entries with the same key.
+* Fix the bug where user comparator was sometimes fed with InternalKey instead of the user key. The bug manifests when during GenerateBottommostFiles.
+* Fix a bug in WritePrepared txns where if the number of old snapshots goes beyond the snapshot cache size (128 default) the rest will not be checked when evicting a commit entry from the commit cache.
+* Fixed Get correctness bug in the presence of range tombstones where merge operands covered by a range tombstone always result in NotFound.
+* Start populating `NO_FILE_CLOSES` ticker statistic, which was always zero previously.
+* The default value of NewBloomFilterPolicy()'s argument use_block_based_builder is changed to false. Note that this new default may cause large temp memory usage when building very large SST files.
+
+## 5.17.0 (10/05/2018)
+### Public API Change
+* `OnTableFileCreated` will now be called for empty files generated during compaction. In that case, `TableFileCreationInfo::file_path` will be "(nil)" and `TableFileCreationInfo::file_size` will be zero.
+* Add `FlushOptions::allow_write_stall`, which controls whether Flush calls start working immediately, even if it causes user writes to stall, or will wait until flush can be performed without causing write stall (similar to `CompactRangeOptions::allow_write_stall`). Note that the default value is false, meaning we add delay to Flush calls until stalling can be avoided when possible. This is behavior change compared to previous RocksDB versions, where Flush calls didn't check if they might cause stall or not.
+* Application using PessimisticTransactionDB is expected to rollback/commit recovered transactions before starting new ones. This assumption is used to skip concurrency control during recovery.
+* Expose column family id to `OnCompactionCompleted`.
+
+### New Features
+* TransactionOptions::skip_concurrency_control allows pessimistic transactions to skip the overhead of concurrency control. Could be used for optimizing certain transactions or during recovery.
+
+### Bug Fixes
+* Avoid creating empty SSTs and subsequently deleting them in certain cases during compaction.
+* Sync CURRENT file contents during checkpoint.
+
+## 5.16.3 (10/1/2018)
+### Bug Fixes
+* Fix crash caused when `CompactFiles` run with `CompactionOptions::compression == CompressionType::kDisableCompressionOption`. Now that setting causes the compression type to be chosen according to the column family-wide compression options.
+
+## 5.16.2 (9/21/2018)
+### Bug Fixes
+* Fix bug in partition filters with format_version=4.
+
+## 5.16.1 (9/17/2018)
+### Bug Fixes
+* Remove trace_analyzer_tool from rocksdb_lib target in TARGETS file.
+* Fix RocksDB Java build and tests.
+* Remove sync point in Block destructor.
+
+## 5.16.0 (8/21/2018)
+### Public API Change
+* The merge operands are passed to `MergeOperator::ShouldMerge` in the reversed order relative to how they were merged (passed to FullMerge or FullMergeV2) for performance reasons
+* GetAllKeyVersions() to take an extra argument of `max_num_ikeys`.
+* Using ZSTD dictionary trainer (i.e., setting `CompressionOptions::zstd_max_train_bytes` to a nonzero value) now requires ZSTD version 1.1.3 or later.
+
+### New Features
+* Changes the format of index blocks by delta encoding the index values, which are the block handles. This saves the encoding of BlockHandle::offset of the non-head index entries in each restart interval. The feature is backward compatible but not forward compatible. It is disabled by default unless format_version 4 or above is used.
+* Add a new tool: trace_analyzer. Trace_analyzer analyzes the trace file generated by using trace_replay API. It can convert the binary format trace file to a human readable txt file, output the statistics of the analyzed query types such as access statistics and size statistics, combining the dumped whole key space file to analyze, support query correlation analyzing, and etc. Current supported query types are: Get, Put, Delete, SingleDelete, DeleteRange, Merge, Iterator (Seek, SeekForPrev only).
+* Add hash index support to data blocks, which helps reducing the cpu utilization of point-lookup operations. This feature is backward compatible with the data block created without the hash index. It is disabled by default unless BlockBasedTableOptions::data_block_index_type is set to data_block_index_type = kDataBlockBinaryAndHash.
+
+### Bug Fixes
+* Fix a bug in misreporting the estimated partition index size in properties block.
+
+## 5.15.0 (7/17/2018)
+### Public API Change
+* Remove managed iterator. ReadOptions.managed is not effective anymore.
+* For bottommost_compression, a compatible CompressionOptions is added via `bottommost_compression_opts`. To keep backward compatible, a new boolean `enabled` is added to CompressionOptions. For compression_opts, it will be always used no matter what value of `enabled` is. For bottommost_compression_opts, it will only be used when user set `enabled=true`, otherwise, compression_opts will be used for bottommost_compression as default.
+* With LRUCache, when high_pri_pool_ratio > 0, midpoint insertion strategy will be enabled to put low-pri items to the tail of low-pri list (the midpoint) when they first inserted into the cache. This is to make cache entries never get hit age out faster, improving cache efficiency when large background scan presents.
+* For users of `Statistics` objects created via `CreateDBStatistics()`, the format of the string returned by its `ToString()` method has changed.
+* The "rocksdb.num.entries" table property no longer counts range deletion tombstones as entries.
+
+### New Features
+* Changes the format of index blocks by storing the key in their raw form rather than converting them to InternalKey. This saves 8 bytes per index key. The feature is backward compatible but not forward compatible. It is disabled by default unless format_version 3 or above is used.
+* Avoid memcpy when reading mmap files with OpenReadOnly and max_open_files==-1.
+* Support dynamically changing `ColumnFamilyOptions::ttl` via `SetOptions()`.
+* Add a new table property, "rocksdb.num.range-deletions", which counts the number of range deletion tombstones in the table.
+* Improve the performance of iterators doing long range scans by using readahead, when using direct IO.
+* pin_top_level_index_and_filter (default true) in BlockBasedTableOptions can be used in combination with cache_index_and_filter_blocks to prefetch and pin the top-level index of partitioned index and filter blocks in cache. It has no impact when cache_index_and_filter_blocks is false.
+* Write properties meta-block at the end of block-based table to save read-ahead IO.
+
+### Bug Fixes
+* Fix deadlock with enable_pipelined_write=true and max_successive_merges > 0
+* Check conflict at output level in CompactFiles.
+* Fix corruption in non-iterator reads when mmap is used for file reads
+* Fix bug with prefix search in partition filters where a shared prefix would be ignored from the later partitions. The bug could report an eixstent key as missing. The bug could be triggered if prefix_extractor is set and partition filters is enabled.
+* Change default value of `bytes_max_delete_chunk` to 0 in NewSstFileManager() as it doesn't work well with checkpoints.
+* Fix a bug caused by not copying the block trailer with compressed SST file, direct IO, prefetcher and no compressed block cache.
+* Fix write can stuck indefinitely if enable_pipelined_write=true. The issue exists since pipelined write was introduced in 5.5.0.
+
+## 5.14.0 (5/16/2018)
+### Public API Change
+* Add a BlockBasedTableOption to align uncompressed data blocks on the smaller of block size or page size boundary, to reduce flash reads by avoiding reads spanning 4K pages.
+* The background thread naming convention changed (on supporting platforms) to "rocksdb:<thread pool priority><thread number>", e.g., "rocksdb:low0".
+* Add a new ticker stat rocksdb.number.multiget.keys.found to count number of keys successfully read in MultiGet calls
+* Touch-up to write-related counters in PerfContext. New counters added: write_scheduling_flushes_compactions_time, write_thread_wait_nanos. Counters whose behavior was fixed or modified: write_memtable_time, write_pre_and_post_process_time, write_delay_time.
+* Posix Env's NewRandomRWFile() will fail if the file doesn't exist.
+* Now, `DBOptions::use_direct_io_for_flush_and_compaction` only applies to background writes, and `DBOptions::use_direct_reads` applies to both user reads and background reads. This conforms with Linux's `open(2)` manpage, which advises against simultaneously reading a file in buffered and direct modes, due to possibly undefined behavior and degraded performance.
+* Iterator::Valid() always returns false if !status().ok(). So, now when doing a Seek() followed by some Next()s, there's no need to check status() after every operation.
+* Iterator::Seek()/SeekForPrev()/SeekToFirst()/SeekToLast() always resets status().
+* Introduced `CompressionOptions::kDefaultCompressionLevel`, which is a generic way to tell RocksDB to use the compression library's default level. It is now the default value for `CompressionOptions::level`. Previously the level defaulted to -1, which gave poor compression ratios in ZSTD.
+
+### New Features
+* Introduce TTL for level compaction so that all files older than ttl go through the compaction process to get rid of old data.
+* TransactionDBOptions::write_policy can be configured to enable WritePrepared 2PC transactions. Read more about them in the wiki.
+* Add DB properties "rocksdb.block-cache-capacity", "rocksdb.block-cache-usage", "rocksdb.block-cache-pinned-usage" to show block cache usage.
+* Add `Env::LowerThreadPoolCPUPriority(Priority)` method, which lowers the CPU priority of background (esp. compaction) threads to minimize interference with foreground tasks.
+* Fsync parent directory after deleting a file in delete scheduler.
+* In level-based compaction, if bottom-pri thread pool was setup via `Env::SetBackgroundThreads()`, compactions to the bottom level will be delegated to that thread pool.
+* `prefix_extractor` has been moved from ImmutableCFOptions to MutableCFOptions, meaning it can be dynamically changed without a DB restart.
+
+### Bug Fixes
+* Fsync after writing global seq number to the ingestion file in ExternalSstFileIngestionJob.
+* Fix WAL corruption caused by race condition between user write thread and FlushWAL when two_write_queue is not set.
+* Fix `BackupableDBOptions::max_valid_backups_to_open` to not delete backup files when refcount cannot be accurately determined.
+* Fix memory leak when pin_l0_filter_and_index_blocks_in_cache is used with partitioned filters
+* Disable rollback of merge operands in WritePrepared transactions to work around an issue in MyRocks. It can be enabled back by setting TransactionDBOptions::rollback_merge_operands to true.
+* Fix wrong results by ReverseBytewiseComparator::FindShortSuccessor()
+
+### Java API Changes
+* Add `BlockBasedTableConfig.setBlockCache` to allow sharing a block cache across DB instances.
+* Added SstFileManager to the Java API to allow managing SST files across DB instances.
+
+## 5.13.0 (3/20/2018)
+### Public API Change
+* RocksDBOptionsParser::Parse()'s `ignore_unknown_options` argument will only be effective if the option file shows it is generated using a higher version of RocksDB than the current version.
+* Remove CompactionEventListener.
+
+### New Features
+* SstFileManager now can cancel compactions if they will result in max space errors. SstFileManager users can also use SetCompactionBufferSize to specify how much space must be leftover during a compaction for auxiliary file functions such as logging and flushing.
+* Avoid unnecessarily flushing in `CompactRange()` when the range specified by the user does not overlap unflushed memtables.
+* If `ColumnFamilyOptions::max_subcompactions` is set greater than one, we now parallelize large manual level-based compactions.
+* Add "rocksdb.live-sst-files-size" DB property to return total bytes of all SST files belong to the latest LSM tree.
+* NewSstFileManager to add an argument bytes_max_delete_chunk with default 64MB. With this argument, a file larger than 64MB will be ftruncated multiple times based on this size.
+
+### Bug Fixes
+* Fix a leak in prepared_section_completed_ where the zeroed entries would not removed from the map.
+* Fix WAL corruption caused by race condition between user write thread and backup/checkpoint thread.
+
+## 5.12.0 (2/14/2018)
+### Public API Change
+* Iterator::SeekForPrev is now a pure virtual method. This is to prevent user who implement the Iterator interface fail to implement SeekForPrev by mistake.
+* Add `include_end` option to make the range end exclusive when `include_end == false` in `DeleteFilesInRange()`.
+* Add `CompactRangeOptions::allow_write_stall`, which makes `CompactRange` start working immediately, even if it causes user writes to stall. The default value is false, meaning we add delay to `CompactRange` calls until stalling can be avoided when possible. Note this delay is not present in previous RocksDB versions.
+* Creating checkpoint with empty directory now returns `Status::InvalidArgument`; previously, it returned `Status::IOError`.
+* Adds a BlockBasedTableOption to turn off index block compression.
+* Close() method now returns a status when closing a db.
+
+### New Features
+* Improve the performance of iterators doing long range scans by using readahead.
+* Add new function `DeleteFilesInRanges()` to delete files in multiple ranges at once for better performance.
+* FreeBSD build support for RocksDB and RocksJava.
+* Improved performance of long range scans with readahead.
+* Updated to and now continuously tested in Visual Studio 2017.
+
+### Bug Fixes
+* Fix `DisableFileDeletions()` followed by `GetSortedWalFiles()` to not return obsolete WAL files that `PurgeObsoleteFiles()` is going to delete.
+* Fix Handle error return from WriteBuffer() during WAL file close and DB close.
+* Fix advance reservation of arena block addresses.
+* Fix handling of empty string as checkpoint directory.
+
+## 5.11.0 (01/08/2018)
+### Public API Change
+* Add `autoTune` and `getBytesPerSecond()` to RocksJava RateLimiter
+
+### New Features
+* Add a new histogram stat called rocksdb.db.flush.micros for memtable flush.
+* Add "--use_txn" option to use transactional API in db_stress.
+* Disable onboard cache for compaction output in Windows platform.
+* Improve the performance of iterators doing long range scans by using readahead.
+
+### Bug Fixes
+* Fix a stack-use-after-scope bug in ForwardIterator.
+* Fix builds on platforms including Linux, Windows, and PowerPC.
+* Fix buffer overrun in backup engine for DBs with huge number of files.
+* Fix a mislabel bug for bottom-pri compaction threads.
+* Fix DB::Flush() keep waiting after flush finish under certain condition.
+
+## 5.10.0 (12/11/2017)
+### Public API Change
+* When running `make` with environment variable `USE_SSE` set and `PORTABLE` unset, will use all machine features available locally. Previously this combination only compiled SSE-related features.
+
+### New Features
+* Provide lifetime hints when writing files on Linux. This reduces hardware write-amp on storage devices supporting multiple streams.
+* Add a DB stat, `NUMBER_ITER_SKIP`, which returns how many internal keys were skipped during iterations (e.g., due to being tombstones or duplicate versions of a key).
+* Add PerfContext counters, `key_lock_wait_count` and `key_lock_wait_time`, which measure the number of times transactions wait on key locks and total amount of time waiting.
+
 ### Bug Fixes
 * Fix IOError on WAL write doesn't propagate to write group follower
+* Make iterator invalid on merge error.
+* Fix performance issue in `IngestExternalFile()` affecting databases with large number of SST files.
+* Fix possible corruption to LSM structure when `DeleteFilesInRange()` deletes a subset of files spanned by a `DeleteRange()` marker.
 
-## 5.7.4 (08/31/2017)
-No significant changes.
+## 5.9.0 (11/1/2017)
+### Public API Change
+* `BackupableDBOptions::max_valid_backups_to_open == 0` now means no backups will be opened during BackupEngine initialization. Previously this condition disabled limiting backups opened.
+* `DBOptions::preserve_deletes` is a new option that allows one to specify that DB should not drop tombstones for regular deletes if they have sequence number larger than what was set by the new API call `DB::SetPreserveDeletesSequenceNumber(SequenceNumber seqnum)`. Disabled by default.
+* API call `DB::SetPreserveDeletesSequenceNumber(SequenceNumber seqnum)` was added, users who wish to preserve deletes are expected to periodically call this function to advance the cutoff seqnum (all deletes made before this seqnum can be dropped by DB). It's user responsibility to figure out how to advance the seqnum in the way so the tombstones are kept for the desired period of time, yet are eventually processed in time and don't eat up too much space.
+* `ReadOptions::iter_start_seqnum` was added;
+if set to something > 0 user will see 2 changes in iterators behavior 1) only keys written with sequence larger than this parameter would be returned and 2) the `Slice` returned by iter->key() now points to the memory that keep User-oriented representation of the internal key, rather than user key. New struct `FullKey` was added to represent internal keys, along with a new helper function `ParseFullKey(const Slice& internal_key, FullKey* result);`.
+* Deprecate trash_dir param in NewSstFileManager, right now we will rename deleted files to <name>.trash instead of moving them to trash directory
+* Allow setting a custom trash/DB size ratio limit in the SstFileManager, after which files that are to be scheduled for deletion are deleted immediately, regardless of any delete ratelimit.
+* Return an error on write if write_options.sync = true and write_options.disableWAL = true to warn user of inconsistent options. Previously we will not write to WAL and not respecting the sync options in this case.
+
+### New Features
+* CRC32C is now using the 3-way pipelined SSE algorithm `crc32c_3way` on supported platforms to improve performance. The system will choose to use this algorithm on supported platforms automatically whenever possible. If PCLMULQDQ is not supported it will fall back to the old Fast_CRC32 algorithm.
+* `DBOptions::writable_file_max_buffer_size` can now be changed dynamically.
+* `DBOptions::bytes_per_sync`, `DBOptions::compaction_readahead_size`, and `DBOptions::wal_bytes_per_sync` can now be changed dynamically, `DBOptions::wal_bytes_per_sync` will flush all memtables and switch to a new WAL file.
+* Support dynamic adjustment of rate limit according to demand for background I/O. It can be enabled by passing `true` to the `auto_tuned` parameter in `NewGenericRateLimiter()`. The value passed as `rate_bytes_per_sec` will still be respected as an upper-bound.
+* Support dynamically changing `ColumnFamilyOptions::compaction_options_fifo`.
+* Introduce `EventListener::OnStallConditionsChanged()` callback. Users can implement it to be notified when user writes are stalled, stopped, or resumed.
+* Add a new db property "rocksdb.estimate-oldest-key-time" to return oldest data timestamp. The property is available only for FIFO compaction with compaction_options_fifo.allow_compaction = false.
+* Upon snapshot release, recompact bottommost files containing deleted/overwritten keys that previously could not be dropped due to the snapshot. This alleviates space-amp caused by long-held snapshots.
+* Support lower bound on iterators specified via `ReadOptions::iterate_lower_bound`.
+* Support for differential snapshots (via iterator emitting the sequence of key-values representing the difference between DB state at two different sequence numbers). Supports preserving and emitting puts and regular deletes, doesn't support SingleDeletes, MergeOperator, Blobs and Range Deletes.
 
-## 5.7.3 (08/29/2017)
 ### Bug Fixes
+* Fix a potential data inconsistency issue during point-in-time recovery. `DB:Open()` will abort if column family inconsistency is found during PIT recovery.
+* Fix possible metadata corruption in databases using `DeleteRange()`.
+
+## 5.8.0 (08/30/2017)
+### Public API Change
+* Users of `Statistics::getHistogramString()` will see fewer histogram buckets and different bucket endpoints.
+* `Slice::compare` and BytewiseComparator `Compare` no longer accept `Slice`s containing nullptr.
+* `Transaction::Get` and `Transaction::GetForUpdate` variants with `PinnableSlice` added.
+
+### New Features
+* Add Iterator::Refresh(), which allows users to update the iterator state so that they can avoid some initialization costs of recreating iterators.
+* Replace dynamic_cast<> (except unit test) so people can choose to build with RTTI off. With make, release mode is by default built with -fno-rtti and debug mode is built without it. Users can override it by setting USE_RTTI=0 or 1.
+* Universal compactions including the bottom level can be executed in a dedicated thread pool. This alleviates head-of-line blocking in the compaction queue, which cause write stalling, particularly in multi-instance use cases. Users can enable this feature via `Env::SetBackgroundThreads(N, Env::Priority::BOTTOM)`, where `N > 0`.
+* Allow merge operator to be called even with a single merge operand during compactions, by appropriately overriding `MergeOperator::AllowSingleOperand`.
+* Add `DB::VerifyChecksum()`, which verifies the checksums in all SST files in a running DB.
+* Block-based table support for disabling checksums by setting `BlockBasedTableOptions::checksum = kNoChecksum`.
+
+### Bug Fixes
+* Fix wrong latencies in `rocksdb.db.get.micros`, `rocksdb.db.write.micros`, and `rocksdb.sst.read.micros`.
+* Fix incorrect dropping of deletions during intra-L0 compaction.
 * Fix transient reappearance of keys covered by range deletions when memtable prefix bloom filter is enabled.
 * Fix potentially wrong file smallest key when range deletions separated by snapshot are written together.
 
-## 5.7.2 (08/15/2017)
-### Bug Fixes
-* Fix incorrect dropping of deletions issue with FIFO compaction.
-* Fix LITE build compiler error with missing abort().
-
-## 5.7.1 (08/13/2017)
-### Bug Fixes
-* Fix incorrect dropping of deletions during intra-L0 compaction.
-
 ## 5.7.0 (07/13/2017)
 ### Public API Change
 * DB property "rocksdb.sstables" now prints keys in hex form.

INSTALL.md

@@ -43,6 +43,8 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
       command line flags processing. You can compile rocksdb library even
       if you don't have gflags installed.
 
+* If you wish to build the RocksJava static target, then cmake is required for building Snappy.
+
 ## Supported platforms
 
 * **Linux - Ubuntu**
@@ -107,12 +109,70 @@ to build a portable binary, add `PORTABLE=1` before your make commands, like thi
             * run `brew tap homebrew/versions; brew install gcc48 --use-llvm` to install gcc 4.8 (or higher).
     * run `brew install rocksdb`
 
+* **FreeBSD** (11.01):
+
+    * You can either install RocksDB from the Ports system using `cd /usr/ports/databases/rocksdb && make install`, or you can follow the details below to install dependencies and compile from source code:
+
+    * Install the dependencies for RocksDB:
+
+        export BATCH=YES
+        cd /usr/ports/devel/gmake && make install
+        cd /usr/ports/devel/gflags && make install
+
+        cd /usr/ports/archivers/snappy && make install
+        cd /usr/ports/archivers/bzip2 && make install
+        cd /usr/ports/archivers/liblz4 && make install
+        cd /usr/ports/archivesrs/zstd && make install
+
+        cd /usr/ports/devel/git && make install
+
+
+    * Install the dependencies for RocksJava (optional):
+
+        export BATCH=yes
+        cd /usr/ports/java/openjdk7 && make install
+
+    * Build RocksDB from source:
+        cd ~
+        git clone https://github.com/facebook/rocksdb.git
+        cd rocksdb
+        gmake static_lib
+
+    * Build RocksJava from source (optional):
+        cd rocksdb
+        export JAVA_HOME=/usr/local/openjdk7
+        gmake rocksdbjava
+
+* **OpenBSD** (6.3/-current):
+
+    * As RocksDB is not available in the ports yet you have to build it on your own:
+
+    * Install the dependencies for RocksDB:
+
+        pkg_add gmake gflags snappy bzip2 lz4 zstd git jdk bash findutils gnuwatch 
+
+    * Build RocksDB from source:
+
+        cd ~
+        git clone https://github.com/facebook/rocksdb.git
+        cd rocksdb
+        gmake static_lib
+
+    * Build RocksJava from source (optional):
+
+        cd rocksdb
+        export JAVA_HOME=/usr/local/jdk-1.8.0
+        export PATH=$PATH:/usr/local/jdk-1.8.0/bin
+        gmake rocksdbjava
+
 * **iOS**:
   * Run: `TARGET_OS=IOS make static_lib`. When building the project which uses rocksdb iOS library, make sure to define two important pre-processing macros: `ROCKSDB_LITE` and `IOS_CROSS_COMPILE`.
 
 * **Windows**:
   * For building with MS Visual Studio 13 you will need Update 4 installed.
   * Read and follow the instructions at CMakeLists.txt
+  * Or install via [vcpkg](https://github.com/microsoft/vcpkg) 
+       * run `vcpkg install rocksdb:x64-windows`
 
 * **AIX 6.1**
     * Install AIX Toolbox rpms with gcc

LANGUAGE-BINDINGS.md

@@ -9,5 +9,12 @@ This is the list of all known third-party language bindings for RocksDB. If some
 * Haskell - https://hackage.haskell.org/package/rocksdb-haskell
 * PHP - https://github.com/Photonios/rocksdb-php
 * C# - https://github.com/warrenfalk/rocksdb-sharp
-* Rust - https://github.com/spacejam/rust-rocksdb
+* Rust
+    * https://github.com/pingcap/rust-rocksdb (used in production fork of https://github.com/spacejam/rust-rocksdb)
+    * https://github.com/spacejam/rust-rocksdb
+    * https://github.com/bh1xuw/rust-rocks
 * D programming language - https://github.com/b1naryth1ef/rocksdb
+* Erlang - https://gitlab.com/barrel-db/erlang-rocksdb
+* Elixir - https://github.com/urbint/rox
+* Nim - https://github.com/status-im/nim-rocksdb
+* Swift and Objective-C (iOS/OSX) - https://github.com/iabudiab/ObjectiveRocks 

README.md

@@ -1,8 +1,8 @@
 ## RocksDB: A Persistent Key-Value Store for Flash and RAM Storage
 
-[![Build Status](https://travis-ci.org/facebook/rocksdb.svg?branch=master)](https://travis-ci.org/facebook/rocksdb)
-[![Build status](https://ci.appveyor.com/api/projects/status/fbgfu0so3afcno78/branch/master?svg=true)](https://ci.appveyor.com/project/Facebook/rocksdb/branch/master)
-
+[![Linux/Mac Build Status](https://travis-ci.org/facebook/rocksdb.svg?branch=master)](https://travis-ci.org/facebook/rocksdb)
+[![Windows Build status](https://ci.appveyor.com/api/projects/status/fbgfu0so3afcno78/branch/master?svg=true)](https://ci.appveyor.com/project/Facebook/rocksdb/branch/master)
+[![PPC64le Build Status](http://140.211.168.68:8080/buildStatus/icon?job=Rocksdb)](http://140.211.168.68:8080/job/Rocksdb)
 
 RocksDB is developed and maintained by Facebook Database Engineering Team.
 It is built on earlier work on LevelDB by Sanjay Ghemawat (sanjay@google.com)
@@ -25,3 +25,7 @@ rely on the details of any other header files in this package.  Those
 internal APIs may be changed without warning.
 
 Design discussions are conducted in https://www.facebook.com/groups/rocksdb.dev/
+
+## License
+
+RocksDB is dual-licensed under both the GPLv2 (found in the COPYING file in the root directory) and Apache 2.0 License (found in the LICENSE.Apache file in the root directory).  You may select, at your option, one of the above-listed licenses.

ROCKSDB_LITE.md

@@ -5,7 +5,7 @@ RocksDBLite is a project focused on mobile use cases, which don't need a lot of
 Some examples of the features disabled by ROCKSDB_LITE:
 * compiled-in support for LDB tool
 * No backupable DB
-* No support for replication (which we provide in form of TrasactionalIterator)
+* No support for replication (which we provide in form of TransactionalIterator)
 * No advanced monitoring tools
 * No special-purpose memtables that are highly optimized for specific use cases
 * No Transactions

USERS.md

@@ -5,12 +5,15 @@ At Facebook, we use RocksDB as storage engines in multiple data management servi
 
 1. MyRocks -- https://github.com/MySQLOnRocksDB/mysql-5.6
 2. MongoRocks -- https://github.com/mongodb-partners/mongo-rocks
-3. ZippyDB --  Facebook's distributed key-value store with Paxos-style replication, built on top of RocksDB.[*] https://www.youtube.com/watch?v=DfiN7pG0D0khtt
-4. Laser -- Laser is a high query throughput, low (millisecond) latency, key-value storage service built on top of RocksDB.[*]
+3. ZippyDB --  Facebook's distributed key-value store with Paxos-style replication, built on top of RocksDB.[1] https://www.youtube.com/watch?v=DfiN7pG0D0khtt
+4. Laser -- Laser is a high query throughput, low (millisecond) latency, key-value storage service built on top of RocksDB.[1]
 4. Dragon -- a distributed graph query engine. https://code.facebook.com/posts/1737605303120405/dragon-a-distributed-graph-query-engine/
-5. Stylus -- a low-level stream processing framework writtenin C++.[*]
+5. Stylus -- a low-level stream processing framework writtenin C++.[1]
+6. LogDevice -- a distributed data store for logs [2]
 
-[*] https://research.facebook.com/publications/realtime-data-processing-at-facebook/
+[1] https://research.facebook.com/publications/realtime-data-processing-at-facebook/
+
+[2] https://code.facebook.com/posts/357056558062811/logdevice-a-distributed-data-store-for-logs/
 
 ## LinkedIn
 Two different use cases at Linkedin are using RocksDB as a storage engine:
@@ -24,7 +27,7 @@ Learn more about those use cases in a Tech Talk by Ankit Gupta and Naveen Somasu
 Yahoo is using RocksDB as a storage engine for their biggest distributed data store Sherpa. Learn more about it here: http://yahooeng.tumblr.com/post/120730204806/sherpa-scales-new-heights
 
 ## CockroachDB
-CockroachDB is an open-source geo-replicated transactional database (still in development). They are using RocksDB as their storage engine. Check out their github: https://github.com/cockroachdb/cockroach
+CockroachDB is an open-source geo-replicated transactional database. They are using RocksDB as their storage engine. Check out their github: https://github.com/cockroachdb/cockroach
 
 ## DNANexus
 DNANexus is using RocksDB to speed up processing of genomics data.
@@ -80,3 +83,12 @@ quasardb uses a heavily tuned RocksDB as its persistence layer.
 
 ## 360 Pika
 [360](http://www.360.cn/) [Pika](https://github.com/Qihoo360/pika) is a nosql compatible with redis. With the huge amount of data stored, redis may suffer for a capacity bottleneck, and pika was born for solving it. It has widely been widely used in many company
+
+## LzLabs
+LzLabs is using RocksDB as a storage engine in their multi-database distributed framework to store application configuration and user data.
+
+## ProfaneDB
+[ProfaneDB](https://profanedb.gitlab.io/) is a database for Protocol Buffers, and uses RocksDB for storage. It is accessible via gRPC, and the schema is defined using directly `.proto` files.
+
+## IOTA Foundation
+ [IOTA Foundation](https://www.iota.org/) is using RocksDB in the [IOTA Reference Implementation (IRI)](https://github.com/iotaledger/iri) to store the local state of the Tangle. The Tangle is the first open-source distributed ledger powering the future of the Internet of Things.

Vagrantfile

@@ -14,6 +14,11 @@ Vagrant.configure("2") do |config|
     box.vm.box = "chef/centos-6.5"
   end
 
+  config.vm.define "centos7" do |box|
+    box.vm.box = "centos/7"
+    box.vm.provision "shell", path: "build_tools/setup_centos7.sh"
+  end
+
   config.vm.define "FreeBSD10" do |box|
     box.vm.guest = :freebsd
     box.vm.box = "robin/freebsd-10"

WINDOWS_PORT.md

@@ -45,7 +45,7 @@ We plan to use this port for our business purposes here at Bing and this provide
 * All posix specific headers were replaced to port/port.h which worked well
 * Replaced `dirent.h` for `port/dirent.h` (very few places) with the implementation of the relevant interfaces within `rocksdb::port` namespace
 * Replaced `sys/time.h` to `port/sys_time.h` (few places) implemented equivalents within `rocksdb::port`
-* `printf %z` specification is not supported on Windows. To imitate existing standards we came up with a string macro `ROCKSDB_PRIszt` which expands to `%z` on posix systems and to Iu on windows.
+* `printf %z` specification is not supported on Windows. To imitate existing standards we came up with a string macro `ROCKSDB_PRIszt` which expands to `zu` on posix systems and to `Iu` on windows.
 * in class member initialization were moved to a __ctors in some cases
 * `constexpr` is not supported. We had to replace `std::numeric_limits<>::max/min()` to its C macros for constants. Sometimes we had to make class members `static const` and place a definition within a .cc file.
 * `constexpr` for functions was replaced to a template specialization (1 place)

appveyor.yml

@@ -1,14 +1,14 @@
 version: 1.0.{build}
-image: Visual Studio 2015
+image: Visual Studio 2017
 before_build:
 - md %APPVEYOR_BUILD_FOLDER%\build
 - cd %APPVEYOR_BUILD_FOLDER%\build
-- cmake -G "Visual Studio 14 2015 Win64" -DOPTDBG=1 -DXPRESS=1 ..
+- cmake -G "Visual Studio 15 Win64" -DOPTDBG=1 -DWITH_XPRESS=1 -DPORTABLE=1 -DJNI=1 ..
 - cd ..
 build:
   project: build\rocksdb.sln
   parallel: true
-  verbosity: minimal
+  verbosity: normal
 test:
 test_script:
 - ps: build_tools\run_ci_db_test.ps1 -SuiteRun db_basic_test,db_test2,db_test,env_basic_test,env_test -Concurrency 8

arcanist_util/__phutil_library_init__.php

@@ -1,3 +0,0 @@
-<?php
-
-phutil_register_library('arcanist_util', __FILE__);

arcanist_util/__phutil_library_map__.php

@@ -1,71 +0,0 @@
-<?php
-
-/**
- * This file is automatically generated. Use 'arc liberate' to rebuild it.
- * @generated
- * @phutil-library-version 2
- */
-
-if (class_exists('ArcanistWorkflow')) {
-  phutil_register_library_map(array(
-    '__library_version__' => 2,
-    'class' =>
-    array(
-      'ArcanistCpplintLinter' => 'cpp_linter/ArcanistCpplintLinter.php',
-      'BaseDirectoryScopedFormatLinter' => 'cpp_linter/BaseDirectoryScopedFormatLinter.php',
-      'FacebookArcanistConfiguration' => 'config/FacebookArcanistConfiguration.php',
-      'FacebookFbcodeLintEngine' => 'lint_engine/FacebookFbcodeLintEngine.php',
-      'FacebookFbcodeUnitTestEngine' => 'unit_engine/FacebookFbcodeUnitTestEngine.php',
-      'FacebookHowtoevenLintEngine' => 'lint_engine/FacebookHowtoevenLintEngine.php',
-      'FacebookHowtoevenLinter' => 'cpp_linter/FacebookHowtoevenLinter.php',
-      'FbcodeClangFormatLinter' => 'cpp_linter/FbcodeClangFormatLinter.php',
-      'FbcodeCppLinter' => 'cpp_linter/FbcodeCppLinter.php',
-    ),
-    'function' =>
-    array(
-    ),
-    'xmap' =>
-    array(
-      'ArcanistCpplintLinter' => 'ArcanistLinter',
-      'BaseDirectoryScopedFormatLinter' => 'ArcanistLinter',
-      'FacebookArcanistConfiguration' => 'ArcanistConfiguration',
-      'FacebookFbcodeLintEngine' => 'ArcanistLintEngine',
-      'FacebookFbcodeUnitTestEngine' => 'ArcanistBaseUnitTestEngine',
-      'FacebookHowtoevenLintEngine' => 'ArcanistLintEngine',
-      'FacebookHowtoevenLinter' => 'ArcanistLinter',
-      'FbcodeClangFormatLinter' => 'BaseDirectoryScopedFormatLinter',
-      'FbcodeCppLinter' => 'ArcanistLinter',
-    ),
-  ));
-} else {
-  phutil_register_library_map(array(
-    '__library_version__' => 2,
-    'class' =>
-    array(
-      'ArcanistCpplintLinter' => 'cpp_linter/ArcanistCpplintLinter.php',
-      'BaseDirectoryScopedFormatLinter' => 'cpp_linter/BaseDirectoryScopedFormatLinter.php',
-      'FacebookArcanistConfiguration' => 'config/FacebookOldArcanistConfiguration.php',
-      'FacebookFbcodeLintEngine' => 'lint_engine/FacebookFbcodeLintEngine.php',
-      'FacebookFbcodeUnitTestEngine' => 'unit_engine/FacebookOldFbcodeUnitTestEngine.php',
-      'FacebookHowtoevenLintEngine' => 'lint_engine/FacebookHowtoevenLintEngine.php',
-      'FacebookHowtoevenLinter' => 'cpp_linter/FacebookHowtoevenLinter.php',
-      'FbcodeClangFormatLinter' => 'cpp_linter/FbcodeClangFormatLinter.php',
-      'FbcodeCppLinter' => 'cpp_linter/FbcodeCppLinter.php',
-    ),
-    'function' =>
-    array(
-    ),
-    'xmap' =>
-    array(
-      'ArcanistCpplintLinter' => 'ArcanistLinter',
-      'BaseDirectoryScopedFormatLinter' => 'ArcanistLinter',
-      'FacebookArcanistConfiguration' => 'ArcanistConfiguration',
-      'FacebookFbcodeLintEngine' => 'ArcanistLintEngine',
-      'FacebookFbcodeUnitTestEngine' => 'ArcanistBaseUnitTestEngine',
-      'FacebookHowtoevenLintEngine' => 'ArcanistLintEngine',
-      'FacebookHowtoevenLinter' => 'ArcanistLinter',
-      'FbcodeClangFormatLinter' => 'BaseDirectoryScopedFormatLinter',
-      'FbcodeCppLinter' => 'ArcanistLinter',
-    ),
-  ));
-}

arcanist_util/config/FacebookArcanistConfiguration.php

@@ -1,43 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-require('RocksDBCommonHelper.php');
-
-define("DIFF_COMMAND", "diff");
-
-class FacebookArcanistConfiguration extends ArcanistConfiguration {
-  public function getCustomArgumentsForCommand($command) {
-    if ($command == "land") {
-      return array(
-          'async' => array('help' => 'Just to make tools happy'));
-    }
-    return array();
-  }
-
-  public function didRunWorkflow($command,
-                                 ArcanistWorkflow $workflow,
-                                 $error_code) {
-    // Default options don't terminate on failure, but that's what we want. In
-    // the current case we use assertions intentionally as "terminate on failure
-    // invariants".
-    assert_options(ASSERT_BAIL, true);
-
-    assert($workflow);
-    assert(strlen($command) > 0);
-
-    if ($command == DIFF_COMMAND && !$workflow->isRawDiffSource()) {
-      $diffID = $workflow->getDiffId();
-
-      // When submitting a diff this code path gets executed multiple times in
-      // a row. We only care about the case when ID for the diff is provided
-      // because that's what we need to apply the diff and trigger the tests.
-      if (strlen($diffID) > 0) {
-        assert(is_numeric($diffID));
-        startTestsInSandcastle(true /* $applyDiff */, $workflow, $diffID);
-      }
-    }
-  }
-}

arcanist_util/config/FacebookOldArcanistConfiguration.php

@@ -1,43 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-require('RocksDBCommonHelper.php');
-
-define("DIFF_COMMAND", "diff");
-
-class FacebookArcanistConfiguration extends ArcanistConfiguration {
-  public function getCustomArgumentsForCommand($command) {
-    if ($command == "land") {
-      return array(
-          'async' => array('help' => 'Just to make tools happy'));
-    }
-    return array();
-  }
-
-  public function didRunWorkflow($command,
-                                 ArcanistBaseWorkflow $workflow,
-                                 $error_code) {
-    // Default options don't terminate on failure, but that's what we want. In
-    // the current case we use assertions intentionally as "terminate on failure
-    // invariants".
-    assert_options(ASSERT_BAIL, true);
-
-    assert($workflow);
-    assert(strlen($command) > 0);
-
-    if ($command == DIFF_COMMAND && !$workflow->isRawDiffSource()) {
-      $diffID = $workflow->getDiffId();
-
-      // When submitting a diff this code path gets executed multiple times in
-      // a row. We only care about the case when ID for the diff is provided
-      // because that's what we need to apply the diff and trigger the tests.
-      if (strlen($diffID) > 0) {
-        assert(is_numeric($diffID));
-        startTestsInSandcastle(true /* $applyDiff */, $workflow, $diffID);
-      }
-    }
-  }
-}

arcanist_util/cpp_linter/ArcanistCpplintLinter.php

@@ -1,88 +0,0 @@
-<?php
-
-/**
- * Uses google's cpplint.py to check code. RocksDB team forked this file from
- * phabricator's /src/lint/linter/ArcanistCpplintLinter.php, and customized it
- * for its own use.
- *
- * You can get it here:
- * http://google-styleguide.googlecode.com/svn/trunk/cpplint/cpplint.py
- * @group linter
- */
-final class ArcanistCpplintLinter extends ArcanistLinter {
-
-  public function willLintPaths(array $paths) {
-    return;
-  }
-
-  public function getLinterName() {
-    return 'cpplint.py';
-  }
-
-  public function getLintPath() {
-    $bin = 'cpplint.py';
-    // Search under current dir
-    list($err) = exec_manual('which %s/%s', $this->linterDir(), $bin);
-    if (!$err) {
-      return $this->linterDir().'/'.$bin;
-    }
-
-    // Look for globally installed cpplint.py
-    list($err) = exec_manual('which %s', $bin);
-    if ($err) {
-      throw new ArcanistUsageException(
-        "cpplint.py does not appear to be installed on this system. Install ".
-        "it (e.g., with 'wget \"http://google-styleguide.googlecode.com/".
-        "svn/trunk/cpplint/cpplint.py\"') ".
-        "in your .arcconfig to point to the directory where it resides. ".
-        "Also don't forget to chmod a+x cpplint.py!");
-    }
-
-    return $bin;
-  }
-
-  public function lintPath($path) {
-    $bin = $this->getLintPath();
-    $path = $this->rocksdbDir().'/'.$path;
-
-    $f = new ExecFuture("%C $path", $bin);
-
-    list($err, $stdout, $stderr) = $f->resolve();
-
-    if ($err === 2) {
-      throw new Exception("cpplint failed to run correctly:\n".$stderr);
-    }
-
-    $lines = explode("\n", $stderr);
-    $messages = array();
-    foreach ($lines as $line) {
-      $line = trim($line);
-      $matches = null;
-      $regex = '/^[^:]+:(\d+):\s*(.*)\s*\[(.*)\] \[(\d+)\]$/';
-      if (!preg_match($regex, $line, $matches)) {
-        continue;
-      }
-      foreach ($matches as $key => $match) {
-        $matches[$key] = trim($match);
-      }
-      $message = new ArcanistLintMessage();
-      $message->setPath($path);
-      $message->setLine($matches[1]);
-      $message->setCode($matches[3]);
-      $message->setName($matches[3]);
-      $message->setDescription($matches[2]);
-      $message->setSeverity(ArcanistLintSeverity::SEVERITY_WARNING);
-      $this->addLintMessage($message);
-    }
-  }
-
-  // The path of this linter
-  private function linterDir() {
-    return dirname(__FILE__);
-  }
-
-  // TODO(kaili) a quick and dirty way to figure out rocksdb's root dir.
-  private function rocksdbDir() {
-    return $this->linterDir()."/../..";
-  }
-}

arcanist_util/cpp_linter/BaseDirectoryScopedFormatLinter.php

@@ -1,74 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-abstract class BaseDirectoryScopedFormatLinter extends ArcanistLinter {
-
-  const LINT_FORMATTING = 1;
-
-  private $changedLines = array();
-  private $rawLintOutput = array();
-
-  abstract protected function getPathsToLint();
-
-  protected function shouldLintPath($path) {
-    foreach ($this->getPathsToLint() as $p) {
-      // check if $path starts with $p
-      if (strncmp($path, $p, strlen($p)) === 0) {
-        return true;
-      }
-    }
-    return false;
-  }
-
-  // API to tell this linter which lines were changed
-  final public function setPathChangedLines($path, $changed) {
-    $this->changedLines[$path] = $changed;
-  }
-
-  final public function willLintPaths(array $paths) {
-    $futures = array();
-    foreach ($paths as $path) {
-      if (!$this->shouldLintPath($path)) {
-        continue;
-      }
-
-      $changed = $this->changedLines[$path];
-      if (!isset($changed)) {
-        // do not run linter if there are no changes
-        continue;
-      }
-
-      $futures[$path] = $this->getFormatFuture($path, $changed);
-    }
-
-    foreach (id(new FutureIterator($futures))->limit(8) as $p => $f) {
-      $this->rawLintOutput[$p] = $f->resolvex();
-    }
-  }
-
-  abstract protected function getFormatFuture($path, array $changed);
-  abstract protected function getLintMessage($diff);
-
-  final public function lintPath($path) {
-    if (!isset($this->rawLintOutput[$path])) {
-      return;
-    }
-
-    list($new_content) = $this->rawLintOutput[$path];
-    $old_content = $this->getData($path);
-
-    if ($new_content != $old_content) {
-      $diff = ArcanistDiffUtils::renderDifferences($old_content, $new_content);
-      $this->raiseLintAtOffset(
-        0,
-        self::LINT_FORMATTING,
-        $this->getLintMessage($diff),
-        $old_content,
-        $new_content);
-    }
-  }
-
-}

arcanist_util/cpp_linter/FacebookHowtoevenLinter.php

@@ -1,223 +0,0 @@
-<?php
-// Copyright 2015-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-final class FacebookHowtoevenLinter extends ArcanistLinter {
-
-  const VERSION = 'fd9192f324c36d28136d14380f0b552a1385b59b';
-
-  private $parsedTargets = array();
-
-  public function getLinterName() {
-    return 'Howtoeven';
-  }
-
-  protected function getSeverity($code) {
-    $severities = array(
-      ArcanistLintSeverity::SEVERITY_DISABLED,
-      ArcanistLintSeverity::SEVERITY_ADVICE,
-      ArcanistLintSeverity::SEVERITY_WARNING,
-      ArcanistLintSeverity::SEVERITY_ERROR,
-    );
-    return idx($severities, $code, ArcanistLintSeverity::SEVERITY_WARNING);
-  }
-
-  public function willLintPaths(array $paths) {
-    // Cleanup previous runs.
-    $this->localExecx("rm -rf _build/_lint");
-
-    // Build compilation database.
-    $lintable_paths = $this->getLintablePaths($paths);
-    $interesting_paths = $this->getInterestingPaths($lintable_paths);
-
-    if (!$lintable_paths) {
-      return;
-    }
-
-    // Run lint.
-    try {
-      $this->localExecx(
-        "%C %C -p _build/dev/ %Ls",
-        $this->getBinaryPath(),
-        $this->getFilteredIssues(),
-        $lintable_paths);
-    } catch (CommandException $exception) {
-      PhutilConsole::getConsole()->writeErr($exception->getMessage());
-    }
-
-    // Load results.
-    $result = id(
-      new SQLite3(
-        $this->getProjectRoot().'/_build/_lint/lint.db',
-        SQLITE3_OPEN_READONLY))
-      ->query("SELECT * FROM raised_issues");
-
-    while ($issue = $result->fetchArray(SQLITE3_ASSOC)) {
-      // Skip issues not part of the linted file.
-      if (in_array($issue['file'], $interesting_paths)) {
-        $this->addLintMessage(id(new ArcanistLintMessage())
-          ->setPath($issue['file'])
-          ->setLine($issue['line'])
-          ->setChar($issue['column'])
-          ->setCode('Howtoeven')
-          ->setSeverity($this->getSeverity($issue['severity']))
-          ->setName('Hte-'.$issue['name'])
-          ->setDescription(
-            sprintf(
-              "%s\n\n%s",
-              ($issue['message']) ? $issue['message'] : $issue['description'],
-              $issue['explanation']))
-          ->setOriginalText(idx($issue, 'original', ''))
-          ->setReplacementText(idx($issue, 'replacement', '')));
-      }
-    }
-  }
-
-  public function lintPath($path) {
-  }
-
-  /**
-   * Get the paths that we know how to lint.
-   *
-   * The strategy is to first look whether there's an existing compilation
-   * database and use that if it's exhaustive. We generate our own only if
-   * necessary.
-   */
-  private function getLintablePaths($paths) {
-    // Replace headers with existing sources.
-    for ($i = 0; $i < count($paths); $i++) {
-      if (preg_match("/\.h$/", $paths[$i])) {
-        $header = preg_replace("/\.h$/", ".cpp", $paths[$i]);
-        if (file_exists($header)) {
-          $paths[$i] = $header;
-        }
-      }
-    }
-
-    // Check if database exists and is exhaustive.
-    $available_paths = $this->getAvailablePaths();
-    $lintable_paths = array_intersect($paths, $available_paths);
-    if ($paths === $lintable_paths) {
-      return $lintable_paths;
-    }
-
-    // Generate our own database.
-    $targets = $this->getTargetsFor($paths);
-    if (!$targets) {
-      PhutilConsole::getConsole()->writeErr(
-        "No build targets found for %s\n",
-        implode(', ', $paths));
-      return array();
-    }
-
-    $this->localExecx("./tools/build/bin/fbconfig.par -r %Ls", $targets);
-    $this->localExecx("./tools/build/bin/fbmake.par gen_cdb");
-
-    $available_paths = $this->getAvailablePaths();
-    $lintable_paths = array_intersect($paths, $available_paths);
-    if ($paths != $lintable_paths) {
-      PhutilConsole::getConsole()->writeErr(
-        "Can't lint %s\n",
-        implode(', ', array_diff($paths, $available_paths)));
-    }
-
-    // Return what we know how to lint.
-    return $lintable_paths;
-  }
-
-  /**
-   * Get the available paths in the current compilation database.
-   */
-  private function getAvailablePaths() {
-    $database_path = $this->getProjectRoot()
-      .'/_build/dev/compile_commands.json';
-    if (!file_exists($database_path)) {
-      return array();
-    }
-
-    $entries = json_decode(file_get_contents($database_path), true);
-    $paths = array();
-    foreach ($entries as $entry) {
-      $paths[] = $entry['file'];
-    }
-    return $paths;
-  }
-
-  /**
-   * Search for the targets directories for the given files.
-   */
-  private static function getTargetsFor($paths) {
-    $targets = array();
-    foreach ($paths as $path) {
-      while (($path = dirname($path)) !== '.') {
-        if (in_array('TARGETS', scandir($path))) {
-          $contents = file_get_contents($path.'/TARGETS');
-          if (strpos($contents, 'cpp_binary') !== false) {
-            $targets[] = $path;
-            break;
-          }
-        }
-      }
-    }
-    return array_unique($targets);
-  }
-
-  /**
-   * The paths that we actually want to report on.
-   */
-  private function getInterestingPaths($paths) {
-    $headers = array();
-    foreach ($paths as $path) {
-      $headers[] = preg_replace("/\.cpp$/", ".h", $path);
-    }
-    return array_merge($paths, $headers);
-  }
-
-  /**
-   * The path where the binary is located. Will return the current dewey binary
-   * unless the `HOWTOEVEN_BUILD` environment variable is set.
-   */
-  private function getBinaryPath() {
-    $path = sprintf(
-      "/mnt/dewey/fbcode/.commits/%s/builds/howtoeven/client",
-      self::VERSION);
-
-    $build = getenv('HOWTOEVEN_BUILD');
-    if ($build) {
-      $path = sprintf(
-        "./_build/%s/tools/howtoeven/client",
-        $build);
-      if (!file_exists($path)) {
-        PhutilConsole::getConsole()->writeErr(">> %s does not exist\n", $path);
-        exit(1);
-      }
-    }
-
-    return $path;
-  }
-
-  /**
-   * Execute the command in the root directory.
-   */
-  private function localExecx($command /* , ... */) {
-    $arguments = func_get_args();
-    return newv('ExecFuture', $arguments)
-      ->setCWD($this->getProjectRoot())
-      ->resolvex();
-  }
-
-  /**
-   * The root of the project.
-   */
-  private function getProjectRoot() {
-    return $this->getEngine()->getWorkingCopy()->getProjectRoot();
-  }
-
-  private function getFilteredIssues() {
-    $issues = getenv('HOWTOEVEN_ISSUES');
-    return ($issues) ? csprintf('-issues %s', $issues) : '';
-  }
-
-}

arcanist_util/cpp_linter/FbcodeClangFormatLinter.php

@@ -1,58 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-final class FbcodeClangFormatLinter extends BaseDirectoryScopedFormatLinter {
-
-  const LINT_FORMATTING = 1;
-  const CLANG_FORMAT_BINARY = '/mnt/vol/engshare/admin/scripts/clang-format';
-
-  protected function getPathsToLint() {
-    return array('');
-  }
-
-  public function getLinterName() {
-    return 'CLANG_FORMAT';
-  }
-
-  public function getLintSeverityMap() {
-    return array(
-      self::LINT_FORMATTING => ArcanistLintSeverity::SEVERITY_ADVICE,
-    );
-  }
-
-  public function getLintNameMap() {
-    return array(
-      self::LINT_FORMATTING => pht('Changes are not clang-formatted'),
-    );
-  }
-
-  protected function getFormatFuture($path, array $changed) {
-    $args = "";
-    foreach ($changed as $key => $value) {
-      $args .= " --lines=$key:$key";
-    }
-
-    $binary = self::CLANG_FORMAT_BINARY;
-    if (!file_exists($binary)) {
-      // trust the $PATH
-      $binary = "clang-format";
-    }
-
-    return new ExecFuture(
-      "%s %s $args",
-      $binary,
-      $this->getEngine()->getFilePathOnDisk($path));
-  }
-
-  protected function getLintMessage($diff) {
-    $link_to_clang_format =
-      "[[ http://fburl.com/clang-format | clang-format ]]";
-    return <<<LINT_MSG
-Changes in this file were not formatted using $link_to_clang_format.
-Please run build_tools/format-diff.sh or `make format`
-LINT_MSG;
-  }
-}

arcanist_util/cpp_linter/FbcodeCppLinter.php

@@ -1,126 +0,0 @@
-<?php
-// Copyright 2004-present Facebook.  All rights reserved.
-
-class FbcodeCppLinter extends ArcanistLinter {
-  const FLINT      = "/home/engshare/tools/flint";
-  const LINT_ERROR   = 1;
-  const LINT_WARNING = 2;
-  const LINT_ADVICE  = 3;
-  const C_FLAG = "--c_mode=true";
-
-  private $rawLintOutput = array();
-
-  public function willLintPaths(array $paths) {
-    if (!file_exists(self::FLINT)) {
-      return;
-    }
-    $futures = array();
-    foreach ($paths as $p) {
-      $lpath = $this->getEngine()->getFilePathOnDisk($p);
-      $lpath_file = file($lpath);
-      if (preg_match('/\.(c)$/', $lpath) ||
-          preg_match('/-\*-.*Mode: C[; ].*-\*-/', $lpath_file[0]) ||
-          preg_match('/vim(:.*)*:\s*(set\s+)?filetype=c\s*:/', $lpath_file[0])
-          ) {
-        $futures[$p] = new ExecFuture("%s %s %s 2>&1",
-                           self::FLINT, self::C_FLAG,
-                           $this->getEngine()->getFilePathOnDisk($p));
-      } else {
-        $futures[$p] = new ExecFuture("%s %s 2>&1",
-          self::FLINT, $this->getEngine()->getFilePathOnDisk($p));
-      }
-    }
-
-    foreach (Futures($futures)->limit(8) as $p => $f) {
-      $this->rawLintOutput[$p] = $f->resolvex();
-    }
-
-    return;
-  }
-
-  public function getLinterName() {
-    return "FBCPP";
-  }
-
-  public function lintPath($path) {
-    $this->runCppLint($path);
-  }
-
-  private function runCppLint($path) {
-    $msgs = $this->getCppLintOutput($path);
-    foreach ($msgs as $m) {
-      $this->raiseLintAtLine($m['line'], 0, $m['severity'], $m['msg']);
-    }
-  }
-
-  private function adviseOnEachPattern(
-    $path,
-    $regex,
-    $message,
-    $lint_type = self::LINT_ADVICE,
-    $match_idx = 0) {
-      $file_data = $this->getData($path);
-      $matches = array();
-      if (!preg_match_all($regex, $file_data, $matches, PREG_OFFSET_CAPTURE)) {
-        return;
-      }
-
-      foreach ($matches[$match_idx] as $match) {
-        list($match_str, $offset) = $match;
-        $this->raiseLintAtOffset($offset, $lint_type, $message, $match_str);
-      }
-  }
-
-  public function getLintSeverityMap() {
-    return array(
-      self::LINT_WARNING => ArcanistLintSeverity::SEVERITY_WARNING,
-      self::LINT_ADVICE  => ArcanistLintSeverity::SEVERITY_ADVICE,
-      self::LINT_ERROR   => ArcanistLintSeverity::SEVERITY_ERROR
-    );
-  }
-
-  public function getLintNameMap() {
-    return array(
-      self::LINT_ADVICE   => "CppLint Advice",
-      self::LINT_WARNING  => "CppLint Warning",
-      self::LINT_ERROR    => "CppLint Error"
-    );
-  }
-
-  private function getCppLintOutput($path) {
-     if (!array_key_exists($path, $this->rawLintOutput)) {
-       return array();
-     }
-    list($output) = $this->rawLintOutput[$path];
-
-    $msgs = array();
-    $current = null;
-    $matches = array();
-    foreach (explode("\n", $output) as $line) {
-      if (preg_match('/.*?:(\d+):(.*)/', $line, $matches)) {
-        if ($current) {
-          $msgs[] = $current;
-        }
-        $line = $matches[1];
-        $text = $matches[2];
-        if (preg_match('/.*Warning.*/', $text)) {
-          $sev = self::LINT_WARNING;
-        } else if (preg_match('/.*Advice.*/', $text)) {
-          $sev = self::LINT_ADVICE;
-        } else {
-          $sev = self::LINT_ERROR;
-        }
-        $current = array('line'     => $line,
-                         'msg'      => $text,
-                         'severity' => $sev);
-      } else if ($current) {
-        $current['msg'] .= ' ' . $line;
-      }
-    }
-    if ($current) {
-      $msgs[] = $current;
-    }
-
-    return $msgs;
-  }
-}

arcanist_util/lint_engine/FacebookFbcodeLintEngine.php

@@ -1,138 +0,0 @@
-<?php
-// Copyright 2004-present Facebook.  All rights reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-class FacebookFbcodeLintEngine extends ArcanistLintEngine {
-
-  public function buildLinters() {
-    $linters = array();
-    $paths = $this->getPaths();
-
-    // Remove all deleted files, which are not checked by the
-    // following linters.
-    foreach ($paths as $key => $path) {
-      if (!Filesystem::pathExists($this->getFilePathOnDisk($path))) {
-        unset($paths[$key]);
-      }
-    }
-
-    $generated_linter = new ArcanistGeneratedLinter();
-    $linters[] = $generated_linter;
-
-    $nolint_linter = new ArcanistNoLintLinter();
-    $linters[] = $nolint_linter;
-
-    $text_linter = new ArcanistTextLinter();
-    $text_linter->setCustomSeverityMap(array(
-      ArcanistTextLinter::LINT_LINE_WRAP
-        => ArcanistLintSeverity::SEVERITY_ADVICE,
-    ));
-    $linters[] = $text_linter;
-
-    $java_text_linter = new ArcanistTextLinter();
-    $java_text_linter->setMaxLineLength(100);
-    $java_text_linter->setCustomSeverityMap(array(
-      ArcanistTextLinter::LINT_LINE_WRAP
-        => ArcanistLintSeverity::SEVERITY_ADVICE,
-    ));
-    $linters[] = $java_text_linter;
-
-    $python_linter = new ArcanistPEP8Linter();
-    $linters[] = $python_linter;
-
-    $cpp_linters = array();
-    $cpp_linters[] = $linters[] = new ArcanistCpplintLinter();
-    $cpp_linters[] = $linters[] = new FbcodeCppLinter();
-
-    $clang_format_linter = new FbcodeClangFormatLinter();
-    $linters[] = $clang_format_linter;
-
-    $spelling_linter = new ArcanistSpellingLinter();
-    $linters[] = $spelling_linter;
-
-    foreach ($paths as $path) {
-      $is_text = false;
-
-      $text_extensions = (
-        '/\.('.
-        'cpp|cxx|c|cc|h|hpp|hxx|tcc|'.
-        'py|rb|hs|pl|pm|tw|'.
-        'php|phpt|css|js|'.
-        'java|'.
-        'thrift|'.
-        'lua|'.
-        'siv|'.
-        'txt'.
-        ')$/'
-      );
-      if (preg_match($text_extensions, $path)) {
-        $is_text = true;
-      }
-      if ($is_text) {
-        $nolint_linter->addPath($path);
-
-        $generated_linter->addPath($path);
-        $generated_linter->addData($path, $this->loadData($path));
-
-        if (preg_match('/\.java$/', $path)) {
-          $java_text_linter->addPath($path);
-          $java_text_linter->addData($path, $this->loadData($path));
-        } else {
-          $text_linter->addPath($path);
-          $text_linter->addData($path, $this->loadData($path));
-        }
-
-        $spelling_linter->addPath($path);
-        $spelling_linter->addData($path, $this->loadData($path));
-      }
-      if (preg_match('/\.(cpp|c|cc|cxx|h|hh|hpp|hxx|tcc)$/', $path)
-          && !preg_match('/third-party/', $path)) {
-        foreach ($cpp_linters as &$linter) {
-          $linter->addPath($path);
-          $linter->addData($path, $this->loadData($path));
-        }
-
-        $clang_format_linter->addPath($path);
-        $clang_format_linter->addData($path, $this->loadData($path));
-        $clang_format_linter->setPathChangedLines(
-          $path, $this->getPathChangedLines($path));
-      }
-
-      // Match *.py and contbuild config files
-      if (preg_match('/(\.(py|tw|smcprops)|^contbuild\/configs\/[^\/]*)$/',
-                    $path)) {
-        $space_count = 4;
-        $real_path = $this->getFilePathOnDisk($path);
-        $dir = dirname($real_path);
-        do {
-          if (file_exists($dir.'/.python2space')) {
-            $space_count = 2;
-            break;
-          }
-          $dir = dirname($dir);
-        } while ($dir != '/' && $dir != '.');
-
-        $cur_path_linter = $python_linter;
-        $cur_path_linter->addPath($path);
-        $cur_path_linter->addData($path, $this->loadData($path));
-
-        if (preg_match('/\.tw$/', $path)) {
-          $cur_path_linter->setCustomSeverityMap(array(
-            'E251' => ArcanistLintSeverity::SEVERITY_DISABLED,
-          ));
-        }
-      }
-    }
-
-    $name_linter = new ArcanistFilenameLinter();
-    $linters[] = $name_linter;
-    foreach ($paths as $path) {
-      $name_linter->addPath($path);
-    }
-
-    return $linters;
-  }
-
-}

arcanist_util/lint_engine/FacebookHowtoevenLintEngine.php

@@ -1,27 +0,0 @@
-<?php
-// Copyright 2015-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-final class FacebookHowtoevenLintEngine extends ArcanistLintEngine {
-
-  public function buildLinters() {
-    $paths = array();
-
-    foreach ($this->getPaths() as $path) {
-      // Don't try to lint deleted files or changed directories.
-      if (!Filesystem::pathExists($path) || is_dir($path)) {
-        continue;
-      }
-
-      if (preg_match('/\.(cpp|c|cc|cxx|h|hh|hpp|hxx|tcc)$/', $path)) {
-        $paths[] = $path;
-      }
-    }
-
-    $howtoeven = new FacebookHowtoevenLinter();
-    $howtoeven->setPaths($paths);
-    return array($howtoeven);
-  }
-}

arcanist_util/unit_engine/FacebookFbcodeUnitTestEngine.php

@@ -1,17 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-class FacebookFbcodeUnitTestEngine extends ArcanistUnitTestEngine {
-
-  public function run() {
-      // For a call to `arc call-conduit differential.updateunitresults` to
-      // succeed we need at least one entry here.
-      $result = new ArcanistUnitTestResult();
-      $result->setName("dummy_placeholder_entry");
-      $result->setResult(ArcanistUnitTestResult::RESULT_PASS);
-      return array($result);
-  }
-}

arcanist_util/unit_engine/FacebookOldFbcodeUnitTestEngine.php

@@ -1,17 +0,0 @@
-<?php
-// Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
-
-class FacebookFbcodeUnitTestEngine extends ArcanistBaseUnitTestEngine {
-
-  public function run() {
-      // For a call to `arc call-conduit differential.updateunitresults` to
-      // succeed we need at least one entry here.
-      $result = new ArcanistUnitTestResult();
-      $result->setName("dummy_placeholder_entry");
-      $result->setResult(ArcanistUnitTestResult::RESULT_PASS);
-      return array($result);
-  }
-}

buckifier/buckify_rocksdb.py

@@ -3,14 +3,11 @@ from __future__ import division
 from __future__ import print_function
 from __future__ import unicode_literals
 from targets_builder import TARGETSBuilder
-from optparse import OptionParser
 import os
 import fnmatch
 import sys
-import tempfile
 
 from util import ColorString
-import util
 
 # tests to export as libraries for inclusion in other projects
 _EXPORTED_TEST_LIBS = ["env_basic_test"]
@@ -36,7 +33,7 @@ def parse_src_mk(repo_path):
 # get all .cc / .c files
 def get_cc_files(repo_path):
     cc_files = []
-    for root, dirnames, filenames in os.walk(repo_path):
+    for root, dirnames, filenames in os.walk(repo_path):  # noqa: B007 T25377293 Grandfathered in
         root = root[(len(repo_path) + 1):]
         if "java" in root:
             # Skip java
@@ -112,12 +109,14 @@ def generate_targets(repo_path):
         "rocksdb_test_lib",
         src_mk.get("MOCK_LIB_SOURCES", []) +
         src_mk.get("TEST_LIB_SOURCES", []) +
-        src_mk.get("EXP_LIB_SOURCES", []),
+        src_mk.get("EXP_LIB_SOURCES", []) +
+        src_mk.get("ANALYZER_LIB_SOURCES", []),
         [":rocksdb_lib"])
     # rocksdb_tools_lib
     TARGETS.add_library(
         "rocksdb_tools_lib",
         src_mk.get("BENCH_LIB_SOURCES", []) +
+        src_mk.get("ANALYZER_LIB_SOURCES", []) +
         ["util/testutil.cc"],
         [":rocksdb_lib"])
 

buckifier/rocks_test_runner.sh

@@ -1,4 +1,5 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # Create a tmp directory for the test to use
 TEST_DIR=$(mktemp -d /dev/shm/fbcode_rocksdb_XXXXXXX)
+# shellcheck disable=SC2068
 TEST_TMPDIR="$TEST_DIR" $@ && rm -rf "$TEST_DIR"

buckifier/targets_builder.py

@@ -3,10 +3,8 @@ from __future__ import division
 from __future__ import print_function
 from __future__ import unicode_literals
 import targets_cfg
-import pprint
 
-# TODO(tec): replace this with PrettyPrinter
-def pretty_list(lst, indent=6):
+def pretty_list(lst, indent=8):
     if lst is None or len(lst) == 0:
         return ""
 
@@ -14,8 +12,8 @@ def pretty_list(lst, indent=6):
         return "\"%s\"" % lst[0]
 
     separator = "\",\n%s\"" % (" " * indent)
-    res = separator.join(lst)
-    res = "\n" + (" " * indent) + "\"" + res + "\",\n" + (" " * (indent - 2))
+    res = separator.join(sorted(lst))
+    res = "\n" + (" " * indent) + "\"" + res + "\",\n" + (" " * (indent - 4))
     return res
 
 
@@ -27,19 +25,22 @@ class TARGETSBuilder:
         self.total_lib = 0
         self.total_bin = 0
         self.total_test = 0
-        self.tests_cfg = []
+        self.tests_cfg = ""
 
     def __del__(self):
         self.targets_file.close()
 
     def add_library(self, name, srcs, deps=None, headers=None):
+        headers_attr_prefix = ""
         if headers is None:
+            headers_attr_prefix = "auto_"
             headers = "AutoHeaders.RECURSIVE_GLOB"
-        self.targets_file.write(targets_cfg.library_template % (
-            name,
-            headers,
-            pretty_list(srcs),
-            pretty_list(deps)))
+        self.targets_file.write(targets_cfg.library_template.format(
+            name=name,
+            srcs=pretty_list(srcs),
+            headers_attr_prefix=headers_attr_prefix,
+            headers=headers,
+            deps=pretty_list(deps)))
         self.total_lib = self.total_lib + 1
 
     def add_binary(self, name, srcs, deps=None):
@@ -53,13 +54,13 @@ class TARGETSBuilder:
         exec_mode = "serial"
         if is_parallel:
             exec_mode = "parallel"
-        self.tests_cfg.append([test_name, str(src), str(exec_mode)])
+        self.tests_cfg += targets_cfg.test_cfg_template % (
+            test_name,
+            str(src),
+            str(exec_mode))
 
         self.total_test = self.total_test + 1
 
     def flush_tests(self):
-        self.targets_file.write(targets_cfg.unittests_template % (
-            pprint.PrettyPrinter().pformat(self.tests_cfg)
-        ))
-
-        self.tests_cfg = []
+        self.targets_file.write(targets_cfg.unittests_template % self.tests_cfg)
+        self.tests_cfg = ""

buckifier/targets_cfg.py

@@ -2,119 +2,132 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 from __future__ import unicode_literals
-rocksdb_target_header = """
-import os
+rocksdb_target_header = """load("@fbcode_macros//build_defs:auto_headers.bzl", "AutoHeaders")
+load(":defs.bzl", "test_binary")
 
-TARGETS_PATH = os.path.dirname(__file__)
-REPO_PATH = TARGETS_PATH[(TARGETS_PATH.find('fbcode/') + len('fbcode/')):] + "/"
-BUCK_BINS = "buck-out/gen/" + REPO_PATH
-TEST_RUNNER = REPO_PATH + "buckifier/rocks_test_runner.sh"
-rocksdb_compiler_flags = [
-  "-msse",
-  "-msse4.2",
-  "-fno-builtin-memcmp",
-  "-DROCKSDB_PLATFORM_POSIX",
-  "-DROCKSDB_LIB_IO_POSIX",
-  "-DROCKSDB_FALLOCATE_PRESENT",
-  "-DROCKSDB_MALLOC_USABLE_SIZE",
-  "-DROCKSDB_RANGESYNC_PRESENT",
-  "-DROCKSDB_SCHED_GETCPU_PRESENT",
-  "-DROCKSDB_SUPPORT_THREAD_LOCAL",
-  "-DHAVE_SSE42",
-  "-DOS_LINUX",
-  # Flags to enable libs we include
-  "-DSNAPPY",
-  "-DZLIB",
-  "-DBZIP2",
-  "-DLZ4",
-  "-DZSTD",
-  "-DGFLAGS=gflags",
-  "-DNUMA",
-  "-DTBB",
-  # Needed to compile in fbcode
-  "-Wno-expansion-to-defined",
+REPO_PATH = package_name() + "/"
+
+ROCKSDB_COMPILER_FLAGS = [
+    "-fno-builtin-memcmp",
+    "-DROCKSDB_PLATFORM_POSIX",
+    "-DROCKSDB_LIB_IO_POSIX",
+    "-DROCKSDB_FALLOCATE_PRESENT",
+    "-DROCKSDB_MALLOC_USABLE_SIZE",
+    "-DROCKSDB_RANGESYNC_PRESENT",
+    "-DROCKSDB_SCHED_GETCPU_PRESENT",
+    "-DROCKSDB_SUPPORT_THREAD_LOCAL",
+    "-DOS_LINUX",
+    # Flags to enable libs we include
+    "-DSNAPPY",
+    "-DZLIB",
+    "-DBZIP2",
+    "-DLZ4",
+    "-DZSTD",
+    "-DGFLAGS=gflags",
+    "-DNUMA",
+    "-DTBB",
+    # Needed to compile in fbcode
+    "-Wno-expansion-to-defined",
+    # Added missing flags from output of build_detect_platform
+    "-DROCKSDB_PTHREAD_ADAPTIVE_MUTEX",
+    "-DROCKSDB_BACKTRACE",
+    "-Wnarrowing",
 ]
 
-rocksdb_external_deps = [
-  ('bzip2', None, 'bz2'),
-  ('snappy', None, "snappy"),
-  ('zlib', None, 'z'),
-  ('gflags', None, 'gflags'),
-  ('lz4', None, 'lz4'),
-  ('zstd', None),
-  ('tbb', None),
-  ("numa", None, "numa"),
-  ("googletest", None, "gtest"),
+ROCKSDB_EXTERNAL_DEPS = [
+    ("bzip2", None, "bz2"),
+    ("snappy", None, "snappy"),
+    ("zlib", None, "z"),
+    ("gflags", None, "gflags"),
+    ("lz4", None, "lz4"),
+    ("zstd", None),
+    ("tbb", None),
+    ("numa", None, "numa"),
+    ("googletest", None, "gtest"),
 ]
 
-rocksdb_preprocessor_flags = [
-  # Directories with files for #include
-  "-I" + REPO_PATH + "include/",
-  "-I" + REPO_PATH,
+ROCKSDB_PREPROCESSOR_FLAGS = [
+    # Directories with files for #include
+    "-I" + REPO_PATH + "include/",
+    "-I" + REPO_PATH,
 ]
+
+ROCKSDB_ARCH_PREPROCESSOR_FLAGS = {
+    "x86_64": [
+        "-DHAVE_SSE42",
+        "-DHAVE_PCLMUL",
+    ],
+}
+
+build_mode = read_config("fbcode", "build_mode")
+
+is_opt_mode = build_mode.startswith("opt")
+
+# -DNDEBUG is added by default in opt mode in fbcode. But adding it twice
+# doesn't harm and avoid forgetting to add it.
+ROCKSDB_COMPILER_FLAGS += (["-DNDEBUG"] if is_opt_mode else [])
+
+sanitizer = read_config("fbcode", "sanitizer")
+
+# Do not enable jemalloc if sanitizer presents. RocksDB will further detect
+# whether the binary is linked with jemalloc at runtime.
+ROCKSDB_COMPILER_FLAGS += (["-DROCKSDB_JEMALLOC"] if sanitizer == "" else [])
+
+ROCKSDB_EXTERNAL_DEPS += ([("jemalloc", None, "headers")] if sanitizer == "" else [])
 """
 
 
 library_template = """
 cpp_library(
-    name = "%s",
-    headers = %s,
-    srcs = [%s],
-    deps = [%s],
-    preprocessor_flags = rocksdb_preprocessor_flags,
-    compiler_flags = rocksdb_compiler_flags,
-    external_deps = rocksdb_external_deps,
+    name = "{name}",
+    srcs = [{srcs}],
+    {headers_attr_prefix}headers = {headers},
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+    deps = [{deps}],
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 """
 
 binary_template = """
 cpp_binary(
-  name = "%s",
-  srcs = [%s],
-  deps = [%s],
-  preprocessor_flags = rocksdb_preprocessor_flags,
-  compiler_flags = rocksdb_compiler_flags,
-  external_deps = rocksdb_external_deps,
+    name = "%s",
+    srcs = [%s],
+    arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+    compiler_flags = ROCKSDB_COMPILER_FLAGS,
+    preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+    deps = [%s],
+    external_deps = ROCKSDB_EXTERNAL_DEPS,
 )
 """
 
+test_cfg_template = """    [
+        "%s",
+        "%s",
+        "%s",
+    ],
+"""
+
 unittests_template = """
 # [test_name, test_src, test_type]
-ROCKS_TESTS = %s
-
+ROCKS_TESTS = [
+%s]
 
 # Generate a test rule for each entry in ROCKS_TESTS
-for test_cfg in ROCKS_TESTS:
-    test_name = test_cfg[0]
-    test_cc = test_cfg[1]
-    ttype = "gtest" if test_cfg[2] == "parallel" else "simple"
-    test_bin = test_name + "_bin"
-
-    cpp_binary (
-      name = test_bin,
-      srcs = [test_cc],
-      deps = [":rocksdb_test_lib"],
-      preprocessor_flags = rocksdb_preprocessor_flags,
-      compiler_flags = rocksdb_compiler_flags,
-      external_deps = rocksdb_external_deps,
+# Do not build the tests in opt mode, since SyncPoint and other test code
+# will not be included.
+[
+    test_binary(
+        parallelism = parallelism,
+        rocksdb_arch_preprocessor_flags = ROCKSDB_ARCH_PREPROCESSOR_FLAGS,
+        rocksdb_compiler_flags = ROCKSDB_COMPILER_FLAGS,
+        rocksdb_external_deps = ROCKSDB_EXTERNAL_DEPS,
+        rocksdb_preprocessor_flags = ROCKSDB_PREPROCESSOR_FLAGS,
+        test_cc = test_cc,
+        test_name = test_name,
     )
-
-    custom_unittest(
-      name = test_name,
-      type = ttype,
-      deps = [":" + test_bin],
-      command = [TEST_RUNNER, BUCK_BINS + test_bin]
-    )
-
-custom_unittest(
-    name = "make_rocksdbjavastatic",
-    type = "simple",
-    command = ["internal_repo_rocksdb/make_rocksdbjavastatic.sh"],
-)
-
-custom_unittest(
-    name = "make_rocksdb_lite_release",
-    type = "simple",
-    command = ["internal_repo_rocksdb/make_rocksdb_lite_release.sh"],
-)
+    for test_name, test_cc, parallelism in ROCKS_TESTS
+    if not is_opt_mode
+]
 """

build_tools/RocksDBCommonHelper.php

@@ -1,18 +1,18 @@
 <?php
 // Copyright 2004-present Facebook. All Rights Reserved.
-// This source code is licensed under the BSD-style license found in the
-// LICENSE file in the root directory of this source tree. An additional grant
-// of patent rights can be found in the PATENTS file in the same directory.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
 
 // Name of the environment variables which need to be set by the entity which
 // triggers continuous runs so that code at the end of the file gets executed
 // and Sandcastle run starts.
-define("ENV_POST_RECEIVE_HOOK", "POST_RECEIVE_HOOK");
-define("ENV_HTTPS_APP_VALUE", "HTTPS_APP_VALUE");
-define("ENV_HTTPS_TOKEN_VALUE", "HTTPS_TOKEN_VALUE");
+const ENV_POST_RECEIVE_HOOK = "POST_RECEIVE_HOOK";
+const ENV_HTTPS_APP_VALUE = "HTTPS_APP_VALUE";
+const ENV_HTTPS_TOKEN_VALUE = "HTTPS_TOKEN_VALUE";
 
-define("PRIMARY_TOKEN_FILE", '/home/krad/.sandcastle');
-define("CONT_RUN_ALIAS", "leveldb");
+const PRIMARY_TOKEN_FILE = '/home/krad/.sandcastle';
+const CONT_RUN_ALIAS = "leveldb";
 
 //////////////////////////////////////////////////////////////////////
 /*  Run tests in sandcastle */
@@ -21,11 +21,17 @@ function postURL($diffID, $url) {
   assert(is_numeric($diffID));
   assert(strlen($url) > 0);
 
-  $cmd = 'echo \'{"diff_id": ' . $diffID . ', '
-         . '"name":"click here for sandcastle tests for D' . $diffID . '", '
-         . '"link":"' . $url . '"}\' | '
-         . 'arc call-conduit '
-         . 'differential.updateunitresults';
+  $cmd_args = array(
+    'diff_id' => (int)$diffID,
+    'name' => sprintf(
+      'click here for sandcastle tests for D%d',
+      (int)$diffID
+    ),
+    'link' => $url
+  );
+  $cmd = 'echo ' . escapeshellarg(json_encode($cmd_args))
+         . ' | arc call-conduit differential.updateunitresults';
+
   shell_exec($cmd);
 }
 
@@ -35,11 +41,15 @@ function buildUpdateTestStatusCmd($diffID, $test, $status) {
   assert(strlen($test) > 0);
   assert(strlen($status) > 0);
 
-  $cmd = 'echo \'{"diff_id": ' . $diffID . ', '
-         . '"name":"' . $test . '", '
-         . '"result":"' . $status . '"}\' | '
-         . 'arc call-conduit '
-         . 'differential.updateunitresults';
+  $cmd_args = array(
+    'diff_id' => (int)$diffID,
+    'name' => $test,
+    'result' => $status
+  );
+
+  $cmd = 'echo ' . escapeshellarg(json_encode($cmd_args))
+         . ' | arc call-conduit differential.updateunitresults';
+
   return $cmd;
 }
 
@@ -68,7 +78,7 @@ function getSteps($applyDiff, $diffID, $username, $test) {
     // and authenticate using that in Sandcastle.
     $setup = array(
       "name" => "Setup arcrc",
-      "shell" => "echo " . $arcrc_content . " | base64 --decode"
+      "shell" => "echo " . escapeshellarg($arcrc_content) . " | base64 --decode"
                  . " | gzip -d > ~/.arcrc",
       "user" => "root"
     );
@@ -87,14 +97,24 @@ function getSteps($applyDiff, $diffID, $username, $test) {
   }
 
   // fbcode is a sub-repo. We cannot patch until we add it to ignore otherwise
-  // Git thinks it is an uncommited change.
+  // Git thinks it is an uncommitted change.
   $fix_git_ignore = array(
     "name" => "Fix git ignore",
     "shell" => "echo fbcode >> .git/info/exclude",
     "user" => "root"
   );
 
+  // This fixes "FATAL: ThreadSanitizer can not mmap the shadow memory"
+  // Source:
+  // https://github.com/google/sanitizers/wiki/ThreadSanitizerCppManual#FAQ
+  $fix_kernel_issue = array(
+    "name" => "Fix kernel issue with tsan",
+    "shell" => "echo 2 >/proc/sys/kernel/randomize_va_space",
+    "user" => "root"
+  );
+
   $steps[] = $fix_git_ignore;
+  $steps[] = $fix_kernel_issue;
 
   // This will be the command used to execute particular type of tests.
   $cmd = "";
@@ -104,7 +124,7 @@ function getSteps($applyDiff, $diffID, $username, $test) {
     $patch = array(
       "name" => "Patch " . $diffID,
       "shell" => "arc --arcrc-file ~/.arcrc "
-                  . "patch --nocommit --diff " . $diffID,
+                  . "patch --nocommit --diff " . escapeshellarg($diffID),
       "user" => "root"
     );
 
@@ -115,8 +135,8 @@ function getSteps($applyDiff, $diffID, $username, $test) {
   }
 
   // Run the actual command.
-  $cmd = $cmd . "J=$(nproc) ./build_tools/precommit_checker.py " . $test
-           . "; exit_code=$?; ";
+  $cmd = $cmd . "J=$(nproc) ./build_tools/precommit_checker.py " .
+           escapeshellarg($test) . "; exit_code=$?; ";
 
   if ($applyDiff) {
     $cmd = $cmd . "([[ \$exit_code -eq 0 ]] &&"
@@ -149,7 +169,7 @@ function getSteps($applyDiff, $diffID, $username, $test) {
     "name" => "Run " . $test,
     "shell" => $cmd,
     "user" => "root",
-    "parser" => "python build_tools/error_filter.py " . $test,
+    "parser" => "python build_tools/error_filter.py " . escapeshellarg($test),
   );
 
   $steps[] = $run_test;
@@ -197,7 +217,7 @@ function getSandcastleConfig() {
     if (file_exists(PRIMARY_TOKEN_FILE)) {
       $cmd = 'cat ' . PRIMARY_TOKEN_FILE;
     } else {
-      $cmd = 'cat ' . $cwd_token_file;
+      $cmd = 'cat ' . escapeshellarg($cwd_token_file);
     }
 
     assert(strlen($cmd) > 0);
@@ -241,7 +261,7 @@ function getSandcastleConfig() {
   // List of tests we want to run in Sandcastle.
   $tests = array("unit", "unit_non_shm", "unit_481", "clang_unit", "tsan",
                  "asan", "lite_test", "valgrind", "release", "release_481",
-                 "clang_release", "punit", "clang_analyze", "code_cov",
+                 "clang_release", "clang_analyze", "code_cov",
                  "java_build", "no_compression", "unity", "ubsan");
 
   $send_email_template = array(
@@ -321,9 +341,11 @@ function getSandcastleConfig() {
   $app = $sandcastle_config[0];
   $token = $sandcastle_config[1];
 
-  $cmd = 'curl -s -k -F app=' . $app . ' '
-          . '-F token=' . $token . ' -F job=\'' . json_encode($job)
-          .'\' "' . $url . '"';
+  $cmd = 'curl -s -k '
+          . ' -F app=' . escapeshellarg($app)
+          . ' -F token=' . escapeshellarg($token)
+          . ' -F job=' . escapeshellarg(json_encode($job))
+          .' ' . escapeshellarg($url);
 
   $output = shell_exec($cmd);
   assert(strlen($output) > 0);

build_tools/build_detect_platform

@@ -1,4 +1,4 @@
-#!/bin/sh
+#!/usr/bin/env bash
 #
 # Detects OS we're compiling on and outputs a file specified by the first
 # argument, which in turn gets read while processing Makefile.
@@ -16,6 +16,8 @@
 #   PLATFORM_CXXFLAGS           C++ compiler flags.  Will contain:
 #   PLATFORM_SHARED_VERSIONED   Set to 'true' if platform supports versioned
 #                               shared libraries, empty otherwise.
+#   FIND			Command for the find utility
+#   WATCH			Command for the watch utility
 #
 # The PLATFORM_CCFLAGS and PLATFORM_CXXFLAGS might include the following:
 #
@@ -54,10 +56,10 @@ if [ -z "$ROCKSDB_NO_FBCODE" -a -d /mnt/gvfs/third-party ]; then
     if [ -n "$ROCKSDB_FBCODE_BUILD_WITH_481" ]; then
       # we need this to build with MySQL. Don't use for other purposes.
       source "$PWD/build_tools/fbcode_config4.8.1.sh"
-    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_5xx" ]; then
-      source "$PWD/build_tools/fbcode_config.sh"
-    else
+    elif [ -n "$ROCKSDB_FBCODE_BUILD_WITH_PLATFORM007" ]; then
       source "$PWD/build_tools/fbcode_config_platform007.sh"
+    else
+      source "$PWD/build_tools/fbcode_config.sh"
     fi
 fi
 
@@ -66,11 +68,23 @@ rm -f "$OUTPUT"
 touch "$OUTPUT"
 
 if test -z "$CC"; then
-   CC=cc
+    if [ -x "$(command -v cc)" ]; then
+        CC=cc
+    elif [ -x "$(command -v clang)" ]; then
+        CC=clang
+    else
+        CC=cc
+    fi
 fi
 
 if test -z "$CXX"; then
-    CXX=g++
+    if [ -x "$(command -v g++)" ]; then
+        CXX=g++
+    elif [ -x "$(command -v clang++)" ]; then
+        CXX=clang++
+    else
+        CXX=g++
+    fi
 fi
 
 # Detect OS
@@ -87,7 +101,15 @@ if test -z "$CLANG_SCAN_BUILD"; then
 fi
 
 if test -z "$CLANG_ANALYZER"; then
-    CLANG_ANALYZER=$(which clang++ 2> /dev/null)
+    CLANG_ANALYZER=$(command -v clang++ 2> /dev/null)
+fi
+
+if test -z "$FIND"; then
+    FIND=find
+fi
+
+if test -z "$WATCH"; then
+    WATCH=watch
 fi
 
 COMMON_FLAGS="$COMMON_FLAGS ${CFLAGS}"
@@ -124,6 +146,8 @@ case "$TARGET_OS" in
         COMMON_FLAGS="$COMMON_FLAGS -DOS_LINUX"
         if [ -z "$USE_CLANG" ]; then
             COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp"
+        else
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -latomic"
         fi
         PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt"
         # PORT_FILES=port/linux/linux_specific.cc
@@ -143,6 +167,7 @@ case "$TARGET_OS" in
         ;;
     FreeBSD)
         PLATFORM=OS_FREEBSD
+        CXX=clang++
         COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_FREEBSD"
         PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread"
         # PORT_FILES=port/freebsd/freebsd_specific.cc
@@ -155,9 +180,12 @@ case "$TARGET_OS" in
         ;;
     OpenBSD)
         PLATFORM=OS_OPENBSD
+	CXX=clang++
         COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp -D_REENTRANT -DOS_OPENBSD"
         PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -pthread"
         # PORT_FILES=port/openbsd/openbsd_specific.cc
+	FIND=gfind
+	WATCH=gnuwatch
         ;;
     DragonFly)
         PLATFORM=OS_DRAGONFLYBSD
@@ -172,6 +200,8 @@ case "$TARGET_OS" in
         COMMON_FLAGS="$COMMON_FLAGS -DCYGWIN"
         if [ -z "$USE_CLANG" ]; then
             COMMON_FLAGS="$COMMON_FLAGS -fno-builtin-memcmp"
+        else
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -latomic"
         fi
         PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lpthread -lrt"
         # PORT_FILES=port/linux/linux_specific.cc
@@ -212,117 +242,147 @@ EOF
         fi
     fi
 
-    # Test whether Snappy library is installed
-    # http://code.google.com/p/snappy/
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <snappy.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_SNAPPY; then
+        # Test whether Snappy library is installed
+        # http://code.google.com/p/snappy/
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <snappy.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DSNAPPY"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lsnappy"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -lsnappy"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DSNAPPY"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lsnappy"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -lsnappy"
+        fi
     fi
 
-    # Test whether gflags library is installed
-    # http://gflags.github.io/gflags/
-    # check if the namespace is gflags
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
-      #include <gflags/gflags.h>
-      using namespace gflags;
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_GFLAGS; then
+        # Test whether gflags library is installed
+        # http://gflags.github.io/gflags/
+        # check if the namespace is gflags
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
+          #include <gflags/gflags.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=gflags"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
-    else
-      # check if namespace is google
-      $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
-        #include <gflags/gflags.h>
-        using namespace google;
-        int main() {}
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=1"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
+        else
+          # check if namespace is google
+          $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null << EOF
+            #include <gflags/gflags.h>
+            using namespace google;
+            int main() {}
 EOF
-      if [ "$?" = 0 ]; then
-          COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=google"
-          PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
-      fi
+          if [ "$?" = 0 ]; then
+              COMMON_FLAGS="$COMMON_FLAGS -DGFLAGS=google"
+              PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lgflags"
+          fi
+        fi
     fi
 
-    # Test whether zlib library is installed
-    $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <zlib.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_ZLIB; then
+        # Test whether zlib library is installed
+        $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <zlib.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DZLIB"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lz"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -lz"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DZLIB"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lz"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -lz"
+        fi
     fi
 
-    # Test whether bzip library is installed
-    $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <bzlib.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_BZIP; then
+        # Test whether bzip library is installed
+        $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <bzlib.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DBZIP2"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lbz2"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -lbz2"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DBZIP2"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lbz2"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -lbz2"
+        fi
     fi
 
-    # Test whether lz4 library is installed
-    $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <lz4.h>
-      #include <lz4hc.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_LZ4; then
+        # Test whether lz4 library is installed
+        $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <lz4.h>
+          #include <lz4hc.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DLZ4"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -llz4"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -llz4"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DLZ4"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -llz4"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -llz4"
+        fi
     fi
 
-    # Test whether zstd library is installed
-    $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <zstd.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_ZSTD; then
+        # Test whether zstd library is installed
+        $CXX $CFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <zstd.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DZSTD"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lzstd"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -lzstd"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DZSTD"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lzstd"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -lzstd"
+        fi
     fi
 
-    # Test whether numa is available
-    $CXX $CFLAGS -x c++ - -o /dev/null -lnuma 2>/dev/null  <<EOF
-      #include <numa.h>
-      #include <numaif.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_NUMA; then
+        # Test whether numa is available
+        $CXX $CFLAGS -x c++ - -o /dev/null -lnuma 2>/dev/null  <<EOF
+          #include <numa.h>
+          #include <numaif.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DNUMA"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lnuma"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -lnuma"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DNUMA"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lnuma"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -lnuma"
+        fi
     fi
 
-    # Test whether tbb is available
-    $CXX $CFLAGS $LDFLAGS -x c++ - -o /dev/null -ltbb 2>/dev/null  <<EOF
-      #include <tbb/tbb.h>
-      int main() {}
+    if ! test $ROCKSDB_DISABLE_TBB; then
+        # Test whether tbb is available
+        $CXX $CFLAGS $LDFLAGS -x c++ - -o /dev/null -ltbb 2>/dev/null  <<EOF
+          #include <tbb/tbb.h>
+          int main() {}
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DTBB"
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ltbb"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -ltbb"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DTBB"
+            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ltbb"
+            JAVA_LDFLAGS="$JAVA_LDFLAGS -ltbb"
+        fi
     fi
 
-    # Test whether jemalloc is available
-    if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null -ljemalloc \
-      2>/dev/null; then
-        PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -ljemalloc"
-        JAVA_LDFLAGS="$JAVA_LDFLAGS -ljemalloc"
-        JEMALLOC=1
-    else
+    if ! test $ROCKSDB_DISABLE_JEMALLOC; then
+        # Test whether jemalloc is available
+        if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null -ljemalloc \
+          2>/dev/null; then
+            # This will enable some preprocessor identifiers in the Makefile
+            JEMALLOC=1
+            # JEMALLOC can be enabled either using the flag (like here) or by
+            # providing direct link to the jemalloc library
+            WITH_JEMALLOC_FLAG=1
+            # check for JEMALLOC installed with HomeBrew
+            if [ "$PLATFORM" == "OS_MACOSX" ]; then
+                if hash brew 2>/dev/null && brew ls --versions jemalloc > /dev/null; then
+                    JEMALLOC_VER=$(brew ls --versions jemalloc | tail -n 1 | cut -f 2 -d ' ')
+                    JEMALLOC_INCLUDE="-I/usr/local/Cellar/jemalloc/${JEMALLOC_VER}/include"
+                    JEMALLOC_LIB="/usr/local/Cellar/jemalloc/${JEMALLOC_VER}/lib/libjemalloc_pic.a"
+                    PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS $JEMALLOC_LIB"
+                    JAVA_STATIC_LDFLAGS="$JAVA_STATIC_LDFLAGS $JEMALLOC_LIB"
+                fi
+            fi
+        fi
+    fi
+    if ! test $JEMALLOC && ! test $ROCKSDB_DISABLE_TCMALLOC; then
         # jemalloc is not available. Let's try tcmalloc
         if echo 'int main() {}' | $CXX $CFLAGS -x c++ - -o /dev/null \
           -ltcmalloc 2>/dev/null; then
@@ -331,88 +391,111 @@ EOF
         fi
     fi
 
-    # Test whether malloc_usable_size is available
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <malloc.h>
-      int main() {
-        size_t res = malloc_usable_size(0);
-        return 0;
-      }
+    if ! test $ROCKSDB_DISABLE_MALLOC_USABLE_SIZE; then
+        # Test whether malloc_usable_size is available
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <malloc.h>
+          int main() {
+            size_t res = malloc_usable_size(0);
+            return 0;
+          }
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_MALLOC_USABLE_SIZE"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_MALLOC_USABLE_SIZE"
+        fi
     fi
 
-    # Test whether PTHREAD_MUTEX_ADAPTIVE_NP mutex type is available
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <pthread.h>
-      int main() {
-        int x = PTHREAD_MUTEX_ADAPTIVE_NP;
-        return 0;
-      }
+    if ! test $ROCKSDB_DISABLE_PTHREAD_MUTEX_ADAPTIVE_NP; then
+        # Test whether PTHREAD_MUTEX_ADAPTIVE_NP mutex type is available
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <pthread.h>
+          int main() {
+            int x = PTHREAD_MUTEX_ADAPTIVE_NP;
+            return 0;
+          }
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_PTHREAD_ADAPTIVE_MUTEX"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_PTHREAD_ADAPTIVE_MUTEX"
+        fi
     fi
 
-    # Test whether backtrace is available
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <execinfo.h>>
-      int main() {
-        void* frames[1];
-        backtrace_symbols(frames, backtrace(frames, 1));
-        return 0;
-      }
-EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_BACKTRACE"
-    else
-        # Test whether execinfo library is installed
-        $CXX $CFLAGS -lexecinfo -x c++ - -o /dev/null 2>/dev/null  <<EOF
-          #include <execinfo.h>
+    if ! test $ROCKSDB_DISABLE_BACKTRACE; then
+        # Test whether backtrace is available
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <execinfo.h>>
           int main() {
             void* frames[1];
             backtrace_symbols(frames, backtrace(frames, 1));
+            return 0;
           }
 EOF
         if [ "$?" = 0 ]; then
             COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_BACKTRACE"
-            PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lexecinfo"
-            JAVA_LDFLAGS="$JAVA_LDFLAGS -lexecinfo"
+        else
+            # Test whether execinfo library is installed
+            $CXX $CFLAGS -lexecinfo -x c++ - -o /dev/null 2>/dev/null  <<EOF
+              #include <execinfo.h>
+              int main() {
+                void* frames[1];
+                backtrace_symbols(frames, backtrace(frames, 1));
+              }
+EOF
+            if [ "$?" = 0 ]; then
+                COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_BACKTRACE"
+                PLATFORM_LDFLAGS="$PLATFORM_LDFLAGS -lexecinfo"
+                JAVA_LDFLAGS="$JAVA_LDFLAGS -lexecinfo"
+            fi
         fi
     fi
 
-    # Test if -pg is supported
-    $CXX $CFLAGS -pg -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      int main() {
-        return 0;
-      }
+    if ! test $ROCKSDB_DISABLE_PG; then
+        # Test if -pg is supported
+        $CXX $CFLAGS -pg -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          int main() {
+            return 0;
+          }
 EOF
-    if [ "$?" = 0 ]; then
-        PROFILING_FLAGS=-pg
+        if [ "$?" = 0 ]; then
+            PROFILING_FLAGS=-pg
+        fi
     fi
 
-    # Test whether sync_file_range is supported for compatibility with an old glibc
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <fcntl.h>
-      int main() {
-        int fd = open("/dev/null", 0);
-        sync_file_range(fd, 0, 1024, SYNC_FILE_RANGE_WRITE);
-      }
+    if ! test $ROCKSDB_DISABLE_SYNC_FILE_RANGE; then
+        # Test whether sync_file_range is supported for compatibility with an old glibc
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <fcntl.h>
+          int main() {
+            int fd = open("/dev/null", 0);
+            sync_file_range(fd, 0, 1024, SYNC_FILE_RANGE_WRITE);
+          }
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_RANGESYNC_PRESENT"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_RANGESYNC_PRESENT"
+        fi
     fi
 
-    # Test whether sched_getcpu is supported
-    $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
-      #include <sched.h>
-      int main() {
-        int cpuid = sched_getcpu();
-      }
+    if ! test $ROCKSDB_DISABLE_SCHED_GETCPU; then
+        # Test whether sched_getcpu is supported
+        $CXX $CFLAGS -x c++ - -o /dev/null 2>/dev/null  <<EOF
+          #include <sched.h>
+          int main() {
+            int cpuid = sched_getcpu();
+          }
 EOF
-    if [ "$?" = 0 ]; then
-        COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_SCHED_GETCPU_PRESENT"
+        if [ "$?" = 0 ]; then
+            COMMON_FLAGS="$COMMON_FLAGS -DROCKSDB_SCHED_GETCPU_PRESENT"
+        fi
+    fi
+
+    if ! test $ROCKSDB_DISABLE_ALIGNED_NEW; then
+        # Test whether c++17 aligned-new is supported
+        $CXX $PLATFORM_CXXFLAGS -faligned-new -x c++ - -o /dev/null 2>/dev/null <<EOF
+            struct alignas(1024) t {int a;};
+            int main() {}
+EOF
+        if [ "$?" = 0 ]; then
+            PLATFORM_CXXFLAGS="$PLATFORM_CXXFLAGS -faligned-new -DHAVE_ALIGNED_NEW"
+        fi
     fi
 fi
 
@@ -444,21 +527,28 @@ if test "$USE_HDFS"; then
   JAVA_LDFLAGS="$JAVA_LDFLAGS $HDFS_LDFLAGS"
 fi
 
-if test "$USE_SSE"; then
-  COMMON_FLAGS="$COMMON_FLAGS -msse4.2"
-elif test -z "$PORTABLE"; then
+if test -z "$PORTABLE"; then
   if test -n "`echo $TARGET_ARCHITECTURE | grep ^ppc64`"; then
     # Tune for this POWER processor, treating '+' models as base models
     POWER=`LD_SHOW_AUXV=1 /bin/true | grep AT_PLATFORM | grep -E -o power[0-9]+`
     COMMON_FLAGS="$COMMON_FLAGS -mcpu=$POWER -mtune=$POWER "
   elif test -n "`echo $TARGET_ARCHITECTURE | grep ^s390x`"; then
     COMMON_FLAGS="$COMMON_FLAGS -march=z10 "
-  elif [ "$TARGET_OS" != AIX ] && [ "$TARGET_OS" != SunOS ]; then
+  elif test -n "`echo $TARGET_ARCHITECTURE | grep ^arm`"; then
+    # TODO: Handle this with approprite options.
+    COMMON_FLAGS="$COMMON_FLAGS"
+  elif [ "$TARGET_OS" == "IOS" ]; then
+    COMMON_FLAGS="$COMMON_FLAGS"
+  elif [ "$TARGET_OS" != "AIX" ] && [ "$TARGET_OS" != "SunOS" ]; then
     COMMON_FLAGS="$COMMON_FLAGS -march=native "
+  elif test "$USE_SSE"; then
+    COMMON_FLAGS="$COMMON_FLAGS -msse4.2 -mpclmul"
   fi
+elif test "$USE_SSE"; then
+  COMMON_FLAGS="$COMMON_FLAGS -msse4.2 -mpclmul"
 fi
 
-$CXX $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
+$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
   #include <cstdint>
   #include <nmmintrin.h>
   int main() {
@@ -469,6 +559,24 @@ if [ "$?" = 0 ]; then
   COMMON_FLAGS="$COMMON_FLAGS -DHAVE_SSE42"
 elif test "$USE_SSE"; then
   echo "warning: USE_SSE specified but compiler could not use SSE intrinsics, disabling"
+  exit 1
+fi
+
+$CXX $PLATFORM_CXXFLAGS $COMMON_FLAGS -x c++ - -o /dev/null 2>/dev/null <<EOF
+  #include <cstdint>
+  #include <wmmintrin.h>
+  int main() {
+    const auto a = _mm_set_epi64x(0, 0);
+    const auto b = _mm_set_epi64x(0, 0);
+    const auto c = _mm_clmulepi64_si128(a, b, 0x00);
+    auto d = _mm_cvtsi128_si64(c);
+  }
+EOF
+if [ "$?" = 0 ]; then
+  COMMON_FLAGS="$COMMON_FLAGS -DHAVE_PCLMUL"
+elif test "$USE_SSE"; then
+  echo "warning: USE_SSE specified but compiler could not use PCLMUL intrinsics, disabling"
+  exit 1
 fi
 
 # iOS doesn't support thread-local storage, but this check would erroneously
@@ -519,7 +627,16 @@ echo "ROCKSDB_PATCH=$ROCKSDB_PATCH" >> "$OUTPUT"
 echo "CLANG_SCAN_BUILD=$CLANG_SCAN_BUILD" >> "$OUTPUT"
 echo "CLANG_ANALYZER=$CLANG_ANALYZER" >> "$OUTPUT"
 echo "PROFILING_FLAGS=$PROFILING_FLAGS" >> "$OUTPUT"
+echo "FIND=$FIND" >> "$OUTPUT"
+echo "WATCH=$WATCH" >> "$OUTPUT"
+# This will enable some related identifiers for the preprocessor
 if test -n "$JEMALLOC"; then
   echo "JEMALLOC=1" >> "$OUTPUT"
 fi
+# Indicates that jemalloc should be enabled using -ljemalloc flag
+# The alternative is to porvide a direct link to the library via JEMALLOC_LIB
+# and JEMALLOC_INCLUDE
+if test -n "$WITH_JEMALLOC_FLAG"; then
+  echo "WITH_JEMALLOC_FLAG=$WITH_JEMALLOC_FLAG" >> "$OUTPUT"
+fi
 echo "LUA_PATH=$LUA_PATH" >> "$OUTPUT"

build_tools/cont_integration.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 #
 # Copyright (c) 2016, Facebook. All rights reserved.
 #
@@ -13,10 +13,12 @@ error=0
 
 function log {
   DATE=`date +%Y-%m-%d:%H:%M:%S`
+  # shellcheck disable=SC2068
   echo $DATE $@
 }
 
 function log_err {
+  # shellcheck disable=SC2145
   log "ERROR: $@ Error code: $error."
 }
 
@@ -67,7 +69,7 @@ function update_repo_status {
 #
 
 # Path to the determinator from the root of the RocksDB repo.
-CONTRUN_DETERMINATOR=./arcanist_util/config/RocksDBCommonHelper.php
+CONTRUN_DETERMINATOR=./build_tools/RocksDBCommonHelper.php
 
 # Value of the previous commit.
 PREV_COMMIT=

build_tools/dependencies.sh

@@ -1,18 +1,18 @@
-GCC_BASE=/mnt/gvfs/third-party2/gcc/2928bb3ed95bf64f5b388ee88c30dc74710c3b35/5.x/centos6-native/f4950a1
-CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/a5fea028cb7ba43498976e1f8054b0b2e790c295/stable/centos6-native/6aaf4de
-LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/7a9099f6587ee4378c0b1fa32bb8934019d30ca4/5.x/gcc-5-glibc-2.23/339d858
-GLIBC_BASE=/mnt/gvfs/third-party2/glibc/3b7c6469854dfc7832a1c3cc5b86919a84e5f865/2.23/gcc-5-glibc-2.23/ca1d1c0
-SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/8c38a4c1e52b4c2cc8a9cdc31b9c947ed7dbfcb4/1.1.3/gcc-5-glibc-2.23/9bc6787
-ZLIB_BASE=/mnt/gvfs/third-party2/zlib/d7861abe6f0e27ab98c9303b95a662f0e4cdedb5/1.2.8/gcc-5-glibc-2.23/9bc6787
-BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/740325875f6729f42d28deaa2147b0854f3a347e/1.0.6/gcc-5-glibc-2.23/9bc6787
-LZ4_BASE=/mnt/gvfs/third-party2/lz4/0815d59804160c96caac5f27ca004f51af893dc6/r131/gcc-5-glibc-2.23/9bc6787
-ZSTD_BASE=/mnt/gvfs/third-party2/zstd/c15a4f5f619a2930478d01e2e34dc1e0652b0873/1.1.4/gcc-5-glibc-2.23/03859b5
-GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/f905a5e1032fb30c05db3d3752319857388c0c49/2.2.0/gcc-5-glibc-2.23/9bc6787
-JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/8d60633d822a2a55849c73db24e74a25e52b71db/master/gcc-5-glibc-2.23/1c32b4b
-NUMA_BASE=/mnt/gvfs/third-party2/numa/17c514c4d102a25ca15f4558be564eeed76f4b6a/2.0.8/gcc-5-glibc-2.23/9bc6787
-LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/8db74270cd6d0212ac92d69e7fc7beefe617d772/trunk/gcc-5-glibc-2.23/b1847cb
-TBB_BASE=/mnt/gvfs/third-party2/tbb/9d9a554877d0c5bef330fe818ab7178806dd316a/4.0_update2/gcc-5-glibc-2.23/9bc6787
-KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/90c9734afc5579c9d1db529fa788d09f97763b85/4.0.9-36_fbk5_2933_gd092e3f/gcc-5-glibc-2.23/da39a3e
-BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/9e829389ef61b92c62de8748c80169aaf25ce1f0/2.26.1/centos6-native/da39a3e
-VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/d7f4d4d86674a57668e3a96f76f0e17dd0eb8765/3.11.0/gcc-5-glibc-2.23/9bc6787
-LUA_BASE=/mnt/gvfs/third-party2/lua/61e4abf5813bbc39bc4f548757ccfcadde175a48/5.2.3/gcc-5-glibc-2.23/65372bd
+GCC_BASE=/mnt/gvfs/third-party2/gcc/112ec378fec7002ad3e09afde022e656049f7191/5.x/centos7-native/c447969
+CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/04999bdb3ce81a11073535dcb00b5e13dc1cbaf5/stable/centos7-native/c9f9104
+LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/92b0c8e5c8eecc71eb042594ce1ab3413799b385/5.x/gcc-5-glibc-2.23/339d858
+GLIBC_BASE=/mnt/gvfs/third-party2/glibc/3d8698d5973ba94f41620a80a67e4457fdf01e90/2.23/gcc-5-glibc-2.23/ca1d1c0
+SNAPPY_BASE=/mnt/gvfs/third-party2/snappy/7f9bdaada18f59bc27ec2b0871eb8a6144343aef/1.1.3/gcc-5-glibc-2.23/9bc6787
+ZLIB_BASE=/mnt/gvfs/third-party2/zlib/22c2d65676fb7c23cfa797c4f6937f38b026f3cf/1.2.8/gcc-5-glibc-2.23/9bc6787
+BZIP2_BASE=/mnt/gvfs/third-party2/bzip2/dc49a21c5fceec6456a7a28a94dcd16690af1337/1.0.6/gcc-5-glibc-2.23/9bc6787
+LZ4_BASE=/mnt/gvfs/third-party2/lz4/907b498203d297947f3bb70b9466f47e100f1873/r131/gcc-5-glibc-2.23/9bc6787
+ZSTD_BASE=/mnt/gvfs/third-party2/zstd/af6628a46758f1a15484a1760cd7294164bc5ba1/1.3.5/gcc-5-glibc-2.23/03859b5
+GFLAGS_BASE=/mnt/gvfs/third-party2/gflags/0b9929d2588991c65a57168bf88aff2db87c5d48/2.2.0/gcc-5-glibc-2.23/9bc6787
+JEMALLOC_BASE=/mnt/gvfs/third-party2/jemalloc/b1a0e56c1e3e6929813a4331ade3a58ff083afbb/master/gcc-5-glibc-2.23/aa64d6b
+NUMA_BASE=/mnt/gvfs/third-party2/numa/9cbf2460284c669ed19c3ccb200a71f7dd7e53c7/2.0.11/gcc-5-glibc-2.23/9bc6787
+LIBUNWIND_BASE=/mnt/gvfs/third-party2/libunwind/bf3d7497fe4e6d007354f0adffa16ce3003f8338/1.3/gcc-5-glibc-2.23/b443de1
+TBB_BASE=/mnt/gvfs/third-party2/tbb/ff4e0b093534704d8abab678a4fd7f5ea7b094c7/2018_U5/gcc-5-glibc-2.23/9bc6787
+KERNEL_HEADERS_BASE=/mnt/gvfs/third-party2/kernel-headers/b5c4a61a5c483ba24722005ae07895971a2ac707/4.0.9-36_fbk5_2933_gd092e3f/gcc-5-glibc-2.23/da39a3e
+BINUTILS_BASE=/mnt/gvfs/third-party2/binutils/55031de95a2b46c82948743419a603b3d6aefe28/2.29.1/centos7-native/da39a3e
+VALGRIND_BASE=/mnt/gvfs/third-party2/valgrind/f3f697a28122e6bcd513273dd9c1ff23852fc59f/3.13.0/gcc-5-glibc-2.23/9bc6787
+LUA_BASE=/mnt/gvfs/third-party2/lua/f0cd714433206d5139df61659eb7b28b1dea6683/5.2.3/gcc-5-glibc-2.23/65372bd

build_tools/dependencies_4.8.1.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC2148
 GCC_BASE=/mnt/gvfs/third-party2/gcc/cf7d14c625ce30bae1a4661c2319c5a283e4dd22/4.8.1/centos6-native/cc6c9dc
 CLANG_BASE=/mnt/gvfs/third-party2/llvm-fb/8598c375b0e94e1448182eb3df034704144a838d/stable/centos6-native/3f16ddd
 LIBGCC_BASE=/mnt/gvfs/third-party2/libgcc/d6e0a7da6faba45f5e5b1638f9edd7afc2f34e7d/4.8.1/gcc-4.8.1-glibc-2.17/8aac7fc

build_tools/dockerbuild.sh

@@ -1,2 +1,2 @@
-#!/bin/bash
+#!/usr/bin/env bash
 docker run -v $PWD:/rocks -w /rocks buildpack-deps make

build_tools/error_filter.py

@@ -64,8 +64,12 @@ class MatchErrorParser(ErrorParserBase):
 
 class CompilerErrorParser(MatchErrorParser):
     def __init__(self):
-        # format: '<filename>:<line #>:<column #>: error: <error msg>'
-        super(CompilerErrorParser, self).__init__(r'\S+:\d+:\d+: error:')
+        # format (compile error):
+        #   '<filename>:<line #>:<column #>: error: <error msg>'
+        # format (link error):
+        #   '<filename>:<line #>: error: <error msg>'
+        # The below regex catches both
+        super(CompilerErrorParser, self).__init__(r'\S+:\d+: error:')
 
 
 class ScanBuildErrorParser(MatchErrorParser):

build_tools/fbcode_config.sh

@@ -43,12 +43,16 @@ if test -z $PIC_BUILD; then
   LZ4_INCLUDE=" -I $LZ4_BASE/include/"
   LZ4_LIBS=" $LZ4_BASE/lib/liblz4.a"
   CFLAGS+=" -DLZ4"
-
-  ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
-  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
-  CFLAGS+=" -DZSTD"
 fi
 
+ZSTD_INCLUDE=" -I $ZSTD_BASE/include/"
+if test -z $PIC_BUILD; then
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd.a"
+else
+  ZSTD_LIBS=" $ZSTD_BASE/lib/libzstd_pic.a"
+fi
+CFLAGS+=" -DZSTD -DZSTD_STATIC_LINKING_ONLY"
+
 # location of gflags headers and libraries
 GFLAGS_INCLUDE=" -I $GFLAGS_BASE/include/"
 if test -z $PIC_BUILD; then
@@ -83,6 +87,7 @@ CFLAGS+=" -DTBB"
 
 # use Intel SSE support for checksum calculations
 export USE_SSE=1
+export PORTABLE=1
 
 BINUTILS="$BINUTILS_BASE/bin"
 AR="$BINUTILS/ar"

build_tools/fbcode_config4.8.1.sh

@@ -54,6 +54,7 @@ TBB_LIBS="$TBB_BASE/lib/libtbb.a"
 
 # use Intel SSE support for checksum calculations
 export USE_SSE=1
+export PORTABLE=1
 
 BINUTILS="$BINUTILS_BASE/bin"
 AR="$BINUTILS/ar"

build_tools/format-diff.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # If clang_format_diff.py command is not specfied, we assume we are able to
 # access directly without any path.
 if [ -z $CLANG_FORMAT_DIFF ]

build_tools/gnu_parallel

@@ -5082,8 +5082,8 @@ sub openoutputfiles {
 	# Set reading FD if using --group (--ungroup does not need)
 	for my $fdno (1,2) {
 	    # Re-open the file for reading
-	    # so fdw can be closed seperately
-	    # and fdr can be seeked seperately (for --line-buffer)
+	    # so fdw can be closed separately
+	    # and fdr can be seeked separately (for --line-buffer)
 	    open(my $fdr,"<", $self->fh($fdno,'name')) ||
 		::die_bug("fdr: Cannot open ".$self->fh($fdno,'name'));
 	    $self->set_fh($fdno,'r',$fdr);

build_tools/make_new_version.sh

@@ -1,53 +0,0 @@
-#!/bin/bash
-#  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
-#  This source code is licensed under the BSD-style license found in the
-#  LICENSE file in the root directory of this source tree. An additional grant
-#  of patent rights can be found in the PATENTS file in the same directory.
-
-set -e
-if [ -z "$GIT" ]
-then
-  GIT="git"
-fi
-
-# Print out the colored progress info so that it can be brainlessly
-# distinguished by users.
-function title() {
-  echo -e "\033[1;32m$*\033[0m"
-}
-
-usage="Create new RocksDB version and prepare it for the release process\n"
-usage+="USAGE: ./make_new_version.sh <version> [<remote>]\n"
-usage+="  version: specify a version without '.fb' suffix (e.g. 5.4).\n"
-usage+="  remote: name of the remote to push the branch to (default: origin)."
-
-# -- Pre-check
-if [[ $# < 1 ]]; then
-  echo -e $usage
-  exit 1
-fi
-
-ROCKSDB_VERSION=$1
-
-REMOTE="origin"
-if [[ $# > 1 ]]; then
-  REMOTE=$2
-fi
-
-GIT_BRANCH=`git rev-parse --abbrev-ref HEAD`
-echo $GIT_BRANCH
-
-if [ $GIT_BRANCH != "master" ]; then
-  echo "Error: Current branch is '$GIT_BRANCH', Please switch to master branch."
-  exit 1
-fi
-
-title "Adding new tag for this release ..."
-BRANCH="$ROCKSDB_VERSION.fb"
-$GIT checkout -b $BRANCH
-
-# Setting up the proxy for remote repo access
-title "Pushing new branch to remote repo ..."
-git push $REMOTE --set-upstream $BRANCH
-
-title "Branch $BRANCH is pushed to github;"

build_tools/make_package.sh

@@ -1,3 +1,4 @@
+# shellcheck disable=SC1113
 #/usr/bin/env bash
 
 set -e
@@ -28,12 +29,14 @@ function package() {
     if dpkg --get-selections | grep --quiet $1; then
       log "$1 is already installed. skipping."
     else
+      # shellcheck disable=SC2068
       apt-get install $@ -y
     fi
   elif [[ $OS = "centos" ]]; then
     if rpm -qa | grep --quiet $1; then
       log "$1 is already installed. skipping."
     else
+      # shellcheck disable=SC2068
       yum install $@ -y
     fi
   fi
@@ -52,6 +55,7 @@ function gem_install() {
   if gem list | grep --quiet $1; then
     log "$1 is already installed. skipping."
   else
+    # shellcheck disable=SC2068
     gem install $@
   fi
 }
@@ -125,4 +129,5 @@ function main() {
     include $LIB_DIR
 }
 
+# shellcheck disable=SC2068
 main $@

build_tools/regression_build_test.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 set -e
 

build_tools/rocksdb-lego-determinator

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 # This script is executed by Sandcastle
 # to determine next steps to run
 
@@ -89,6 +89,8 @@ CLANG="USE_CLANG=1"
 # using platform007 gives us gcc-8 or higher which has that bug fixed.
 TSAN="ROCKSDB_FBCODE_BUILD_WITH_PLATFORM007=1 COMPILE_WITH_TSAN=1"
 UBSAN="COMPILE_WITH_UBSAN=1"
+TSAN_CRASH='CRASH_TEST_EXT_ARGS="--compression_type=zstd --log2_keys_per_lock=22"'
+NON_TSAN_CRASH="CRASH_TEST_EXT_ARGS=--compression_type=zstd"
 DISABLE_JEMALLOC="DISABLE_JEMALLOC=1"
 HTTP_PROXY="https_proxy=http://fwdproxy.29.prn1:8080 http_proxy=http://fwdproxy.29.prn1:8080 ftp_proxy=http://fwdproxy.29.prn1:8080"
 SETUP_JAVA_ENV="export $HTTP_PROXY; export JAVA_HOME=/usr/local/jdk-8u60-64/; export PATH=\$JAVA_HOME/bin:\$PATH"
@@ -344,7 +346,7 @@ LITE_BUILD_COMMANDS="[
             $CLEANUP_ENV,
             {
                 'name':'Build RocksDB debug version',
-                'shell':'$LITE make J=1 static_lib || $CONTRUN_NAME=lite_static_lib $TASK_CREATION_TOOL',
+                'shell':'make J=1 LITE=1 all check || $CONTRUN_NAME=lite $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -354,23 +356,19 @@ LITE_BUILD_COMMANDS="[
 ]"
 
 #
-# RocksDB lite tests
-#
-LITE_UNIT_TEST_COMMANDS="[
+# Report RocksDB lite binary size to scuba
+REPORT_LITE_BINARY_SIZE_COMMANDS="[
     {
-        'name':'Rocksdb Lite Unit Test',
+        'name':'Rocksdb Lite Binary Size',
         'oncall':'$ONCALL',
         'steps': [
             $CLEANUP_ENV,
             {
-                'name':'Build RocksDB debug version',
-                'shell':'$SHM $LITE make J=1 check || $CONTRUN_NAME=lite_check $TASK_CREATION_TOOL',
+                'name':'Report RocksDB Lite binary size to scuba',
+                'shell':'tools/report_lite_binary_size.sh',
                 'user':'root',
-                $PARSER
             },
         ],
-        $REPORT
-    }
 ]"
 
 #
@@ -385,14 +383,14 @@ STRESS_CRASH_TEST_COMMANDS="[
             $CLEANUP_ENV,
             {
                 'name':'Build and run RocksDB debug stress tests',
-                'shell':'$SHM $DEBUG make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 db_stress || $CONTRUN_NAME=db_stress $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
             {
                 'name':'Build and run RocksDB debug crash tests',
                 'timeout': 86400,
-                'shell':'$SHM $DEBUG make J=1 crash_test || $CONTRUN_NAME=crash_test $TASK_CREATION_TOOL',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 crash_test || $CONTRUN_NAME=crash_test $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             }
@@ -457,7 +455,7 @@ ASAN_CRASH_TEST_COMMANDS="[
             {
                 'name':'Build and run RocksDB debug asan_crash_test',
                 'timeout': 86400,
-                'shell':'$SHM $DEBUG make J=1 asan_crash_test || $CONTRUN_NAME=asan_crash_test $TASK_CREATION_TOOL',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 asan_crash_test || $CONTRUN_NAME=asan_crash_test $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -499,7 +497,7 @@ UBSAN_CRASH_TEST_COMMANDS="[
             {
                 'name':'Build and run RocksDB debug ubsan_crash_test',
                 'timeout': 86400,
-                'shell':'$SHM $DEBUG make J=1 ubsan_crash_test || $CONTRUN_NAME=ubsan_crash_test $TASK_CREATION_TOOL',
+                'shell':'$SHM $DEBUG $NON_TSAN_CRASH make J=1 ubsan_crash_test || $CONTRUN_NAME=ubsan_crash_test $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -565,7 +563,7 @@ TSAN_CRASH_TEST_COMMANDS="[
             {
                 'name':'Compile and run',
                 'timeout': 86400,
-                'shell':'set -o pipefail && $SHM $DEBUG $TSAN CRASH_TEST_KILL_ODD=1887 make J=1 crash_test || $CONTRUN_NAME=tsan_crash_test $TASK_CREATION_TOOL',
+                'shell':'set -o pipefail && $SHM $DEBUG $TSAN $TSAN_CRASH CRASH_TEST_KILL_ODD=1887 make J=1 crash_test || $CONTRUN_NAME=tsan_crash_test $TASK_CREATION_TOOL',
                 'user':'root',
                 $PARSER
             },
@@ -584,39 +582,7 @@ run_format_compatible()
   rm -rf /dev/shm/rocksdb
   mkdir /dev/shm/rocksdb
 
-  echo '
-  if [ -e "build_tools/build_detect_platform" ]
-  then
-    sed "s/tcmalloc/nothingnothingnothing/g" build_tools/build_detect_platform > $TEST_TMPDIR/temp_build_file
-    rm -rf build_tools/build_detect_platform
-    cp $TEST_TMPDIR/temp_build_file build_tools/build_detect_platform
-    chmod +x build_tools/build_detect_platform
-  fi
-
-  if [ -e "build_detect_platform" ]
-  then
-    sed "s/tcmalloc/nothingnothingnothing/g" build_detect_platform > $TEST_TMPDIR/temp_build_file
-    rm -rf build_detect_platform
-    cp $TEST_TMPDIR/temp_build_file build_detect_platform
-    chmod +x build_detect_platform
-  fi
-
-  make ldb -j32
-
-  if [ -e "build_detect_platform" ]
-  then
-    git checkout -- build_detect_platform
-  fi
-
-  if [ -e "build_tools/build_detect_platform" ]
-  then
-    git checkout -- build_tools/build_detect_platform
-  fi
-  ' > temp_build_ldb.sh
-
-  sed "s/make ldb -j32/source temp_build_ldb.sh/g" tools/check_format_compatible.sh > tools/temp_check_format_compatible.sh
-  chmod +x tools/temp_check_format_compatible.sh
-  tools/temp_check_format_compatible.sh
+  tools/check_format_compatible.sh
 }
 
 FORMAT_COMPATIBLE_COMMANDS="[
@@ -698,12 +664,12 @@ run_regression()
 
   # === lite build ===
   make clean
-  OPT=-DROCKSDB_LITE make -j$(nproc) static_lib
+  make LITE=1 -j$(nproc) static_lib
   send_size_to_ods static_lib_lite $(stat --printf="%s" librocksdb.a)
   strip librocksdb.a
   send_size_to_ods static_lib_lite_stripped $(stat --printf="%s" librocksdb.a)
 
-  OPT=-DROCKSDB_LITE make -j$(nproc) shared_lib
+  make LITE=1 -j$(nproc) shared_lib
   send_size_to_ods shared_lib_lite $(stat --printf="%s" `readlink -f librocksdb.so`)
   strip `readlink -f librocksdb.so`
   send_size_to_ods shared_lib_lite_stripped $(stat --printf="%s" `readlink -f librocksdb.so`)
@@ -781,8 +747,8 @@ case $1 in
   lite)
     echo $LITE_BUILD_COMMANDS
     ;;
-  lite_test)
-    echo $LITE_UNIT_TEST_COMMANDS
+  report_lite_binary_size)
+    echo $REPORT_LITE_BINARY_SIZE_COMMANDS
     ;;
   stress_crash)
     echo $STRESS_CRASH_TEST_COMMANDS

build_tools/run_ci_db_test.ps1

@@ -1,13 +1,16 @@
 # This script enables you running RocksDB tests by running
 # All the tests concurrently and utilizing all the cores
 Param(
-  [switch]$EnableJE = $false,  # Look for and use _je executable, append _je to listed exclusions
+  [switch]$EnableJE = $false,  # Look for and use test executable, append _je to listed exclusions
   [switch]$RunAll = $false,    # Will attempt discover all *_test[_je].exe binaries and run all
                                # of them as Google suites. I.e. It will run test cases concurrently
                                # except those mentioned as $Run, those will run as individual test cases
                                # And any execlued with $ExcludeExes or $ExcludeCases
                                # It will also not run any individual test cases
                                # excluded but $ExcludeCasese
+  [switch]$RunAllExe = $false, # Look for and use test exdcutables, append _je to exclusions automatically
+                               # It will attempt to run them in parallel w/o breaking them up on individual
+                               # test cases. Those listed with $ExcludeExes will be excluded
   [string]$SuiteRun = "",      # Split test suites in test cases and run in parallel, not compatible with $RunAll
   [string]$Run = "",           # Run specified executables in parallel but do not split to test cases
   [string]$ExcludeCases = "",  # Exclude test cases, expects a comma separated list, no spaces
@@ -39,13 +42,18 @@ $RunOnly.Add("compact_on_deletion_collector_test") | Out-Null
 $RunOnly.Add("merge_test") | Out-Null
 $RunOnly.Add("stringappend_test") | Out-Null # Apparently incorrectly written
 $RunOnly.Add("backupable_db_test") | Out-Null # Disabled
-
+$RunOnly.Add("timer_queue_test") | Out-Null # Not a gtest
 
 if($RunAll -and $SuiteRun -ne "") {
     Write-Error "$RunAll and $SuiteRun are not compatible"
     exit 1
 }
 
+if($RunAllExe -and $Run -ne "") {
+    Write-Error "$RunAllExe and $Run are not compatible"
+    exit 1
+}
+
 # If running under Appveyor assume that root
 [string]$Appveyor = $Env:APPVEYOR_BUILD_FOLDER
 if($Appveyor -ne "") {
@@ -131,12 +139,8 @@ function ExtractTestCases([string]$GTestExe, $HashTable) {
 
       # Leading whitespace is fine
       $l = $l -replace '^\s+',''
-      # but no whitespace any other place
-      if($l -match "\s+") {
-        continue
-      }
       # Trailing dot is a test group but no whitespace
-      elseif ( $l -match "\.$" ) {
+      if ($l -match "\.$" -and $l -notmatch "\s+") {
         $Group = $l
       }  else {
         # Otherwise it is a test name, remove leading space
@@ -223,13 +227,11 @@ $TestExes = [ordered]@{}
 if($Run -ne "") {
 
   $test_list = $Run -split ' '
-
   ForEach($t in $test_list) {
 
     if($EnableJE) {
       $t += "_je"
     }
-
     MakeAndAdd -token $t -HashTable $TestExes
   }
 
@@ -237,6 +239,38 @@ if($Run -ne "") {
      Write-Error "Failed to extract tests from $Run"
      exit 1
   }
+} elseif($RunAllExe) {
+  # Discover all the test binaries
+  if($EnableJE) {
+    $pattern = "*_test_je.exe"
+  } else {
+    $pattern = "*_test.exe"
+  }
+
+  $search_path = -join ($BinariesFolder, $pattern)
+  Write-Host "Binaries Search Path: $search_path"
+
+  $DiscoveredExe = @()
+  dir -Path $search_path | ForEach-Object {
+     $DiscoveredExe += ($_.Name)     
+  }
+
+  # Remove exclusions
+  ForEach($e in $DiscoveredExe) {
+    $e = $e -replace '.exe$', ''
+    $bare_name = $e -replace '_je$', ''
+
+    if($ExcludeExesSet.Contains($bare_name)) {
+      Write-Warning "Test $e is excluded"
+      continue
+    }
+    MakeAndAdd -token $e -HashTable $TestExes
+  }
+
+  if($TestExes.Count -lt 1) {
+     Write-Error "Failed to discover test executables"
+     exit 1
+  }
 }
 
 # Ordered by exe @{ Exe = @{ TestCase = LogName }}
@@ -245,9 +279,7 @@ $CasesToRun = [ordered]@{}
 if($SuiteRun -ne "") {
   $suite_list = $SuiteRun -split ' '
   ProcessSuites -ListOfSuites $suite_list -HashOfHashes $CasesToRun
-}
-
-if($RunAll) {
+} elseif ($RunAll) {
 # Discover all the test binaries
   if($EnableJE) {
     $pattern = "*_test_je.exe"
@@ -255,7 +287,6 @@ if($RunAll) {
     $pattern = "*_test.exe"
   }
 
-
   $search_path = -join ($BinariesFolder, $pattern)
   Write-Host "Binaries Search Path: $search_path"
 
@@ -287,8 +318,6 @@ if($RunAll) {
 }
 
 
-Write-Host "Attempting to start: $NumTestsToStart tests"
-
 # Invoke a test with a filter and redirect all output
 $InvokeTestCase = {
     param($exe, $test, $log);
@@ -307,7 +336,7 @@ $InvokeTestAsync = {
 # Test limiting factor here
 [int]$count = 0
 # Overall status
-[bool]$success = $true;
+[bool]$script:success = $true;
 
 function RunJobs($Suites, $TestCmds, [int]$ConcurrencyVal)
 {
@@ -365,6 +394,7 @@ function RunJobs($Suites, $TestCmds, [int]$ConcurrencyVal)
                  break
                }
 
+              Write-Host "Starting $exe_name"
               [string]$Exe =  -Join ($BinariesFolder, $exe_name)
               $job = Start-Job -Name $exe_name -ScriptBlock $InvokeTestAsync -ArgumentList @($Exe,$log_path)
               $JobToLog.Add($job, $log_path)
@@ -395,7 +425,7 @@ function RunJobs($Suites, $TestCmds, [int]$ConcurrencyVal)
         $log_content = @(Get-Content $log)
 
         if($completed.State -ne "Completed") {
-            $success = $false
+            $script:success = $false
             Write-Warning $message
             $log_content | Write-Warning
         } else {
@@ -419,7 +449,7 @@ function RunJobs($Suites, $TestCmds, [int]$ConcurrencyVal)
             }
 
             if(!$pass_found) {
-                $success = $false;
+                $script:success = $false;
                 Write-Warning $message
                 $log_content | Write-Warning
             } else {
@@ -443,7 +473,7 @@ New-TimeSpan -Start $StartDate -End $EndDate |
   }
 
 
-if(!$success) {
+if(!$script:success) {
 # This does not succeed killing off jobs quick
 # So we simply exit
 #    Remove-Job -Job $jobs -Force

build_tools/setup_centos7.sh

@@ -0,0 +1,43 @@
+#!/bin/bash
+set -e
+
+ROCKSDB_VERSION="5.10.3"
+ZSTD_VERSION="1.1.3"
+
+echo "This script configures CentOS with everything needed to build and run RocksDB"
+
+yum update -y && yum install epel-release -y
+
+yum install -y \
+  wget \
+  gcc-c++ \
+  snappy snappy-devel \
+  zlib zlib-devel \
+  bzip2 bzip2-devel \
+  lz4-devel \
+  libasan \
+  gflags
+
+mkdir -pv /usr/local/rocksdb-${ROCKSDB_VERSION}
+ln -sfT /usr/local/rocksdb-${ROCKSDB_VERSION} /usr/local/rocksdb
+
+wget -qO /tmp/zstd-${ZSTD_VERSION}.tar.gz https://github.com/facebook/zstd/archive/v${ZSTD_VERSION}.tar.gz
+wget -qO /tmp/rocksdb-${ROCKSDB_VERSION}.tar.gz https://github.com/facebook/rocksdb/archive/v${ROCKSDB_VERSION}.tar.gz
+
+cd /tmp
+
+tar xzvf zstd-${ZSTD_VERSION}.tar.gz
+tar xzvf rocksdb-${ROCKSDB_VERSION}.tar.gz -C /usr/local/
+
+echo "Installing ZSTD..."
+pushd zstd-${ZSTD_VERSION}
+make && make install
+popd
+
+echo "Compiling RocksDB..."
+cd /usr/local/rocksdb
+chown -R vagrant:vagrant /usr/local/rocksdb/
+sudo -u vagrant make static_lib
+cd examples/
+sudo -u vagrant make all
+sudo -u vagrant ./c_simple_example

build_tools/update_dependencies.sh

@@ -38,6 +38,7 @@ function get_lib_base()
     # platform is not provided, use latest gcc
     result=`ls -dr1v $result/gcc-*[^fb]/ | head -n1`
   else
+    echo $lib_platform
     result="$result/$lib_platform/"
   fi
   
@@ -103,29 +104,29 @@ touch "$OUTPUT"
 echo "Writing dependencies to $OUTPUT"
 
 # Compilers locations
-GCC_BASE=`readlink -f $TP2_LATEST/gcc/5.x/centos6-native/*/`
-CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos6-native/*/`
+GCC_BASE=`readlink -f $TP2_LATEST/gcc/5.x/centos7-native/*/`
+CLANG_BASE=`readlink -f $TP2_LATEST/llvm-fb/stable/centos7-native/*/`
 
 log_variable GCC_BASE
 log_variable CLANG_BASE
 
 # Libraries locations
-get_lib_base libgcc     5.x
-get_lib_base glibc      2.23
-get_lib_base snappy     LATEST gcc-5-glibc-2.23
-get_lib_base zlib       LATEST
-get_lib_base bzip2      LATEST
-get_lib_base lz4        LATEST
-get_lib_base zstd       LATEST
-get_lib_base gflags     LATEST
-get_lib_base jemalloc   LATEST
-get_lib_base numa       LATEST
-get_lib_base libunwind  LATEST
-get_lib_base tbb        4.0_update2 gcc-5-glibc-2.23
+get_lib_base libgcc     5.x     gcc-5-glibc-2.23
+get_lib_base glibc      2.23    gcc-5-glibc-2.23
+get_lib_base snappy     LATEST  gcc-5-glibc-2.23
+get_lib_base zlib       LATEST  gcc-5-glibc-2.23
+get_lib_base bzip2      LATEST  gcc-5-glibc-2.23
+get_lib_base lz4        LATEST  gcc-5-glibc-2.23
+get_lib_base zstd       LATEST  gcc-5-glibc-2.23
+get_lib_base gflags     LATEST  gcc-5-glibc-2.23
+get_lib_base jemalloc   LATEST  gcc-5-glibc-2.23
+get_lib_base numa       LATEST  gcc-5-glibc-2.23
+get_lib_base libunwind  LATEST  gcc-5-glibc-2.23
+get_lib_base tbb        LATEST  gcc-5-glibc-2.23
 
-get_lib_base kernel-headers LATEST 
-get_lib_base binutils   LATEST centos6-native 
-get_lib_base valgrind   3.10.0 gcc-5-glibc-2.23
+get_lib_base kernel-headers 4.0.9-36_fbk5_2933_gd092e3f gcc-5-glibc-2.23
+get_lib_base binutils   LATEST centos7-native
+get_lib_base valgrind   LATEST gcc-5-glibc-2.23
 get_lib_base lua        5.2.3 gcc-5-glibc-2.23
 
 git diff $OUTPUT

build_tools/version.sh

@@ -1,4 +1,4 @@
-#!/bin/sh
+#!/usr/bin/env bash
 if [ "$#" = "0" ]; then
   echo "Usage: $0 major|minor|patch|full"
   exit 1

cache/cache_bench.cc

@@ -17,16 +17,16 @@ int main() {
 #include <inttypes.h>
 #include <sys/types.h>
 #include <stdio.h>
-#include <gflags/gflags.h>
 
-#include "rocksdb/db.h"
-#include "rocksdb/cache.h"
-#include "rocksdb/env.h"
 #include "port/port.h"
+#include "rocksdb/cache.h"
+#include "rocksdb/db.h"
+#include "rocksdb/env.h"
+#include "util/gflags_compat.h"
 #include "util/mutexlock.h"
 #include "util/random.h"
 
-using GFLAGS::ParseCommandLineFlags;
+using GFLAGS_NAMESPACE::ParseCommandLineFlags;
 
 static const uint32_t KB = 1024;
 
@@ -52,7 +52,7 @@ namespace rocksdb {
 
 class CacheBench;
 namespace {
-void deleter(const Slice& key, void* value) {
+void deleter(const Slice& /*key*/, void* value) {
     delete reinterpret_cast<char *>(value);
 }
 

cache/cache_test.cc

@@ -40,9 +40,9 @@ static int DecodeValue(void* v) {
 const std::string kLRU = "lru";
 const std::string kClock = "clock";
 
-void dumbDeleter(const Slice& key, void* value) {}
+void dumbDeleter(const Slice& /*key*/, void* /*value*/) {}
 
-void eraseDeleter(const Slice& key, void* value) {
+void eraseDeleter(const Slice& /*key*/, void* value) {
   Cache* cache = reinterpret_cast<Cache*>(value);
   cache->Erase("foo");
 }
@@ -64,8 +64,8 @@ class CacheTest : public testing::TestWithParam<std::string> {
 
   std::vector<int> deleted_keys_;
   std::vector<int> deleted_values_;
-  shared_ptr<Cache> cache_;
-  shared_ptr<Cache> cache2_;
+  std::shared_ptr<Cache> cache_;
+  std::shared_ptr<Cache> cache2_;
 
   CacheTest()
       : cache_(NewCache(kCacheSize, kNumShardBits, false)),
@@ -145,7 +145,7 @@ class CacheTest : public testing::TestWithParam<std::string> {
 CacheTest* CacheTest::current_;
 
 TEST_P(CacheTest, UsageTest) {
-  // cache is shared_ptr and will be automatically cleaned up.
+  // cache is std::shared_ptr and will be automatically cleaned up.
   const uint64_t kCapacity = 100000;
   auto cache = NewCache(kCapacity, 8, false);
 
@@ -173,7 +173,7 @@ TEST_P(CacheTest, UsageTest) {
 }
 
 TEST_P(CacheTest, PinnedUsageTest) {
-  // cache is shared_ptr and will be automatically cleaned up.
+  // cache is std::shared_ptr and will be automatically cleaned up.
   const uint64_t kCapacity = 100000;
   auto cache = NewCache(kCapacity, 8, false);
 
@@ -470,7 +470,7 @@ class Value {
 };
 
 namespace {
-void deleter(const Slice& key, void* value) {
+void deleter(const Slice& /*key*/, void* value) {
   delete static_cast<Value *>(value);
 }
 }  // namespace

cache/clock_cache.cc

@@ -13,8 +13,8 @@
 
 namespace rocksdb {
 
-std::shared_ptr<Cache> NewClockCache(size_t capacity, int num_shard_bits,
-                                     bool strict_capacity_limit) {
+std::shared_ptr<Cache> NewClockCache(size_t /*capacity*/, int /*num_shard_bits*/,
+                                     bool /*strict_capacity_limit*/) {
   // Clock cache not supported.
   return nullptr;
 }
@@ -27,6 +27,11 @@ std::shared_ptr<Cache> NewClockCache(size_t capacity, int num_shard_bits,
 #include <atomic>
 #include <deque>
 
+// "tbb/concurrent_hash_map.h" requires RTTI if exception is enabled.
+// Disable it so users can chooose to disable RTTI.
+#ifndef ROCKSDB_USE_RTTI
+#define TBB_USE_EXCEPTIONS 0
+#endif
 #include "tbb/concurrent_hash_map.h"
 
 #include "cache/sharded_cache.h"
@@ -362,7 +367,9 @@ ClockCacheShard::~ClockCacheShard() {
   for (auto& handle : list_) {
     uint32_t flags = handle.flags.load(std::memory_order_relaxed);
     if (InCache(flags) || CountRefs(flags) > 0) {
-      (*handle.deleter)(handle.key, handle.value);
+      if (handle.deleter != nullptr) {
+        (*handle.deleter)(handle.key, handle.value);
+      }
       delete[] handle.key.data();
     }
   }
@@ -581,7 +588,7 @@ Status ClockCacheShard::Insert(const Slice& key, uint32_t hash, void* value,
                                size_t charge,
                                void (*deleter)(const Slice& key, void* value),
                                Cache::Handle** out_handle,
-                               Cache::Priority priority) {
+                               Cache::Priority /*priority*/) {
   CleanupContext context;
   HashTable::accessor accessor;
   char* key_data = new char[key.size()];

cache/lru_cache.cc

@@ -22,7 +22,7 @@
 
 namespace rocksdb {
 
-LRUHandleTable::LRUHandleTable() : length_(0), elems_(0), list_(nullptr) {
+LRUHandleTable::LRUHandleTable() : list_(nullptr), length_(0), elems_(0) {
   Resize();
 }
 
@@ -99,12 +99,20 @@ void LRUHandleTable::Resize() {
   length_ = new_length;
 }
 
-LRUCacheShard::LRUCacheShard()
-    : usage_(0), lru_usage_(0), high_pri_pool_usage_(0) {
+LRUCacheShard::LRUCacheShard(size_t capacity, bool strict_capacity_limit,
+                             double high_pri_pool_ratio)
+    : capacity_(0),
+      high_pri_pool_usage_(0),
+      strict_capacity_limit_(strict_capacity_limit),
+      high_pri_pool_ratio_(high_pri_pool_ratio),
+      high_pri_pool_capacity_(0),
+      usage_(0),
+      lru_usage_(0) {
   // Make empty circular linked list
   lru_.next = &lru_;
   lru_.prev = &lru_;
   lru_low_pri_ = &lru_;
+  SetCapacity(capacity);
 }
 
 LRUCacheShard::~LRUCacheShard() {}
@@ -157,6 +165,21 @@ void LRUCacheShard::TEST_GetLRUList(LRUHandle** lru, LRUHandle** lru_low_pri) {
   *lru_low_pri = lru_low_pri_;
 }
 
+size_t LRUCacheShard::TEST_GetLRUSize() {
+  LRUHandle* lru_handle = lru_.next;
+  size_t lru_size = 0;
+  while (lru_handle != &lru_) {
+    lru_size++;
+    lru_handle = lru_handle->next;
+  }
+  return lru_size;
+}
+
+double LRUCacheShard::GetHighPriPoolRatio() {
+  MutexLock l(&mutex_);
+  return high_pri_pool_ratio_;
+}
+
 void LRUCacheShard::LRU_Remove(LRUHandle* e) {
   assert(e->next != nullptr);
   assert(e->prev != nullptr);
@@ -176,7 +199,7 @@ void LRUCacheShard::LRU_Remove(LRUHandle* e) {
 void LRUCacheShard::LRU_Insert(LRUHandle* e) {
   assert(e->next == nullptr);
   assert(e->prev == nullptr);
-  if (high_pri_pool_ratio_ > 0 && e->IsHighPri()) {
+  if (high_pri_pool_ratio_ > 0 && (e->IsHighPri() || e->HasHit())) {
     // Inset "e" to head of LRU list.
     e->next = &lru_;
     e->prev = lru_.prev;
@@ -252,6 +275,7 @@ Cache::Handle* LRUCacheShard::Lookup(const Slice& key, uint32_t hash) {
       LRU_Remove(e);
     }
     e->refs++;
+    e->SetHit();
   }
   return reinterpret_cast<Cache::Handle*>(e);
 }
@@ -327,6 +351,7 @@ Status LRUCacheShard::Insert(const Slice& key, uint32_t hash, void* value,
   e->deleter = deleter;
   e->charge = charge;
   e->key_length = key.size();
+  e->flags = 0;
   e->hash = hash;
   e->refs = (handle == nullptr
                  ? 1
@@ -436,18 +461,29 @@ std::string LRUCacheShard::GetPrintableOptions() const {
 }
 
 LRUCache::LRUCache(size_t capacity, int num_shard_bits,
-                   bool strict_capacity_limit, double high_pri_pool_ratio)
-    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit) {
-  int num_shards = 1 << num_shard_bits;
-  shards_ = new LRUCacheShard[num_shards];
-  SetCapacity(capacity);
-  SetStrictCapacityLimit(strict_capacity_limit);
-  for (int i = 0; i < num_shards; i++) {
-    shards_[i].SetHighPriorityPoolRatio(high_pri_pool_ratio);
+                   bool strict_capacity_limit, double high_pri_pool_ratio,
+                   std::shared_ptr<MemoryAllocator> allocator)
+    : ShardedCache(capacity, num_shard_bits, strict_capacity_limit,
+                   std::move(allocator)) {
+  num_shards_ = 1 << num_shard_bits;
+  shards_ = reinterpret_cast<LRUCacheShard*>(
+      port::cacheline_aligned_alloc(sizeof(LRUCacheShard) * num_shards_));
+  size_t per_shard = (capacity + (num_shards_ - 1)) / num_shards_;
+  for (int i = 0; i < num_shards_; i++) {
+    new (&shards_[i])
+        LRUCacheShard(per_shard, strict_capacity_limit, high_pri_pool_ratio);
   }
 }
 
-LRUCache::~LRUCache() { delete[] shards_; }
+LRUCache::~LRUCache() {
+  if (shards_ != nullptr) {
+    assert(num_shards_ > 0);
+    for (int i = 0; i < num_shards_; i++) {
+      shards_[i].~LRUCacheShard();
+    }
+    port::cacheline_aligned_free(shards_);
+  }
+}
 
 CacheShard* LRUCache::GetShard(int shard) {
   return reinterpret_cast<CacheShard*>(&shards_[shard]);
@@ -469,11 +505,48 @@ uint32_t LRUCache::GetHash(Handle* handle) const {
   return reinterpret_cast<const LRUHandle*>(handle)->hash;
 }
 
-void LRUCache::DisownData() { shards_ = nullptr; }
+void LRUCache::DisownData() {
+// Do not drop data if compile with ASAN to suppress leak warning.
+#if defined(__clang__)
+#if !defined(__has_feature) || !__has_feature(address_sanitizer)
+  shards_ = nullptr;
+  num_shards_ = 0;
+#endif
+#else  // __clang__
+#ifndef __SANITIZE_ADDRESS__
+  shards_ = nullptr;
+  num_shards_ = 0;
+#endif  // !__SANITIZE_ADDRESS__
+#endif  // __clang__
+}
 
-std::shared_ptr<Cache> NewLRUCache(size_t capacity, int num_shard_bits,
-                                   bool strict_capacity_limit,
-                                   double high_pri_pool_ratio) {
+size_t LRUCache::TEST_GetLRUSize() {
+  size_t lru_size_of_all_shards = 0;
+  for (int i = 0; i < num_shards_; i++) {
+    lru_size_of_all_shards += shards_[i].TEST_GetLRUSize();
+  }
+  return lru_size_of_all_shards;
+}
+
+double LRUCache::GetHighPriPoolRatio() {
+  double result = 0.0;
+  if (num_shards_ > 0) {
+    result = shards_[0].GetHighPriPoolRatio();
+  }
+  return result;
+}
+
+std::shared_ptr<Cache> NewLRUCache(const LRUCacheOptions& cache_opts) {
+  return NewLRUCache(cache_opts.capacity, cache_opts.num_shard_bits,
+                     cache_opts.strict_capacity_limit,
+                     cache_opts.high_pri_pool_ratio,
+                     cache_opts.memory_allocator);
+}
+
+std::shared_ptr<Cache> NewLRUCache(
+    size_t capacity, int num_shard_bits, bool strict_capacity_limit,
+    double high_pri_pool_ratio,
+    std::shared_ptr<MemoryAllocator> memory_allocator) {
   if (num_shard_bits >= 20) {
     return nullptr;  // the cache cannot be sharded into too many fine pieces
   }
@@ -485,7 +558,8 @@ std::shared_ptr<Cache> NewLRUCache(size_t capacity, int num_shard_bits,
     num_shard_bits = GetDefaultCacheShardBits(capacity);
   }
   return std::make_shared<LRUCache>(capacity, num_shard_bits,
-                                    strict_capacity_limit, high_pri_pool_ratio);
+                                    strict_capacity_limit, high_pri_pool_ratio,
+                                    std::move(memory_allocator));
 }
 
 }  // namespace rocksdb

cache/lru_cache.h

@@ -57,7 +57,7 @@ struct LRUHandle {
   // Include the following flags:
   //   in_cache:    whether this entry is referenced by the hash table.
   //   is_high_pri: whether this entry is high priority entry.
-  //   in_high_pro_pool: whether this entry is in high-pri pool.
+  //   in_high_pri_pool: whether this entry is in high-pri pool.
   char flags;
 
   uint32_t hash;     // Hash of key(); used for fast sharding and comparisons
@@ -77,6 +77,7 @@ struct LRUHandle {
   bool InCache() { return flags & 1; }
   bool IsHighPri() { return flags & 2; }
   bool InHighPriPool() { return flags & 4; }
+  bool HasHit() { return flags & 8; }
 
   void SetInCache(bool in_cache) {
     if (in_cache) {
@@ -102,6 +103,8 @@ struct LRUHandle {
     }
   }
 
+  void SetHit() { flags |= 8; }
+
   void Free() {
     assert((refs == 1 && InCache()) || (refs == 0 && !InCache()));
     if (deleter) {
@@ -148,15 +151,16 @@ class LRUHandleTable {
 
   // The table consists of an array of buckets where each bucket is
   // a linked list of cache entries that hash into the bucket.
+  LRUHandle** list_;
   uint32_t length_;
   uint32_t elems_;
-  LRUHandle** list_;
 };
 
 // A single shard of sharded cache.
-class LRUCacheShard : public CacheShard {
+class ALIGN_AS(CACHE_LINE_SIZE) LRUCacheShard : public CacheShard {
  public:
-  LRUCacheShard();
+  LRUCacheShard(size_t capacity, bool strict_capacity_limit,
+                double high_pri_pool_ratio);
   virtual ~LRUCacheShard();
 
   // Separate from constructor so caller can easily make an array of LRUCache
@@ -198,6 +202,13 @@ class LRUCacheShard : public CacheShard {
 
   void TEST_GetLRUList(LRUHandle** lru, LRUHandle** lru_low_pri);
 
+  //  Retrieves number of elements in LRU, for unit test purpose only
+  //  not threadsafe
+  size_t TEST_GetLRUSize();
+
+  //  Retrives high pri pool ratio
+  double GetHighPriPoolRatio();
+
  private:
   void LRU_Remove(LRUHandle* e);
   void LRU_Insert(LRUHandle* e);
@@ -219,12 +230,6 @@ class LRUCacheShard : public CacheShard {
   // Initialized before use.
   size_t capacity_;
 
-  // Memory size for entries residing in the cache
-  size_t usage_;
-
-  // Memory size for entries residing only in the LRU list
-  size_t lru_usage_;
-
   // Memory size for entries in high-pri pool.
   size_t high_pri_pool_usage_;
 
@@ -238,11 +243,6 @@ class LRUCacheShard : public CacheShard {
   // Remember the value to avoid recomputing each time.
   double high_pri_pool_capacity_;
 
-  // mutex_ protects the following state.
-  // We don't count mutex_ as the cache's internal state so semantically we
-  // don't mind mutex_ invoking the non-const actions.
-  mutable port::Mutex mutex_;
-
   // Dummy head of LRU list.
   // lru.prev is newest entry, lru.next is oldest entry.
   // LRU contains items which can be evicted, ie reference only by cache
@@ -251,13 +251,36 @@ class LRUCacheShard : public CacheShard {
   // Pointer to head of low-pri pool in LRU list.
   LRUHandle* lru_low_pri_;
 
+  // ------------^^^^^^^^^^^^^-----------
+  // Not frequently modified data members
+  // ------------------------------------
+  //
+  // We separate data members that are updated frequently from the ones that
+  // are not frequently updated so that they don't share the same cache line
+  // which will lead into false cache sharing
+  //
+  // ------------------------------------
+  // Frequently modified data members
+  // ------------vvvvvvvvvvvvv-----------
   LRUHandleTable table_;
+
+  // Memory size for entries residing in the cache
+  size_t usage_;
+
+  // Memory size for entries residing only in the LRU list
+  size_t lru_usage_;
+
+  // mutex_ protects the following state.
+  // We don't count mutex_ as the cache's internal state so semantically we
+  // don't mind mutex_ invoking the non-const actions.
+  mutable port::Mutex mutex_;
 };
 
 class LRUCache : public ShardedCache {
  public:
   LRUCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit,
-           double high_pri_pool_ratio);
+           double high_pri_pool_ratio,
+           std::shared_ptr<MemoryAllocator> memory_allocator = nullptr);
   virtual ~LRUCache();
   virtual const char* Name() const override { return "LRUCache"; }
   virtual CacheShard* GetShard(int shard) override;
@@ -267,8 +290,14 @@ class LRUCache : public ShardedCache {
   virtual uint32_t GetHash(Handle* handle) const override;
   virtual void DisownData() override;
 
+  //  Retrieves number of elements in LRU, for unit test purpose only
+  size_t TEST_GetLRUSize();
+  //  Retrives high pri pool ratio
+  double GetHighPriPoolRatio();
+
  private:
-  LRUCacheShard* shards_;
+  LRUCacheShard* shards_ = nullptr;
+  int num_shards_ = 0;
 };
 
 }  // namespace rocksdb

cache/lru_cache_test.cc

@@ -7,6 +7,7 @@
 
 #include <string>
 #include <vector>
+#include "port/port.h"
 #include "util/testharness.h"
 
 namespace rocksdb {
@@ -14,13 +15,22 @@ namespace rocksdb {
 class LRUCacheTest : public testing::Test {
  public:
   LRUCacheTest() {}
-  ~LRUCacheTest() {}
+  ~LRUCacheTest() { DeleteCache(); }
+
+  void DeleteCache() {
+    if (cache_ != nullptr) {
+      cache_->~LRUCacheShard();
+      port::cacheline_aligned_free(cache_);
+      cache_ = nullptr;
+    }
+  }
 
   void NewCache(size_t capacity, double high_pri_pool_ratio = 0.0) {
-    cache_.reset(new LRUCacheShard());
-    cache_->SetCapacity(capacity);
-    cache_->SetStrictCapacityLimit(false);
-    cache_->SetHighPriorityPoolRatio(high_pri_pool_ratio);
+    DeleteCache();
+    cache_ = reinterpret_cast<LRUCacheShard*>(
+        port::cacheline_aligned_alloc(sizeof(LRUCacheShard)));
+    new (cache_) LRUCacheShard(capacity, false /*strict_capcity_limit*/,
+                               high_pri_pool_ratio);
   }
 
   void Insert(const std::string& key,
@@ -76,7 +86,7 @@ class LRUCacheTest : public testing::Test {
   }
 
  private:
-  std::unique_ptr<LRUCacheShard> cache_;
+  LRUCacheShard* cache_ = nullptr;
 };
 
 TEST_F(LRUCacheTest, BasicLRU) {
@@ -105,7 +115,30 @@ TEST_F(LRUCacheTest, BasicLRU) {
   ValidateLRUList({"e", "z", "d", "u", "v"});
 }
 
-TEST_F(LRUCacheTest, MidPointInsertion) {
+TEST_F(LRUCacheTest, MidpointInsertion) {
+  // Allocate 2 cache entries to high-pri pool.
+  NewCache(5, 0.45);
+
+  Insert("a", Cache::Priority::LOW);
+  Insert("b", Cache::Priority::LOW);
+  Insert("c", Cache::Priority::LOW);
+  Insert("x", Cache::Priority::HIGH);
+  Insert("y", Cache::Priority::HIGH);
+  ValidateLRUList({"a", "b", "c", "x", "y"}, 2);
+
+  // Low-pri entries inserted to the tail of low-pri list (the midpoint).
+  // After lookup, it will move to the tail of the full list.
+  Insert("d", Cache::Priority::LOW);
+  ValidateLRUList({"b", "c", "d", "x", "y"}, 2);
+  ASSERT_TRUE(Lookup("d"));
+  ValidateLRUList({"b", "c", "x", "y", "d"}, 2);
+
+  // High-pri entries will be inserted to the tail of full list.
+  Insert("z", Cache::Priority::HIGH);
+  ValidateLRUList({"c", "x", "y", "d", "z"}, 2);
+}
+
+TEST_F(LRUCacheTest, EntriesWithPriority) {
   // Allocate 2 cache entries to high-pri pool.
   NewCache(5, 0.45);
 
@@ -131,15 +164,15 @@ TEST_F(LRUCacheTest, MidPointInsertion) {
   Insert("a", Cache::Priority::LOW);
   ValidateLRUList({"v", "X", "a", "Y", "Z"}, 2);
 
-  // Low-pri entries will be inserted to head of low-pri pool after lookup.
+  // Low-pri entries will be inserted to head of high-pri pool after lookup.
   ASSERT_TRUE(Lookup("v"));
-  ValidateLRUList({"X", "a", "v", "Y", "Z"}, 2);
+  ValidateLRUList({"X", "a", "Y", "Z", "v"}, 2);
 
   // High-pri entries will be inserted to the head of the list after lookup.
   ASSERT_TRUE(Lookup("X"));
-  ValidateLRUList({"a", "v", "Y", "Z", "X"}, 2);
+  ValidateLRUList({"a", "Y", "Z", "v", "X"}, 2);
   ASSERT_TRUE(Lookup("Z"));
-  ValidateLRUList({"a", "v", "Y", "X", "Z"}, 2);
+  ValidateLRUList({"a", "Y", "v", "X", "Z"}, 2);
 
   Erase("Y");
   ValidateLRUList({"a", "v", "X", "Z"}, 2);
@@ -152,7 +185,7 @@ TEST_F(LRUCacheTest, MidPointInsertion) {
   Insert("g", Cache::Priority::LOW);
   ValidateLRUList({"d", "e", "f", "g", "Z"}, 1);
   ASSERT_TRUE(Lookup("d"));
-  ValidateLRUList({"e", "f", "g", "d", "Z"}, 1);
+  ValidateLRUList({"e", "f", "g", "Z", "d"}, 2);
 }
 
 }  // namespace rocksdb

cache/sharded_cache.cc

@@ -20,8 +20,10 @@
 namespace rocksdb {
 
 ShardedCache::ShardedCache(size_t capacity, int num_shard_bits,
-                           bool strict_capacity_limit)
-    : num_shard_bits_(num_shard_bits),
+                           bool strict_capacity_limit,
+                           std::shared_ptr<MemoryAllocator> allocator)
+    : Cache(std::move(allocator)),
+      num_shard_bits_(num_shard_bits),
       capacity_(capacity),
       strict_capacity_limit_(strict_capacity_limit),
       last_id_(1) {}
@@ -53,7 +55,7 @@ Status ShardedCache::Insert(const Slice& key, void* value, size_t charge,
       ->Insert(key, hash, value, charge, deleter, handle, priority);
 }
 
-Cache::Handle* ShardedCache::Lookup(const Slice& key, Statistics* stats) {
+Cache::Handle* ShardedCache::Lookup(const Slice& key, Statistics* /*stats*/) {
   uint32_t hash = HashSlice(key);
   return GetShard(Shard(hash))->Lookup(key, hash);
 }
@@ -142,6 +144,9 @@ std::string ShardedCache::GetPrintableOptions() const {
              strict_capacity_limit_);
     ret.append(buffer);
   }
+  snprintf(buffer, kBufferSize, "    memory_allocator : %s\n",
+           memory_allocator() ? memory_allocator()->Name() : "None");
+  ret.append(buffer);
   ret.append(GetShard(0)->GetPrintableOptions());
   return ret;
 }

cache/sharded_cache.h

@@ -47,7 +47,8 @@ class CacheShard {
 // Keys are sharded by the highest num_shard_bits bits of hash value.
 class ShardedCache : public Cache {
  public:
-  ShardedCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit);
+  ShardedCache(size_t capacity, int num_shard_bits, bool strict_capacity_limit,
+               std::shared_ptr<MemoryAllocator> memory_allocator = nullptr);
   virtual ~ShardedCache() = default;
   virtual const char* Name() const override = 0;
   virtual CacheShard* GetShard(int shard) = 0;

cmake/RocksDBConfig.cmake.in

@@ -0,0 +1,3 @@
+@PACKAGE_INIT@
+include("${CMAKE_CURRENT_LIST_DIR}/RocksDBTargets.cmake")
+check_required_components(RocksDB)

cmake/modules/FindNUMA.cmake

@@ -0,0 +1,21 @@
+# - Find NUMA
+# Find the NUMA library and includes
+#
+# NUMA_INCLUDE_DIR - where to find numa.h, etc.
+# NUMA_LIBRARIES - List of libraries when using NUMA.
+# NUMA_FOUND - True if NUMA found.
+
+find_path(NUMA_INCLUDE_DIR
+  NAMES numa.h numaif.h
+  HINTS ${NUMA_ROOT_DIR}/include)
+
+find_library(NUMA_LIBRARIES
+  NAMES numa
+  HINTS ${NUMA_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(NUMA DEFAULT_MSG NUMA_LIBRARIES NUMA_INCLUDE_DIR)
+
+mark_as_advanced(
+  NUMA_LIBRARIES
+  NUMA_INCLUDE_DIR)

cmake/modules/FindTBB.cmake

@@ -0,0 +1,25 @@
+# - Find TBB
+# Find the Thread Building Blocks library and includes
+#
+# TBB_INCLUDE_DIR - where to find tbb.h, etc.
+# TBB_LIBRARIES - List of libraries when using TBB.
+# TBB_FOUND - True if TBB found.
+
+if(NOT DEFINED TBB_ROOT_DIR)
+  set(TBB_ROOT_DIR "$ENV{TBBROOT}")
+endif()
+
+find_path(TBB_INCLUDE_DIR
+NAMES tbb/tbb.h
+HINTS ${TBB_ROOT_DIR}/include)
+
+find_library(TBB_LIBRARIES
+NAMES tbb
+HINTS ${TBB_ROOT_DIR}/lib ENV LIBRARY_PATH)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(TBB DEFAULT_MSG TBB_LIBRARIES TBB_INCLUDE_DIR)
+
+mark_as_advanced(
+TBB_LIBRARIES
+TBB_INCLUDE_DIR)

cmake/modules/Findbzip2.cmake

@@ -0,0 +1,21 @@
+# - Find Bzip2
+# Find the bzip2 compression library and includes
+#
+# BZIP2_INCLUDE_DIR - where to find bzlib.h, etc.
+# BZIP2_LIBRARIES - List of libraries when using bzip2.
+# BZIP2_FOUND - True if bzip2 found.
+
+find_path(BZIP2_INCLUDE_DIR
+  NAMES bzlib.h
+  HINTS ${BZIP2_ROOT_DIR}/include)
+
+find_library(BZIP2_LIBRARIES
+  NAMES bz2
+  HINTS ${BZIP2_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(bzip2 DEFAULT_MSG BZIP2_LIBRARIES BZIP2_INCLUDE_DIR)
+
+mark_as_advanced(
+  BZIP2_LIBRARIES
+  BZIP2_INCLUDE_DIR)

cmake/modules/Findlz4.cmake

@@ -0,0 +1,21 @@
+# - Find Lz4
+# Find the lz4 compression library and includes
+#
+# LZ4_INCLUDE_DIR - where to find lz4.h, etc.
+# LZ4_LIBRARIES - List of libraries when using lz4.
+# LZ4_FOUND - True if lz4 found.
+
+find_path(LZ4_INCLUDE_DIR
+  NAMES lz4.h
+  HINTS ${LZ4_ROOT_DIR}/include)
+
+find_library(LZ4_LIBRARIES
+  NAMES lz4
+  HINTS ${LZ4_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(lz4 DEFAULT_MSG LZ4_LIBRARIES LZ4_INCLUDE_DIR)
+
+mark_as_advanced(
+  LZ4_LIBRARIES
+  LZ4_INCLUDE_DIR)

cmake/modules/Findzstd.cmake

@@ -0,0 +1,21 @@
+# - Find zstd
+# Find the zstd compression library and includes
+#
+# ZSTD_INCLUDE_DIR - where to find zstd.h, etc.
+# ZSTD_LIBRARIES - List of libraries when using zstd.
+# ZSTD_FOUND - True if zstd found.
+
+find_path(ZSTD_INCLUDE_DIR
+  NAMES zstd.h
+  HINTS ${ZSTD_ROOT_DIR}/include)
+
+find_library(ZSTD_LIBRARIES
+  NAMES zstd
+  HINTS ${ZSTD_ROOT_DIR}/lib)
+
+include(FindPackageHandleStandardArgs)
+find_package_handle_standard_args(zstd DEFAULT_MSG ZSTD_LIBRARIES ZSTD_INCLUDE_DIR)
+
+mark_as_advanced(
+  ZSTD_LIBRARIES
+  ZSTD_INCLUDE_DIR)

coverage/coverage_test.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Exit on error.
 set -e

coverage/parse_gcov_output.py

@@ -1,4 +1,3 @@
-import optparse
 import re
 import sys
 
@@ -72,7 +71,7 @@ def display_file_coverage(per_file_coverage, total_coverage):
     header_template = \
         "%" + str(max_file_name_length) + "s\t%s\t%s"
     separator = "-" * (max_file_name_length + 10 + 20)
-    print header_template % ("Filename", "Coverage", "Lines")
+    print header_template % ("Filename", "Coverage", "Lines")  # noqa: E999 T25377293 Grandfathered in
     print separator
 
     # -- Print body

db/builder.cc

@@ -18,6 +18,7 @@
 #include "db/event_helpers.h"
 #include "db/internal_stats.h"
 #include "db/merge_helper.h"
+#include "db/range_del_aggregator.h"
 #include "db/table_cache.h"
 #include "db/version_edit.h"
 #include "monitoring/iostats_context_imp.h"
@@ -28,6 +29,7 @@
 #include "rocksdb/options.h"
 #include "rocksdb/table.h"
 #include "table/block_based_table_builder.h"
+#include "table/format.h"
 #include "table/internal_iterator.h"
 #include "util/file_reader_writer.h"
 #include "util/filename.h"
@@ -39,7 +41,7 @@ namespace rocksdb {
 class TableFactory;
 
 TableBuilder* NewTableBuilder(
-    const ImmutableCFOptions& ioptions,
+    const ImmutableCFOptions& ioptions, const MutableCFOptions& moptions,
     const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
@@ -47,15 +49,16 @@ TableBuilder* NewTableBuilder(
     WritableFileWriter* file, const CompressionType compression_type,
     const CompressionOptions& compression_opts, int level,
     const std::string* compression_dict, const bool skip_filters,
-    const uint64_t creation_time) {
+    const uint64_t creation_time, const uint64_t oldest_key_time) {
   assert((column_family_id ==
           TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
          column_family_name.empty());
   return ioptions.table_factory->NewTableBuilder(
-      TableBuilderOptions(ioptions, internal_comparator,
+      TableBuilderOptions(ioptions, moptions, internal_comparator,
                           int_tbl_prop_collector_factories, compression_type,
                           compression_opts, compression_dict, skip_filters,
-                          column_family_name, level, creation_time),
+                          column_family_name, level, creation_time,
+                          oldest_key_time),
       column_family_id, file);
 }
 
@@ -63,19 +66,20 @@ Status BuildTable(
     const std::string& dbname, Env* env, const ImmutableCFOptions& ioptions,
     const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
     TableCache* table_cache, InternalIterator* iter,
-    std::unique_ptr<InternalIterator> range_del_iter, FileMetaData* meta,
-    const InternalKeyComparator& internal_comparator,
+    std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
+        range_del_iters,
+    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
     uint32_t column_family_id, const std::string& column_family_name,
     std::vector<SequenceNumber> snapshots,
     SequenceNumber earliest_write_conflict_snapshot,
-    const CompressionType compression,
+    SnapshotChecker* snapshot_checker, const CompressionType compression,
     const CompressionOptions& compression_opts, bool paranoid_file_checks,
     InternalStats* internal_stats, TableFileCreationReason reason,
     EventLogger* event_logger, int job_id, const Env::IOPriority io_priority,
-    TableProperties* table_properties, int level,
-    const uint64_t creation_time) {
+    TableProperties* table_properties, int level, const uint64_t creation_time,
+    const uint64_t oldest_key_time, Env::WriteLifeTimeHint write_hint) {
   assert((column_family_id ==
           TablePropertiesCollectorFactory::Context::kUnknownColumnFamily) ==
          column_family_name.empty());
@@ -84,15 +88,13 @@ Status BuildTable(
   Status s;
   meta->fd.file_size = 0;
   iter->SeekToFirst();
-  std::unique_ptr<RangeDelAggregator> range_del_agg(
-      new RangeDelAggregator(internal_comparator, snapshots));
-  s = range_del_agg->AddTombstones(std::move(range_del_iter));
-  if (!s.ok()) {
-    // may be non-ok if a range tombstone key is unparsable
-    return s;
+  std::unique_ptr<CompactionRangeDelAggregator> range_del_agg(
+      new CompactionRangeDelAggregator(&internal_comparator, snapshots));
+  for (auto& range_del_iter : range_del_iters) {
+    range_del_agg->AddTombstones(std::move(range_del_iter));
   }
 
-  std::string fname = TableFileName(ioptions.db_paths, meta->fd.GetNumber(),
+  std::string fname = TableFileName(ioptions.cf_paths, meta->fd.GetNumber(),
                                     meta->fd.GetPathId());
 #ifndef ROCKSDB_LITE
   EventHelpers::NotifyTableFileCreationStarted(
@@ -100,11 +102,11 @@ Status BuildTable(
 #endif  // !ROCKSDB_LITE
   TableProperties tp;
 
-  if (iter->Valid() || range_del_agg->ShouldAddTombstones()) {
+  if (iter->Valid() || !range_del_agg->IsEmpty()) {
     TableBuilder* builder;
-    unique_ptr<WritableFileWriter> file_writer;
+    std::unique_ptr<WritableFileWriter> file_writer;
     {
-      unique_ptr<WritableFile> file;
+      std::unique_ptr<WritableFile> file;
 #ifndef NDEBUG
       bool use_direct_writes = env_options.use_direct_writes;
       TEST_SYNC_POINT_CALLBACK("BuildTable:create_file", &use_direct_writes);
@@ -117,25 +119,29 @@ Status BuildTable(
         return s;
       }
       file->SetIOPriority(io_priority);
+      file->SetWriteLifeTimeHint(write_hint);
 
-      file_writer.reset(new WritableFileWriter(std::move(file), env_options,
-                                               ioptions.statistics));
-
+      file_writer.reset(
+          new WritableFileWriter(std::move(file), fname, env_options, env,
+                                 ioptions.statistics, ioptions.listeners));
       builder = NewTableBuilder(
-          ioptions, internal_comparator, int_tbl_prop_collector_factories,
-          column_family_id, column_family_name, file_writer.get(), compression,
-          compression_opts, level, nullptr /* compression_dict */,
-          false /* skip_filters */, creation_time);
+          ioptions, mutable_cf_options, internal_comparator,
+          int_tbl_prop_collector_factories, column_family_id,
+          column_family_name, file_writer.get(), compression, compression_opts,
+          level, nullptr /* compression_dict */, false /* skip_filters */,
+          creation_time, oldest_key_time);
     }
 
     MergeHelper merge(env, internal_comparator.user_comparator(),
                       ioptions.merge_operator, nullptr, ioptions.info_log,
                       true /* internal key corruption is not ok */,
-                      snapshots.empty() ? 0 : snapshots.back());
+                      snapshots.empty() ? 0 : snapshots.back(),
+                      snapshot_checker);
 
     CompactionIterator c_iter(
         iter, internal_comparator.user_comparator(), &merge, kMaxSequenceNumber,
-        &snapshots, earliest_write_conflict_snapshot, env,
+        &snapshots, earliest_write_conflict_snapshot, snapshot_checker, env,
+        ShouldReportDetailedTime(env, ioptions.statistics),
         true /* internal key corruption is not ok */, range_del_agg.get());
     c_iter.SeekToFirst();
     for (; c_iter.Valid(); c_iter.Next()) {
@@ -151,12 +157,20 @@ Status BuildTable(
             ThreadStatus::FLUSH_BYTES_WRITTEN, IOSTATS(bytes_written));
       }
     }
-    // nullptr for table_{min,max} so all range tombstones will be flushed
-    range_del_agg->AddToBuilder(builder, nullptr /* lower_bound */,
-                                nullptr /* upper_bound */, meta);
+
+    auto range_del_it = range_del_agg->NewIterator();
+    for (range_del_it->SeekToFirst(); range_del_it->Valid();
+         range_del_it->Next()) {
+      auto tombstone = range_del_it->Tombstone();
+      auto kv = tombstone.Serialize();
+      builder->Add(kv.first.Encode(), kv.second);
+      meta->UpdateBoundariesForRange(kv.first, tombstone.SerializeEndKey(),
+                                     tombstone.seq_, internal_comparator);
+    }
 
     // Finish and check for builder errors
-    bool empty = builder->NumEntries() == 0;
+    tp = builder->GetTableProperties();
+    bool empty = builder->NumEntries() == 0 && tp.num_range_deletions == 0;
     s = c_iter.status();
     if (!s.ok() || empty) {
       builder->Abandon();
@@ -169,7 +183,7 @@ Status BuildTable(
       meta->fd.file_size = file_size;
       meta->marked_for_compaction = builder->NeedCompact();
       assert(meta->fd.GetFileSize() > 0);
-      tp = builder->GetTableProperties();
+      tp = builder->GetTableProperties(); // refresh now that builder is finished
       if (table_properties) {
         *table_properties = tp;
       }
@@ -193,8 +207,9 @@ Status BuildTable(
       // we will regrad this verification as user reads since the goal is
       // to cache it here for further user reads
       std::unique_ptr<InternalIterator> it(table_cache->NewIterator(
-          ReadOptions(), env_options, internal_comparator, meta->fd,
-          nullptr /* range_del_agg */, nullptr,
+          ReadOptions(), env_options, internal_comparator, *meta,
+          nullptr /* range_del_agg */,
+          mutable_cf_options.prefix_extractor.get(), nullptr,
           (internal_stats == nullptr) ? nullptr
                                       : internal_stats->GetFileReadHist(0),
           false /* for_compaction */, nullptr /* arena */,

db/builder.h

@@ -9,6 +9,7 @@
 #include <string>
 #include <utility>
 #include <vector>
+#include "db/range_tombstone_fragmenter.h"
 #include "db/table_properties_collector.h"
 #include "options/cf_options.h"
 #include "rocksdb/comparator.h"
@@ -29,12 +30,12 @@ struct FileMetaData;
 class Env;
 struct EnvOptions;
 class Iterator;
+class SnapshotChecker;
 class TableCache;
 class VersionEdit;
 class TableBuilder;
 class WritableFileWriter;
 class InternalStats;
-class InternalIterator;
 
 // @param column_family_name Name of the column family that is also identified
 //    by column_family_id, or empty string if unknown. It must outlive the
@@ -42,7 +43,7 @@ class InternalIterator;
 // @param compression_dict Data for presetting the compression library's
 //    dictionary, or nullptr.
 TableBuilder* NewTableBuilder(
-    const ImmutableCFOptions& options,
+    const ImmutableCFOptions& options, const MutableCFOptions& moptions,
     const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
@@ -50,7 +51,8 @@ TableBuilder* NewTableBuilder(
     WritableFileWriter* file, const CompressionType compression_type,
     const CompressionOptions& compression_opts, int level,
     const std::string* compression_dict = nullptr,
-    const bool skip_filters = false, const uint64_t creation_time = 0);
+    const bool skip_filters = false, const uint64_t creation_time = 0,
+    const uint64_t oldest_key_time = 0);
 
 // Build a Table file from the contents of *iter.  The generated file
 // will be named according to number specified in meta. On success, the rest of
@@ -64,19 +66,21 @@ extern Status BuildTable(
     const std::string& dbname, Env* env, const ImmutableCFOptions& options,
     const MutableCFOptions& mutable_cf_options, const EnvOptions& env_options,
     TableCache* table_cache, InternalIterator* iter,
-    std::unique_ptr<InternalIterator> range_del_iter, FileMetaData* meta,
-    const InternalKeyComparator& internal_comparator,
+    std::vector<std::unique_ptr<FragmentedRangeTombstoneIterator>>
+        range_del_iters,
+    FileMetaData* meta, const InternalKeyComparator& internal_comparator,
     const std::vector<std::unique_ptr<IntTblPropCollectorFactory>>*
         int_tbl_prop_collector_factories,
     uint32_t column_family_id, const std::string& column_family_name,
     std::vector<SequenceNumber> snapshots,
     SequenceNumber earliest_write_conflict_snapshot,
-    const CompressionType compression,
+    SnapshotChecker* snapshot_checker, const CompressionType compression,
     const CompressionOptions& compression_opts, bool paranoid_file_checks,
     InternalStats* internal_stats, TableFileCreationReason reason,
     EventLogger* event_logger = nullptr, int job_id = 0,
     const Env::IOPriority io_priority = Env::IO_HIGH,
     TableProperties* table_properties = nullptr, int level = -1,
-    const uint64_t creation_time = 0);
+    const uint64_t creation_time = 0, const uint64_t oldest_key_time = 0,
+    Env::WriteLifeTimeHint write_hint = Env::WLTH_NOT_SET);
 
 }  // namespace rocksdb

db/c_test.c

@@ -19,11 +19,8 @@
 
 // Can not use port/port.h macros as this is a c file
 #ifdef OS_WIN
-
 #include <windows.h>
 
-#define snprintf _snprintf
-
 // Ok for uniqueness
 int geteuid() {
   int result = 0;
@@ -34,6 +31,11 @@ int geteuid() {
   return result;
 }
 
+// VS < 2015
+#if defined(_MSC_VER) && (_MSC_VER < 1900)
+#define snprintf _snprintf
+#endif
+
 #endif
 
 const char* phase = "";
@@ -41,17 +43,25 @@ static char dbname[200];
 static char sstfilename[200];
 static char dbbackupname[200];
 static char dbcheckpointname[200];
+static char dbpathname[200];
 
 static void StartPhase(const char* name) {
   fprintf(stderr, "=== Test %s\n", name);
   phase = name;
 }
+#ifdef _MSC_VER
+#pragma warning(push)
+#pragma warning (disable: 4996) // getenv security warning
+#endif
 static const char* GetTempDir(void) {
     const char* ret = getenv("TEST_TMPDIR");
     if (ret == NULL || ret[0] == '\0')
         ret = "/tmp";
     return ret;
 }
+#ifdef _MSC_VER
+#pragma warning(pop)
+#endif
 
 #define CheckNoError(err)                                               \
   if ((err) != NULL) {                                                  \
@@ -191,10 +201,11 @@ static void CheckDel(void* ptr, const char* k, size_t klen) {
   (*state)++;
 }
 
-static void CmpDestroy(void* arg) { }
+static void CmpDestroy(void* arg) { (void)arg; }
 
 static int CmpCompare(void* arg, const char* a, size_t alen,
                       const char* b, size_t blen) {
+  (void)arg;
   size_t n = (alen < blen) ? alen : blen;
   int r = memcmp(a, b, n);
   if (r == 0) {
@@ -205,13 +216,15 @@ static int CmpCompare(void* arg, const char* a, size_t alen,
 }
 
 static const char* CmpName(void* arg) {
+  (void)arg;
   return "foo";
 }
 
 // Custom filter policy
 static unsigned char fake_filter_result = 1;
-static void FilterDestroy(void* arg) { }
+static void FilterDestroy(void* arg) { (void)arg; }
 static const char* FilterName(void* arg) {
+  (void)arg;
   return "TestFilter";
 }
 static char* FilterCreate(
@@ -219,6 +232,10 @@ static char* FilterCreate(
     const char* const* key_array, const size_t* key_length_array,
     int num_keys,
     size_t* filter_length) {
+  (void)arg;
+  (void)key_array;
+  (void)key_length_array;
+  (void)num_keys;
   *filter_length = 4;
   char* result = malloc(4);
   memcpy(result, "fake", 4);
@@ -228,20 +245,30 @@ static unsigned char FilterKeyMatch(
     void* arg,
     const char* key, size_t length,
     const char* filter, size_t filter_length) {
+  (void)arg;
+  (void)key;
+  (void)length;
   CheckCondition(filter_length == 4);
   CheckCondition(memcmp(filter, "fake", 4) == 0);
   return fake_filter_result;
 }
 
 // Custom compaction filter
-static void CFilterDestroy(void* arg) {}
-static const char* CFilterName(void* arg) { return "foo"; }
+static void CFilterDestroy(void* arg) { (void)arg; }
+static const char* CFilterName(void* arg) {
+  (void)arg;
+  return "foo";
+}
 static unsigned char CFilterFilter(void* arg, int level, const char* key,
                                    size_t key_length,
                                    const char* existing_value,
                                    size_t value_length, char** new_value,
                                    size_t* new_value_length,
                                    unsigned char* value_changed) {
+  (void)arg;
+  (void)level;
+  (void)existing_value;
+  (void)value_length;
   if (key_length == 3) {
     if (memcmp(key, "bar", key_length) == 0) {
       return 1;
@@ -255,10 +282,15 @@ static unsigned char CFilterFilter(void* arg, int level, const char* key,
   return 0;
 }
 
-static void CFilterFactoryDestroy(void* arg) {}
-static const char* CFilterFactoryName(void* arg) { return "foo"; }
+static void CFilterFactoryDestroy(void* arg) { (void)arg; }
+static const char* CFilterFactoryName(void* arg) {
+  (void)arg;
+  return "foo";
+}
 static rocksdb_compactionfilter_t* CFilterCreate(
     void* arg, rocksdb_compactionfiltercontext_t* context) {
+  (void)arg;
+  (void)context;
   return rocksdb_compactionfilter_create(NULL, CFilterDestroy, CFilterFilter,
                                          CFilterName);
 }
@@ -289,8 +321,9 @@ static rocksdb_t* CheckCompaction(rocksdb_t* db, rocksdb_options_t* options,
 }
 
 // Custom merge operator
-static void MergeOperatorDestroy(void* arg) { }
+static void MergeOperatorDestroy(void* arg) { (void)arg; }
 static const char* MergeOperatorName(void* arg) {
+  (void)arg;
   return "TestMergeOperator";
 }
 static char* MergeOperatorFullMerge(
@@ -300,6 +333,14 @@ static char* MergeOperatorFullMerge(
     const char* const* operands_list, const size_t* operands_list_length,
     int num_operands,
     unsigned char* success, size_t* new_value_length) {
+  (void)arg;
+  (void)key;
+  (void)key_length;
+  (void)existing_value;
+  (void)existing_value_length;
+  (void)operands_list;
+  (void)operands_list_length;
+  (void)num_operands;
   *new_value_length = 4;
   *success = 1;
   char* result = malloc(4);
@@ -312,6 +353,12 @@ static char* MergeOperatorPartialMerge(
     const char* const* operands_list, const size_t* operands_list_length,
     int num_operands,
     unsigned char* success, size_t* new_value_length) {
+  (void)arg;
+  (void)key;
+  (void)key_length;
+  (void)operands_list;
+  (void)operands_list_length;
+  (void)num_operands;
   *new_value_length = 4;
   *success = 1;
   char* result = malloc(4);
@@ -333,6 +380,20 @@ static void CheckTxnGet(
         Free(&val);
 }
 
+static void CheckTxnGetCF(rocksdb_transaction_t* txn,
+                          const rocksdb_readoptions_t* options,
+                          rocksdb_column_family_handle_t* column_family,
+                          const char* key, const char* expected) {
+  char* err = NULL;
+  size_t val_len;
+  char* val;
+  val = rocksdb_transaction_get_cf(txn, options, column_family, key,
+                                   strlen(key), &val_len, &err);
+  CheckNoError(err);
+  CheckEqual(expected, val, val_len);
+  Free(&val);
+}
+
 static void CheckTxnDBGet(
         rocksdb_transactiondb_t* txn_db,
         const rocksdb_readoptions_t* options,
@@ -347,10 +408,27 @@ static void CheckTxnDBGet(
         Free(&val);
 }
 
+static void CheckTxnDBGetCF(rocksdb_transactiondb_t* txn_db,
+                            const rocksdb_readoptions_t* options,
+                            rocksdb_column_family_handle_t* column_family,
+                            const char* key, const char* expected) {
+  char* err = NULL;
+  size_t val_len;
+  char* val;
+  val = rocksdb_transactiondb_get_cf(txn_db, options, column_family, key,
+                                     strlen(key), &val_len, &err);
+  CheckNoError(err);
+  CheckEqual(expected, val, val_len);
+  Free(&val);
+}
+
 int main(int argc, char** argv) {
+  (void)argc;
+  (void)argv;
   rocksdb_t* db;
   rocksdb_comparator_t* cmp;
   rocksdb_cache_t* cache;
+  rocksdb_dbpath_t *dbpath;
   rocksdb_env_t* env;
   rocksdb_options_t* options;
   rocksdb_compactoptions_t* coptions;
@@ -362,6 +440,8 @@ int main(int argc, char** argv) {
   rocksdb_transactiondb_options_t* txn_db_options;
   rocksdb_transaction_t* txn;
   rocksdb_transaction_options_t* txn_options;
+  rocksdb_optimistictransactiondb_t* otxn_db;
+  rocksdb_optimistictransaction_options_t* otxn_options;
   char* err = NULL;
   int run = -1;
 
@@ -385,8 +465,14 @@ int main(int argc, char** argv) {
            GetTempDir(),
            ((int)geteuid()));
 
+  snprintf(dbpathname, sizeof(dbpathname),
+           "%s/rocksdb_c_test-%d-dbpath",
+           GetTempDir(),
+           ((int) geteuid()));
+
   StartPhase("create_objects");
   cmp = rocksdb_comparator_create(NULL, CmpDestroy, CmpCompare, CmpName);
+  dbpath = rocksdb_dbpath_create(dbpathname, 1024 * 1024);
   env = rocksdb_create_default_env();
   cache = rocksdb_cache_create_lru(100000);
 
@@ -566,7 +652,7 @@ int main(int argc, char** argv) {
     rocksdb_sstfilewriter_t* writer =
         rocksdb_sstfilewriter_create(env_opt, io_options);
 
-    unlink(sstfilename);
+    remove(sstfilename);
     rocksdb_sstfilewriter_open(writer, sstfilename, &err);
     CheckNoError(err);
     rocksdb_sstfilewriter_put(writer, "sstk1", 5, "v1", 2, &err);
@@ -587,7 +673,7 @@ int main(int argc, char** argv) {
     CheckGet(db, roptions, "sstk2", "v2");
     CheckGet(db, roptions, "sstk3", "v3");
 
-    unlink(sstfilename);
+    remove(sstfilename);
     rocksdb_sstfilewriter_open(writer, sstfilename, &err);
     CheckNoError(err);
     rocksdb_sstfilewriter_put(writer, "sstk2", 5, "v4", 2, &err);
@@ -831,7 +917,8 @@ int main(int argc, char** argv) {
     rocksdb_writebatch_wi_t* wbi = rocksdb_writebatch_wi_create(0, 1);
     rocksdb_writebatch_wi_put(wbi, "bar", 3, "b", 1);
     rocksdb_writebatch_wi_delete(wbi, "foo", 3);
-    rocksdb_iterator_t* iter = rocksdb_writebatch_wi_create_iterator_with_base(wbi, base_iter);
+    rocksdb_iterator_t* iter =
+        rocksdb_writebatch_wi_create_iterator_with_base(wbi, base_iter);
     CheckCondition(!rocksdb_iter_valid(iter));
     rocksdb_iter_seek_to_first(iter);
     CheckCondition(rocksdb_iter_valid(iter));
@@ -1257,6 +1344,47 @@ int main(int argc, char** argv) {
     rocksdb_destroy_db(options, dbname, &err);
   }
 
+  // Check memory usage stats
+  StartPhase("approximate_memory_usage");
+  {
+    // Create database
+    db = rocksdb_open(options, dbname, &err);
+    CheckNoError(err);
+
+    rocksdb_memory_consumers_t* consumers;
+    consumers = rocksdb_memory_consumers_create();
+    rocksdb_memory_consumers_add_db(consumers, db);
+    rocksdb_memory_consumers_add_cache(consumers, cache);
+
+    // take memory usage report before write-read operation
+    rocksdb_memory_usage_t* mu1;
+    mu1 = rocksdb_approximate_memory_usage_create(consumers, &err);
+    CheckNoError(err);
+
+    // Put data (this should affect memtables)
+    rocksdb_put(db, woptions, "memory", 6, "test", 4, &err);
+    CheckNoError(err);
+    CheckGet(db, roptions, "memory", "test");
+
+    // take memory usage report after write-read operation
+    rocksdb_memory_usage_t* mu2;
+    mu2 = rocksdb_approximate_memory_usage_create(consumers, &err);
+    CheckNoError(err);
+
+    // amount of memory used within memtables should grow
+    CheckCondition(rocksdb_approximate_memory_usage_get_mem_table_total(mu2) >=
+                   rocksdb_approximate_memory_usage_get_mem_table_total(mu1));
+    CheckCondition(rocksdb_approximate_memory_usage_get_mem_table_unflushed(mu2) >=
+                   rocksdb_approximate_memory_usage_get_mem_table_unflushed(mu1));
+
+    rocksdb_memory_consumers_destroy(consumers);
+    rocksdb_approximate_memory_usage_destroy(mu1);
+    rocksdb_approximate_memory_usage_destroy(mu2);
+    rocksdb_close(db);
+    rocksdb_destroy_db(options, dbname, &err);
+    CheckNoError(err);
+  }
+
   StartPhase("cuckoo_options");
   {
     rocksdb_cuckoo_table_options_t* cuckoo_options;
@@ -1357,6 +1485,18 @@ int main(int argc, char** argv) {
     CheckNoError(err);
     CheckTxnDBGet(txn_db, roptions, "foo", NULL);
 
+    // write batch into TransactionDB
+    rocksdb_writebatch_t* wb = rocksdb_writebatch_create();
+    rocksdb_writebatch_put(wb, "foo", 3, "a", 1);
+    rocksdb_writebatch_clear(wb);
+    rocksdb_writebatch_put(wb, "bar", 3, "b", 1);
+    rocksdb_writebatch_put(wb, "box", 3, "c", 1);
+    rocksdb_writebatch_delete(wb, "bar", 3);
+    rocksdb_transactiondb_write(txn_db, woptions, wb, &err);
+    rocksdb_writebatch_destroy(wb);
+    CheckTxnDBGet(txn_db, roptions, "box", "c");
+    CheckNoError(err);
+
     // begin a transaction
     txn = rocksdb_transaction_begin(txn_db, woptions, txn_options, NULL);
     // put
@@ -1388,7 +1528,7 @@ int main(int argc, char** argv) {
     const rocksdb_snapshot_t* snapshot;
     snapshot = rocksdb_transactiondb_create_snapshot(txn_db);
     rocksdb_readoptions_set_snapshot(roptions, snapshot);
-  
+
     rocksdb_transactiondb_put(txn_db, woptions, "foo", 3, "hey", 3,  &err);
     CheckNoError(err);
 
@@ -1413,6 +1553,42 @@ int main(int argc, char** argv) {
     CheckNoError(err);
     CheckTxnDBGet(txn_db, roptions, "bar", NULL);
 
+    // save point
+    rocksdb_transaction_put(txn, "foo1", 4, "hi1", 3, &err);
+    rocksdb_transaction_set_savepoint(txn);
+    CheckTxnGet(txn, roptions, "foo1", "hi1");
+    rocksdb_transaction_put(txn, "foo2", 4, "hi2", 3, &err);
+    CheckTxnGet(txn, roptions, "foo2", "hi2");
+
+    // rollback to savepoint
+    rocksdb_transaction_rollback_to_savepoint(txn, &err);
+    CheckNoError(err);
+    CheckTxnGet(txn, roptions, "foo2", NULL);
+    CheckTxnGet(txn, roptions, "foo1", "hi1");
+    CheckTxnDBGet(txn_db, roptions, "foo1", NULL);
+    CheckTxnDBGet(txn_db, roptions, "foo2", NULL);
+    rocksdb_transaction_commit(txn, &err);
+    CheckNoError(err);
+    CheckTxnDBGet(txn_db, roptions, "foo1", "hi1");
+    CheckTxnDBGet(txn_db, roptions, "foo2", NULL);
+
+    // Column families.
+    rocksdb_column_family_handle_t* cfh;
+    cfh = rocksdb_transactiondb_create_column_family(txn_db, options,
+                                                     "txn_db_cf", &err);
+    CheckNoError(err);
+
+    rocksdb_transactiondb_put_cf(txn_db, woptions, cfh, "cf_foo", 6, "cf_hello",
+                                 8, &err);
+    CheckNoError(err);
+    CheckTxnDBGetCF(txn_db, roptions, cfh, "cf_foo", "cf_hello");
+
+    rocksdb_transactiondb_delete_cf(txn_db, woptions, cfh, "cf_foo", 6, &err);
+    CheckNoError(err);
+    CheckTxnDBGetCF(txn_db, roptions, cfh, "cf_foo", NULL);
+
+    rocksdb_column_family_handle_destroy(cfh);
+
     // close and destroy
     rocksdb_transaction_destroy(txn);
     rocksdb_transactiondb_close(txn_db);
@@ -1422,6 +1598,105 @@ int main(int argc, char** argv) {
     rocksdb_transactiondb_options_destroy(txn_db_options);
   }
 
+  StartPhase("optimistic_transactions");
+  {
+    rocksdb_options_t* db_options = rocksdb_options_create();
+    rocksdb_options_set_create_if_missing(db_options, 1);
+    rocksdb_options_set_allow_concurrent_memtable_write(db_options, 1);
+    otxn_db = rocksdb_optimistictransactiondb_open(db_options, dbname, &err);
+    otxn_options = rocksdb_optimistictransaction_options_create();
+    rocksdb_transaction_t* txn1 = rocksdb_optimistictransaction_begin(
+        otxn_db, woptions, otxn_options, NULL);
+    rocksdb_transaction_t* txn2 = rocksdb_optimistictransaction_begin(
+        otxn_db, woptions, otxn_options, NULL);
+    rocksdb_transaction_put(txn1, "key", 3, "value", 5, &err);
+    CheckNoError(err);
+    rocksdb_transaction_put(txn2, "key1", 4, "value1", 6, &err);
+    CheckNoError(err);
+    CheckTxnGet(txn1, roptions, "key", "value");
+    rocksdb_transaction_commit(txn1, &err);
+    CheckNoError(err);
+    rocksdb_transaction_commit(txn2, &err);
+    CheckNoError(err);
+    rocksdb_transaction_destroy(txn1);
+    rocksdb_transaction_destroy(txn2);
+
+    // Check column family
+    db = rocksdb_optimistictransactiondb_get_base_db(otxn_db);
+    rocksdb_put(db, woptions, "key", 3, "value", 5, &err);
+    CheckNoError(err);
+    rocksdb_column_family_handle_t *cfh1, *cfh2;
+    cfh1 = rocksdb_create_column_family(db, db_options, "txn_db_cf1", &err);
+    cfh2 = rocksdb_create_column_family(db, db_options, "txn_db_cf2", &err);
+    txn = rocksdb_optimistictransaction_begin(otxn_db, woptions, otxn_options,
+                                              NULL);
+    rocksdb_transaction_put_cf(txn, cfh1, "key_cf1", 7, "val_cf1", 7, &err);
+    CheckNoError(err);
+    rocksdb_transaction_put_cf(txn, cfh2, "key_cf2", 7, "val_cf2", 7, &err);
+    CheckNoError(err);
+    rocksdb_transaction_commit(txn, &err);
+    CheckNoError(err);
+    txn = rocksdb_optimistictransaction_begin(otxn_db, woptions, otxn_options,
+                                              txn);
+    CheckGetCF(db, roptions, cfh1, "key_cf1", "val_cf1");
+    CheckTxnGetCF(txn, roptions, cfh1, "key_cf1", "val_cf1");
+
+    // Check iterator with column family
+    rocksdb_transaction_put_cf(txn, cfh1, "key1_cf", 7, "val1_cf", 7, &err);
+    CheckNoError(err);
+    rocksdb_iterator_t* iter =
+        rocksdb_transaction_create_iterator_cf(txn, roptions, cfh1);
+    CheckCondition(!rocksdb_iter_valid(iter));
+    rocksdb_iter_seek_to_first(iter);
+    CheckCondition(rocksdb_iter_valid(iter));
+    CheckIter(iter, "key1_cf", "val1_cf");
+    rocksdb_iter_get_error(iter, &err);
+    CheckNoError(err);
+    rocksdb_iter_destroy(iter);
+
+    rocksdb_transaction_destroy(txn);
+    rocksdb_column_family_handle_destroy(cfh1);
+    rocksdb_column_family_handle_destroy(cfh2);
+    rocksdb_optimistictransactiondb_close_base_db(db);
+    rocksdb_optimistictransactiondb_close(otxn_db);
+
+    // Check open optimistic transaction db with column families
+    size_t cf_len;
+    char** column_fams =
+        rocksdb_list_column_families(db_options, dbname, &cf_len, &err);
+    CheckNoError(err);
+    CheckEqual("default", column_fams[0], 7);
+    CheckEqual("txn_db_cf1", column_fams[1], 10);
+    CheckEqual("txn_db_cf2", column_fams[2], 10);
+    CheckCondition(cf_len == 3);
+    rocksdb_list_column_families_destroy(column_fams, cf_len);
+
+    const char* cf_names[3] = {"default", "txn_db_cf1", "txn_db_cf2"};
+    rocksdb_options_t* cf_options = rocksdb_options_create();
+    const rocksdb_options_t* cf_opts[3] = {cf_options, cf_options, cf_options};
+
+    rocksdb_options_set_error_if_exists(cf_options, 0);
+    rocksdb_column_family_handle_t* cf_handles[3];
+    otxn_db = rocksdb_optimistictransactiondb_open_column_families(
+        db_options, dbname, 3, cf_names, cf_opts, cf_handles, &err);
+    CheckNoError(err);
+    rocksdb_transaction_t* txn_cf = rocksdb_optimistictransaction_begin(
+        otxn_db, woptions, otxn_options, NULL);
+    CheckTxnGetCF(txn_cf, roptions, cf_handles[0], "key", "value");
+    CheckTxnGetCF(txn_cf, roptions, cf_handles[1], "key_cf1", "val_cf1");
+    CheckTxnGetCF(txn_cf, roptions, cf_handles[2], "key_cf2", "val_cf2");
+    rocksdb_transaction_destroy(txn_cf);
+    rocksdb_options_destroy(cf_options);
+    rocksdb_column_family_handle_destroy(cf_handles[0]);
+    rocksdb_column_family_handle_destroy(cf_handles[1]);
+    rocksdb_column_family_handle_destroy(cf_handles[2]);
+    rocksdb_optimistictransactiondb_close(otxn_db);
+    rocksdb_destroy_db(db_options, dbname, &err);
+    rocksdb_options_destroy(db_options);
+    rocksdb_optimistictransaction_options_destroy(otxn_options);
+    CheckNoError(err);
+  }
+
   // Simple sanity check that setting memtable rep works.
   StartPhase("memtable_reps");
   {
@@ -1440,6 +1715,18 @@ int main(int argc, char** argv) {
     CheckNoError(err);
   }
 
+  // Simple sanity check that options setting db_paths work.
+  StartPhase("open_db_paths");
+  {
+    rocksdb_close(db);
+    rocksdb_destroy_db(options, dbname, &err);
+
+    const rocksdb_dbpath_t* paths[1] = {dbpath};
+    rocksdb_options_set_db_paths(options, paths, 1);
+    db = rocksdb_open(options, dbname, &err);
+    CheckNoError(err);
+  }
+
   StartPhase("cleanup");
   rocksdb_close(db);
   rocksdb_options_destroy(options);
@@ -1449,6 +1736,7 @@ int main(int argc, char** argv) {
   rocksdb_compactoptions_destroy(coptions);
   rocksdb_cache_destroy(cache);
   rocksdb_comparator_destroy(cmp);
+  rocksdb_dbpath_destroy(dbpath);
   rocksdb_env_destroy(env);
 
   fprintf(stderr, "PASS\n");

db/column_family.cc

@@ -20,10 +20,12 @@
 #include <limits>
 
 #include "db/compaction_picker.h"
+#include "db/compaction_picker_fifo.h"
 #include "db/compaction_picker_universal.h"
 #include "db/db_impl.h"
 #include "db/internal_stats.h"
 #include "db/job_context.h"
+#include "db/range_del_aggregator.h"
 #include "db/table_properties_collector.h"
 #include "db/version_set.h"
 #include "db/write_controller.h"
@@ -31,8 +33,10 @@
 #include "monitoring/thread_status_util.h"
 #include "options/options_helper.h"
 #include "table/block_based_table_factory.h"
+#include "table/merging_iterator.h"
 #include "util/autovector.h"
 #include "util/compression.h"
+#include "util/sst_file_manager_impl.h"
 
 namespace rocksdb {
 
@@ -53,6 +57,9 @@ ColumnFamilyHandleImpl::~ColumnFamilyHandleImpl() {
 #endif  // ROCKSDB_LITE
     // Job id == 0 means that this is not our background process, but rather
     // user thread
+    // Need to hold some shared pointers owned by the initial_cf_options
+    // before final cleaning up finishes.
+    ColumnFamilyOptions initial_cf_options_copy = cfd_->initial_cf_options();
     JobContext job_context(0);
     mutex_->Lock();
     if (cfd_->Unref()) {
@@ -80,6 +87,7 @@ Status ColumnFamilyHandleImpl::GetDescriptor(ColumnFamilyDescriptor* desc) {
   *desc = ColumnFamilyDescriptor(cfd()->GetName(), cfd()->GetLatestCFOptions());
   return Status::OK();
 #else
+  (void)desc;
   return Status::NotSupported();
 #endif  // !ROCKSDB_LITE
 }
@@ -99,9 +107,6 @@ void GetIntTblPropCollectorFactory(
     int_tbl_prop_collector_factories->emplace_back(
         new UserKeyTablePropertiesCollectorFactory(collector_factories[i]));
   }
-  // Add collector to collect internal key statistics
-  int_tbl_prop_collector_factories->emplace_back(
-      new InternalKeyPropertiesCollectorFactory);
 }
 
 Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
@@ -123,6 +128,18 @@ Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options) {
           " is not linked with the binary.");
     }
   }
+  if (cf_options.compression_opts.zstd_max_train_bytes > 0) {
+    if (!ZSTD_TrainDictionarySupported()) {
+      return Status::InvalidArgument(
+          "zstd dictionary trainer cannot be used because ZSTD 1.1.3+ "
+          "is not linked with the binary.");
+    }
+    if (cf_options.compression_opts.max_dict_bytes == 0) {
+      return Status::InvalidArgument(
+          "The dictionary size limit (`CompressionOptions::max_dict_bytes`) "
+          "should be nonzero if we're using zstd's dictionary generator.");
+    }
+  }
   return Status::OK();
 }
 
@@ -139,6 +156,28 @@ Status CheckConcurrentWritesSupported(const ColumnFamilyOptions& cf_options) {
   return Status::OK();
 }
 
+Status CheckCFPathsSupported(const DBOptions& db_options,
+                             const ColumnFamilyOptions& cf_options) {
+  // More than one cf_paths are supported only in universal
+  // and level compaction styles. This function also checks the case
+  // in which cf_paths is not specified, which results in db_paths
+  // being used.
+  if ((cf_options.compaction_style != kCompactionStyleUniversal) &&
+      (cf_options.compaction_style != kCompactionStyleLevel)) {
+    if (cf_options.cf_paths.size() > 1) {
+      return Status::NotSupported(
+          "More than one CF paths are only supported in "
+          "universal and level compaction styles. ");
+    } else if (cf_options.cf_paths.empty() &&
+               db_options.db_paths.size() > 1) {
+      return Status::NotSupported(
+          "More than one DB paths are only supported in "
+          "universal and level compaction styles. ");
+    }
+  }
+  return Status::OK();
+}
+
 ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
                                     const ColumnFamilyOptions& src) {
   ColumnFamilyOptions result = src;
@@ -257,9 +296,24 @@ ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
         result.hard_pending_compaction_bytes_limit;
   }
 
+#ifndef ROCKSDB_LITE
+  // When the DB is stopped, it's possible that there are some .trash files that
+  // were not deleted yet, when we open the DB we will find these .trash files
+  // and schedule them to be deleted (or delete immediately if SstFileManager
+  // was not used)
+  auto sfm = static_cast<SstFileManagerImpl*>(db_options.sst_file_manager.get());
+  for (size_t i = 0; i < result.cf_paths.size(); i++) {
+    DeleteScheduler::CleanupDirectory(db_options.env, sfm, result.cf_paths[i].path);
+  }
+#endif
+
+  if (result.cf_paths.empty()) {
+    result.cf_paths = db_options.db_paths;
+  }
+
   if (result.level_compaction_dynamic_level_bytes) {
     if (result.compaction_style != kCompactionStyleLevel ||
-        db_options.db_paths.size() > 1U) {
+        result.cf_paths.size() > 1U) {
       // 1. level_compaction_dynamic_level_bytes only makes sense for
       //    level-based compaction.
       // 2. we don't yet know how to make both of this feature and multiple
@@ -328,12 +382,13 @@ void SuperVersionUnrefHandle(void* ptr) {
   // When latter happens, we are in ~ColumnFamilyData(), no get should happen as
   // well.
   SuperVersion* sv = static_cast<SuperVersion*>(ptr);
-  if (sv->Unref()) {
-    sv->db_mutex->Lock();
-    sv->Cleanup();
-    sv->db_mutex->Unlock();
-    delete sv;
-  }
+  bool was_last_ref __attribute__((__unused__));
+  was_last_ref = sv->Unref();
+  // Thread-local SuperVersions can't outlive ColumnFamilyData::super_version_.
+  // This is important because we can't do SuperVersion cleanup here.
+  // That would require locking DB mutex, which would deadlock because
+  // SuperVersionUnrefHandle is called with locked ThreadLocalPtr mutex.
+  assert(!was_last_ref);
 }
 }  // anonymous namespace
 
@@ -365,11 +420,13 @@ ColumnFamilyData::ColumnFamilyData(
       next_(nullptr),
       prev_(nullptr),
       log_number_(0),
+      flush_reason_(FlushReason::kOthers),
       column_family_set_(column_family_set),
-      pending_flush_(false),
-      pending_compaction_(false),
+      queued_for_flush_(false),
+      queued_for_compaction_(false),
       prev_compaction_needed_bytes_(0),
-      allow_2pc_(db_options.allow_2pc) {
+      allow_2pc_(db_options.allow_2pc),
+      last_memtable_id_(0) {
   Ref();
 
   // Convert user defined table properties collector factories to internal ones.
@@ -442,8 +499,8 @@ ColumnFamilyData::~ColumnFamilyData() {
 
   // It would be wrong if this ColumnFamilyData is in flush_queue_ or
   // compaction_queue_ and we destroyed it
-  assert(!pending_flush_);
-  assert(!pending_compaction_);
+  assert(!queued_for_flush_);
+  assert(!queued_for_compaction_);
 
   if (super_version_ != nullptr) {
     // Release SuperVersion reference kept in ThreadLocalPtr.
@@ -452,7 +509,7 @@ ColumnFamilyData::~ColumnFamilyData() {
     local_sv_.reset();
     super_version_->db_mutex->Lock();
 
-    bool is_last_reference __attribute__((unused));
+    bool is_last_reference __attribute__((__unused__));
     is_last_reference = super_version_->Unref();
     assert(is_last_reference);
     super_version_->Cleanup();
@@ -463,7 +520,8 @@ ColumnFamilyData::~ColumnFamilyData() {
   if (dummy_versions_ != nullptr) {
     // List must be empty
     assert(dummy_versions_->TEST_Next() == dummy_versions_);
-    bool deleted __attribute__((unused)) = dummy_versions_->Unref();
+    bool deleted __attribute__((__unused__));
+    deleted = dummy_versions_->Unref();
     assert(deleted);
   }
 
@@ -495,7 +553,9 @@ uint64_t ColumnFamilyData::OldestLogToKeep() {
   auto current_log = GetLogNumber();
 
   if (allow_2pc_) {
-    auto imm_prep_log = imm()->GetMinLogContainingPrepSection();
+    autovector<MemTable*> empty_list;
+    auto imm_prep_log =
+        imm()->PrecomputeMinLogContainingPrepSection(empty_list);
     auto mem_prep_log = mem()->GetMinLogContainingPrepSection();
 
     if (imm_prep_log > 0 && imm_prep_log < current_log) {
@@ -613,58 +673,97 @@ int GetL0ThresholdSpeedupCompaction(int level0_file_num_compaction_trigger,
 }
 }  // namespace
 
-void ColumnFamilyData::RecalculateWriteStallConditions(
+std::pair<WriteStallCondition, ColumnFamilyData::WriteStallCause>
+ColumnFamilyData::GetWriteStallConditionAndCause(
+    int num_unflushed_memtables, int num_l0_files,
+    uint64_t num_compaction_needed_bytes,
+    const MutableCFOptions& mutable_cf_options) {
+  if (num_unflushed_memtables >= mutable_cf_options.max_write_buffer_number) {
+    return {WriteStallCondition::kStopped, WriteStallCause::kMemtableLimit};
+  } else if (!mutable_cf_options.disable_auto_compactions &&
+             num_l0_files >= mutable_cf_options.level0_stop_writes_trigger) {
+    return {WriteStallCondition::kStopped, WriteStallCause::kL0FileCountLimit};
+  } else if (!mutable_cf_options.disable_auto_compactions &&
+             mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
+             num_compaction_needed_bytes >=
+                 mutable_cf_options.hard_pending_compaction_bytes_limit) {
+    return {WriteStallCondition::kStopped,
+            WriteStallCause::kPendingCompactionBytes};
+  } else if (mutable_cf_options.max_write_buffer_number > 3 &&
+             num_unflushed_memtables >=
+                 mutable_cf_options.max_write_buffer_number - 1) {
+    return {WriteStallCondition::kDelayed, WriteStallCause::kMemtableLimit};
+  } else if (!mutable_cf_options.disable_auto_compactions &&
+             mutable_cf_options.level0_slowdown_writes_trigger >= 0 &&
+             num_l0_files >=
+                 mutable_cf_options.level0_slowdown_writes_trigger) {
+    return {WriteStallCondition::kDelayed, WriteStallCause::kL0FileCountLimit};
+  } else if (!mutable_cf_options.disable_auto_compactions &&
+             mutable_cf_options.soft_pending_compaction_bytes_limit > 0 &&
+             num_compaction_needed_bytes >=
+                 mutable_cf_options.soft_pending_compaction_bytes_limit) {
+    return {WriteStallCondition::kDelayed,
+            WriteStallCause::kPendingCompactionBytes};
+  }
+  return {WriteStallCondition::kNormal, WriteStallCause::kNone};
+}
+
+WriteStallCondition ColumnFamilyData::RecalculateWriteStallConditions(
       const MutableCFOptions& mutable_cf_options) {
+  auto write_stall_condition = WriteStallCondition::kNormal;
   if (current_ != nullptr) {
     auto* vstorage = current_->storage_info();
     auto write_controller = column_family_set_->write_controller_;
     uint64_t compaction_needed_bytes =
         vstorage->estimated_compaction_needed_bytes();
 
+    auto write_stall_condition_and_cause = GetWriteStallConditionAndCause(
+        imm()->NumNotFlushed(), vstorage->l0_delay_trigger_count(),
+        vstorage->estimated_compaction_needed_bytes(), mutable_cf_options);
+    write_stall_condition = write_stall_condition_and_cause.first;
+    auto write_stall_cause = write_stall_condition_and_cause.second;
+
     bool was_stopped = write_controller->IsStopped();
     bool needed_delay = write_controller->NeedsDelay();
 
-    if (imm()->NumNotFlushed() >= mutable_cf_options.max_write_buffer_number) {
+    if (write_stall_condition == WriteStallCondition::kStopped &&
+        write_stall_cause == WriteStallCause::kMemtableLimit) {
       write_controller_token_ = write_controller->GetStopToken();
-      internal_stats_->AddCFStats(InternalStats::MEMTABLE_COMPACTION, 1);
+      internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_STOPS, 1);
       ROCKS_LOG_WARN(
           ioptions_.info_log,
           "[%s] Stopping writes because we have %d immutable memtables "
           "(waiting for flush), max_write_buffer_number is set to %d",
           name_.c_str(), imm()->NumNotFlushed(),
           mutable_cf_options.max_write_buffer_number);
-    } else if (!mutable_cf_options.disable_auto_compactions &&
-               vstorage->l0_delay_trigger_count() >=
-                   mutable_cf_options.level0_stop_writes_trigger) {
+    } else if (write_stall_condition == WriteStallCondition::kStopped &&
+               write_stall_cause == WriteStallCause::kL0FileCountLimit) {
       write_controller_token_ = write_controller->GetStopToken();
-      internal_stats_->AddCFStats(InternalStats::LEVEL0_NUM_FILES_TOTAL, 1);
+      internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_STOPS, 1);
       if (compaction_picker_->IsLevel0CompactionInProgress()) {
         internal_stats_->AddCFStats(
-            InternalStats::LEVEL0_NUM_FILES_WITH_COMPACTION, 1);
+            InternalStats::LOCKED_L0_FILE_COUNT_LIMIT_STOPS, 1);
       }
       ROCKS_LOG_WARN(ioptions_.info_log,
                      "[%s] Stopping writes because we have %d level-0 files",
                      name_.c_str(), vstorage->l0_delay_trigger_count());
-    } else if (!mutable_cf_options.disable_auto_compactions &&
-               mutable_cf_options.hard_pending_compaction_bytes_limit > 0 &&
-               compaction_needed_bytes >=
-                   mutable_cf_options.hard_pending_compaction_bytes_limit) {
+    } else if (write_stall_condition == WriteStallCondition::kStopped &&
+               write_stall_cause == WriteStallCause::kPendingCompactionBytes) {
       write_controller_token_ = write_controller->GetStopToken();
       internal_stats_->AddCFStats(
-          InternalStats::HARD_PENDING_COMPACTION_BYTES_LIMIT, 1);
+          InternalStats::PENDING_COMPACTION_BYTES_LIMIT_STOPS, 1);
       ROCKS_LOG_WARN(
           ioptions_.info_log,
           "[%s] Stopping writes because of estimated pending compaction "
           "bytes %" PRIu64,
           name_.c_str(), compaction_needed_bytes);
-    } else if (mutable_cf_options.max_write_buffer_number > 3 &&
-               imm()->NumNotFlushed() >=
-                   mutable_cf_options.max_write_buffer_number - 1) {
+    } else if (write_stall_condition == WriteStallCondition::kDelayed &&
+               write_stall_cause == WriteStallCause::kMemtableLimit) {
       write_controller_token_ =
           SetupDelay(write_controller, compaction_needed_bytes,
                      prev_compaction_needed_bytes_, was_stopped,
                      mutable_cf_options.disable_auto_compactions);
-      internal_stats_->AddCFStats(InternalStats::MEMTABLE_SLOWDOWN, 1);
+      internal_stats_->AddCFStats(InternalStats::MEMTABLE_LIMIT_SLOWDOWNS, 1);
       ROCKS_LOG_WARN(
           ioptions_.info_log,
           "[%s] Stalling writes because we have %d immutable memtables "
@@ -673,10 +772,8 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
           name_.c_str(), imm()->NumNotFlushed(),
           mutable_cf_options.max_write_buffer_number,
           write_controller->delayed_write_rate());
-    } else if (!mutable_cf_options.disable_auto_compactions &&
-               mutable_cf_options.level0_slowdown_writes_trigger >= 0 &&
-               vstorage->l0_delay_trigger_count() >=
-                   mutable_cf_options.level0_slowdown_writes_trigger) {
+    } else if (write_stall_condition == WriteStallCondition::kDelayed &&
+               write_stall_cause == WriteStallCause::kL0FileCountLimit) {
       // L0 is the last two files from stopping.
       bool near_stop = vstorage->l0_delay_trigger_count() >=
                        mutable_cf_options.level0_stop_writes_trigger - 2;
@@ -684,20 +781,19 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
           SetupDelay(write_controller, compaction_needed_bytes,
                      prev_compaction_needed_bytes_, was_stopped || near_stop,
                      mutable_cf_options.disable_auto_compactions);
-      internal_stats_->AddCFStats(InternalStats::LEVEL0_SLOWDOWN_TOTAL, 1);
+      internal_stats_->AddCFStats(InternalStats::L0_FILE_COUNT_LIMIT_SLOWDOWNS,
+                                  1);
       if (compaction_picker_->IsLevel0CompactionInProgress()) {
         internal_stats_->AddCFStats(
-            InternalStats::LEVEL0_SLOWDOWN_WITH_COMPACTION, 1);
+            InternalStats::LOCKED_L0_FILE_COUNT_LIMIT_SLOWDOWNS, 1);
       }
       ROCKS_LOG_WARN(ioptions_.info_log,
                      "[%s] Stalling writes because we have %d level-0 files "
                      "rate %" PRIu64,
                      name_.c_str(), vstorage->l0_delay_trigger_count(),
                      write_controller->delayed_write_rate());
-    } else if (!mutable_cf_options.disable_auto_compactions &&
-               mutable_cf_options.soft_pending_compaction_bytes_limit > 0 &&
-               vstorage->estimated_compaction_needed_bytes() >=
-                   mutable_cf_options.soft_pending_compaction_bytes_limit) {
+    } else if (write_stall_condition == WriteStallCondition::kDelayed &&
+               write_stall_cause == WriteStallCause::kPendingCompactionBytes) {
       // If the distance to hard limit is less than 1/4 of the gap between soft
       // and
       // hard bytes limit, we think it is near stop and speed up the slowdown.
@@ -714,7 +810,7 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
                      prev_compaction_needed_bytes_, was_stopped || near_stop,
                      mutable_cf_options.disable_auto_compactions);
       internal_stats_->AddCFStats(
-          InternalStats::SOFT_PENDING_COMPACTION_BYTES_LIMIT, 1);
+          InternalStats::PENDING_COMPACTION_BYTES_LIMIT_SLOWDOWNS, 1);
       ROCKS_LOG_WARN(
           ioptions_.info_log,
           "[%s] Stalling writes because of estimated pending compaction "
@@ -722,6 +818,7 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
           name_.c_str(), vstorage->estimated_compaction_needed_bytes(),
           write_controller->delayed_write_rate());
     } else {
+      assert(write_stall_condition == WriteStallCondition::kNormal);
       if (vstorage->l0_delay_trigger_count() >=
           GetL0ThresholdSpeedupCompaction(
               mutable_cf_options.level0_file_num_compaction_trigger,
@@ -769,6 +866,7 @@ void ColumnFamilyData::RecalculateWriteStallConditions(
     }
     prev_compaction_needed_bytes_ = compaction_needed_bytes;
   }
+  return write_stall_condition;
 }
 
 const EnvOptions* ColumnFamilyData::soptions() const {
@@ -787,6 +885,10 @@ uint64_t ColumnFamilyData::GetTotalSstFilesSize() const {
   return VersionSet::GetTotalSstFilesSize(dummy_versions_);
 }
 
+uint64_t ColumnFamilyData::GetLiveSstFilesSize() const {
+  return current_->GetSstFilesSize();
+}
+
 MemTable* ColumnFamilyData::ConstructNewMemtable(
     const MutableCFOptions& mutable_cf_options, SequenceNumber earliest_seq) {
   return new MemTable(internal_comparator_, ioptions_, mutable_cf_options,
@@ -823,16 +925,70 @@ bool ColumnFamilyData::RangeOverlapWithCompaction(
       smallest_user_key, largest_user_key, level);
 }
 
+Status ColumnFamilyData::RangesOverlapWithMemtables(
+    const autovector<Range>& ranges, SuperVersion* super_version,
+    bool* overlap) {
+  assert(overlap != nullptr);
+  *overlap = false;
+  // Create an InternalIterator over all unflushed memtables
+  Arena arena;
+  ReadOptions read_opts;
+  read_opts.total_order_seek = true;
+  MergeIteratorBuilder merge_iter_builder(&internal_comparator_, &arena);
+  merge_iter_builder.AddIterator(
+      super_version->mem->NewIterator(read_opts, &arena));
+  super_version->imm->AddIterators(read_opts, &merge_iter_builder);
+  ScopedArenaIterator memtable_iter(merge_iter_builder.Finish());
+
+  auto read_seq = super_version->current->version_set()->LastSequence();
+  ReadRangeDelAggregator range_del_agg(&internal_comparator_, read_seq);
+  auto* active_range_del_iter =
+      super_version->mem->NewRangeTombstoneIterator(read_opts, read_seq);
+  range_del_agg.AddTombstones(
+      std::unique_ptr<FragmentedRangeTombstoneIterator>(active_range_del_iter));
+  super_version->imm->AddRangeTombstoneIterators(read_opts, nullptr /* arena */,
+                                                 &range_del_agg);
+
+  Status status;
+  for (size_t i = 0; i < ranges.size() && status.ok() && !*overlap; ++i) {
+    auto* vstorage = super_version->current->storage_info();
+    auto* ucmp = vstorage->InternalComparator()->user_comparator();
+    InternalKey range_start(ranges[i].start, kMaxSequenceNumber,
+                            kValueTypeForSeek);
+    memtable_iter->Seek(range_start.Encode());
+    status = memtable_iter->status();
+    ParsedInternalKey seek_result;
+    if (status.ok()) {
+      if (memtable_iter->Valid() &&
+          !ParseInternalKey(memtable_iter->key(), &seek_result)) {
+        status = Status::Corruption("DB have corrupted keys");
+      }
+    }
+    if (status.ok()) {
+      if (memtable_iter->Valid() &&
+          ucmp->Compare(seek_result.user_key, ranges[i].limit) <= 0) {
+        *overlap = true;
+      } else if (range_del_agg.IsRangeOverlapped(ranges[i].start,
+                                                 ranges[i].limit)) {
+        *overlap = true;
+      }
+    }
+  }
+  return status;
+}
+
 const int ColumnFamilyData::kCompactAllLevels = -1;
 const int ColumnFamilyData::kCompactToBaseLevel = -2;
 
 Compaction* ColumnFamilyData::CompactRange(
     const MutableCFOptions& mutable_cf_options, int input_level,
-    int output_level, uint32_t output_path_id, const InternalKey* begin,
-    const InternalKey* end, InternalKey** compaction_end, bool* conflict) {
+    int output_level, uint32_t output_path_id, uint32_t max_subcompactions,
+    const InternalKey* begin, const InternalKey* end,
+    InternalKey** compaction_end, bool* conflict) {
   auto* result = compaction_picker_->CompactRange(
       GetName(), mutable_cf_options, current_->storage_info(), input_level,
-      output_level, output_path_id, begin, end, compaction_end, conflict);
+      output_level, output_path_id, max_subcompactions, begin, end,
+      compaction_end, conflict);
   if (result != nullptr) {
     result->SetInputVersion(current_);
   }
@@ -841,10 +997,13 @@ Compaction* ColumnFamilyData::CompactRange(
 
 SuperVersion* ColumnFamilyData::GetReferencedSuperVersion(
     InstrumentedMutex* db_mutex) {
-  SuperVersion* sv = nullptr;
-  sv = GetThreadLocalSuperVersion(db_mutex);
+  SuperVersion* sv = GetThreadLocalSuperVersion(db_mutex);
   sv->Ref();
   if (!ReturnThreadLocalSuperVersion(sv)) {
+    // This Unref() corresponds to the Ref() in GetThreadLocalSuperVersion()
+    // when the thread-local pointer was populated. So, the Ref() earlier in
+    // this function still prevents the returned SuperVersion* from being
+    // deleted out from under the caller.
     sv->Unref();
   }
   return sv;
@@ -852,7 +1011,6 @@ SuperVersion* ColumnFamilyData::GetReferencedSuperVersion(
 
 SuperVersion* ColumnFamilyData::GetThreadLocalSuperVersion(
     InstrumentedMutex* db_mutex) {
-  SuperVersion* sv = nullptr;
   // The SuperVersion is cached in thread local storage to avoid acquiring
   // mutex when SuperVersion does not change since the last use. When a new
   // SuperVersion is installed, the compaction or flush thread cleans up
@@ -871,7 +1029,7 @@ SuperVersion* ColumnFamilyData::GetThreadLocalSuperVersion(
   // should only keep kSVInUse before ReturnThreadLocalSuperVersion call
   // (if no Scrape happens).
   assert(ptr != SuperVersion::kSVInUse);
-  sv = static_cast<SuperVersion*>(ptr);
+  SuperVersion* sv = static_cast<SuperVersion*>(ptr);
   if (sv == SuperVersion::kSVObsolete ||
       sv->version_number != super_version_number_.load()) {
     RecordTick(ioptions_.statistics, NUMBER_SUPERVERSION_ACQUIRES);
@@ -914,15 +1072,16 @@ bool ColumnFamilyData::ReturnThreadLocalSuperVersion(SuperVersion* sv) {
   return false;
 }
 
-SuperVersion* ColumnFamilyData::InstallSuperVersion(
-    SuperVersion* new_superversion, InstrumentedMutex* db_mutex) {
+void ColumnFamilyData::InstallSuperVersion(
+    SuperVersionContext* sv_context, InstrumentedMutex* db_mutex) {
   db_mutex->AssertHeld();
-  return InstallSuperVersion(new_superversion, db_mutex, mutable_cf_options_);
+  return InstallSuperVersion(sv_context, db_mutex, mutable_cf_options_);
 }
 
-SuperVersion* ColumnFamilyData::InstallSuperVersion(
-    SuperVersion* new_superversion, InstrumentedMutex* db_mutex,
+void ColumnFamilyData::InstallSuperVersion(
+    SuperVersionContext* sv_context, InstrumentedMutex* db_mutex,
     const MutableCFOptions& mutable_cf_options) {
+  SuperVersion* new_superversion = sv_context->new_superversion.release();
   new_superversion->db_mutex = db_mutex;
   new_superversion->mutable_cf_options = mutable_cf_options;
   new_superversion->Init(mem_, imm_.current(), current_);
@@ -930,16 +1089,31 @@ SuperVersion* ColumnFamilyData::InstallSuperVersion(
   super_version_ = new_superversion;
   ++super_version_number_;
   super_version_->version_number = super_version_number_;
-  // Reset SuperVersions cached in thread local storage
-  ResetThreadLocalSuperVersions();
+  super_version_->write_stall_condition =
+      RecalculateWriteStallConditions(mutable_cf_options);
 
-  RecalculateWriteStallConditions(mutable_cf_options);
+  if (old_superversion != nullptr) {
+    // Reset SuperVersions cached in thread local storage.
+    // This should be done before old_superversion->Unref(). That's to ensure
+    // that local_sv_ never holds the last reference to SuperVersion, since
+    // it has no means to safely do SuperVersion cleanup.
+    ResetThreadLocalSuperVersions();
 
-  if (old_superversion != nullptr && old_superversion->Unref()) {
-    old_superversion->Cleanup();
-    return old_superversion;  // will let caller delete outside of mutex
+    if (old_superversion->mutable_cf_options.write_buffer_size !=
+        mutable_cf_options.write_buffer_size) {
+      mem_->UpdateWriteBufferSize(mutable_cf_options.write_buffer_size);
+    }
+    if (old_superversion->write_stall_condition !=
+        new_superversion->write_stall_condition) {
+      sv_context->PushWriteStallNotification(
+          old_superversion->write_stall_condition,
+          new_superversion->write_stall_condition, GetName(), ioptions());
+    }
+    if (old_superversion->Unref()) {
+      old_superversion->Cleanup();
+      sv_context->superversions_to_free.push_back(old_superversion);
+    }
   }
-  return nullptr;
 }
 
 void ColumnFamilyData::ResetThreadLocalSuperVersions() {
@@ -951,10 +1125,12 @@ void ColumnFamilyData::ResetThreadLocalSuperVersions() {
       continue;
     }
     auto sv = static_cast<SuperVersion*>(ptr);
-    if (sv->Unref()) {
-      sv->Cleanup();
-      delete sv;
-    }
+    bool was_last_ref __attribute__((__unused__));
+    was_last_ref = sv->Unref();
+    // sv couldn't have been the last reference because
+    // ResetThreadLocalSuperVersions() is called before
+    // unref'ing super_version_.
+    assert(!was_last_ref);
   }
 }
 
@@ -962,8 +1138,9 @@ void ColumnFamilyData::ResetThreadLocalSuperVersions() {
 Status ColumnFamilyData::SetOptions(
       const std::unordered_map<std::string, std::string>& options_map) {
   MutableCFOptions new_mutable_cf_options;
-  Status s = GetMutableOptionsFromStrings(mutable_cf_options_, options_map,
-                                          &new_mutable_cf_options);
+  Status s =
+      GetMutableOptionsFromStrings(mutable_cf_options_, options_map,
+                                   ioptions_.info_log, &new_mutable_cf_options);
   if (s.ok()) {
     mutable_cf_options_ = new_mutable_cf_options;
     mutable_cf_options_.RefreshDerivedOptions(ioptions_);
@@ -972,6 +1149,49 @@ Status ColumnFamilyData::SetOptions(
 }
 #endif  // ROCKSDB_LITE
 
+// REQUIRES: DB mutex held
+Env::WriteLifeTimeHint ColumnFamilyData::CalculateSSTWriteHint(int level) {
+  if (initial_cf_options_.compaction_style != kCompactionStyleLevel) {
+    return Env::WLTH_NOT_SET;
+  }
+  if (level == 0) {
+    return Env::WLTH_MEDIUM;
+  }
+  int base_level = current_->storage_info()->base_level();
+
+  // L1: medium, L2: long, ...
+  if (level - base_level >= 2) {
+    return Env::WLTH_EXTREME;
+  }
+  return static_cast<Env::WriteLifeTimeHint>(level - base_level +
+                            static_cast<int>(Env::WLTH_MEDIUM));
+}
+
+Status ColumnFamilyData::AddDirectories() {
+  Status s;
+  assert(data_dirs_.empty());
+  for (auto& p : ioptions_.cf_paths) {
+    std::unique_ptr<Directory> path_directory;
+    s = DBImpl::CreateAndNewDirectory(ioptions_.env, p.path, &path_directory);
+    if (!s.ok()) {
+      return s;
+    }
+    assert(path_directory != nullptr);
+    data_dirs_.emplace_back(path_directory.release());
+  }
+  assert(data_dirs_.size() == ioptions_.cf_paths.size());
+  return s;
+}
+
+Directory* ColumnFamilyData::GetDataDir(size_t path_id) const {
+  if (data_dirs_.empty()) {
+    return nullptr;
+  }
+
+  assert(path_id < data_dirs_.size());
+  return data_dirs_[path_id].get();
+}
+
 ColumnFamilySet::ColumnFamilySet(const std::string& dbname,
                                  const ImmutableDBOptions* db_options,
                                  const EnvOptions& env_options,
@@ -998,10 +1218,14 @@ ColumnFamilySet::~ColumnFamilySet() {
   while (column_family_data_.size() > 0) {
     // cfd destructor will delete itself from column_family_data_
     auto cfd = column_family_data_.begin()->second;
-    cfd->Unref();
+    bool last_ref __attribute__((__unused__));
+    last_ref = cfd->Unref();
+    assert(last_ref);
     delete cfd;
   }
-  dummy_cfd_->Unref();
+  bool dummy_last_ref __attribute__((__unused__));
+  dummy_last_ref = dummy_cfd_->Unref();
+  assert(dummy_last_ref);
   delete dummy_cfd_;
 }
 

db/column_family.h

@@ -30,6 +30,7 @@ namespace rocksdb {
 
 class Version;
 class VersionSet;
+class VersionStorageInfo;
 class MemTable;
 class MemTableListVersion;
 class CompactionPicker;
@@ -41,6 +42,7 @@ class DBImpl;
 class LogBuffer;
 class InstrumentedMutex;
 class InstrumentedMutexLock;
+struct SuperVersionContext;
 
 extern const double kIncSlowdownRatio;
 
@@ -76,7 +78,7 @@ class ColumnFamilyHandleImpl : public ColumnFamilyHandle {
 class ColumnFamilyHandleInternal : public ColumnFamilyHandleImpl {
  public:
   ColumnFamilyHandleInternal()
-      : ColumnFamilyHandleImpl(nullptr, nullptr, nullptr) {}
+      : ColumnFamilyHandleImpl(nullptr, nullptr, nullptr), internal_cfd_(nullptr) {}
 
   void SetCFD(ColumnFamilyData* _cfd) { internal_cfd_ = _cfd; }
   virtual ColumnFamilyData* cfd() const override { return internal_cfd_; }
@@ -95,6 +97,7 @@ struct SuperVersion {
   MutableCFOptions mutable_cf_options;
   // Version number of the current SuperVersion
   uint64_t version_number;
+  WriteStallCondition write_stall_condition;
 
   InstrumentedMutex* db_mutex;
 
@@ -136,6 +139,9 @@ extern Status CheckCompressionSupported(const ColumnFamilyOptions& cf_options);
 extern Status CheckConcurrentWritesSupported(
     const ColumnFamilyOptions& cf_options);
 
+extern Status CheckCFPathsSupported(const DBOptions& db_options,
+                                    const ColumnFamilyOptions& cf_options);
+
 extern ColumnFamilyOptions SanitizeOptions(const ImmutableDBOptions& db_options,
                                            const ColumnFamilyOptions& src);
 // Wrap user defined table proproties collector factories `from cf_options`
@@ -192,7 +198,7 @@ class ColumnFamilyData {
   // *) delete all memory associated with that column family
   // *) delete all the files associated with that column family
   void SetDropped();
-  bool IsDropped() const { return dropped_; }
+  bool IsDropped() const { return dropped_.load(std::memory_order_relaxed); }
 
   // thread-safe
   int NumberLevels() const { return ioptions_.num_levels; }
@@ -200,6 +206,10 @@ class ColumnFamilyData {
   void SetLogNumber(uint64_t log_number) { log_number_ = log_number; }
   uint64_t GetLogNumber() const { return log_number_; }
 
+  void SetFlushReason(FlushReason flush_reason) {
+    flush_reason_ = flush_reason;
+  }
+  FlushReason GetFlushReason() const { return flush_reason_; }
   // thread-safe
   const EnvOptions* soptions() const;
   const ImmutableCFOptions* ioptions() const { return &ioptions_; }
@@ -237,7 +247,12 @@ class ColumnFamilyData {
   void SetCurrent(Version* _current);
   uint64_t GetNumLiveVersions() const;  // REQUIRE: DB mutex held
   uint64_t GetTotalSstFilesSize() const;  // REQUIRE: DB mutex held
-  void SetMemtable(MemTable* new_mem) { mem_ = new_mem; }
+  uint64_t GetLiveSstFilesSize() const;   // REQUIRE: DB mutex held
+  void SetMemtable(MemTable* new_mem) {
+    uint64_t memtable_id = last_memtable_id_.fetch_add(1) + 1;
+    new_mem->SetID(memtable_id);
+    mem_ = new_mem;
+  }
 
   // calculate the oldest log needed for the durability of this column family
   uint64_t OldestLogToKeep();
@@ -263,17 +278,27 @@ class ColumnFamilyData {
                                   const Slice& largest_user_key,
                                   int level) const;
 
+  // Check if the passed ranges overlap with any unflushed memtables
+  // (immutable or mutable).
+  //
+  // @param super_version A referenced SuperVersion that will be held for the
+  //    duration of this function.
+  //
+  // Thread-safe
+  Status RangesOverlapWithMemtables(const autovector<Range>& ranges,
+                                    SuperVersion* super_version, bool* overlap);
+
   // A flag to tell a manual compaction is to compact all levels together
-  // instad of for specific level.
+  // instead of a specific level.
   static const int kCompactAllLevels;
   // A flag to tell a manual compaction's output is base level.
   static const int kCompactToBaseLevel;
   // REQUIRES: DB mutex held
   Compaction* CompactRange(const MutableCFOptions& mutable_cf_options,
                            int input_level, int output_level,
-                           uint32_t output_path_id, const InternalKey* begin,
-                           const InternalKey* end, InternalKey** compaction_end,
-                           bool* manual_conflict);
+                           uint32_t output_path_id, uint32_t max_subcompactions,
+                           const InternalKey* begin, const InternalKey* end,
+                           InternalKey** compaction_end, bool* manual_conflict);
 
   CompactionPicker* compaction_picker() { return compaction_picker_.get(); }
   // thread-safe
@@ -311,32 +336,55 @@ class ColumnFamilyData {
   // As argument takes a pointer to allocated SuperVersion to enable
   // the clients to allocate SuperVersion outside of mutex.
   // IMPORTANT: Only call this from DBImpl::InstallSuperVersion()
-  SuperVersion* InstallSuperVersion(SuperVersion* new_superversion,
-                                    InstrumentedMutex* db_mutex,
-                                    const MutableCFOptions& mutable_cf_options);
-  SuperVersion* InstallSuperVersion(SuperVersion* new_superversion,
-                                    InstrumentedMutex* db_mutex);
+  void InstallSuperVersion(SuperVersionContext* sv_context,
+                           InstrumentedMutex* db_mutex,
+                           const MutableCFOptions& mutable_cf_options);
+  void InstallSuperVersion(SuperVersionContext* sv_context,
+                           InstrumentedMutex* db_mutex);
 
   void ResetThreadLocalSuperVersions();
 
   // Protected by DB mutex
-  void set_pending_flush(bool value) { pending_flush_ = value; }
-  void set_pending_compaction(bool value) { pending_compaction_ = value; }
-  bool pending_flush() { return pending_flush_; }
-  bool pending_compaction() { return pending_compaction_; }
+  void set_queued_for_flush(bool value) { queued_for_flush_ = value; }
+  void set_queued_for_compaction(bool value) { queued_for_compaction_ = value; }
+  bool queued_for_flush() { return queued_for_flush_; }
+  bool queued_for_compaction() { return queued_for_compaction_; }
+
+  enum class WriteStallCause {
+    kNone,
+    kMemtableLimit,
+    kL0FileCountLimit,
+    kPendingCompactionBytes,
+  };
+  static std::pair<WriteStallCondition, WriteStallCause>
+  GetWriteStallConditionAndCause(int num_unflushed_memtables, int num_l0_files,
+                                 uint64_t num_compaction_needed_bytes,
+                                 const MutableCFOptions& mutable_cf_options);
 
   // Recalculate some small conditions, which are changed only during
   // compaction, adding new memtable and/or
   // recalculation of compaction score. These values are used in
   // DBImpl::MakeRoomForWrite function to decide, if it need to make
   // a write stall
-  void RecalculateWriteStallConditions(
+  WriteStallCondition RecalculateWriteStallConditions(
       const MutableCFOptions& mutable_cf_options);
 
   void set_initialized() { initialized_.store(true); }
 
   bool initialized() const { return initialized_.load(); }
 
+  const ColumnFamilyOptions& initial_cf_options() {
+    return initial_cf_options_;
+  }
+
+  Env::WriteLifeTimeHint CalculateSSTWriteHint(int level);
+
+  Status AddDirectories();
+
+  Directory* GetDataDir(size_t path_id) const;
+
+  ThreadLocalPtr* TEST_GetLocalSV() { return local_sv_.get(); }
+
  private:
   friend class ColumnFamilySet;
   ColumnFamilyData(uint32_t id, const std::string& name,
@@ -354,7 +402,7 @@ class ColumnFamilyData {
 
   std::atomic<int> refs_;      // outstanding references to ColumnFamilyData
   std::atomic<bool> initialized_;
-  bool dropped_;               // true if client dropped it
+  std::atomic<bool> dropped_;  // true if client dropped it
 
   const InternalKeyComparator internal_comparator_;
   std::vector<std::unique_ptr<IntTblPropCollectorFactory>>
@@ -396,6 +444,8 @@ class ColumnFamilyData {
   // recovered from
   uint64_t log_number_;
 
+  std::atomic<FlushReason> flush_reason_;
+
   // An object that keeps all the compaction stats
   // and picks the next compaction
   std::unique_ptr<CompactionPicker> compaction_picker_;
@@ -405,16 +455,22 @@ class ColumnFamilyData {
   std::unique_ptr<WriteControllerToken> write_controller_token_;
 
   // If true --> this ColumnFamily is currently present in DBImpl::flush_queue_
-  bool pending_flush_;
+  bool queued_for_flush_;
 
   // If true --> this ColumnFamily is currently present in
   // DBImpl::compaction_queue_
-  bool pending_compaction_;
+  bool queued_for_compaction_;
 
   uint64_t prev_compaction_needed_bytes_;
 
   // if the database was opened with 2pc enabled
   bool allow_2pc_;
+
+  // Memtable id to track flush.
+  std::atomic<uint64_t> last_memtable_id_;
+
+  // Directories corresponding to cf_paths.
+  std::vector<std::unique_ptr<Directory>> data_dirs_;
 };
 
 // ColumnFamilySet has interesting thread-safety requirements

db/compact_files_test.cc

@@ -24,7 +24,7 @@ class CompactFilesTest : public testing::Test {
  public:
   CompactFilesTest() {
     env_ = Env::Default();
-    db_name_ = test::TmpDir(env_) + "/compact_files_test";
+    db_name_ = test::PerThreadDBPath("compact_files_test");
   }
 
   std::string db_name_;
@@ -37,8 +37,7 @@ class FlushedFileCollector : public EventListener {
   FlushedFileCollector() {}
   ~FlushedFileCollector() {}
 
-  virtual void OnFlushCompleted(
-      DB* db, const FlushJobInfo& info) override {
+  virtual void OnFlushCompleted(DB* /*db*/, const FlushJobInfo& info) override {
     std::lock_guard<std::mutex> lock(mutex_);
     flushed_files_.push_back(info.file_path);
   }
@@ -257,9 +256,9 @@ TEST_F(CompactFilesTest, CapturingPendingFiles) {
 TEST_F(CompactFilesTest, CompactionFilterWithGetSv) {
   class FilterWithGet : public CompactionFilter {
    public:
-    virtual bool Filter(int level, const Slice& key, const Slice& value,
-                        std::string* new_value,
-                        bool* value_changed) const override {
+    virtual bool Filter(int /*level*/, const Slice& /*key*/,
+                        const Slice& /*value*/, std::string* /*new_value*/,
+                        bool* /*value_changed*/) const override {
       if (db_ == nullptr) {
         return true;
       }
@@ -309,6 +308,100 @@ TEST_F(CompactFilesTest, CompactionFilterWithGetSv) {
   delete db;
 }
 
+TEST_F(CompactFilesTest, SentinelCompressionType) {
+  if (!Zlib_Supported()) {
+    fprintf(stderr, "zlib compression not supported, skip this test\n");
+    return;
+  }
+  if (!Snappy_Supported()) {
+    fprintf(stderr, "snappy compression not supported, skip this test\n");
+    return;
+  }
+  // Check that passing `CompressionType::kDisableCompressionOption` to
+  // `CompactFiles` causes it to use the column family compression options.
+  for (auto compaction_style :
+       {CompactionStyle::kCompactionStyleLevel,
+        CompactionStyle::kCompactionStyleUniversal,
+        CompactionStyle::kCompactionStyleNone}) {
+    DestroyDB(db_name_, Options());
+    Options options;
+    options.compaction_style = compaction_style;
+    // L0: Snappy, L1: ZSTD, L2: Snappy
+    options.compression_per_level = {CompressionType::kSnappyCompression,
+                                     CompressionType::kZlibCompression,
+                                     CompressionType::kSnappyCompression};
+    options.create_if_missing = true;
+    FlushedFileCollector* collector = new FlushedFileCollector();
+    options.listeners.emplace_back(collector);
+    DB* db = nullptr;
+    ASSERT_OK(DB::Open(options, db_name_, &db));
+
+    db->Put(WriteOptions(), "key", "val");
+    db->Flush(FlushOptions());
+
+    auto l0_files = collector->GetFlushedFiles();
+    ASSERT_EQ(1, l0_files.size());
+
+    // L0->L1 compaction, so output should be ZSTD-compressed
+    CompactionOptions compaction_opts;
+    compaction_opts.compression = CompressionType::kDisableCompressionOption;
+    ASSERT_OK(db->CompactFiles(compaction_opts, l0_files, 1));
+
+    rocksdb::TablePropertiesCollection all_tables_props;
+    ASSERT_OK(db->GetPropertiesOfAllTables(&all_tables_props));
+    for (const auto& name_and_table_props : all_tables_props) {
+      ASSERT_EQ(CompressionTypeToString(CompressionType::kZlibCompression),
+                name_and_table_props.second->compression_name);
+    }
+    delete db;
+  }
+}
+
+TEST_F(CompactFilesTest, GetCompactionJobInfo) {
+  Options options;
+  options.create_if_missing = true;
+  // Disable RocksDB background compaction.
+  options.compaction_style = kCompactionStyleNone;
+  options.level0_slowdown_writes_trigger = 1000;
+  options.level0_stop_writes_trigger = 1000;
+  options.write_buffer_size = 65536;
+  options.max_write_buffer_number = 2;
+  options.compression = kNoCompression;
+  options.max_compaction_bytes = 5000;
+
+  // Add listener
+  FlushedFileCollector* collector = new FlushedFileCollector();
+  options.listeners.emplace_back(collector);
+
+  DB* db = nullptr;
+  DestroyDB(db_name_, options);
+  Status s = DB::Open(options, db_name_, &db);
+  assert(s.ok());
+  assert(db);
+
+  // create couple files
+  for (int i = 0; i < 500; ++i) {
+    db->Put(WriteOptions(), ToString(i), std::string(1000, 'a' + (i % 26)));
+  }
+  reinterpret_cast<DBImpl*>(db)->TEST_WaitForFlushMemTable();
+  auto l0_files_1 = collector->GetFlushedFiles();
+  CompactionOptions co;
+  co.compression = CompressionType::kLZ4Compression;
+  CompactionJobInfo compaction_job_info;
+  ASSERT_OK(
+      db->CompactFiles(co, l0_files_1, 0, -1, nullptr, &compaction_job_info));
+  ASSERT_EQ(compaction_job_info.base_input_level, 0);
+  ASSERT_EQ(compaction_job_info.cf_id, db->DefaultColumnFamily()->GetID());
+  ASSERT_EQ(compaction_job_info.cf_name, db->DefaultColumnFamily()->GetName());
+  ASSERT_EQ(compaction_job_info.compaction_reason,
+            CompactionReason::kManualCompaction);
+  ASSERT_EQ(compaction_job_info.compression, CompressionType::kLZ4Compression);
+  ASSERT_EQ(compaction_job_info.output_level, 0);
+  ASSERT_OK(compaction_job_info.status);
+  // no assertion failure
+  delete db;
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {
@@ -319,7 +412,7 @@ int main(int argc, char** argv) {
 #else
 #include <stdio.h>
 
-int main(int argc, char** argv) {
+int main(int /*argc*/, char** /*argv*/) {
   fprintf(stderr,
           "SKIPPED as DBImpl::CompactFiles is not supported in ROCKSDB_LITE\n");
   return 0;

db/compacted_db_impl.cc

@@ -17,29 +17,20 @@ extern bool SaveValue(void* arg, const ParsedInternalKey& parsed_key,
 
 CompactedDBImpl::CompactedDBImpl(
   const DBOptions& options, const std::string& dbname)
-  : DBImpl(options, dbname) {
+  : DBImpl(options, dbname), cfd_(nullptr), version_(nullptr),
+    user_comparator_(nullptr) {
 }
 
 CompactedDBImpl::~CompactedDBImpl() {
 }
 
 size_t CompactedDBImpl::FindFile(const Slice& key) {
-  size_t left = 0;
   size_t right = files_.num_files - 1;
-  while (left < right) {
-    size_t mid = (left + right) >> 1;
-    const FdWithKeyRange& f = files_.files[mid];
-    if (user_comparator_->Compare(ExtractUserKey(f.largest_key), key) < 0) {
-      // Key at "mid.largest" is < "target".  Therefore all
-      // files at or before "mid" are uninteresting.
-      left = mid + 1;
-    } else {
-      // Key at "mid.largest" is >= "target".  Therefore all files
-      // after "mid" are uninteresting.
-      right = mid;
-    }
-  }
-  return right;
+  auto cmp = [&](const FdWithKeyRange& f, const Slice& k) -> bool {
+    return user_comparator_->Compare(ExtractUserKey(f.largest_key), k) < 0;
+  };
+  return static_cast<size_t>(std::lower_bound(files_.files,
+                            files_.files + right, key, cmp) - files_.files);
 }
 
 Status CompactedDBImpl::Get(const ReadOptions& options, ColumnFamilyHandle*,
@@ -48,8 +39,8 @@ Status CompactedDBImpl::Get(const ReadOptions& options, ColumnFamilyHandle*,
                          GetContext::kNotFound, key, value, nullptr, nullptr,
                          nullptr, nullptr);
   LookupKey lkey(key, kMaxSequenceNumber);
-  files_.files[FindFile(key)].fd.table_reader->Get(
-      options, lkey.internal_key(), &get_context);
+  files_.files[FindFile(key)].fd.table_reader->Get(options, lkey.internal_key(),
+                                                   &get_context, nullptr);
   if (get_context.State() == GetContext::kFound) {
     return Status::OK();
   }
@@ -81,7 +72,7 @@ std::vector<Status> CompactedDBImpl::MultiGet(const ReadOptions& options,
                              GetContext::kNotFound, keys[idx], &pinnable_val,
                              nullptr, nullptr, nullptr, nullptr);
       LookupKey lkey(keys[idx], kMaxSequenceNumber);
-      r->Get(options, lkey.internal_key(), &get_context);
+      r->Get(options, lkey.internal_key(), &get_context, nullptr);
       value.assign(pinnable_val.data(), pinnable_val.size());
       if (get_context.State() == GetContext::kFound) {
         statuses[idx] = Status::OK();
@@ -93,6 +84,7 @@ std::vector<Status> CompactedDBImpl::MultiGet(const ReadOptions& options,
 }
 
 Status CompactedDBImpl::Init(const Options& options) {
+  SuperVersionContext sv_context(/* create_superversion */ true);
   mutex_.Lock();
   ColumnFamilyDescriptor cf(kDefaultColumnFamilyName,
                             ColumnFamilyOptions(options));
@@ -100,9 +92,10 @@ Status CompactedDBImpl::Init(const Options& options) {
   if (s.ok()) {
     cfd_ = reinterpret_cast<ColumnFamilyHandleImpl*>(
               DefaultColumnFamily())->cfd();
-    delete cfd_->InstallSuperVersion(new SuperVersion(), &mutex_);
+    cfd_->InstallSuperVersion(&sv_context, &mutex_);
   }
   mutex_.Unlock();
+  sv_context.Clean();
   if (!s.ok()) {
     return s;
   }
@@ -154,6 +147,7 @@ Status CompactedDBImpl::Open(const Options& options,
   std::unique_ptr<CompactedDBImpl> db(new CompactedDBImpl(db_options, dbname));
   Status s = db->Init(options);
   if (s.ok()) {
+    db->StartTimedTasks();
     ROCKS_LOG_INFO(db->immutable_db_options_.info_log,
                    "Opened the db as fully compacted mode");
     LogFlush(db->immutable_db_options_.info_log);

db/compacted_db_impl.h

@@ -32,55 +32,57 @@ class CompactedDBImpl : public DBImpl {
     override;
 
   using DBImpl::Put;
-  virtual Status Put(const WriteOptions& options,
-                     ColumnFamilyHandle* column_family, const Slice& key,
-                     const Slice& value) override {
+  virtual Status Put(const WriteOptions& /*options*/,
+                     ColumnFamilyHandle* /*column_family*/,
+                     const Slice& /*key*/, const Slice& /*value*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
   using DBImpl::Merge;
-  virtual Status Merge(const WriteOptions& options,
-                       ColumnFamilyHandle* column_family, const Slice& key,
-                       const Slice& value) override {
+  virtual Status Merge(const WriteOptions& /*options*/,
+                       ColumnFamilyHandle* /*column_family*/,
+                       const Slice& /*key*/, const Slice& /*value*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
   using DBImpl::Delete;
-  virtual Status Delete(const WriteOptions& options,
-                        ColumnFamilyHandle* column_family,
-                        const Slice& key) override {
+  virtual Status Delete(const WriteOptions& /*options*/,
+                        ColumnFamilyHandle* /*column_family*/,
+                        const Slice& /*key*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
-  virtual Status Write(const WriteOptions& options,
-                       WriteBatch* updates) override {
+  virtual Status Write(const WriteOptions& /*options*/,
+                       WriteBatch* /*updates*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
   using DBImpl::CompactRange;
-  virtual Status CompactRange(const CompactRangeOptions& options,
-                              ColumnFamilyHandle* column_family,
-                              const Slice* begin, const Slice* end) override {
+  virtual Status CompactRange(const CompactRangeOptions& /*options*/,
+                              ColumnFamilyHandle* /*column_family*/,
+                              const Slice* /*begin*/,
+                              const Slice* /*end*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
 
   virtual Status DisableFileDeletions() override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
-  virtual Status EnableFileDeletions(bool force) override {
+  virtual Status EnableFileDeletions(bool /*force*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
-  virtual Status GetLiveFiles(std::vector<std::string>&,
+  virtual Status GetLiveFiles(std::vector<std::string>& ret,
                               uint64_t* manifest_file_size,
-                              bool flush_memtable = true) override {
-    return Status::NotSupported("Not supported in compacted db mode.");
+                              bool /*flush_memtable*/) override {
+    return DBImpl::GetLiveFiles(ret, manifest_file_size,
+                                false /* flush_memtable */);
   }
   using DBImpl::Flush;
-  virtual Status Flush(const FlushOptions& options,
-                       ColumnFamilyHandle* column_family) override {
+  virtual Status Flush(const FlushOptions& /*options*/,
+                       ColumnFamilyHandle* /*column_family*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
   using DB::IngestExternalFile;
   virtual Status IngestExternalFile(
-      ColumnFamilyHandle* column_family,
-      const std::vector<std::string>& external_files,
-      const IngestExternalFileOptions& ingestion_options) override {
+      ColumnFamilyHandle* /*column_family*/,
+      const std::vector<std::string>& /*external_files*/,
+      const IngestExternalFileOptions& /*ingestion_options*/) override {
     return Status::NotSupported("Not supported in compacted db mode.");
   }
 

db/compaction.cc

@@ -23,6 +23,43 @@
 
 namespace rocksdb {
 
+const uint64_t kRangeTombstoneSentinel =
+    PackSequenceAndType(kMaxSequenceNumber, kTypeRangeDeletion);
+
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey& a,
+                      const InternalKey& b) {
+  auto c = user_cmp->Compare(a.user_key(), b.user_key());
+  if (c != 0) {
+    return c;
+  }
+  auto a_footer = ExtractInternalKeyFooter(a.Encode());
+  auto b_footer = ExtractInternalKeyFooter(b.Encode());
+  if (a_footer == kRangeTombstoneSentinel) {
+    if (b_footer != kRangeTombstoneSentinel) {
+      return -1;
+    }
+  } else if (b_footer == kRangeTombstoneSentinel) {
+    return 1;
+  }
+  return 0;
+}
+
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey* a,
+                      const InternalKey& b) {
+  if (a == nullptr) {
+    return -1;
+  }
+  return sstableKeyCompare(user_cmp, *a, b);
+}
+
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey& a,
+                      const InternalKey* b) {
+  if (b == nullptr) {
+    return -1;
+  }
+  return sstableKeyCompare(user_cmp, a, *b);
+}
+
 uint64_t TotalFileSize(const std::vector<FileMetaData*>& files) {
   uint64_t sum = 0;
   for (size_t i = 0; i < files.size() && files[i]; i++) {
@@ -81,40 +118,71 @@ void Compaction::GetBoundaryKeys(
   }
 }
 
+std::vector<CompactionInputFiles> Compaction::PopulateWithAtomicBoundaries(
+    VersionStorageInfo* vstorage, std::vector<CompactionInputFiles> inputs) {
+  const Comparator* ucmp = vstorage->InternalComparator()->user_comparator();
+  for (size_t i = 0; i < inputs.size(); i++) {
+    if (inputs[i].level == 0 || inputs[i].files.empty()) {
+      continue;
+    }
+    inputs[i].atomic_compaction_unit_boundaries.reserve(inputs[i].files.size());
+    AtomicCompactionUnitBoundary cur_boundary;
+    size_t first_atomic_idx = 0;
+    auto add_unit_boundary = [&](size_t to) {
+      if (first_atomic_idx == to) return;
+      for (size_t k = first_atomic_idx; k < to; k++) {
+        inputs[i].atomic_compaction_unit_boundaries.push_back(cur_boundary);
+      }
+      first_atomic_idx = to;
+    };
+    for (size_t j = 0; j < inputs[i].files.size(); j++) {
+      const auto* f = inputs[i].files[j];
+      if (j == 0) {
+        // First file in a level.
+        cur_boundary.smallest = &f->smallest;
+        cur_boundary.largest = &f->largest;
+      } else if (sstableKeyCompare(ucmp, *cur_boundary.largest, f->smallest) ==
+                 0) {
+        // SSTs overlap but the end key of the previous file was not
+        // artificially extended by a range tombstone. Extend the current
+        // boundary.
+        cur_boundary.largest = &f->largest;
+      } else {
+        // Atomic compaction unit has ended.
+        add_unit_boundary(j);
+        cur_boundary.smallest = &f->smallest;
+        cur_boundary.largest = &f->largest;
+      }
+    }
+    add_unit_boundary(inputs[i].files.size());
+    assert(inputs[i].files.size() ==
+           inputs[i].atomic_compaction_unit_boundaries.size());
+  }
+  return inputs;
+}
+
 // helper function to determine if compaction is creating files at the
 // bottommost level
 bool Compaction::IsBottommostLevel(
     int output_level, VersionStorageInfo* vstorage,
     const std::vector<CompactionInputFiles>& inputs) {
-  if (inputs[0].level == 0 &&
-      inputs[0].files.back() != vstorage->LevelFiles(0).back()) {
-    return false;
+  int output_l0_idx;
+  if (output_level == 0) {
+    output_l0_idx = 0;
+    for (const auto* file : vstorage->LevelFiles(0)) {
+      if (inputs[0].files.back() == file) {
+        break;
+      }
+      ++output_l0_idx;
+    }
+    assert(static_cast<size_t>(output_l0_idx) < vstorage->LevelFiles(0).size());
+  } else {
+    output_l0_idx = -1;
   }
-
   Slice smallest_key, largest_key;
   GetBoundaryKeys(vstorage, inputs, &smallest_key, &largest_key);
-
-  // Checks whether there are files living beyond the output_level.
-  // If lower levels have files, it checks for overlap between files
-  // if the compaction process and those files.
-  // Bottomlevel optimizations can be made if there are no files in
-  // lower levels or if there is no overlap with the files in
-  // the lower levels.
-  for (int i = output_level + 1; i < vstorage->num_levels(); i++) {
-    // It is not the bottommost level if there are files in higher
-    // levels when the output level is 0 or if there are files in
-    // higher levels which overlap with files to be compacted.
-    // output_level == 0 means that we want it to be considered
-    // s the bottommost level only if the last file on the level
-    // is a part of the files to be compacted - this is verified by
-    // the first if condition in this function
-    if (vstorage->NumLevelFiles(i) > 0 &&
-        (output_level == 0 ||
-         vstorage->OverlapInLevel(i, &smallest_key, &largest_key))) {
-      return false;
-    }
-  }
-  return true;
+  return !vstorage->RangeMightExistAfterSortedRun(smallest_key, largest_key,
+                                                  output_level, output_l0_idx);
 }
 
 // test function to validate the functionality of IsBottommostLevel()
@@ -146,6 +214,8 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
                        int _output_level, uint64_t _target_file_size,
                        uint64_t _max_compaction_bytes, uint32_t _output_path_id,
                        CompressionType _compression,
+                       CompressionOptions _compression_opts,
+                       uint32_t _max_subcompactions,
                        std::vector<FileMetaData*> _grandparents,
                        bool _manual_compaction, double _score,
                        bool _deletion_compaction,
@@ -155,6 +225,7 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
       output_level_(_output_level),
       max_output_file_size_(_target_file_size),
       max_compaction_bytes_(_max_compaction_bytes),
+      max_subcompactions_(_max_subcompactions),
       immutable_cf_options_(_immutable_cf_options),
       mutable_cf_options_(_mutable_cf_options),
       input_version_(nullptr),
@@ -162,18 +233,23 @@ Compaction::Compaction(VersionStorageInfo* vstorage,
       cfd_(nullptr),
       output_path_id_(_output_path_id),
       output_compression_(_compression),
+      output_compression_opts_(_compression_opts),
       deletion_compaction_(_deletion_compaction),
-      inputs_(std::move(_inputs)),
+      inputs_(PopulateWithAtomicBoundaries(vstorage, std::move(_inputs))),
       grandparents_(std::move(_grandparents)),
       score_(_score),
       bottommost_level_(IsBottommostLevel(output_level_, vstorage, inputs_)),
       is_full_compaction_(IsFullCompaction(vstorage, inputs_)),
       is_manual_compaction_(_manual_compaction),
+      is_trivial_move_(false),
       compaction_reason_(_compaction_reason) {
   MarkFilesBeingCompacted(true);
   if (is_manual_compaction_) {
     compaction_reason_ = CompactionReason::kManualCompaction;
   }
+  if (max_subcompactions_ == 0) {
+    max_subcompactions_ = immutable_cf_options_.max_subcompactions;
+  }
 
 #ifndef NDEBUG
   for (size_t i = 1; i < inputs_.size(); ++i) {
@@ -240,7 +316,7 @@ bool Compaction::IsTrivialMove() const {
 
   // Used in universal compaction, where trivial move can be done if the
   // input files are non overlapping
-  if ((immutable_cf_options_.compaction_options_universal.allow_trivial_move) &&
+  if ((mutable_cf_options_.compaction_options_universal.allow_trivial_move) &&
       (output_level_ != 0)) {
     return is_trivial_move_;
   }
@@ -283,10 +359,10 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
   assert(input_version_ != nullptr);
   assert(level_ptrs != nullptr);
   assert(level_ptrs->size() == static_cast<size_t>(number_levels_));
-  if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel) {
-    if (output_level_ == 0) {
-      return false;
-    }
+  if (bottommost_level_) {
+    return true;
+  } else if (output_level_ != 0 &&
+             cfd_->ioptions()->compaction_style == kCompactionStyleLevel) {
     // Maybe use binary search to find right entry instead of linear search?
     const Comparator* user_cmp = cfd_->user_comparator();
     for (int lvl = output_level_ + 1; lvl < number_levels_; lvl++) {
@@ -297,8 +373,8 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
         if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
           // We've advanced far enough
           if (user_cmp->Compare(user_key, f->smallest.user_key()) >= 0) {
-            // Key falls in this file's range, so definitely
-            // exists beyond output level
+            // Key falls in this file's range, so it may
+            // exist beyond output level
             return false;
           }
           break;
@@ -306,9 +382,8 @@ bool Compaction::KeyNotExistsBeyondOutputLevel(
       }
     }
     return true;
-  } else {
-    return bottommost_level_;
   }
+  return false;
 }
 
 // Mark (or clear) each file that is being compacted
@@ -336,12 +411,14 @@ const char* Compaction::InputLevelSummary(
     if (!is_first) {
       len +=
           snprintf(scratch->buffer + len, sizeof(scratch->buffer) - len, " + ");
+      len = std::min(len, static_cast<int>(sizeof(scratch->buffer)));
     } else {
       is_first = false;
     }
     len += snprintf(scratch->buffer + len, sizeof(scratch->buffer) - len,
                     "%" ROCKSDB_PRIszt "@%d", input_level.size(),
                     input_level.level);
+    len = std::min(len, static_cast<int>(sizeof(scratch->buffer)));
   }
   snprintf(scratch->buffer + len, sizeof(scratch->buffer) - len,
            " files to L%d", output_level());
@@ -417,20 +494,23 @@ void Compaction::Summary(char* output, int len) {
 uint64_t Compaction::OutputFilePreallocationSize() const {
   uint64_t preallocation_size = 0;
 
-  if (max_output_file_size_ != port::kMaxUint64 &&
-      (cfd_->ioptions()->compaction_style == kCompactionStyleLevel ||
-       output_level() > 0)) {
-    preallocation_size = max_output_file_size_;
-  } else {
-    for (const auto& level_files : inputs_) {
-      for (const auto& file : level_files.files) {
-        preallocation_size += file->fd.GetFileSize();
-      }
+  for (const auto& level_files : inputs_) {
+    for (const auto& file : level_files.files) {
+      preallocation_size += file->fd.GetFileSize();
     }
   }
+
+  if (max_output_file_size_ != port::kMaxUint64 &&
+      (immutable_cf_options_.compaction_style == kCompactionStyleLevel ||
+       output_level() > 0)) {
+    preallocation_size = std::min(max_output_file_size_, preallocation_size);
+  }
+
   // Over-estimate slightly so we don't end up just barely crossing
   // the threshold
-  return preallocation_size + (preallocation_size / 10);
+  // No point to prellocate more than 1GB.
+  return std::min(uint64_t{1073741824},
+                  preallocation_size + (preallocation_size / 10));
 }
 
 std::unique_ptr<CompactionFilter> Compaction::CreateCompactionFilter() const {
@@ -451,11 +531,12 @@ bool Compaction::IsOutputLevelEmpty() const {
 }
 
 bool Compaction::ShouldFormSubcompactions() const {
-  if (immutable_cf_options_.max_subcompactions <= 1 || cfd_ == nullptr) {
+  if (max_subcompactions_ <= 1 || cfd_ == nullptr) {
     return false;
   }
   if (cfd_->ioptions()->compaction_style == kCompactionStyleLevel) {
-    return start_level_ == 0 && output_level_ > 0 && !IsOutputLevelEmpty();
+    return (start_level_ == 0 || is_manual_compaction_) && output_level_ > 0 &&
+           !IsOutputLevelEmpty();
   } else if (cfd_->ioptions()->compaction_style == kCompactionStyleUniversal) {
     return number_levels_ > 1 && output_level_ > 0;
   } else {
@@ -476,4 +557,8 @@ uint64_t Compaction::MaxInputFileCreationTime() const {
   return max_creation_time;
 }
 
+int Compaction::GetInputBaseLevel() const {
+  return input_vstorage_->base_level();
+}
+
 }  // namespace rocksdb

db/compaction.h

@@ -15,11 +15,43 @@
 
 namespace rocksdb {
 
+// Utility for comparing sstable boundary keys. Returns -1 if either a or b is
+// null which provides the property that a==null indicates a key that is less
+// than any key and b==null indicates a key that is greater than any key. Note
+// that the comparison is performed primarily on the user-key portion of the
+// key. If the user-keys compare equal, an additional test is made to sort
+// range tombstone sentinel keys before other keys with the same user-key. The
+// result is that 2 user-keys will compare equal if they differ purely on
+// their sequence number and value, but the range tombstone sentinel for that
+// user-key will compare not equal. This is necessary because the range
+// tombstone sentinel key is set as the largest key for an sstable even though
+// that key never appears in the database. We don't want adjacent sstables to
+// be considered overlapping if they are separated by the range tombstone
+// sentinel.
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey& a,
+                      const InternalKey& b);
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey* a,
+                      const InternalKey& b);
+int sstableKeyCompare(const Comparator* user_cmp, const InternalKey& a,
+                      const InternalKey* b);
+
+// An AtomicCompactionUnitBoundary represents a range of keys [smallest,
+// largest] that exactly spans one ore more neighbouring SSTs on the same
+// level. Every pair of  SSTs in this range "overlap" (i.e., the largest
+// user key of one file is the smallest user key of the next file). These
+// boundaries are propagated down to RangeDelAggregator during compaction
+// to provide safe truncation boundaries for range tombstones.
+struct AtomicCompactionUnitBoundary {
+  const InternalKey* smallest = nullptr;
+  const InternalKey* largest = nullptr;
+};
+
 // The structure that manages compaction input files associated
 // with the same physical level.
 struct CompactionInputFiles {
   int level;
   std::vector<FileMetaData*> files;
+  std::vector<AtomicCompactionUnitBoundary> atomic_compaction_unit_boundaries;
   inline bool empty() const { return files.empty(); }
   inline size_t size() const { return files.size(); }
   inline void clear() { files.clear(); }
@@ -40,6 +72,7 @@ class Compaction {
              std::vector<CompactionInputFiles> inputs, int output_level,
              uint64_t target_file_size, uint64_t max_compaction_bytes,
              uint32_t output_path_id, CompressionType compression,
+             CompressionOptions compression_opts, uint32_t max_subcompactions,
              std::vector<FileMetaData*> grandparents,
              bool manual_compaction = false, double score = -1,
              bool deletion_compaction = false,
@@ -95,6 +128,12 @@ class Compaction {
     return inputs_[compaction_input_level][i];
   }
 
+  const std::vector<AtomicCompactionUnitBoundary>* boundaries(
+      size_t compaction_input_level) const {
+    assert(compaction_input_level < inputs_.size());
+    return &inputs_[compaction_input_level].atomic_compaction_unit_boundaries;
+  }
+
   // Returns the list of file meta data of the specified compaction
   // input level.
   // REQUIREMENT: "compaction_input_level" must be >= 0 and
@@ -118,6 +157,11 @@ class Compaction {
   // What compression for output
   CompressionType output_compression() const { return output_compression_; }
 
+  // What compression options for output
+  CompressionOptions output_compression_opts() const {
+    return output_compression_opts_;
+  }
+
   // Whether need to write output file to second DB path.
   uint32_t output_path_id() const { return output_path_id_; }
 
@@ -233,6 +277,8 @@ class Compaction {
 
   Slice GetLargestUserKey() const { return largest_user_key_; }
 
+  int GetInputBaseLevel() const;
+
   CompactionReason compaction_reason() { return compaction_reason_; }
 
   const std::vector<FileMetaData*>& grandparents() const {
@@ -241,6 +287,8 @@ class Compaction {
 
   uint64_t max_compaction_bytes() const { return max_compaction_bytes_; }
 
+  uint32_t max_subcompactions() const { return max_subcompactions_; }
+
   uint64_t MaxInputFileCreationTime() const;
 
  private:
@@ -252,6 +300,13 @@ class Compaction {
                               const std::vector<CompactionInputFiles>& inputs,
                               Slice* smallest_key, Slice* largest_key);
 
+  // Get the atomic file boundaries for all files in the compaction. Necessary
+  // in order to avoid the scenario described in
+  // https://github.com/facebook/rocksdb/pull/4432#discussion_r221072219 and plumb
+  // down appropriate key boundaries to RangeDelAggregator during compaction.
+  static std::vector<CompactionInputFiles> PopulateWithAtomicBoundaries(
+      VersionStorageInfo* vstorage, std::vector<CompactionInputFiles> inputs);
+
   // helper function to determine if compaction with inputs and storage is
   // bottommost
   static bool IsBottommostLevel(
@@ -267,6 +322,7 @@ class Compaction {
   const int output_level_;  // levels to which output files are stored
   uint64_t max_output_file_size_;
   uint64_t max_compaction_bytes_;
+  uint32_t max_subcompactions_;
   const ImmutableCFOptions immutable_cf_options_;
   const MutableCFOptions mutable_cf_options_;
   Version* input_version_;
@@ -277,6 +333,7 @@ class Compaction {
 
   const uint32_t output_path_id_;
   CompressionType output_compression_;
+  CompressionOptions output_compression_opts_;
   // If true, then the comaction can be done by simply deleting input files.
   const bool deletion_compaction_;
 

db/compaction_iteration_stats.h

@@ -16,6 +16,8 @@ struct CompactionIterationStats {
   int64_t num_record_drop_obsolete = 0;
   int64_t num_record_drop_range_del = 0;
   int64_t num_range_del_drop_obsolete = 0;
+  // Deletions obsoleted before bottom level due to file gap optimization.
+  int64_t num_optimized_del_drop_obsolete = 0;
   uint64_t total_filter_time = 0;
 
   // Input statistics

db/compaction_iterator.cc

@@ -4,92 +4,104 @@
 //  (found in the LICENSE.Apache file in the root directory).
 
 #include "db/compaction_iterator.h"
+
+#include "db/snapshot_checker.h"
+#include "port/likely.h"
 #include "rocksdb/listener.h"
 #include "table/internal_iterator.h"
+#include "util/sync_point.h"
+
+#define DEFINITELY_IN_SNAPSHOT(seq, snapshot)                       \
+  ((seq) <= (snapshot) &&                                           \
+   (snapshot_checker_ == nullptr ||                                 \
+    LIKELY(snapshot_checker_->CheckInSnapshot((seq), (snapshot)) == \
+           SnapshotCheckerResult::kInSnapshot)))
+
+#define DEFINITELY_NOT_IN_SNAPSHOT(seq, snapshot)                     \
+  ((seq) > (snapshot) ||                                              \
+   (snapshot_checker_ != nullptr &&                                   \
+    UNLIKELY(snapshot_checker_->CheckInSnapshot((seq), (snapshot)) == \
+             SnapshotCheckerResult::kNotInSnapshot)))
+
+#define IN_EARLIEST_SNAPSHOT(seq) \
+  ((seq) <= earliest_snapshot_ && \
+   (snapshot_checker_ == nullptr || LIKELY(IsInEarliestSnapshot(seq))))
 
 namespace rocksdb {
 
-#ifndef ROCKSDB_LITE
-CompactionEventListener::CompactionListenerValueType fromInternalValueType(
-    ValueType vt) {
-  switch (vt) {
-    case kTypeDeletion:
-      return CompactionEventListener::CompactionListenerValueType::kDelete;
-    case kTypeValue:
-      return CompactionEventListener::CompactionListenerValueType::kValue;
-    case kTypeMerge:
-      return CompactionEventListener::CompactionListenerValueType::
-          kMergeOperand;
-    case kTypeSingleDeletion:
-      return CompactionEventListener::CompactionListenerValueType::
-          kSingleDelete;
-    case kTypeRangeDeletion:
-      return CompactionEventListener::CompactionListenerValueType::kRangeDelete;
-    default:
-      assert(false);
-      return CompactionEventListener::CompactionListenerValueType::kInvalid;
-  }
-}
-#endif  // ROCKSDB_LITE
-
 CompactionIterator::CompactionIterator(
     InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
     SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
-    SequenceNumber earliest_write_conflict_snapshot, Env* env,
-    bool expect_valid_internal_key, RangeDelAggregator* range_del_agg,
-    const Compaction* compaction, const CompactionFilter* compaction_filter,
-    CompactionEventListener* compaction_listener,
-    const std::atomic<bool>* shutting_down)
+    SequenceNumber earliest_write_conflict_snapshot,
+    const SnapshotChecker* snapshot_checker, Env* env,
+    bool report_detailed_time, bool expect_valid_internal_key,
+    CompactionRangeDelAggregator* range_del_agg, const Compaction* compaction,
+    const CompactionFilter* compaction_filter,
+    const std::atomic<bool>* shutting_down,
+    const SequenceNumber preserve_deletes_seqnum)
     : CompactionIterator(
           input, cmp, merge_helper, last_sequence, snapshots,
-          earliest_write_conflict_snapshot, env, expect_valid_internal_key,
-          range_del_agg,
+          earliest_write_conflict_snapshot, snapshot_checker, env,
+          report_detailed_time, expect_valid_internal_key, range_del_agg,
           std::unique_ptr<CompactionProxy>(
               compaction ? new CompactionProxy(compaction) : nullptr),
-          compaction_filter, compaction_listener, shutting_down) {}
+          compaction_filter, shutting_down, preserve_deletes_seqnum) {}
 
 CompactionIterator::CompactionIterator(
     InternalIterator* input, const Comparator* cmp, MergeHelper* merge_helper,
-    SequenceNumber last_sequence, std::vector<SequenceNumber>* snapshots,
-    SequenceNumber earliest_write_conflict_snapshot, Env* env,
-    bool expect_valid_internal_key, RangeDelAggregator* range_del_agg,
+    SequenceNumber /*last_sequence*/, std::vector<SequenceNumber>* snapshots,
+    SequenceNumber earliest_write_conflict_snapshot,
+    const SnapshotChecker* snapshot_checker, Env* env,
+    bool report_detailed_time, bool expect_valid_internal_key,
+    CompactionRangeDelAggregator* range_del_agg,
     std::unique_ptr<CompactionProxy> compaction,
     const CompactionFilter* compaction_filter,
-    CompactionEventListener* compaction_listener,
-    const std::atomic<bool>* shutting_down)
+    const std::atomic<bool>* shutting_down,
+    const SequenceNumber preserve_deletes_seqnum)
     : input_(input),
       cmp_(cmp),
       merge_helper_(merge_helper),
       snapshots_(snapshots),
       earliest_write_conflict_snapshot_(earliest_write_conflict_snapshot),
+      snapshot_checker_(snapshot_checker),
       env_(env),
+      report_detailed_time_(report_detailed_time),
       expect_valid_internal_key_(expect_valid_internal_key),
       range_del_agg_(range_del_agg),
       compaction_(std::move(compaction)),
       compaction_filter_(compaction_filter),
-#ifndef ROCKSDB_LITE
-      compaction_listener_(compaction_listener),
-#endif  // ROCKSDB_LITE
       shutting_down_(shutting_down),
+      preserve_deletes_seqnum_(preserve_deletes_seqnum),
       ignore_snapshots_(false),
-      merge_out_iter_(merge_helper_) {
+      current_user_key_sequence_(0),
+      current_user_key_snapshot_(0),
+      merge_out_iter_(merge_helper_),
+      current_key_committed_(false) {
   assert(compaction_filter_ == nullptr || compaction_ != nullptr);
+  assert(snapshots_ != nullptr);
   bottommost_level_ =
       compaction_ == nullptr ? false : compaction_->bottommost_level();
   if (compaction_ != nullptr) {
     level_ptrs_ = std::vector<size_t>(compaction_->number_levels(), 0);
   }
-
   if (snapshots_->size() == 0) {
     // optimize for fast path if there are no snapshots
     visible_at_tip_ = true;
+    earliest_snapshot_iter_ = snapshots_->end();
     earliest_snapshot_ = kMaxSequenceNumber;
     latest_snapshot_ = 0;
   } else {
     visible_at_tip_ = false;
+    earliest_snapshot_iter_ = snapshots_->begin();
     earliest_snapshot_ = snapshots_->at(0);
     latest_snapshot_ = snapshots_->back();
   }
+#ifndef NDEBUG
+  // findEarliestVisibleSnapshot assumes this ordering.
+  for (size_t i = 1; i < snapshots_->size(); ++i) {
+    assert(snapshots_->at(i - 1) < snapshots_->at(i));
+  }
+#endif
   if (compaction_filter_ != nullptr) {
     if (compaction_filter_->IgnoreSnapshots()) {
       ignore_snapshots_ = true;
@@ -98,6 +110,7 @@ CompactionIterator::CompactionIterator(
     ignore_snapshots_ = false;
   }
   input_->SetPinnedItersMgr(&pinned_iters_mgr_);
+  TEST_SYNC_POINT_CALLBACK("CompactionIterator:AfterInit", compaction_.get());
 }
 
 CompactionIterator::~CompactionIterator() {
@@ -111,6 +124,7 @@ void CompactionIterator::ResetRecordCounts() {
   iter_stats_.num_record_drop_obsolete = 0;
   iter_stats_.num_record_drop_range_del = 0;
   iter_stats_.num_range_del_drop_obsolete = 0;
+  iter_stats_.num_optimized_del_drop_obsolete = 0;
 }
 
 void CompactionIterator::SeekToFirst() {
@@ -128,8 +142,8 @@ void CompactionIterator::Next() {
     if (merge_out_iter_.Valid()) {
       key_ = merge_out_iter_.key();
       value_ = merge_out_iter_.value();
-      bool valid_key __attribute__((__unused__)) =
-          ParseInternalKey(key_, &ikey_);
+      bool valid_key __attribute__((__unused__));
+      valid_key =  ParseInternalKey(key_, &ikey_);
       // MergeUntil stops when it encounters a corrupt key and does not
       // include them in the result, so we expect the keys here to be valid.
       assert(valid_key);
@@ -163,6 +177,61 @@ void CompactionIterator::Next() {
   PrepareOutput();
 }
 
+void CompactionIterator::InvokeFilterIfNeeded(bool* need_skip,
+                                              Slice* skip_until) {
+  if (compaction_filter_ != nullptr &&
+      (ikey_.type == kTypeValue || ikey_.type == kTypeBlobIndex) &&
+      (visible_at_tip_ || ignore_snapshots_ ||
+       DEFINITELY_NOT_IN_SNAPSHOT(ikey_.sequence, latest_snapshot_))) {
+    // If the user has specified a compaction filter and the sequence
+    // number is greater than any external snapshot, then invoke the
+    // filter. If the return value of the compaction filter is true,
+    // replace the entry with a deletion marker.
+    CompactionFilter::Decision filter;
+    compaction_filter_value_.clear();
+    compaction_filter_skip_until_.Clear();
+    CompactionFilter::ValueType value_type =
+        ikey_.type == kTypeValue ? CompactionFilter::ValueType::kValue
+                                 : CompactionFilter::ValueType::kBlobIndex;
+    // Hack: pass internal key to BlobIndexCompactionFilter since it needs
+    // to get sequence number.
+    Slice& filter_key = ikey_.type == kTypeValue ? ikey_.user_key : key_;
+    {
+      StopWatchNano timer(env_, report_detailed_time_);
+      filter = compaction_filter_->FilterV2(
+          compaction_->level(), filter_key, value_type, value_,
+          &compaction_filter_value_, compaction_filter_skip_until_.rep());
+      iter_stats_.total_filter_time +=
+          env_ != nullptr && report_detailed_time_ ? timer.ElapsedNanos() : 0;
+    }
+
+    if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil &&
+        cmp_->Compare(*compaction_filter_skip_until_.rep(), ikey_.user_key) <=
+            0) {
+      // Can't skip to a key smaller than the current one.
+      // Keep the key as per FilterV2 documentation.
+      filter = CompactionFilter::Decision::kKeep;
+    }
+
+    if (filter == CompactionFilter::Decision::kRemove) {
+      // convert the current key to a delete; key_ is pointing into
+      // current_key_ at this point, so updating current_key_ updates key()
+      ikey_.type = kTypeDeletion;
+      current_key_.UpdateInternalKey(ikey_.sequence, kTypeDeletion);
+      // no value associated with delete
+      value_.clear();
+      iter_stats_.num_record_drop_user++;
+    } else if (filter == CompactionFilter::Decision::kChangeValue) {
+      value_ = compaction_filter_value_;
+    } else if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil) {
+      *need_skip = true;
+      compaction_filter_skip_until_.ConvertFromUserKey(kMaxSequenceNumber,
+                                                       kValueTypeForSeek);
+      *skip_until = compaction_filter_skip_until_.Encode();
+    }
+  }
+}
+
 void CompactionIterator::NextFromInput() {
   at_next_ = false;
   valid_ = false;
@@ -189,6 +258,7 @@ void CompactionIterator::NextFromInput() {
       valid_ = true;
       break;
     }
+    TEST_SYNC_POINT_CALLBACK("CompactionIterator:ProcessKV", &ikey_);
 
     // Update input statistics
     if (ikey_.type == kTypeDeletion || ikey_.type == kTypeSingleDeletion) {
@@ -217,70 +287,14 @@ void CompactionIterator::NextFromInput() {
       has_outputted_key_ = false;
       current_user_key_sequence_ = kMaxSequenceNumber;
       current_user_key_snapshot_ = 0;
+      current_key_committed_ = KeyCommitted(ikey_.sequence);
 
-#ifndef ROCKSDB_LITE
-      if (compaction_listener_) {
-        compaction_listener_->OnCompaction(compaction_->level(), ikey_.user_key,
-                                           fromInternalValueType(ikey_.type),
-                                           value_, ikey_.sequence, true);
-      }
-#endif  // ROCKSDB_LITE
-
-      // apply the compaction filter to the first occurrence of the user key
-      if (compaction_filter_ != nullptr && ikey_.type == kTypeValue &&
-          (visible_at_tip_ || ikey_.sequence > latest_snapshot_ ||
-           ignore_snapshots_)) {
-        // If the user has specified a compaction filter and the sequence
-        // number is greater than any external snapshot, then invoke the
-        // filter. If the return value of the compaction filter is true,
-        // replace the entry with a deletion marker.
-        CompactionFilter::Decision filter;
-        compaction_filter_value_.clear();
-        compaction_filter_skip_until_.Clear();
-        {
-          StopWatchNano timer(env_, true);
-          filter = compaction_filter_->FilterV2(
-              compaction_->level(), ikey_.user_key,
-              CompactionFilter::ValueType::kValue, value_,
-              &compaction_filter_value_, compaction_filter_skip_until_.rep());
-          iter_stats_.total_filter_time +=
-              env_ != nullptr ? timer.ElapsedNanos() : 0;
-        }
-
-        if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil &&
-            cmp_->Compare(*compaction_filter_skip_until_.rep(),
-                          ikey_.user_key) <= 0) {
-          // Can't skip to a key smaller than the current one.
-          // Keep the key as per FilterV2 documentation.
-          filter = CompactionFilter::Decision::kKeep;
-        }
-
-        if (filter == CompactionFilter::Decision::kRemove) {
-          // convert the current key to a delete; key_ is pointing into
-          // current_key_ at this point, so updating current_key_ updates key()
-          ikey_.type = kTypeDeletion;
-          current_key_.UpdateInternalKey(ikey_.sequence, kTypeDeletion);
-          // no value associated with delete
-          value_.clear();
-          iter_stats_.num_record_drop_user++;
-        } else if (filter == CompactionFilter::Decision::kChangeValue) {
-          value_ = compaction_filter_value_;
-        } else if (filter == CompactionFilter::Decision::kRemoveAndSkipUntil) {
-          need_skip = true;
-          compaction_filter_skip_until_.ConvertFromUserKey(kMaxSequenceNumber,
-                                                           kValueTypeForSeek);
-          skip_until = compaction_filter_skip_until_.Encode();
-        }
+      // Apply the compaction filter to the first committed version of the user
+      // key.
+      if (current_key_committed_) {
+        InvokeFilterIfNeeded(&need_skip, &skip_until);
       }
     } else {
-#ifndef ROCKSDB_LITE
-      if (compaction_listener_) {
-        compaction_listener_->OnCompaction(compaction_->level(), ikey_.user_key,
-                                           fromInternalValueType(ikey_.type),
-                                           value_, ikey_.sequence, false);
-      }
-#endif  // ROCKSDB_LITE
-
       // Update the current key to reflect the new sequence number/type without
       // copying the user key.
       // TODO(rven): Compaction filter does not process keys in this path
@@ -289,13 +303,32 @@ void CompactionIterator::NextFromInput() {
       current_key_.UpdateInternalKey(ikey_.sequence, ikey_.type);
       key_ = current_key_.GetInternalKey();
       ikey_.user_key = current_key_.GetUserKey();
+
+      // Note that newer version of a key is ordered before older versions. If a
+      // newer version of a key is committed, so as the older version. No need
+      // to query snapshot_checker_ in that case.
+      if (UNLIKELY(!current_key_committed_)) {
+        assert(snapshot_checker_ != nullptr);
+        current_key_committed_ = KeyCommitted(ikey_.sequence);
+        // Apply the compaction filter to the first committed version of the
+        // user key.
+        if (current_key_committed_) {
+          InvokeFilterIfNeeded(&need_skip, &skip_until);
+        }
+      }
+    }
+
+    if (UNLIKELY(!current_key_committed_)) {
+      assert(snapshot_checker_ != nullptr);
+      valid_ = true;
+      break;
     }
 
     // If there are no snapshots, then this kv affect visibility at tip.
     // Otherwise, search though all existing snapshots to find the earliest
     // snapshot that is affected by this kv.
-    SequenceNumber last_sequence __attribute__((__unused__)) =
-        current_user_key_sequence_;
+    SequenceNumber last_sequence __attribute__((__unused__));
+    last_sequence = current_user_key_sequence_;
     current_user_key_sequence_ = ikey_.sequence;
     SequenceNumber last_snapshot = current_user_key_snapshot_;
     SequenceNumber prev_snapshot = 0;  // 0 means no previous snapshot
@@ -360,7 +393,8 @@ void CompactionIterator::NextFromInput() {
           cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
         // Check whether the next key belongs to the same snapshot as the
         // SingleDelete.
-        if (prev_snapshot == 0 || next_ikey.sequence > prev_snapshot) {
+        if (prev_snapshot == 0 ||
+            DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot)) {
           if (next_ikey.type == kTypeSingleDeletion) {
             // We encountered two SingleDeletes in a row.  This could be due to
             // unexpected user input.
@@ -371,8 +405,9 @@ void CompactionIterator::NextFromInput() {
             // input_->Next().
             ++iter_stats_.num_record_drop_obsolete;
             ++iter_stats_.num_single_del_mismatch;
-          } else if ((ikey_.sequence <= earliest_write_conflict_snapshot_) ||
-                     has_outputted_key_) {
+          } else if (has_outputted_key_ ||
+                     DEFINITELY_IN_SNAPSHOT(
+                         ikey_.sequence, earliest_write_conflict_snapshot_)) {
             // Found a matching value, we can drop the single delete and the
             // value.  It is safe to drop both records since we've already
             // outputted a key in this snapshot, or there is no earlier
@@ -382,7 +417,8 @@ void CompactionIterator::NextFromInput() {
             // is an unexpected Merge or Delete.  We will compact it out
             // either way. We will maintain counts of how many mismatches
             // happened
-            if (next_ikey.type != kTypeValue) {
+            if (next_ikey.type != kTypeValue &&
+                next_ikey.type != kTypeBlobIndex) {
               ++iter_stats_.num_single_del_mismatch;
             }
 
@@ -419,13 +455,16 @@ void CompactionIterator::NextFromInput() {
         // iteration. If the next key is corrupt, we return before the
         // comparison, so the value of has_current_user_key does not matter.
         has_current_user_key_ = false;
-        if (compaction_ != nullptr && ikey_.sequence <= earliest_snapshot_ &&
+        if (compaction_ != nullptr && IN_EARLIEST_SNAPSHOT(ikey_.sequence) &&
             compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
                                                        &level_ptrs_)) {
           // Key doesn't exist outside of this range.
           // Can compact out this SingleDelete.
           ++iter_stats_.num_record_drop_obsolete;
           ++iter_stats_.num_single_del_fallthru;
+          if (!bottommost_level_) {
+            ++iter_stats_.num_optimized_del_drop_obsolete;
+          }
         } else {
           // Output SingleDelete
           valid_ = true;
@@ -435,21 +474,35 @@ void CompactionIterator::NextFromInput() {
       if (valid_) {
         at_next_ = true;
       }
-    } else if (last_snapshot == current_user_key_snapshot_) {
+    } else if (last_snapshot == current_user_key_snapshot_ ||
+               (last_snapshot > 0 &&
+                last_snapshot < current_user_key_snapshot_)) {
       // If the earliest snapshot is which this key is visible in
       // is the same as the visibility of a previous instance of the
       // same key, then this kv is not visible in any snapshot.
       // Hidden by an newer entry for same user key
-      // TODO(noetzli): why not > ?
       //
       // Note: Dropping this key will not affect TransactionDB write-conflict
       // checking since there has already been a record returned for this key
       // in this snapshot.
       assert(last_sequence >= current_user_key_sequence_);
+
+      // Note2: if last_snapshot < current_user_key_snapshot, it can only
+      // mean last_snapshot is released between we process last value and
+      // this value, and findEarliestVisibleSnapshot returns the next snapshot
+      // as current_user_key_snapshot. In this case last value and current
+      // value are both in current_user_key_snapshot currently.
+      assert(last_snapshot == current_user_key_snapshot_ ||
+             (snapshot_checker_ != nullptr &&
+              snapshot_checker_->CheckInSnapshot(current_user_key_sequence_,
+                                                 last_snapshot) ==
+                  SnapshotCheckerResult::kSnapshotReleased));
+
       ++iter_stats_.num_record_drop_hidden;  // (A)
       input_->Next();
     } else if (compaction_ != nullptr && ikey_.type == kTypeDeletion &&
-               ikey_.sequence <= earliest_snapshot_ &&
+               IN_EARLIEST_SNAPSHOT(ikey_.sequence) &&
+               ikeyNotNeededForIncrementalSnapshot() &&
                compaction_->KeyNotExistsBeyondOutputLevel(ikey_.user_key,
                                                           &level_ptrs_)) {
       // TODO(noetzli): This is the only place where we use compaction_
@@ -466,8 +519,38 @@ void CompactionIterator::NextFromInput() {
       //
       // Note:  Dropping this Delete will not affect TransactionDB
       // write-conflict checking since it is earlier than any snapshot.
+      //
+      // It seems that we can also drop deletion later than earliest snapshot
+      // given that:
+      // (1) The deletion is earlier than earliest_write_conflict_snapshot, and
+      // (2) No value exist earlier than the deletion.
       ++iter_stats_.num_record_drop_obsolete;
+      if (!bottommost_level_) {
+        ++iter_stats_.num_optimized_del_drop_obsolete;
+      }
       input_->Next();
+    } else if ((ikey_.type == kTypeDeletion) && bottommost_level_ &&
+               ikeyNotNeededForIncrementalSnapshot()) {
+      // Handle the case where we have a delete key at the bottom most level
+      // We can skip outputting the key iff there are no subsequent puts for this
+      // key
+      ParsedInternalKey next_ikey;
+      input_->Next();
+      // Skip over all versions of this key that happen to occur in the same snapshot
+      // range as the delete
+      while (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
+             cmp_->Equal(ikey_.user_key, next_ikey.user_key) &&
+             (prev_snapshot == 0 ||
+              DEFINITELY_NOT_IN_SNAPSHOT(next_ikey.sequence, prev_snapshot))) {
+        input_->Next();
+      }
+      // If you find you still need to output a row with this key, we need to output the
+      // delete too
+      if (input_->Valid() && ParseInternalKey(input_->key(), &next_ikey) &&
+          cmp_->Equal(ikey_.user_key, next_ikey.user_key)) {
+        valid_ = true;
+        at_next_ = true;
+      }
     } else if (ikey_.type == kTypeMerge) {
       if (!merge_helper_->HasOperator()) {
         status_ = Status::InvalidArgument(
@@ -492,8 +575,8 @@ void CompactionIterator::NextFromInput() {
         //       These will be correctly set below.
         key_ = merge_out_iter_.key();
         value_ = merge_out_iter_.value();
-        bool valid_key __attribute__((__unused__)) =
-            ParseInternalKey(key_, &ikey_);
+        bool valid_key __attribute__((__unused__));
+        valid_key = ParseInternalKey(key_, &ikey_);
         // MergeUntil stops when it encounters a corrupt key and does not
         // include them in the result, so we expect the keys here to valid.
         assert(valid_key);
@@ -517,7 +600,7 @@ void CompactionIterator::NextFromInput() {
       // 1. new user key -OR-
       // 2. different snapshot stripe
       bool should_delete = range_del_agg_->ShouldDelete(
-          key_, RangeDelAggregator::RangePositioningMode::kForwardTraversal);
+          key_, RangeDelPositioningMode::kForwardTraversal);
       if (should_delete) {
         ++iter_stats_.num_record_drop_hidden;
         ++iter_stats_.num_record_drop_range_del;
@@ -543,12 +626,15 @@ void CompactionIterator::PrepareOutput() {
   // and the earliest snapshot is larger than this seqno
   // and the userkey differs from the last userkey in compaction
   // then we can squash the seqno to zero.
-
+  //
   // This is safe for TransactionDB write-conflict checking since transactions
   // only care about sequence number larger than any active snapshots.
+  //
+  // Can we do the same for levels above bottom level as long as
+  // KeyNotExistsBeyondOutputLevel() return true?
   if ((compaction_ != nullptr && !compaction_->allow_ingest_behind()) &&
-      bottommost_level_ && valid_ && ikey_.sequence <= earliest_snapshot_ &&
-      ikey_.type != kTypeMerge &&
+      ikeyNotNeededForIncrementalSnapshot() && bottommost_level_ && valid_ &&
+      IN_EARLIEST_SNAPSHOT(ikey_.sequence) && ikey_.type != kTypeMerge &&
       !cmp_->Equal(compaction_->GetLargestUserKey(), ikey_.user_key)) {
     assert(ikey_.type != kTypeDeletion && ikey_.type != kTypeSingleDeletion);
     ikey_.sequence = 0;
@@ -559,18 +645,68 @@ void CompactionIterator::PrepareOutput() {
 inline SequenceNumber CompactionIterator::findEarliestVisibleSnapshot(
     SequenceNumber in, SequenceNumber* prev_snapshot) {
   assert(snapshots_->size());
-  SequenceNumber prev __attribute__((__unused__)) = kMaxSequenceNumber;
-  for (const auto cur : *snapshots_) {
-    assert(prev == kMaxSequenceNumber || prev <= cur);
-    if (cur >= in) {
-      *prev_snapshot = prev == kMaxSequenceNumber ? 0 : prev;
-      return cur;
-    }
-    prev = cur;
-    assert(prev < kMaxSequenceNumber);
+  auto snapshots_iter = std::lower_bound(
+      snapshots_->begin(), snapshots_->end(), in);
+  if (snapshots_iter == snapshots_->begin()) {
+    *prev_snapshot = 0;
+  } else {
+    *prev_snapshot = *std::prev(snapshots_iter);
+    assert(*prev_snapshot < in);
+  }
+  if (snapshot_checker_ == nullptr) {
+    return snapshots_iter != snapshots_->end()
+      ? *snapshots_iter : kMaxSequenceNumber;
+  }
+  bool has_released_snapshot = !released_snapshots_.empty();
+  for (; snapshots_iter != snapshots_->end(); ++snapshots_iter) {
+    auto cur = *snapshots_iter;
+    assert(in <= cur);
+    // Skip if cur is in released_snapshots.
+    if (has_released_snapshot && released_snapshots_.count(cur) > 0) {
+      continue;
+    }
+    auto res = snapshot_checker_->CheckInSnapshot(in, cur);
+    if (res == SnapshotCheckerResult::kInSnapshot) {
+      return cur;
+    } else if (res == SnapshotCheckerResult::kSnapshotReleased) {
+      released_snapshots_.insert(cur);
+    }
+    *prev_snapshot = cur;
   }
-  *prev_snapshot = prev;
   return kMaxSequenceNumber;
 }
 
+// used in 2 places - prevents deletion markers to be dropped if they may be
+// needed and disables seqnum zero-out in PrepareOutput for recent keys.
+inline bool CompactionIterator::ikeyNotNeededForIncrementalSnapshot() {
+  return (!compaction_->preserve_deletes()) ||
+         (ikey_.sequence < preserve_deletes_seqnum_);
+}
+
+bool CompactionIterator::IsInEarliestSnapshot(SequenceNumber sequence) {
+  assert(snapshot_checker_ != nullptr);
+  assert(earliest_snapshot_ == kMaxSequenceNumber ||
+         (earliest_snapshot_iter_ != snapshots_->end() &&
+          *earliest_snapshot_iter_ == earliest_snapshot_));
+  auto in_snapshot =
+      snapshot_checker_->CheckInSnapshot(sequence, earliest_snapshot_);
+  while (UNLIKELY(in_snapshot == SnapshotCheckerResult::kSnapshotReleased)) {
+    // Avoid the the current earliest_snapshot_ being return as
+    // earliest visible snapshot for the next value. So if a value's sequence
+    // is zero-ed out by PrepareOutput(), the next value will be compact out.
+    released_snapshots_.insert(earliest_snapshot_);
+    earliest_snapshot_iter_++;
+
+    if (earliest_snapshot_iter_ == snapshots_->end()) {
+      earliest_snapshot_ = kMaxSequenceNumber;
+    } else {
+      earliest_snapshot_ = *earliest_snapshot_iter_;
+    }
+    in_snapshot =
+        snapshot_checker_->CheckInSnapshot(sequence, earliest_snapshot_);
+  }
+  assert(in_snapshot != SnapshotCheckerResult::kSnapshotReleased);
+  return in_snapshot == SnapshotCheckerResult::kInSnapshot;
+}
+
 }  // namespace rocksdb

db/compaction_iterator.h

@@ -7,6 +7,7 @@
 #include <algorithm>
 #include <deque>
 #include <string>
+#include <unordered_set>
 #include <vector>
 
 #include "db/compaction.h"
@@ -14,13 +15,12 @@
 #include "db/merge_helper.h"
 #include "db/pinned_iterators_manager.h"
 #include "db/range_del_aggregator.h"
+#include "db/snapshot_checker.h"
 #include "options/cf_options.h"
 #include "rocksdb/compaction_filter.h"
 
 namespace rocksdb {
 
-class CompactionEventListener;
-
 class CompactionIterator {
  public:
   // A wrapper around Compaction. Has a much smaller interface, only what
@@ -31,7 +31,7 @@ class CompactionIterator {
         : compaction_(compaction) {}
 
     virtual ~CompactionProxy() = default;
-    virtual int level(size_t compaction_input_level = 0) const {
+    virtual int level(size_t /*compaction_input_level*/ = 0) const {
       return compaction_->level();
     }
     virtual bool KeyNotExistsBeyondOutputLevel(
@@ -48,6 +48,9 @@ class CompactionIterator {
     virtual bool allow_ingest_behind() const {
       return compaction_->immutable_cf_options()->allow_ingest_behind;
     }
+    virtual bool preserve_deletes() const {
+      return compaction_->immutable_cf_options()->preserve_deletes;
+    }
 
    protected:
     CompactionProxy() = default;
@@ -59,25 +62,27 @@ class CompactionIterator {
   CompactionIterator(InternalIterator* input, const Comparator* cmp,
                      MergeHelper* merge_helper, SequenceNumber last_sequence,
                      std::vector<SequenceNumber>* snapshots,
-                     SequenceNumber earliest_write_conflict_snapshot, Env* env,
-                     bool expect_valid_internal_key,
-                     RangeDelAggregator* range_del_agg,
+                     SequenceNumber earliest_write_conflict_snapshot,
+                     const SnapshotChecker* snapshot_checker, Env* env,
+                     bool report_detailed_time, bool expect_valid_internal_key,
+                     CompactionRangeDelAggregator* range_del_agg,
                      const Compaction* compaction = nullptr,
                      const CompactionFilter* compaction_filter = nullptr,
-                     CompactionEventListener* compaction_listener = nullptr,
-                     const std::atomic<bool>* shutting_down = nullptr);
+                     const std::atomic<bool>* shutting_down = nullptr,
+                     const SequenceNumber preserve_deletes_seqnum = 0);
 
   // Constructor with custom CompactionProxy, used for tests.
   CompactionIterator(InternalIterator* input, const Comparator* cmp,
                      MergeHelper* merge_helper, SequenceNumber last_sequence,
                      std::vector<SequenceNumber>* snapshots,
-                     SequenceNumber earliest_write_conflict_snapshot, Env* env,
-                     bool expect_valid_internal_key,
-                     RangeDelAggregator* range_del_agg,
+                     SequenceNumber earliest_write_conflict_snapshot,
+                     const SnapshotChecker* snapshot_checker, Env* env,
+                     bool report_detailed_time, bool expect_valid_internal_key,
+                     CompactionRangeDelAggregator* range_del_agg,
                      std::unique_ptr<CompactionProxy> compaction,
                      const CompactionFilter* compaction_filter = nullptr,
-                     CompactionEventListener* compaction_listener = nullptr,
-                     const std::atomic<bool>* shutting_down = nullptr);
+                     const std::atomic<bool>* shutting_down = nullptr,
+                     const SequenceNumber preserve_deletes_seqnum = 0);
 
   ~CompactionIterator();
 
@@ -111,6 +116,9 @@ class CompactionIterator {
   // compression.
   void PrepareOutput();
 
+  // Invoke compaction filter if needed.
+  void InvokeFilterIfNeeded(bool* need_skip, Slice* skip_until);
+
   // Given a sequence number, return the sequence number of the
   // earliest snapshot that this sequence number is visible in.
   // The snapshots themselves are arranged in ascending order of
@@ -120,25 +128,46 @@ class CompactionIterator {
   inline SequenceNumber findEarliestVisibleSnapshot(
       SequenceNumber in, SequenceNumber* prev_snapshot);
 
+  // Checks whether the currently seen ikey_ is needed for
+  // incremental (differential) snapshot and hence can't be dropped
+  // or seqnum be zero-ed out even if all other conditions for it are met.
+  inline bool ikeyNotNeededForIncrementalSnapshot();
+
+  inline bool KeyCommitted(SequenceNumber sequence) {
+    return snapshot_checker_ == nullptr ||
+           snapshot_checker_->CheckInSnapshot(sequence, kMaxSequenceNumber) ==
+               SnapshotCheckerResult::kInSnapshot;
+  }
+
+  bool IsInEarliestSnapshot(SequenceNumber sequence);
+
   InternalIterator* input_;
   const Comparator* cmp_;
   MergeHelper* merge_helper_;
   const std::vector<SequenceNumber>* snapshots_;
+  // List of snapshots released during compaction.
+  // findEarliestVisibleSnapshot() find them out from return of
+  // snapshot_checker, and make sure they will not be returned as
+  // earliest visible snapshot of an older value.
+  // See WritePreparedTransactionTest::ReleaseSnapshotDuringCompaction3.
+  std::unordered_set<SequenceNumber> released_snapshots_;
+  std::vector<SequenceNumber>::const_iterator earliest_snapshot_iter_;
   const SequenceNumber earliest_write_conflict_snapshot_;
+  const SnapshotChecker* const snapshot_checker_;
   Env* env_;
+  bool report_detailed_time_;
   bool expect_valid_internal_key_;
-  RangeDelAggregator* range_del_agg_;
+  CompactionRangeDelAggregator* range_del_agg_;
   std::unique_ptr<CompactionProxy> compaction_;
   const CompactionFilter* compaction_filter_;
-#ifndef ROCKSDB_LITE
-  CompactionEventListener* compaction_listener_;
-#endif  // ROCKSDB_LITE
   const std::atomic<bool>* shutting_down_;
+  const SequenceNumber preserve_deletes_seqnum_;
   bool bottommost_level_;
   bool valid_ = false;
   bool visible_at_tip_;
   SequenceNumber earliest_snapshot_;
   SequenceNumber latest_snapshot_;
+
   bool ignore_snapshots_;
 
   // State
@@ -189,6 +218,10 @@ class CompactionIterator {
   std::vector<size_t> level_ptrs_;
   CompactionIterationStats iter_stats_;
 
+  // Used to avoid purging uncommitted values. The application can specify
+  // uncommitted values by providing a SnapshotChecker object.
+  bool current_key_committed_;
+
   bool IsShuttingDown() {
     // This is a best-effort facility, so memory_order_relaxed is sufficient.
     return shutting_down_ && shutting_down_->load(std::memory_order_relaxed);

db/compaction_iterator_test.cc

@@ -9,23 +9,25 @@
 #include <vector>
 
 #include "port/port.h"
+#include "util/string_util.h"
 #include "util/testharness.h"
 #include "util/testutil.h"
+#include "utilities/merge_operators.h"
 
 namespace rocksdb {
 
 // Expects no merging attempts.
 class NoMergingMergeOp : public MergeOperator {
  public:
-  bool FullMergeV2(const MergeOperationInput& merge_in,
-                   MergeOperationOutput* merge_out) const override {
+  bool FullMergeV2(const MergeOperationInput& /*merge_in*/,
+                   MergeOperationOutput* /*merge_out*/) const override {
     ADD_FAILURE();
     return false;
   }
-  bool PartialMergeMulti(const Slice& key,
-                         const std::deque<Slice>& operand_list,
-                         std::string* new_value,
-                         Logger* logger) const override {
+  bool PartialMergeMulti(const Slice& /*key*/,
+                         const std::deque<Slice>& /*operand_list*/,
+                         std::string* /*new_value*/,
+                         Logger* /*logger*/) const override {
     ADD_FAILURE();
     return false;
   }
@@ -39,9 +41,9 @@ class NoMergingMergeOp : public MergeOperator {
 // Always returns Decition::kRemove.
 class StallingFilter : public CompactionFilter {
  public:
-  virtual Decision FilterV2(int level, const Slice& key, ValueType t,
-                            const Slice& existing_value, std::string* new_value,
-                            std::string* skip_until) const override {
+  Decision FilterV2(int /*level*/, const Slice& key, ValueType /*type*/,
+                    const Slice& /*existing_value*/, std::string* /*new_value*/,
+                    std::string* /*skip_until*/) const override {
     int k = std::atoi(key.ToString().c_str());
     last_seen.store(k);
     while (k >= stall_at.load()) {
@@ -72,6 +74,18 @@ class StallingFilter : public CompactionFilter {
   mutable std::atomic<int> last_seen{0};
 };
 
+// Compaction filter that filter out all keys.
+class FilterAllKeysCompactionFilter : public CompactionFilter {
+ public:
+  Decision FilterV2(int /*level*/, const Slice& /*key*/, ValueType /*type*/,
+                    const Slice& /*existing_value*/, std::string* /*new_value*/,
+                    std::string* /*skip_until*/) const override {
+    return Decision::kRemove;
+  }
+
+  const char* Name() const override { return "AllKeysCompactionFilter"; }
+};
+
 class LoggingForwardVectorIterator : public InternalIterator {
  public:
   struct Action {
@@ -112,7 +126,7 @@ class LoggingForwardVectorIterator : public InternalIterator {
                keys_.begin();
   }
 
-  virtual void SeekForPrev(const Slice& target) override { assert(false); }
+  virtual void SeekForPrev(const Slice& /*target*/) override { assert(false); }
 
   virtual void Next() override {
     assert(Valid());
@@ -144,71 +158,158 @@ class FakeCompaction : public CompactionIterator::CompactionProxy {
  public:
   FakeCompaction() = default;
 
-  virtual int level(size_t compaction_input_level) const { return 0; }
-  virtual bool KeyNotExistsBeyondOutputLevel(
-      const Slice& user_key, std::vector<size_t>* level_ptrs) const {
-    return key_not_exists_beyond_output_level;
+  virtual int level(size_t /*compaction_input_level*/) const override {
+    return 0;
   }
-  virtual bool bottommost_level() const { return false; }
-  virtual int number_levels() const { return 1; }
-  virtual Slice GetLargestUserKey() const {
+  virtual bool KeyNotExistsBeyondOutputLevel(
+      const Slice& /*user_key*/,
+      std::vector<size_t>* /*level_ptrs*/) const override {
+    return is_bottommost_level || key_not_exists_beyond_output_level;
+  }
+  virtual bool bottommost_level() const override { return is_bottommost_level; }
+  virtual int number_levels() const override { return 1; }
+  virtual Slice GetLargestUserKey() const override {
     return "\xff\xff\xff\xff\xff\xff\xff\xff\xff";
   }
-  virtual bool allow_ingest_behind() const { return false; }
+  virtual bool allow_ingest_behind() const override { return false; }
+
+  virtual bool preserve_deletes() const override { return false; }
 
   bool key_not_exists_beyond_output_level = false;
+
+  bool is_bottommost_level = false;
 };
 
-class CompactionIteratorTest : public testing::Test {
+// A simplifed snapshot checker which assumes each snapshot has a global
+// last visible sequence.
+class TestSnapshotChecker : public SnapshotChecker {
+ public:
+  explicit TestSnapshotChecker(
+      SequenceNumber last_committed_sequence,
+      const std::unordered_map<SequenceNumber, SequenceNumber>& snapshots = {})
+      : last_committed_sequence_(last_committed_sequence),
+        snapshots_(snapshots) {}
+
+  SnapshotCheckerResult CheckInSnapshot(
+      SequenceNumber seq, SequenceNumber snapshot_seq) const override {
+    if (snapshot_seq == kMaxSequenceNumber) {
+      return seq <= last_committed_sequence_
+                 ? SnapshotCheckerResult::kInSnapshot
+                 : SnapshotCheckerResult::kNotInSnapshot;
+    }
+    assert(snapshots_.count(snapshot_seq) > 0);
+    return seq <= snapshots_.at(snapshot_seq)
+               ? SnapshotCheckerResult::kInSnapshot
+               : SnapshotCheckerResult::kNotInSnapshot;
+  }
+
+ private:
+  SequenceNumber last_committed_sequence_;
+  // A map of valid snapshot to last visible sequence to the snapshot.
+  std::unordered_map<SequenceNumber, SequenceNumber> snapshots_;
+};
+
+// Test param:
+//   bool: whether to pass snapshot_checker to compaction iterator.
+class CompactionIteratorTest : public testing::TestWithParam<bool> {
  public:
   CompactionIteratorTest()
       : cmp_(BytewiseComparator()), icmp_(cmp_), snapshots_({}) {}
 
-  void InitIterators(const std::vector<std::string>& ks,
-                     const std::vector<std::string>& vs,
-                     const std::vector<std::string>& range_del_ks,
-                     const std::vector<std::string>& range_del_vs,
-                     SequenceNumber last_sequence,
-                     MergeOperator* merge_op = nullptr,
-                     CompactionFilter* filter = nullptr) {
-    std::unique_ptr<InternalIterator> range_del_iter(
+  void InitIterators(
+      const std::vector<std::string>& ks, const std::vector<std::string>& vs,
+      const std::vector<std::string>& range_del_ks,
+      const std::vector<std::string>& range_del_vs,
+      SequenceNumber last_sequence,
+      SequenceNumber last_committed_sequence = kMaxSequenceNumber,
+      MergeOperator* merge_op = nullptr, CompactionFilter* filter = nullptr,
+      bool bottommost_level = false,
+      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
+    std::unique_ptr<InternalIterator> unfragmented_range_del_iter(
         new test::VectorIterator(range_del_ks, range_del_vs));
-    range_del_agg_.reset(new RangeDelAggregator(icmp_, snapshots_));
-    ASSERT_OK(range_del_agg_->AddTombstones(std::move(range_del_iter)));
+    auto tombstone_list = std::make_shared<FragmentedRangeTombstoneList>(
+        std::move(unfragmented_range_del_iter), icmp_);
+    std::unique_ptr<FragmentedRangeTombstoneIterator> range_del_iter(
+        new FragmentedRangeTombstoneIterator(tombstone_list, icmp_,
+                                             kMaxSequenceNumber));
+    range_del_agg_.reset(new CompactionRangeDelAggregator(&icmp_, snapshots_));
+    range_del_agg_->AddTombstones(std::move(range_del_iter));
 
     std::unique_ptr<CompactionIterator::CompactionProxy> compaction;
-    if (filter) {
+    if (filter || bottommost_level) {
       compaction_proxy_ = new FakeCompaction();
+      compaction_proxy_->is_bottommost_level = bottommost_level;
       compaction.reset(compaction_proxy_);
     }
+    bool use_snapshot_checker = UseSnapshotChecker() || GetParam();
+    if (use_snapshot_checker || last_committed_sequence < kMaxSequenceNumber) {
+      snapshot_checker_.reset(
+          new TestSnapshotChecker(last_committed_sequence, snapshot_map_));
+    }
+    merge_helper_.reset(
+        new MergeHelper(Env::Default(), cmp_, merge_op, filter, nullptr, false,
+                        0 /*latest_snapshot*/, snapshot_checker_.get(),
+                        0 /*level*/, nullptr /*statistics*/, &shutting_down_));
 
-    merge_helper_.reset(new MergeHelper(Env::Default(), cmp_, merge_op, filter,
-                                        nullptr, false, 0, 0, nullptr,
-                                        &shutting_down_));
     iter_.reset(new LoggingForwardVectorIterator(ks, vs));
     iter_->SeekToFirst();
     c_iter_.reset(new CompactionIterator(
         iter_.get(), cmp_, merge_helper_.get(), last_sequence, &snapshots_,
-        kMaxSequenceNumber, Env::Default(), false, range_del_agg_.get(),
-        std::move(compaction), filter, nullptr, &shutting_down_));
+        earliest_write_conflict_snapshot, snapshot_checker_.get(),
+        Env::Default(), false /* report_detailed_time */, false,
+        range_del_agg_.get(), std::move(compaction), filter, &shutting_down_));
   }
 
-  void AddSnapshot(SequenceNumber snapshot) { snapshots_.push_back(snapshot); }
+  void AddSnapshot(SequenceNumber snapshot,
+                   SequenceNumber last_visible_seq = kMaxSequenceNumber) {
+    snapshots_.push_back(snapshot);
+    snapshot_map_[snapshot] = last_visible_seq;
+  }
+
+  virtual bool UseSnapshotChecker() const { return false; }
+
+  void RunTest(
+      const std::vector<std::string>& input_keys,
+      const std::vector<std::string>& input_values,
+      const std::vector<std::string>& expected_keys,
+      const std::vector<std::string>& expected_values,
+      SequenceNumber last_committed_seq = kMaxSequenceNumber,
+      MergeOperator* merge_operator = nullptr,
+      CompactionFilter* compaction_filter = nullptr,
+      bool bottommost_level = false,
+      SequenceNumber earliest_write_conflict_snapshot = kMaxSequenceNumber) {
+    InitIterators(input_keys, input_values, {}, {}, kMaxSequenceNumber,
+                  last_committed_seq, merge_operator, compaction_filter,
+                  bottommost_level, earliest_write_conflict_snapshot);
+    c_iter_->SeekToFirst();
+    for (size_t i = 0; i < expected_keys.size(); i++) {
+      std::string info = "i = " + ToString(i);
+      ASSERT_TRUE(c_iter_->Valid()) << info;
+      ASSERT_OK(c_iter_->status()) << info;
+      ASSERT_EQ(expected_keys[i], c_iter_->key().ToString()) << info;
+      ASSERT_EQ(expected_values[i], c_iter_->value().ToString()) << info;
+      c_iter_->Next();
+    }
+    ASSERT_FALSE(c_iter_->Valid());
+  }
 
   const Comparator* cmp_;
   const InternalKeyComparator icmp_;
   std::vector<SequenceNumber> snapshots_;
+  // A map of valid snapshot to last visible sequence to the snapshot.
+  std::unordered_map<SequenceNumber, SequenceNumber> snapshot_map_;
   std::unique_ptr<MergeHelper> merge_helper_;
   std::unique_ptr<LoggingForwardVectorIterator> iter_;
   std::unique_ptr<CompactionIterator> c_iter_;
-  std::unique_ptr<RangeDelAggregator> range_del_agg_;
+  std::unique_ptr<CompactionRangeDelAggregator> range_del_agg_;
+  std::unique_ptr<SnapshotChecker> snapshot_checker_;
   std::atomic<bool> shutting_down_{false};
   FakeCompaction* compaction_proxy_;
 };
 
 // It is possible that the output of the compaction iterator is empty even if
 // the input is not.
-TEST_F(CompactionIteratorTest, EmptyResult) {
+TEST_P(CompactionIteratorTest, EmptyResult) {
   InitIterators({test::KeyStr("a", 5, kTypeSingleDeletion),
                  test::KeyStr("a", 3, kTypeValue)},
                 {"", "val"}, {}, {}, 5);
@@ -218,7 +319,7 @@ TEST_F(CompactionIteratorTest, EmptyResult) {
 
 // If there is a corruption after a single deletion, the corrupted key should
 // be preserved.
-TEST_F(CompactionIteratorTest, CorruptionAfterSingleDeletion) {
+TEST_P(CompactionIteratorTest, CorruptionAfterSingleDeletion) {
   InitIterators({test::KeyStr("a", 5, kTypeSingleDeletion),
                  test::KeyStr("a", 3, kTypeValue, true),
                  test::KeyStr("b", 10, kTypeValue)},
@@ -237,7 +338,7 @@ TEST_F(CompactionIteratorTest, CorruptionAfterSingleDeletion) {
   ASSERT_FALSE(c_iter_->Valid());
 }
 
-TEST_F(CompactionIteratorTest, SimpleRangeDeletion) {
+TEST_P(CompactionIteratorTest, SimpleRangeDeletion) {
   InitIterators({test::KeyStr("morning", 5, kTypeValue),
                  test::KeyStr("morning", 2, kTypeValue),
                  test::KeyStr("night", 3, kTypeValue)},
@@ -253,7 +354,7 @@ TEST_F(CompactionIteratorTest, SimpleRangeDeletion) {
   ASSERT_FALSE(c_iter_->Valid());
 }
 
-TEST_F(CompactionIteratorTest, RangeDeletionWithSnapshots) {
+TEST_P(CompactionIteratorTest, RangeDeletionWithSnapshots) {
   AddSnapshot(10);
   std::vector<std::string> ks1;
   ks1.push_back(test::KeyStr("ma", 28, kTypeRangeDeletion));
@@ -274,11 +375,11 @@ TEST_F(CompactionIteratorTest, RangeDeletionWithSnapshots) {
   ASSERT_FALSE(c_iter_->Valid());
 }
 
-TEST_F(CompactionIteratorTest, CompactionFilterSkipUntil) {
+TEST_P(CompactionIteratorTest, CompactionFilterSkipUntil) {
   class Filter : public CompactionFilter {
-    virtual Decision FilterV2(int level, const Slice& key, ValueType t,
+    virtual Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
                               const Slice& existing_value,
-                              std::string* new_value,
+                              std::string* /*new_value*/,
                               std::string* skip_until) const override {
       std::string k = key.ToString();
       std::string v = existing_value.ToString();
@@ -349,7 +450,7 @@ TEST_F(CompactionIteratorTest, CompactionFilterSkipUntil) {
        test::KeyStr("j", 99, kTypeValue)},
       {"av50", "am45", "bv60", "bv40", "cv35", "dm70", "em71", "fm65", "fm30",
        "fv25", "gv90", "hv91", "im95", "jv99"},
-      {}, {}, kMaxSequenceNumber, &merge_op, &filter);
+      {}, {}, kMaxSequenceNumber, kMaxSequenceNumber, &merge_op, &filter);
 
   // Compaction should output just "a", "e" and "h" keys.
   c_iter_->SeekToFirst();
@@ -384,13 +485,14 @@ TEST_F(CompactionIteratorTest, CompactionFilterSkipUntil) {
   ASSERT_EQ(expected_actions, iter_->log);
 }
 
-TEST_F(CompactionIteratorTest, ShuttingDownInFilter) {
+TEST_P(CompactionIteratorTest, ShuttingDownInFilter) {
   NoMergingMergeOp merge_op;
   StallingFilter filter;
   InitIterators(
       {test::KeyStr("1", 1, kTypeValue), test::KeyStr("2", 2, kTypeValue),
        test::KeyStr("3", 3, kTypeValue), test::KeyStr("4", 4, kTypeValue)},
-      {"v1", "v2", "v3", "v4"}, {}, {}, kMaxSequenceNumber, &merge_op, &filter);
+      {"v1", "v2", "v3", "v4"}, {}, {}, kMaxSequenceNumber, kMaxSequenceNumber,
+      &merge_op, &filter);
   // Don't leave tombstones (kTypeDeletion) for filtered keys.
   compaction_proxy_->key_not_exists_beyond_output_level = true;
 
@@ -421,13 +523,14 @@ TEST_F(CompactionIteratorTest, ShuttingDownInFilter) {
 
 // Same as ShuttingDownInFilter, but shutdown happens during filter call for
 // a merge operand, not for a value.
-TEST_F(CompactionIteratorTest, ShuttingDownInMerge) {
+TEST_P(CompactionIteratorTest, ShuttingDownInMerge) {
   NoMergingMergeOp merge_op;
   StallingFilter filter;
   InitIterators(
       {test::KeyStr("1", 1, kTypeValue), test::KeyStr("2", 2, kTypeMerge),
        test::KeyStr("3", 3, kTypeMerge), test::KeyStr("4", 4, kTypeValue)},
-      {"v1", "v2", "v3", "v4"}, {}, {}, kMaxSequenceNumber, &merge_op, &filter);
+      {"v1", "v2", "v3", "v4"}, {}, {}, kMaxSequenceNumber, kMaxSequenceNumber,
+      &merge_op, &filter);
   compaction_proxy_->key_not_exists_beyond_output_level = true;
 
   std::atomic<bool> seek_done{false};
@@ -455,6 +558,420 @@ TEST_F(CompactionIteratorTest, ShuttingDownInMerge) {
   EXPECT_EQ(2, filter.last_seen.load());
 }
 
+TEST_P(CompactionIteratorTest, SingleMergeOperand) {
+  class Filter : public CompactionFilter {
+    virtual Decision FilterV2(int /*level*/, const Slice& key, ValueType t,
+                              const Slice& existing_value,
+                              std::string* /*new_value*/,
+                              std::string* /*skip_until*/) const override {
+      std::string k = key.ToString();
+      std::string v = existing_value.ToString();
+
+      // See InitIterators() call below for the sequence of keys and their
+      // filtering decisions. Here we closely assert that compaction filter is
+      // called with the expected keys and only them, and with the right values.
+      if (k == "a") {
+        EXPECT_EQ(ValueType::kMergeOperand, t);
+        EXPECT_EQ("av1", v);
+        return Decision::kKeep;
+      } else if (k == "b") {
+        EXPECT_EQ(ValueType::kMergeOperand, t);
+        return Decision::kKeep;
+      } else if (k == "c") {
+        return Decision::kKeep;
+      }
+
+      ADD_FAILURE();
+      return Decision::kKeep;
+    }
+
+    const char* Name() const override {
+      return "CompactionIteratorTest.SingleMergeOperand::Filter";
+    }
+  };
+
+  class SingleMergeOp : public MergeOperator {
+   public:
+    bool FullMergeV2(const MergeOperationInput& merge_in,
+                     MergeOperationOutput* merge_out) const override {
+      // See InitIterators() call below for why "c" is the only key for which
+      // FullMergeV2 should be called.
+      EXPECT_EQ("c", merge_in.key.ToString());
+
+      std::string temp_value;
+      if (merge_in.existing_value != nullptr) {
+        temp_value = merge_in.existing_value->ToString();
+      }
+
+      for (auto& operand : merge_in.operand_list) {
+        temp_value.append(operand.ToString());
+      }
+      merge_out->new_value = temp_value;
+
+      return true;
+    }
+
+    bool PartialMergeMulti(const Slice& key,
+                           const std::deque<Slice>& operand_list,
+                           std::string* new_value,
+                           Logger* /*logger*/) const override {
+      std::string string_key = key.ToString();
+      EXPECT_TRUE(string_key == "a" || string_key == "b");
+
+      if (string_key == "a") {
+        EXPECT_EQ(1, operand_list.size());
+      } else if (string_key == "b") {
+        EXPECT_EQ(2, operand_list.size());
+      }
+
+      std::string temp_value;
+      for (auto& operand : operand_list) {
+        temp_value.append(operand.ToString());
+      }
+      swap(temp_value, *new_value);
+
+      return true;
+    }
+
+    const char* Name() const override {
+      return "CompactionIteratorTest SingleMergeOp";
+    }
+
+    bool AllowSingleOperand() const override { return true; }
+  };
+
+  SingleMergeOp merge_op;
+  Filter filter;
+  InitIterators(
+      // a should invoke PartialMergeMulti with a single merge operand.
+      {test::KeyStr("a", 50, kTypeMerge),
+       // b should invoke PartialMergeMulti with two operands.
+       test::KeyStr("b", 70, kTypeMerge), test::KeyStr("b", 60, kTypeMerge),
+       // c should invoke FullMerge due to kTypeValue at the beginning.
+       test::KeyStr("c", 90, kTypeMerge), test::KeyStr("c", 80, kTypeValue)},
+      {"av1", "bv2", "bv1", "cv2", "cv1"}, {}, {}, kMaxSequenceNumber,
+      kMaxSequenceNumber, &merge_op, &filter);
+
+  c_iter_->SeekToFirst();
+  ASSERT_TRUE(c_iter_->Valid());
+  ASSERT_EQ(test::KeyStr("a", 50, kTypeMerge), c_iter_->key().ToString());
+  ASSERT_EQ("av1", c_iter_->value().ToString());
+  c_iter_->Next();
+  ASSERT_TRUE(c_iter_->Valid());
+  ASSERT_EQ("bv1bv2", c_iter_->value().ToString());
+  c_iter_->Next();
+  ASSERT_EQ("cv1cv2", c_iter_->value().ToString());
+}
+
+// In bottommost level, values earlier than earliest snapshot can be output
+// with sequence = 0.
+TEST_P(CompactionIteratorTest, ZeroOutSequenceAtBottomLevel) {
+  AddSnapshot(1);
+  RunTest({test::KeyStr("a", 1, kTypeValue), test::KeyStr("b", 2, kTypeValue)},
+          {"v1", "v2"},
+          {test::KeyStr("a", 0, kTypeValue), test::KeyStr("b", 2, kTypeValue)},
+          {"v1", "v2"}, kMaxSequenceNumber /*last_commited_seq*/,
+          nullptr /*merge_operator*/, nullptr /*compaction_filter*/,
+          true /*bottommost_level*/);
+}
+
+// In bottommost level, deletions earlier than earliest snapshot can be removed
+// permanently.
+TEST_P(CompactionIteratorTest, RemoveDeletionAtBottomLevel) {
+  AddSnapshot(1);
+  RunTest({test::KeyStr("a", 1, kTypeDeletion),
+           test::KeyStr("b", 3, kTypeDeletion),
+           test::KeyStr("b", 1, kTypeValue)},
+          {"", "", ""},
+          {test::KeyStr("b", 3, kTypeDeletion),
+           test::KeyStr("b", 0, kTypeValue)},
+          {"", ""},
+          kMaxSequenceNumber /*last_commited_seq*/, nullptr /*merge_operator*/,
+          nullptr /*compaction_filter*/, true /*bottommost_level*/);
+}
+
+// In bottommost level, single deletions earlier than earliest snapshot can be
+// removed permanently.
+TEST_P(CompactionIteratorTest, RemoveSingleDeletionAtBottomLevel) {
+  AddSnapshot(1);
+  RunTest({test::KeyStr("a", 1, kTypeSingleDeletion),
+           test::KeyStr("b", 2, kTypeSingleDeletion)},
+          {"", ""}, {test::KeyStr("b", 2, kTypeSingleDeletion)}, {""},
+          kMaxSequenceNumber /*last_commited_seq*/, nullptr /*merge_operator*/,
+          nullptr /*compaction_filter*/, true /*bottommost_level*/);
+}
+
+INSTANTIATE_TEST_CASE_P(CompactionIteratorTestInstance, CompactionIteratorTest,
+                        testing::Values(true, false));
+
+// Tests how CompactionIterator work together with SnapshotChecker.
+class CompactionIteratorWithSnapshotCheckerTest
+    : public CompactionIteratorTest {
+ public:
+  bool UseSnapshotChecker() const override { return true; }
+};
+
+// Uncommitted keys (keys with seq > last_committed_seq) should be output as-is
+// while committed version of these keys should get compacted as usual.
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       PreserveUncommittedKeys_Value) {
+  RunTest(
+      {test::KeyStr("foo", 3, kTypeValue), test::KeyStr("foo", 2, kTypeValue),
+       test::KeyStr("foo", 1, kTypeValue)},
+      {"v3", "v2", "v1"},
+      {test::KeyStr("foo", 3, kTypeValue), test::KeyStr("foo", 2, kTypeValue)},
+      {"v3", "v2"}, 2 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       PreserveUncommittedKeys_Deletion) {
+  RunTest({test::KeyStr("foo", 2, kTypeDeletion),
+           test::KeyStr("foo", 1, kTypeValue)},
+          {"", "v1"},
+          {test::KeyStr("foo", 2, kTypeDeletion),
+           test::KeyStr("foo", 1, kTypeValue)},
+          {"", "v1"}, 1 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       PreserveUncommittedKeys_Merge) {
+  auto merge_op = MergeOperators::CreateStringAppendOperator();
+  RunTest(
+      {test::KeyStr("foo", 3, kTypeMerge), test::KeyStr("foo", 2, kTypeMerge),
+       test::KeyStr("foo", 1, kTypeValue)},
+      {"v3", "v2", "v1"},
+      {test::KeyStr("foo", 3, kTypeMerge), test::KeyStr("foo", 2, kTypeValue)},
+      {"v3", "v1,v2"}, 2 /*last_committed_seq*/, merge_op.get());
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       PreserveUncommittedKeys_SingleDelete) {
+  RunTest({test::KeyStr("foo", 2, kTypeSingleDeletion),
+           test::KeyStr("foo", 1, kTypeValue)},
+          {"", "v1"},
+          {test::KeyStr("foo", 2, kTypeSingleDeletion),
+           test::KeyStr("foo", 1, kTypeValue)},
+          {"", "v1"}, 1 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       PreserveUncommittedKeys_BlobIndex) {
+  RunTest({test::KeyStr("foo", 3, kTypeBlobIndex),
+           test::KeyStr("foo", 2, kTypeBlobIndex),
+           test::KeyStr("foo", 1, kTypeBlobIndex)},
+          {"v3", "v2", "v1"},
+          {test::KeyStr("foo", 3, kTypeBlobIndex),
+           test::KeyStr("foo", 2, kTypeBlobIndex)},
+          {"v3", "v2"}, 2 /*last_committed_seq*/);
+}
+
+// Test compaction iterator dedup keys visible to the same snapshot.
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, DedupSameSnapshot_Value) {
+  AddSnapshot(2, 1);
+  RunTest(
+      {test::KeyStr("foo", 4, kTypeValue), test::KeyStr("foo", 3, kTypeValue),
+       test::KeyStr("foo", 2, kTypeValue), test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "v3", "v2", "v1"},
+      {test::KeyStr("foo", 4, kTypeValue), test::KeyStr("foo", 3, kTypeValue),
+       test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "v3", "v1"}, 3 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, DedupSameSnapshot_Deletion) {
+  AddSnapshot(2, 1);
+  RunTest(
+      {test::KeyStr("foo", 4, kTypeValue),
+       test::KeyStr("foo", 3, kTypeDeletion),
+       test::KeyStr("foo", 2, kTypeValue), test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "", "v2", "v1"},
+      {test::KeyStr("foo", 4, kTypeValue),
+       test::KeyStr("foo", 3, kTypeDeletion),
+       test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "", "v1"}, 3 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, DedupSameSnapshot_Merge) {
+  AddSnapshot(2, 1);
+  AddSnapshot(4, 3);
+  auto merge_op = MergeOperators::CreateStringAppendOperator();
+  RunTest(
+      {test::KeyStr("foo", 5, kTypeMerge), test::KeyStr("foo", 4, kTypeMerge),
+       test::KeyStr("foo", 3, kTypeMerge), test::KeyStr("foo", 2, kTypeMerge),
+       test::KeyStr("foo", 1, kTypeValue)},
+      {"v5", "v4", "v3", "v2", "v1"},
+      {test::KeyStr("foo", 5, kTypeMerge), test::KeyStr("foo", 4, kTypeMerge),
+       test::KeyStr("foo", 3, kTypeMerge), test::KeyStr("foo", 1, kTypeValue)},
+      {"v5", "v4", "v2,v3", "v1"}, 4 /*last_committed_seq*/, merge_op.get());
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       DedupSameSnapshot_SingleDeletion) {
+  AddSnapshot(2, 1);
+  RunTest(
+      {test::KeyStr("foo", 4, kTypeValue),
+       test::KeyStr("foo", 3, kTypeSingleDeletion),
+       test::KeyStr("foo", 2, kTypeValue), test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "", "v2", "v1"},
+      {test::KeyStr("foo", 4, kTypeValue), test::KeyStr("foo", 1, kTypeValue)},
+      {"v4", "v1"}, 3 /*last_committed_seq*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, DedupSameSnapshot_BlobIndex) {
+  AddSnapshot(2, 1);
+  RunTest({test::KeyStr("foo", 4, kTypeBlobIndex),
+           test::KeyStr("foo", 3, kTypeBlobIndex),
+           test::KeyStr("foo", 2, kTypeBlobIndex),
+           test::KeyStr("foo", 1, kTypeBlobIndex)},
+          {"v4", "v3", "v2", "v1"},
+          {test::KeyStr("foo", 4, kTypeBlobIndex),
+           test::KeyStr("foo", 3, kTypeBlobIndex),
+           test::KeyStr("foo", 1, kTypeBlobIndex)},
+          {"v4", "v3", "v1"}, 3 /*last_committed_seq*/);
+}
+
+// At bottom level, sequence numbers can be zero out, and deletions can be
+// removed, but only when they are visible to earliest snapshot.
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       NotZeroOutSequenceIfNotVisibleToEarliestSnapshot) {
+  AddSnapshot(2, 1);
+  RunTest({test::KeyStr("a", 1, kTypeValue), test::KeyStr("b", 2, kTypeValue),
+           test::KeyStr("c", 3, kTypeValue)},
+          {"v1", "v2", "v3"},
+          {test::KeyStr("a", 0, kTypeValue), test::KeyStr("b", 2, kTypeValue),
+           test::KeyStr("c", 3, kTypeValue)},
+          {"v1", "v2", "v3"}, kMaxSequenceNumber /*last_commited_seq*/,
+          nullptr /*merge_operator*/, nullptr /*compaction_filter*/,
+          true /*bottommost_level*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       NotRemoveDeletionIfNotVisibleToEarliestSnapshot) {
+  AddSnapshot(2, 1);
+  RunTest(
+      {test::KeyStr("a", 1, kTypeDeletion), test::KeyStr("b", 2, kTypeDeletion),
+       test::KeyStr("c", 3, kTypeDeletion)},
+      {"", "", ""},
+      {},
+      {"", ""}, kMaxSequenceNumber /*last_commited_seq*/,
+      nullptr /*merge_operator*/, nullptr /*compaction_filter*/,
+      true /*bottommost_level*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       NotRemoveDeletionIfValuePresentToEarlierSnapshot) {
+  AddSnapshot(2,1);
+  RunTest(
+      {test::KeyStr("a", 4, kTypeDeletion), test::KeyStr("a", 1, kTypeValue),
+          test::KeyStr("b", 3, kTypeValue)},
+      {"", "", ""},
+      {test::KeyStr("a", 4, kTypeDeletion), test::KeyStr("a", 0, kTypeValue),
+            test::KeyStr("b", 3, kTypeValue)},
+      {"", "", ""}, kMaxSequenceNumber /*last_commited_seq*/,
+      nullptr /*merge_operator*/, nullptr /*compaction_filter*/,
+      true /*bottommost_level*/);
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       NotRemoveSingleDeletionIfNotVisibleToEarliestSnapshot) {
+  AddSnapshot(2, 1);
+  RunTest({test::KeyStr("a", 1, kTypeSingleDeletion),
+           test::KeyStr("b", 2, kTypeSingleDeletion),
+           test::KeyStr("c", 3, kTypeSingleDeletion)},
+          {"", "", ""},
+          {test::KeyStr("b", 2, kTypeSingleDeletion),
+           test::KeyStr("c", 3, kTypeSingleDeletion)},
+          {"", ""}, kMaxSequenceNumber /*last_commited_seq*/,
+          nullptr /*merge_operator*/, nullptr /*compaction_filter*/,
+          true /*bottommost_level*/);
+}
+
+// Single delete should not cancel out values that not visible to the
+// same set of snapshots
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       SingleDeleteAcrossSnapshotBoundary) {
+  AddSnapshot(2, 1);
+  RunTest({test::KeyStr("a", 2, kTypeSingleDeletion),
+           test::KeyStr("a", 1, kTypeValue)},
+          {"", "v1"},
+          {test::KeyStr("a", 2, kTypeSingleDeletion),
+           test::KeyStr("a", 1, kTypeValue)},
+          {"", "v1"}, 2 /*last_committed_seq*/);
+}
+
+// Single delete should be kept in case it is not visible to the
+// earliest write conflict snapshot. If a single delete is kept for this reason,
+// corresponding value can be trimmed to save space.
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       KeepSingleDeletionForWriteConflictChecking) {
+  AddSnapshot(2, 0);
+  RunTest({test::KeyStr("a", 2, kTypeSingleDeletion),
+           test::KeyStr("a", 1, kTypeValue)},
+          {"", "v1"},
+          {test::KeyStr("a", 2, kTypeSingleDeletion),
+           test::KeyStr("a", 1, kTypeValue)},
+          {"", ""}, 2 /*last_committed_seq*/, nullptr /*merge_operator*/,
+          nullptr /*compaction_filter*/, false /*bottommost_level*/,
+          2 /*earliest_write_conflict_snapshot*/);
+}
+
+// Compaction filter should keep uncommitted key as-is, and
+//   * Convert the latest velue to deletion, and/or
+//   * if latest value is a merge, apply filter to all suequent merges.
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, CompactionFilter_Value) {
+  std::unique_ptr<CompactionFilter> compaction_filter(
+      new FilterAllKeysCompactionFilter());
+  RunTest(
+      {test::KeyStr("a", 2, kTypeValue), test::KeyStr("a", 1, kTypeValue),
+       test::KeyStr("b", 3, kTypeValue), test::KeyStr("c", 1, kTypeValue)},
+      {"v2", "v1", "v3", "v4"},
+      {test::KeyStr("a", 2, kTypeValue), test::KeyStr("a", 1, kTypeDeletion),
+       test::KeyStr("b", 3, kTypeValue), test::KeyStr("c", 1, kTypeDeletion)},
+      {"v2", "", "v3", ""}, 1 /*last_committed_seq*/,
+      nullptr /*merge_operator*/, compaction_filter.get());
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, CompactionFilter_Deletion) {
+  std::unique_ptr<CompactionFilter> compaction_filter(
+      new FilterAllKeysCompactionFilter());
+  RunTest(
+      {test::KeyStr("a", 2, kTypeDeletion), test::KeyStr("a", 1, kTypeValue)},
+      {"", "v1"},
+      {test::KeyStr("a", 2, kTypeDeletion),
+       test::KeyStr("a", 1, kTypeDeletion)},
+      {"", ""}, 1 /*last_committed_seq*/, nullptr /*merge_operator*/,
+      compaction_filter.get());
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest,
+       CompactionFilter_PartialMerge) {
+  std::shared_ptr<MergeOperator> merge_op =
+      MergeOperators::CreateStringAppendOperator();
+  std::unique_ptr<CompactionFilter> compaction_filter(
+      new FilterAllKeysCompactionFilter());
+  RunTest({test::KeyStr("a", 3, kTypeMerge), test::KeyStr("a", 2, kTypeMerge),
+           test::KeyStr("a", 1, kTypeMerge)},
+          {"v3", "v2", "v1"}, {test::KeyStr("a", 3, kTypeMerge)}, {"v3"},
+          2 /*last_committed_seq*/, merge_op.get(), compaction_filter.get());
+}
+
+TEST_F(CompactionIteratorWithSnapshotCheckerTest, CompactionFilter_FullMerge) {
+  std::shared_ptr<MergeOperator> merge_op =
+      MergeOperators::CreateStringAppendOperator();
+  std::unique_ptr<CompactionFilter> compaction_filter(
+      new FilterAllKeysCompactionFilter());
+  RunTest(
+      {test::KeyStr("a", 3, kTypeMerge), test::KeyStr("a", 2, kTypeMerge),
+       test::KeyStr("a", 1, kTypeValue)},
+      {"v3", "v2", "v1"},
+      {test::KeyStr("a", 3, kTypeMerge), test::KeyStr("a", 1, kTypeDeletion)},
+      {"v3", ""}, 2 /*last_committed_seq*/, merge_op.get(),
+      compaction_filter.get());
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/compaction_job.h

@@ -29,6 +29,7 @@
 #include "db/version_edit.h"
 #include "db/write_controller.h"
 #include "db/write_thread.h"
+#include "options/cf_options.h"
 #include "options/db_options.h"
 #include "port/port.h"
 #include "rocksdb/compaction_filter.h"
@@ -45,24 +46,29 @@
 
 namespace rocksdb {
 
+class Arena;
+class ErrorHandler;
 class MemTable;
+class SnapshotChecker;
 class TableCache;
 class Version;
 class VersionEdit;
 class VersionSet;
-class Arena;
 
 class CompactionJob {
  public:
   CompactionJob(int job_id, Compaction* compaction,
                 const ImmutableDBOptions& db_options,
-                const EnvOptions& env_options, VersionSet* versions,
-                const std::atomic<bool>* shutting_down, LogBuffer* log_buffer,
+                const EnvOptions env_options, VersionSet* versions,
+                const std::atomic<bool>* shutting_down,
+                const SequenceNumber preserve_deletes_seqnum,
+                LogBuffer* log_buffer,
                 Directory* db_directory, Directory* output_directory,
                 Statistics* stats, InstrumentedMutex* db_mutex,
-                Status* db_bg_error,
+                ErrorHandler* db_error_handler,
                 std::vector<SequenceNumber> existing_snapshots,
                 SequenceNumber earliest_write_conflict_snapshot,
+                const SnapshotChecker* snapshot_checker,
                 std::shared_ptr<Cache> table_cache, EventLogger* event_logger,
                 bool paranoid_file_checks, bool measure_io_stats,
                 const std::string& dbname,
@@ -98,7 +104,7 @@ class CompactionJob {
 
   Status FinishCompactionOutputFile(
       const Status& input_status, SubcompactionState* sub_compact,
-      RangeDelAggregator* range_del_agg,
+      CompactionRangeDelAggregator* range_del_agg,
       CompactionIterationStats* range_del_out_stats,
       const Slice* next_table_min_key = nullptr);
   Status InstallCompactionResults(const MutableCFOptions& mutable_cf_options);
@@ -127,17 +133,20 @@ class CompactionJob {
   // DBImpl state
   const std::string& dbname_;
   const ImmutableDBOptions& db_options_;
-  const EnvOptions& env_options_;
+  const EnvOptions env_options_;
 
   Env* env_;
+  // env_option optimized for compaction table reads
+  EnvOptions env_optiosn_for_read_;
   VersionSet* versions_;
   const std::atomic<bool>* shutting_down_;
+  const SequenceNumber preserve_deletes_seqnum_;
   LogBuffer* log_buffer_;
   Directory* db_directory_;
   Directory* output_directory_;
   Statistics* stats_;
   InstrumentedMutex* db_mutex_;
-  Status* db_bg_error_;
+  ErrorHandler* db_error_handler_;
   // If there were two snapshots with seq numbers s1 and
   // s2 and s1 < s2, and if we find two instances of a key k1 then lies
   // entirely within s1 and s2, then the earlier version of k1 can be safely
@@ -149,6 +158,8 @@ class CompactionJob {
   // should make sure not to remove evidence that a write occurred.
   SequenceNumber earliest_write_conflict_snapshot_;
 
+  const SnapshotChecker* const snapshot_checker_;
+
   std::shared_ptr<Cache> table_cache_;
 
   EventLogger* event_logger_;
@@ -160,6 +171,7 @@ class CompactionJob {
   std::vector<Slice> boundaries_;
   // Stores the approx size of keys covered in the range of each subcompaction
   std::vector<uint64_t> sizes_;
+  Env::WriteLifeTimeHint write_hint_;
 };
 
 }  // namespace rocksdb

db/compaction_job_stats_test.cc

@@ -98,7 +98,7 @@ class CompactionJobStatsTest : public testing::Test,
   CompactionJobStatsTest() : env_(Env::Default()) {
     env_->SetBackgroundThreads(1, Env::LOW);
     env_->SetBackgroundThreads(1, Env::HIGH);
-    dbname_ = test::TmpDir(env_) + "/compaction_job_stats_test";
+    dbname_ = test::PerThreadDBPath("compaction_job_stats_test");
     alternative_wal_dir_ = dbname_ + "/wal";
     Options options;
     options.create_if_missing = true;
@@ -426,7 +426,7 @@ class CompactionJobStatsChecker : public EventListener {
   // Once a compaction completed, this function will verify the returned
   // CompactionJobInfo with the oldest CompactionJobInfo added earlier
   // in "expected_stats_" which has not yet being used for verification.
-  virtual void OnCompactionCompleted(DB *db, const CompactionJobInfo& ci) {
+  virtual void OnCompactionCompleted(DB* /*db*/, const CompactionJobInfo& ci) {
     if (verify_next_comp_io_stats_) {
       ASSERT_GT(ci.stats.file_write_nanos, 0);
       ASSERT_GT(ci.stats.file_range_sync_nanos, 0);
@@ -806,7 +806,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
     stats_checker->set_verify_next_comp_io_stats(true);
     std::atomic<bool> first_prepare_write(true);
     rocksdb::SyncPoint::GetInstance()->SetCallBack(
-        "WritableFileWriter::Append:BeforePrepareWrite", [&](void* arg) {
+        "WritableFileWriter::Append:BeforePrepareWrite", [&](void* /*arg*/) {
           if (first_prepare_write.load()) {
             options.env->SleepForMicroseconds(3);
             first_prepare_write.store(false);
@@ -815,7 +815,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
 
     std::atomic<bool> first_flush(true);
     rocksdb::SyncPoint::GetInstance()->SetCallBack(
-        "WritableFileWriter::Flush:BeforeAppend", [&](void* arg) {
+        "WritableFileWriter::Flush:BeforeAppend", [&](void* /*arg*/) {
           if (first_flush.load()) {
             options.env->SleepForMicroseconds(3);
             first_flush.store(false);
@@ -824,7 +824,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
 
     std::atomic<bool> first_sync(true);
     rocksdb::SyncPoint::GetInstance()->SetCallBack(
-        "WritableFileWriter::SyncInternal:0", [&](void* arg) {
+        "WritableFileWriter::SyncInternal:0", [&](void* /*arg*/) {
           if (first_sync.load()) {
             options.env->SleepForMicroseconds(3);
             first_sync.store(false);
@@ -833,7 +833,7 @@ TEST_P(CompactionJobStatsTest, CompactionJobStatsTest) {
 
     std::atomic<bool> first_range_sync(true);
     rocksdb::SyncPoint::GetInstance()->SetCallBack(
-        "WritableFileWriter::RangeSync:0", [&](void* arg) {
+        "WritableFileWriter::RangeSync:0", [&](void* /*arg*/) {
           if (first_range_sync.load()) {
             options.env->SleepForMicroseconds(3);
             first_range_sync.store(false);
@@ -1034,7 +1034,7 @@ int main(int argc, char** argv) {
 #else
 #include <stdio.h>
 
-int main(int argc, char** argv) {
+int main(int /*argc*/, char** /*argv*/) {
   fprintf(stderr, "SKIPPED, not supported in ROCKSDB_LITE\n");
   return 0;
 }
@@ -1043,5 +1043,5 @@ int main(int argc, char** argv) {
 
 #else
 
-int main(int argc, char** argv) { return 0; }
+int main(int /*argc*/, char** /*argv*/) { return 0; }
 #endif  // !defined(IOS_CROSS_COMPILE)

db/compaction_job_test.cc

@@ -12,6 +12,7 @@
 
 #include "db/column_family.h"
 #include "db/compaction_job.h"
+#include "db/error_handler.h"
 #include "db/version_set.h"
 #include "rocksdb/cache.h"
 #include "rocksdb/db.h"
@@ -67,7 +68,7 @@ class CompactionJobTest : public testing::Test {
  public:
   CompactionJobTest()
       : env_(Env::Default()),
-        dbname_(test::TmpDir() + "/compaction_job_test"),
+        dbname_(test::PerThreadDBPath("compaction_job_test")),
         db_options_(),
         mutable_cf_options_(cf_options_),
         table_cache_(NewLRUCache(50000, 16)),
@@ -76,7 +77,9 @@ class CompactionJobTest : public testing::Test {
                                  table_cache_.get(), &write_buffer_manager_,
                                  &write_controller_)),
         shutting_down_(false),
-        mock_table_factory_(new mock::MockTableFactory()) {
+        preserve_deletes_seqnum_(0),
+        mock_table_factory_(new mock::MockTableFactory()),
+        error_handler_(nullptr, db_options_, &mutex_) {
     EXPECT_OK(env_->CreateDirIfMissing(dbname_));
     db_options_.db_paths.emplace_back(dbname_,
                                       std::numeric_limits<uint64_t>::max());
@@ -142,7 +145,8 @@ class CompactionJobTest : public testing::Test {
   }
 
   void SetLastSequence(const SequenceNumber sequence_number) {
-    versions_->SetLastToBeWrittenSequence(sequence_number + 1);
+    versions_->SetLastAllocatedSequence(sequence_number + 1);
+    versions_->SetLastPublishedSequence(sequence_number + 1);
     versions_->SetLastSequence(sequence_number + 1);
   }
 
@@ -196,12 +200,12 @@ class CompactionJobTest : public testing::Test {
     new_db.SetLastSequence(0);
 
     const std::string manifest = DescriptorFileName(dbname_, 1);
-    unique_ptr<WritableFile> file;
+    std::unique_ptr<WritableFile> file;
     Status s = env_->NewWritableFile(
         manifest, &file, env_->OptimizeForManifestWrite(env_options_));
     ASSERT_OK(s);
-    unique_ptr<WritableFileWriter> file_writer(
-        new WritableFileWriter(std::move(file), env_options_));
+    std::unique_ptr<WritableFileWriter> file_writer(
+        new WritableFileWriter(std::move(file), manifest, env_options_));
     {
       log::Writer log(std::move(file_writer), 0, false);
       std::string record;
@@ -244,18 +248,21 @@ class CompactionJobTest : public testing::Test {
     Compaction compaction(cfd->current()->storage_info(), *cfd->ioptions(),
                           *cfd->GetLatestMutableCFOptions(),
                           compaction_input_files, 1, 1024 * 1024,
-                          10 * 1024 * 1024, 0, kNoCompression, {}, true);
+                          10 * 1024 * 1024, 0, kNoCompression,
+                          cfd->ioptions()->compression_opts, 0, {}, true);
     compaction.SetInputVersion(cfd->current());
 
     LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, db_options_.info_log.get());
     mutex_.Lock();
     EventLogger event_logger(db_options_.info_log.get());
+    // TODO(yiwu) add a mock snapshot checker and add test for it.
+    SnapshotChecker* snapshot_checker = nullptr;
     CompactionJob compaction_job(
         0, &compaction, db_options_, env_options_, versions_.get(),
-        &shutting_down_, &log_buffer, nullptr, nullptr, nullptr, &mutex_,
-        &bg_error_, snapshots, earliest_write_conflict_snapshot, table_cache_,
+        &shutting_down_, preserve_deletes_seqnum_, &log_buffer, nullptr,
+        nullptr, nullptr, &mutex_, &error_handler_, snapshots,
+        earliest_write_conflict_snapshot, snapshot_checker, table_cache_,
         &event_logger, false, false, dbname_, &compaction_job_stats_);
-
     VerifyInitializationOfCompactionJobStats(compaction_job_stats_);
 
     compaction_job.Prepare();
@@ -291,12 +298,13 @@ class CompactionJobTest : public testing::Test {
   std::unique_ptr<VersionSet> versions_;
   InstrumentedMutex mutex_;
   std::atomic<bool> shutting_down_;
+  SequenceNumber preserve_deletes_seqnum_;
   std::shared_ptr<mock::MockTableFactory> mock_table_factory_;
   CompactionJobStats compaction_job_stats_;
   ColumnFamilyData* cfd_;
   std::unique_ptr<CompactionFilter> compaction_filter_;
   std::shared_ptr<MergeOperator> merge_op_;
-  Status bg_error_;
+  ErrorHandler error_handler_;
 };
 
 TEST_F(CompactionJobTest, Simple) {
@@ -448,8 +456,7 @@ TEST_F(CompactionJobTest, NonAssocMerge) {
 
   auto expected_results =
       mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), "3,4,5"},
-                          {KeyStr("b", 2U, kTypeMerge), "2"},
-                          {KeyStr("b", 1U, kTypeMerge), "1"}});
+                          {KeyStr("b", 2U, kTypeValue), "1,2"}});
 
   SetLastSequence(5U);
   auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -476,7 +483,7 @@ TEST_F(CompactionJobTest, MergeOperandFilter) {
 
   auto expected_results =
       mock::MakeMockFile({{KeyStr("a", 0U, kTypeValue), test::EncodeInt(8U)},
-                          {KeyStr("b", 2U, kTypeMerge), test::EncodeInt(2U)}});
+                          {KeyStr("b", 2U, kTypeValue), test::EncodeInt(2U)}});
 
   SetLastSequence(5U);
   auto files = cfd_->current()->storage_info()->LevelFiles(0);
@@ -940,7 +947,7 @@ int main(int argc, char** argv) {
 #else
 #include <stdio.h>
 
-int main(int argc, char** argv) {
+int main(int /*argc*/, char** /*argv*/) {
   fprintf(stderr,
           "SKIPPED as CompactionJobStats is not supported in ROCKSDB_LITE\n");
   return 0;

db/compaction_picker.cc

@@ -18,6 +18,7 @@
 #include <queue>
 #include <string>
 #include <utility>
+#include <vector>
 #include "db/column_family.h"
 #include "monitoring/statistics.h"
 #include "util/filename.h"
@@ -36,12 +37,13 @@ uint64_t TotalCompensatedFileSize(const std::vector<FileMetaData*>& files) {
   }
   return sum;
 }
+}  // anonymous namespace
 
 bool FindIntraL0Compaction(const std::vector<FileMetaData*>& level_files,
                            size_t min_files_to_compact,
                            uint64_t max_compact_bytes_per_del_file,
                            CompactionInputFiles* comp_inputs) {
-  size_t compact_bytes = level_files[0]->fd.file_size;
+  size_t compact_bytes = static_cast<size_t>(level_files[0]->fd.file_size);
   size_t compact_bytes_per_del_file = port::kMaxSizet;
   // compaction range will be [0, span_len).
   size_t span_len;
@@ -49,7 +51,7 @@ bool FindIntraL0Compaction(const std::vector<FileMetaData*>& level_files,
   // increasing.
   size_t new_compact_bytes_per_del_file = 0;
   for (span_len = 1; span_len < level_files.size(); ++span_len) {
-    compact_bytes += level_files[span_len]->fd.file_size;
+    compact_bytes += static_cast<size_t>(level_files[span_len]->fd.file_size);
     new_compact_bytes_per_del_file = compact_bytes / span_len;
     if (level_files[span_len]->being_compacted ||
         new_compact_bytes_per_del_file > compact_bytes_per_del_file) {
@@ -59,7 +61,7 @@ bool FindIntraL0Compaction(const std::vector<FileMetaData*>& level_files,
   }
 
   if (span_len >= min_files_to_compact &&
-      new_compact_bytes_per_del_file < max_compact_bytes_per_del_file) {
+      compact_bytes_per_del_file < max_compact_bytes_per_del_file) {
     assert(comp_inputs != nullptr);
     comp_inputs->level = 0;
     for (size_t i = 0; i < span_len; ++i) {
@@ -69,7 +71,6 @@ bool FindIntraL0Compaction(const std::vector<FileMetaData*>& level_files,
   }
   return false;
 }
-}  // anonymous namespace
 
 // Determine compression type, based on user options, level of the output
 // file and whether compression is disabled.
@@ -89,7 +90,7 @@ CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
   // If bottommost_compression is set and we are compacting to the
   // bottommost level then we should use it.
   if (ioptions.bottommost_compression != kDisableCompressionOption &&
-      level > base_level && level >= (vstorage->num_non_empty_levels() - 1)) {
+      level >= (vstorage->num_non_empty_levels() - 1)) {
     return ioptions.bottommost_compression;
   }
   // If the user has specified a different compression level for each level,
@@ -110,6 +111,24 @@ CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
   }
 }
 
+CompressionOptions GetCompressionOptions(const ImmutableCFOptions& ioptions,
+                                         const VersionStorageInfo* vstorage,
+                                         int level,
+                                         const bool enable_compression) {
+  if (!enable_compression) {
+    return ioptions.compression_opts;
+  }
+  // If bottommost_compression is set and we are compacting to the
+  // bottommost level then we should use the specified compression options
+  // for the bottmomost_compression.
+  if (ioptions.bottommost_compression != kDisableCompressionOption &&
+      level >= (vstorage->num_non_empty_levels() - 1) &&
+      ioptions.bottommost_compression_opts.enabled) {
+    return ioptions.bottommost_compression_opts;
+  }
+  return ioptions.compression_opts;
+}
+
 CompactionPicker::CompactionPicker(const ImmutableCFOptions& ioptions,
                                    const InternalKeyComparator* icmp)
     : ioptions_(ioptions), icmp_(icmp) {}
@@ -199,9 +218,10 @@ void CompactionPicker::GetRange(const std::vector<CompactionInputFiles>& inputs,
   assert(initialized);
 }
 
-bool CompactionPicker::ExpandInputsToCleanCut(const std::string& cf_name,
+bool CompactionPicker::ExpandInputsToCleanCut(const std::string& /*cf_name*/,
                                               VersionStorageInfo* vstorage,
-                                              CompactionInputFiles* inputs) {
+                                              CompactionInputFiles* inputs,
+                                              InternalKey** next_smallest) {
   // This isn't good compaction
   assert(!inputs->empty());
 
@@ -224,7 +244,8 @@ bool CompactionPicker::ExpandInputsToCleanCut(const std::string& cf_name,
     GetRange(*inputs, &smallest, &largest);
     inputs->clear();
     vstorage->GetOverlappingInputs(level, &smallest, &largest, &inputs->files,
-                                   hint_index, &hint_index);
+                                   hint_index, &hint_index, true,
+                                   next_smallest);
   } while (inputs->size() > old_size);
 
   // we started off with inputs non-empty and the previous loop only grew
@@ -234,11 +255,6 @@ bool CompactionPicker::ExpandInputsToCleanCut(const std::string& cf_name,
   // If, after the expansion, there are files that are already under
   // compaction, then we must drop/cancel this compaction.
   if (AreFilesInCompaction(inputs->files)) {
-    ROCKS_LOG_WARN(
-        ioptions_.info_log,
-        "[%s] ExpandWhileOverlapping() failure because some of the necessary"
-        " compaction input files are currently being compacted.",
-        cf_name.c_str());
     return false;
   }
   return true;
@@ -297,25 +313,34 @@ Compaction* CompactionPicker::CompactFiles(
     VersionStorageInfo* vstorage, const MutableCFOptions& mutable_cf_options,
     uint32_t output_path_id) {
   assert(input_files.size());
+  // This compaction output should not overlap with a running compaction as
+  // `SanitizeCompactionInputFiles` should've checked earlier and db mutex
+  // shouldn't have been released since.
+  assert(!FilesRangeOverlapWithCompaction(input_files, output_level));
 
-  // TODO(rven ): we might be able to run concurrent level 0 compaction
-  // if the key ranges of the two compactions do not overlap, but for now
-  // we do not allow it.
-  if ((input_files[0].level == 0) && !level0_compactions_in_progress_.empty()) {
-    return nullptr;
+  CompressionType compression_type;
+  if (compact_options.compression == kDisableCompressionOption) {
+    int base_level;
+    if (ioptions_.compaction_style == kCompactionStyleLevel) {
+      base_level = vstorage->base_level();
+    } else {
+      base_level = 1;
+    }
+    compression_type =
+        GetCompressionType(ioptions_, vstorage, mutable_cf_options,
+                           output_level, base_level);
+  } else {
+    // TODO(ajkr): `CompactionOptions` offers configurable `CompressionType`
+    // without configurable `CompressionOptions`, which is inconsistent.
+    compression_type = compact_options.compression;
   }
-  // This compaction output could overlap with a running compaction
-  if (FilesRangeOverlapWithCompaction(input_files, output_level)) {
-    return nullptr;
-  }
-  auto c =
-      new Compaction(vstorage, ioptions_, mutable_cf_options, input_files,
-                     output_level, compact_options.output_file_size_limit,
-                     mutable_cf_options.max_compaction_bytes, output_path_id,
-                     compact_options.compression, /* grandparents */ {}, true);
-
-  // If it's level 0 compaction, make sure we don't execute any other level 0
-  // compactions in parallel
+  auto c = new Compaction(
+      vstorage, ioptions_, mutable_cf_options, input_files, output_level,
+      compact_options.output_file_size_limit,
+      mutable_cf_options.max_compaction_bytes, output_path_id, compression_type,
+      GetCompressionOptions(ioptions_, vstorage, output_level),
+      compact_options.max_subcompactions,
+      /* grandparents */ {}, true);
   RegisterCompaction(c);
   return c;
 }
@@ -323,7 +348,7 @@ Compaction* CompactionPicker::CompactFiles(
 Status CompactionPicker::GetCompactionInputsFromFileNumbers(
     std::vector<CompactionInputFiles>* input_files,
     std::unordered_set<uint64_t>* input_set, const VersionStorageInfo* vstorage,
-    const CompactionOptions& compact_options) const {
+    const CompactionOptions& /*compact_options*/) const {
   if (input_set->size() == 0U) {
     return Status::InvalidArgument(
         "Compaction must include at least one file.");
@@ -378,7 +403,7 @@ bool CompactionPicker::IsRangeInCompaction(VersionStorageInfo* vstorage,
   assert(level < NumberLevels());
 
   vstorage->GetOverlappingInputs(level, smallest, largest, &inputs,
-                                 *level_index, level_index);
+                                 level_index ? *level_index : 0, level_index);
   return AreFilesInCompaction(inputs);
 }
 
@@ -401,7 +426,10 @@ bool CompactionPicker::SetupOtherInputs(
   assert(output_level_inputs->empty());
   const int input_level = inputs->level;
   const int output_level = output_level_inputs->level;
-  assert(input_level != output_level);
+  if (input_level == output_level) {
+    // no possibility of conflict
+    return true;
+  }
 
   // For now, we only support merging two levels, start level and output level.
   // We need to assert other levels are empty.
@@ -486,7 +514,7 @@ bool CompactionPicker::SetupOtherInputs(
       ROCKS_LOG_INFO(ioptions_.info_log,
                      "[%s] Expanding@%d %" ROCKSDB_PRIszt "+%" ROCKSDB_PRIszt
                      "(%" PRIu64 "+%" PRIu64 " bytes) to %" ROCKSDB_PRIszt
-                     "+%" ROCKSDB_PRIszt " (%" PRIu64 "+%" PRIu64 "bytes)\n",
+                     "+%" ROCKSDB_PRIszt " (%" PRIu64 "+%" PRIu64 " bytes)\n",
                      cf_name.c_str(), input_level, inputs->size(),
                      output_level_inputs->size(), inputs_size,
                      output_level_inputs_size, expanded_inputs.size(),
@@ -515,7 +543,8 @@ void CompactionPicker::GetGrandparents(
 Compaction* CompactionPicker::CompactRange(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
     VersionStorageInfo* vstorage, int input_level, int output_level,
-    uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
+    uint32_t output_path_id, uint32_t max_subcompactions,
+    const InternalKey* begin, const InternalKey* end,
     InternalKey** compaction_end, bool* manual_conflict) {
   // CompactionPickerFIFO has its own implementation of compact range
   assert(ioptions_.compaction_style != kCompactionStyleFIFO);
@@ -577,11 +606,14 @@ Compaction* CompactionPicker::CompactRange(
 
     Compaction* c = new Compaction(
         vstorage, ioptions_, mutable_cf_options, std::move(inputs),
-        output_level, mutable_cf_options.MaxFileSizeForLevel(output_level),
+        output_level,
+        MaxFileSizeForLevel(mutable_cf_options, output_level,
+                            ioptions_.compaction_style),
         /* max_compaction_bytes */ LLONG_MAX, output_path_id,
         GetCompressionType(ioptions_, vstorage, mutable_cf_options,
                            output_level, 1),
-        /* grandparents */ {}, /* is manual */ true);
+        GetCompressionOptions(ioptions_, vstorage, output_level),
+        max_subcompactions, /* grandparents */ {}, /* is manual */ true);
     RegisterCompaction(c);
     return c;
   }
@@ -620,16 +652,18 @@ Compaction* CompactionPicker::CompactRange(
       uint64_t s = inputs[i]->compensated_file_size;
       total += s;
       if (total >= limit) {
-        **compaction_end = inputs[i + 1]->smallest;
         covering_the_whole_range = false;
         inputs.files.resize(i + 1);
         break;
       }
     }
   }
-  assert(output_path_id < static_cast<uint32_t>(ioptions_.db_paths.size()));
+  assert(output_path_id < static_cast<uint32_t>(ioptions_.cf_paths.size()));
 
-  if (ExpandInputsToCleanCut(cf_name, vstorage, &inputs) == false) {
+  InternalKey key_storage;
+  InternalKey* next_smallest = &key_storage;
+  if (ExpandInputsToCleanCut(cf_name, vstorage, &inputs, &next_smallest) ==
+      false) {
     // manual compaction is now multi-threaded, so it can
     // happen that ExpandWhileOverlapping fails
     // we handle it higher in RunManualCompaction
@@ -637,8 +671,10 @@ Compaction* CompactionPicker::CompactRange(
     return nullptr;
   }
 
-  if (covering_the_whole_range) {
+  if (covering_the_whole_range || !next_smallest) {
     *compaction_end = nullptr;
+  } else {
+    **compaction_end = *next_smallest;
   }
 
   CompactionInputFiles output_level_inputs;
@@ -684,11 +720,16 @@ Compaction* CompactionPicker::CompactRange(
   GetGrandparents(vstorage, inputs, output_level_inputs, &grandparents);
   Compaction* compaction = new Compaction(
       vstorage, ioptions_, mutable_cf_options, std::move(compaction_inputs),
-      output_level, mutable_cf_options.MaxFileSizeForLevel(output_level),
+      output_level,
+      MaxFileSizeForLevel(mutable_cf_options, output_level,
+                          ioptions_.compaction_style, vstorage->base_level(),
+                          ioptions_.level_compaction_dynamic_level_bytes),
       mutable_cf_options.max_compaction_bytes, output_path_id,
       GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
                          vstorage->base_level()),
-      std::move(grandparents), /* is manual compaction */ true);
+      GetCompressionOptions(ioptions_, vstorage, output_level),
+      max_subcompactions, std::move(grandparents),
+      /* is manual compaction */ true);
 
   TEST_SYNC_POINT_CALLBACK("CompactionPicker::CompactRange:Return", compaction);
   RegisterCompaction(compaction);
@@ -735,10 +776,6 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
   auto& levels = cf_meta.levels;
   auto comparator = icmp_->user_comparator();
 
-  // TODO(yhchiang): If there is any input files of L1 or up and there
-  // is at least one L0 files. All L0 files older than the L0 file needs
-  // to be included. Otherwise, it is a false conditoin
-
   // TODO(yhchiang): add is_adjustable to CompactionOptions
 
   // the smallest and largest key of the current compaction input
@@ -799,6 +836,8 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
         }
         last_included++;
       }
+    } else if (output_level > 0) {
+      last_included = static_cast<int>(current_files.size() - 1);
     }
 
     // include all files between the first and the last compaction input files.
@@ -858,6 +897,11 @@ Status CompactionPicker::SanitizeCompactionInputFilesForAllLevels(
       }
     }
   }
+  if (RangeOverlapWithCompaction(smallestkey, largestkey, output_level)) {
+    return Status::Aborted(
+        "A running compaction is writing to the same output level in an "
+        "overlapping key range");
+  }
   return Status::OK();
 }
 
@@ -900,8 +944,8 @@ Status CompactionPicker::SanitizeCompactionInputFiles(
   // any currently-existing files.
   for (auto file_num : *input_files) {
     bool found = false;
-    for (auto level_meta : cf_meta.levels) {
-      for (auto file_meta : level_meta.files) {
+    for (const auto& level_meta : cf_meta.levels) {
+      for (const auto& file_meta : level_meta.files) {
         if (file_num == TableFileNameToNumber(file_meta.name)) {
           if (file_meta.being_compacted) {
             return Status::Aborted("Specified compaction input file " +
@@ -952,8 +996,86 @@ void CompactionPicker::UnregisterCompaction(Compaction* c) {
   compactions_in_progress_.erase(c);
 }
 
+void CompactionPicker::PickFilesMarkedForCompaction(
+    const std::string& cf_name, VersionStorageInfo* vstorage, int* start_level,
+    int* output_level, CompactionInputFiles* start_level_inputs) {
+  if (vstorage->FilesMarkedForCompaction().empty()) {
+    return;
+  }
+
+  auto continuation = [&, cf_name](std::pair<int, FileMetaData*> level_file) {
+    // If it's being compacted it has nothing to do here.
+    // If this assert() fails that means that some function marked some
+    // files as being_compacted, but didn't call ComputeCompactionScore()
+    assert(!level_file.second->being_compacted);
+    *start_level = level_file.first;
+    *output_level =
+        (*start_level == 0) ? vstorage->base_level() : *start_level + 1;
+
+    if (*start_level == 0 && !level0_compactions_in_progress()->empty()) {
+      return false;
+    }
+
+    start_level_inputs->files = {level_file.second};
+    start_level_inputs->level = *start_level;
+    return ExpandInputsToCleanCut(cf_name, vstorage, start_level_inputs);
+  };
+
+  // take a chance on a random file first
+  Random64 rnd(/* seed */ reinterpret_cast<uint64_t>(vstorage));
+  size_t random_file_index = static_cast<size_t>(rnd.Uniform(
+      static_cast<uint64_t>(vstorage->FilesMarkedForCompaction().size())));
+
+  if (continuation(vstorage->FilesMarkedForCompaction()[random_file_index])) {
+    // found the compaction!
+    return;
+  }
+
+  for (auto& level_file : vstorage->FilesMarkedForCompaction()) {
+    if (continuation(level_file)) {
+      // found the compaction!
+      return;
+    }
+  }
+  start_level_inputs->files.clear();
+}
+
+bool CompactionPicker::GetOverlappingL0Files(
+    VersionStorageInfo* vstorage, CompactionInputFiles* start_level_inputs,
+    int output_level, int* parent_index) {
+  // Two level 0 compaction won't run at the same time, so don't need to worry
+  // about files on level 0 being compacted.
+  assert(level0_compactions_in_progress()->empty());
+  InternalKey smallest, largest;
+  GetRange(*start_level_inputs, &smallest, &largest);
+  // Note that the next call will discard the file we placed in
+  // c->inputs_[0] earlier and replace it with an overlapping set
+  // which will include the picked file.
+  start_level_inputs->files.clear();
+  vstorage->GetOverlappingInputs(0, &smallest, &largest,
+                                 &(start_level_inputs->files));
+
+  // If we include more L0 files in the same compaction run it can
+  // cause the 'smallest' and 'largest' key to get extended to a
+  // larger range. So, re-invoke GetRange to get the new key range
+  GetRange(*start_level_inputs, &smallest, &largest);
+  if (IsRangeInCompaction(vstorage, &smallest, &largest, output_level,
+                          parent_index)) {
+    return false;
+  }
+  assert(!start_level_inputs->files.empty());
+
+  return true;
+}
+
 bool LevelCompactionPicker::NeedsCompaction(
     const VersionStorageInfo* vstorage) const {
+  if (!vstorage->ExpiredTtlFiles().empty()) {
+    return true;
+  }
+  if (!vstorage->BottommostFilesMarkedForCompaction().empty()) {
+    return true;
+  }
   if (!vstorage->FilesMarkedForCompaction().empty()) {
     return true;
   }
@@ -1017,8 +1139,7 @@ class LevelCompactionBuilder {
   // otherwise, returns false.
   bool PickIntraL0Compaction();
 
-  // If there is any file marked for compaction, put put it into inputs.
-  void PickFilesMarkedForCompaction();
+  void PickExpiredTtlFiles();
 
   const std::string& cf_name_;
   VersionStorageInfo* vstorage_;
@@ -1046,8 +1167,8 @@ class LevelCompactionBuilder {
   static const int kMinFilesForIntraL0Compaction = 4;
 };
 
-void LevelCompactionBuilder::PickFilesMarkedForCompaction() {
-  if (vstorage_->FilesMarkedForCompaction().empty()) {
+void LevelCompactionBuilder::PickExpiredTtlFiles() {
+  if (vstorage_->ExpiredTtlFiles().empty()) {
     return;
   }
 
@@ -1060,8 +1181,9 @@ void LevelCompactionBuilder::PickFilesMarkedForCompaction() {
     output_level_ =
         (start_level_ == 0) ? vstorage_->base_level() : start_level_ + 1;
 
-    if (start_level_ == 0 &&
-        !compaction_picker_->level0_compactions_in_progress()->empty()) {
+    if ((start_level_ == vstorage_->num_non_empty_levels() - 1) ||
+        (start_level_ == 0 &&
+         !compaction_picker_->level0_compactions_in_progress()->empty())) {
       return false;
     }
 
@@ -1071,22 +1193,13 @@ void LevelCompactionBuilder::PickFilesMarkedForCompaction() {
                                                       &start_level_inputs_);
   };
 
-  // take a chance on a random file first
-  Random64 rnd(/* seed */ reinterpret_cast<uint64_t>(vstorage_));
-  size_t random_file_index = static_cast<size_t>(rnd.Uniform(
-      static_cast<uint64_t>(vstorage_->FilesMarkedForCompaction().size())));
-
-  if (continuation(vstorage_->FilesMarkedForCompaction()[random_file_index])) {
-    // found the compaction!
-    return;
-  }
-
-  for (auto& level_file : vstorage_->FilesMarkedForCompaction()) {
+  for (auto& level_file : vstorage_->ExpiredTtlFiles()) {
     if (continuation(level_file)) {
       // found the compaction!
       return;
     }
   }
+
   start_level_inputs_.files.clear();
 }
 
@@ -1106,11 +1219,7 @@ void LevelCompactionBuilder::SetupInitialFiles() {
       }
       output_level_ =
           (start_level_ == 0) ? vstorage_->base_level() : start_level_ + 1;
-      if (PickFileToCompact() &&
-          compaction_picker_->ExpandInputsToCleanCut(cf_name_, vstorage_,
-                                                     &start_level_inputs_) &&
-          !compaction_picker_->FilesRangeOverlapWithCompaction(
-              {start_level_inputs_}, output_level_)) {
+      if (PickFileToCompact()) {
         // found the compaction!
         if (start_level_ == 0) {
           // L0 score = `num L0 files` / `level0_file_num_compaction_trigger`
@@ -1145,40 +1254,51 @@ void LevelCompactionBuilder::SetupInitialFiles() {
   // if we didn't find a compaction, check if there are any files marked for
   // compaction
   if (start_level_inputs_.empty()) {
-    is_manual_ = true;
     parent_index_ = base_index_ = -1;
-    PickFilesMarkedForCompaction();
+
+    // PickFilesMarkedForCompaction();
+    compaction_picker_->PickFilesMarkedForCompaction(
+        cf_name_, vstorage_, &start_level_, &output_level_, &start_level_inputs_);
     if (!start_level_inputs_.empty()) {
+      is_manual_ = true;
       compaction_reason_ = CompactionReason::kFilesMarkedForCompaction;
+      return;
+    }
+
+    size_t i;
+    for (i = 0; i < vstorage_->BottommostFilesMarkedForCompaction().size();
+         ++i) {
+      auto& level_and_file = vstorage_->BottommostFilesMarkedForCompaction()[i];
+      assert(!level_and_file.second->being_compacted);
+      start_level_inputs_.level = output_level_ = start_level_ =
+          level_and_file.first;
+      start_level_inputs_.files = {level_and_file.second};
+      if (compaction_picker_->ExpandInputsToCleanCut(cf_name_, vstorage_,
+                                                     &start_level_inputs_)) {
+        break;
+      }
+    }
+    if (i == vstorage_->BottommostFilesMarkedForCompaction().size()) {
+      start_level_inputs_.clear();
+    } else {
+      assert(!start_level_inputs_.empty());
+      compaction_reason_ = CompactionReason::kBottommostFiles;
+      return;
+    }
+
+    assert(start_level_inputs_.empty());
+    PickExpiredTtlFiles();
+    if (!start_level_inputs_.empty()) {
+      compaction_reason_ = CompactionReason::kTtl;
     }
   }
 }
 
 bool LevelCompactionBuilder::SetupOtherL0FilesIfNeeded() {
   if (start_level_ == 0 && output_level_ != 0) {
-    // Two level 0 compaction won't run at the same time, so don't need to worry
-    // about files on level 0 being compacted.
-    assert(compaction_picker_->level0_compactions_in_progress()->empty());
-    InternalKey smallest, largest;
-    compaction_picker_->GetRange(start_level_inputs_, &smallest, &largest);
-    // Note that the next call will discard the file we placed in
-    // c->inputs_[0] earlier and replace it with an overlapping set
-    // which will include the picked file.
-    start_level_inputs_.files.clear();
-    vstorage_->GetOverlappingInputs(0, &smallest, &largest,
-                                    &start_level_inputs_.files);
-
-    // If we include more L0 files in the same compaction run it can
-    // cause the 'smallest' and 'largest' key to get extended to a
-    // larger range. So, re-invoke GetRange to get the new key range
-    compaction_picker_->GetRange(start_level_inputs_, &smallest, &largest);
-    if (compaction_picker_->IsRangeInCompaction(
-            vstorage_, &smallest, &largest, output_level_, &parent_index_)) {
-      return false;
-    }
+    return compaction_picker_->GetOverlappingL0Files(
+        vstorage_, &start_level_inputs_, output_level_, &parent_index_);
   }
-  assert(!start_level_inputs_.files.empty());
-
   return true;
 }
 
@@ -1251,13 +1371,17 @@ Compaction* LevelCompactionBuilder::PickCompaction() {
 Compaction* LevelCompactionBuilder::GetCompaction() {
   auto c = new Compaction(
       vstorage_, ioptions_, mutable_cf_options_, std::move(compaction_inputs_),
-      output_level_, mutable_cf_options_.MaxFileSizeForLevel(output_level_),
+      output_level_,
+      MaxFileSizeForLevel(mutable_cf_options_, output_level_,
+                          ioptions_.compaction_style, vstorage_->base_level(),
+                          ioptions_.level_compaction_dynamic_level_bytes),
       mutable_cf_options_.max_compaction_bytes,
       GetPathId(ioptions_, mutable_cf_options_, output_level_),
       GetCompressionType(ioptions_, vstorage_, mutable_cf_options_,
                          output_level_, vstorage_->base_level()),
-      std::move(grandparents_), is_manual_, start_level_score_,
-      false /* deletion_compaction */, compaction_reason_);
+      GetCompressionOptions(ioptions_, vstorage_, output_level_),
+      /* max_subcompactions */ 0, std::move(grandparents_), is_manual_,
+      start_level_score_, false /* deletion_compaction */, compaction_reason_);
 
   // If it's level 0 compaction, make sure we don't execute any other level 0
   // compactions in parallel
@@ -1280,32 +1404,47 @@ uint32_t LevelCompactionBuilder::GetPathId(
     const ImmutableCFOptions& ioptions,
     const MutableCFOptions& mutable_cf_options, int level) {
   uint32_t p = 0;
-  assert(!ioptions.db_paths.empty());
+  assert(!ioptions.cf_paths.empty());
 
   // size remaining in the most recent path
-  uint64_t current_path_size = ioptions.db_paths[0].target_size;
+  uint64_t current_path_size = ioptions.cf_paths[0].target_size;
 
   uint64_t level_size;
   int cur_level = 0;
 
+  // max_bytes_for_level_base denotes L1 size.
+  // We estimate L0 size to be the same as L1.
   level_size = mutable_cf_options.max_bytes_for_level_base;
 
   // Last path is the fallback
-  while (p < ioptions.db_paths.size() - 1) {
+  while (p < ioptions.cf_paths.size() - 1) {
     if (level_size <= current_path_size) {
       if (cur_level == level) {
         // Does desired level fit in this path?
         return p;
       } else {
         current_path_size -= level_size;
-        level_size = static_cast<uint64_t>(
-            level_size * mutable_cf_options.max_bytes_for_level_multiplier);
+        if (cur_level > 0) {
+          if (ioptions.level_compaction_dynamic_level_bytes) {
+            // Currently, level_compaction_dynamic_level_bytes is ignored when
+            // multiple db paths are specified. https://github.com/facebook/
+            // rocksdb/blob/master/db/column_family.cc.
+            // Still, adding this check to avoid accidentally using
+            // max_bytes_for_level_multiplier_additional
+            level_size = static_cast<uint64_t>(
+                level_size * mutable_cf_options.max_bytes_for_level_multiplier);
+          } else {
+            level_size = static_cast<uint64_t>(
+                level_size * mutable_cf_options.max_bytes_for_level_multiplier *
+                mutable_cf_options.MaxBytesMultiplerAdditional(cur_level));
+          }
+        }
         cur_level++;
         continue;
       }
     }
     p++;
-    current_path_size = ioptions.db_paths[p].target_size;
+    current_path_size = ioptions.cf_paths[p].target_size;
   }
   return p;
 }
@@ -1332,12 +1471,10 @@ bool LevelCompactionBuilder::PickFileToCompact() {
   const std::vector<FileMetaData*>& level_files =
       vstorage_->LevelFiles(start_level_);
 
-  // record the first file that is not yet compacted
-  int nextIndex = -1;
-
-  for (unsigned int i = vstorage_->NextCompactionIndex(start_level_);
-       i < file_size.size(); i++) {
-    int index = file_size[i];
+  unsigned int cmp_idx;
+  for (cmp_idx = vstorage_->NextCompactionIndex(start_level_);
+       cmp_idx < file_size.size(); cmp_idx++) {
+    int index = file_size[cmp_idx];
     auto* f = level_files[index];
 
     // do not pick a file to compact if it is being compacted
@@ -1346,27 +1483,42 @@ bool LevelCompactionBuilder::PickFileToCompact() {
       continue;
     }
 
-    // remember the startIndex for the next call to PickCompaction
-    if (nextIndex == -1) {
-      nextIndex = i;
-    }
-
-    // Do not pick this file if its parents at level+1 are being compacted.
-    // Maybe we can avoid redoing this work in SetupOtherInputs
-    parent_index_ = -1;
-    if (compaction_picker_->IsRangeInCompaction(vstorage_, &f->smallest,
-                                                &f->largest, output_level_,
-                                                &parent_index_)) {
-      continue;
-    }
     start_level_inputs_.files.push_back(f);
     start_level_inputs_.level = start_level_;
+    if (!compaction_picker_->ExpandInputsToCleanCut(cf_name_, vstorage_,
+                                                    &start_level_inputs_) ||
+        compaction_picker_->FilesRangeOverlapWithCompaction(
+            {start_level_inputs_}, output_level_)) {
+      // A locked (pending compaction) input-level file was pulled in due to
+      // user-key overlap.
+      start_level_inputs_.clear();
+      continue;
+    }
+
+    // Now that input level is fully expanded, we check whether any output files
+    // are locked due to pending compaction.
+    //
+    // Note we rely on ExpandInputsToCleanCut() to tell us whether any output-
+    // level files are locked, not just the extra ones pulled in for user-key
+    // overlap.
+    InternalKey smallest, largest;
+    compaction_picker_->GetRange(start_level_inputs_, &smallest, &largest);
+    CompactionInputFiles output_level_inputs;
+    output_level_inputs.level = output_level_;
+    vstorage_->GetOverlappingInputs(output_level_, &smallest, &largest,
+                                    &output_level_inputs.files);
+    if (!output_level_inputs.empty() &&
+        !compaction_picker_->ExpandInputsToCleanCut(cf_name_, vstorage_,
+                                                    &output_level_inputs)) {
+      start_level_inputs_.clear();
+      continue;
+    }
     base_index_ = index;
     break;
   }
 
   // store where to start the iteration in the next call to PickCompaction
-  vstorage_->SetNextCompactionIndex(start_level_, nextIndex);
+  vstorage_->SetNextCompactionIndex(start_level_, cmp_idx);
 
   return start_level_inputs_.size() > 0;
 }
@@ -1396,192 +1548,4 @@ Compaction* LevelCompactionPicker::PickCompaction(
   return builder.PickCompaction();
 }
 
-#ifndef ROCKSDB_LITE
-bool FIFOCompactionPicker::NeedsCompaction(
-    const VersionStorageInfo* vstorage) const {
-  const int kLevel0 = 0;
-  return vstorage->CompactionScore(kLevel0) >= 1;
-}
-
-namespace {
-uint64_t GetTotalFilesSize(
-    const std::vector<FileMetaData*>& files) {
-  uint64_t total_size = 0;
-  for (const auto& f : files) {
-    total_size += f->fd.file_size;
-  }
-  return total_size;
-}
-}  // anonymous namespace
-
-Compaction* FIFOCompactionPicker::PickTTLCompaction(
-    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
-  assert(ioptions_.compaction_options_fifo.ttl > 0);
-
-  const int kLevel0 = 0;
-  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
-  uint64_t total_size = GetTotalFilesSize(level_files);
-
-  int64_t _current_time;
-  auto status = ioptions_.env->GetCurrentTime(&_current_time);
-  if (!status.ok()) {
-    ROCKS_LOG_BUFFER(log_buffer,
-                     "[%s] FIFO compaction: Couldn't get current time: %s. "
-                     "Not doing compactions based on TTL. ",
-                     cf_name.c_str(), status.ToString().c_str());
-    return nullptr;
-  }
-  const uint64_t current_time = static_cast<uint64_t>(_current_time);
-
-  std::vector<CompactionInputFiles> inputs;
-  inputs.emplace_back();
-  inputs[0].level = 0;
-
-  for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
-    auto f = *ritr;
-    if (f->fd.table_reader != nullptr &&
-        f->fd.table_reader->GetTableProperties() != nullptr) {
-      auto creation_time =
-          f->fd.table_reader->GetTableProperties()->creation_time;
-      if (creation_time == 0 ||
-          creation_time >=
-              (current_time - ioptions_.compaction_options_fifo.ttl)) {
-        break;
-      }
-      total_size -= f->compensated_file_size;
-      inputs[0].files.push_back(f);
-    }
-  }
-
-  // Return a nullptr and proceed to size-based FIFO compaction if:
-  // 1. there are no files older than ttl OR
-  // 2. there are a few files older than ttl, but deleting them will not bring
-  //    the total size to be less than max_table_files_size threshold.
-  if (inputs[0].files.empty() ||
-      total_size > ioptions_.compaction_options_fifo.max_table_files_size) {
-    return nullptr;
-  }
-
-  for (const auto& f : inputs[0].files) {
-    ROCKS_LOG_BUFFER(log_buffer,
-                     "[%s] FIFO compaction: picking file %" PRIu64
-                     " with creation time %" PRIu64 " for deletion",
-                     cf_name.c_str(), f->fd.GetNumber(),
-                     f->fd.table_reader->GetTableProperties()->creation_time);
-  }
-
-  Compaction* c = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
-      kNoCompression, {}, /* is manual */ false, vstorage->CompactionScore(0),
-      /* is deletion compaction */ true, CompactionReason::kFIFOTtl);
-  return c;
-}
-
-Compaction* FIFOCompactionPicker::PickSizeCompaction(
-    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
-  const int kLevel0 = 0;
-  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
-  uint64_t total_size = GetTotalFilesSize(level_files);
-
-  if (total_size <= ioptions_.compaction_options_fifo.max_table_files_size ||
-      level_files.size() == 0) {
-    // total size not exceeded
-    if (ioptions_.compaction_options_fifo.allow_compaction &&
-        level_files.size() > 0) {
-      CompactionInputFiles comp_inputs;
-      if (FindIntraL0Compaction(
-              level_files,
-              mutable_cf_options
-                  .level0_file_num_compaction_trigger /* min_files_to_compact */,
-              mutable_cf_options.write_buffer_size, &comp_inputs)) {
-        Compaction* c = new Compaction(
-            vstorage, ioptions_, mutable_cf_options, {comp_inputs}, 0,
-            16 * 1024 * 1024 /* output file size limit */,
-            0 /* max compaction bytes, not applicable */,
-            0 /* output path ID */, mutable_cf_options.compression, {},
-            /* is manual */ false, vstorage->CompactionScore(0),
-            /* is deletion compaction */ false,
-            CompactionReason::kFIFOReduceNumFiles);
-        return c;
-      }
-    }
-
-    ROCKS_LOG_BUFFER(log_buffer,
-                     "[%s] FIFO compaction: nothing to do. Total size %" PRIu64
-                     ", max size %" PRIu64 "\n",
-                     cf_name.c_str(), total_size,
-                     ioptions_.compaction_options_fifo.max_table_files_size);
-    return nullptr;
-  }
-
-  if (!level0_compactions_in_progress_.empty()) {
-    ROCKS_LOG_BUFFER(
-        log_buffer,
-        "[%s] FIFO compaction: Already executing compaction. No need "
-        "to run parallel compactions since compactions are very fast",
-        cf_name.c_str());
-    return nullptr;
-  }
-
-  std::vector<CompactionInputFiles> inputs;
-  inputs.emplace_back();
-  inputs[0].level = 0;
-
-  for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
-    auto f = *ritr;
-    total_size -= f->compensated_file_size;
-    inputs[0].files.push_back(f);
-    char tmp_fsize[16];
-    AppendHumanBytes(f->fd.GetFileSize(), tmp_fsize, sizeof(tmp_fsize));
-    ROCKS_LOG_BUFFER(log_buffer,
-                     "[%s] FIFO compaction: picking file %" PRIu64
-                     " with size %s for deletion",
-                     cf_name.c_str(), f->fd.GetNumber(), tmp_fsize);
-    if (total_size <= ioptions_.compaction_options_fifo.max_table_files_size) {
-      break;
-    }
-  }
-
-  Compaction* c = new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
-      kNoCompression, {}, /* is manual */ false, vstorage->CompactionScore(0),
-      /* is deletion compaction */ true, CompactionReason::kFIFOMaxSize);
-  return c;
-}
-
-Compaction* FIFOCompactionPicker::PickCompaction(
-    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
-  assert(vstorage->num_levels() == 1);
-
-  Compaction* c = nullptr;
-  if (ioptions_.compaction_options_fifo.ttl > 0) {
-    c = PickTTLCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
-  }
-  if (c == nullptr) {
-    c = PickSizeCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
-  }
-  RegisterCompaction(c);
-  return c;
-}
-
-Compaction* FIFOCompactionPicker::CompactRange(
-    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-    VersionStorageInfo* vstorage, int input_level, int output_level,
-    uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
-    InternalKey** compaction_end, bool* manual_conflict) {
-  assert(input_level == 0);
-  assert(output_level == 0);
-  *compaction_end = nullptr;
-  LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, ioptions_.info_log);
-  Compaction* c =
-      PickCompaction(cf_name, mutable_cf_options, vstorage, &log_buffer);
-  log_buffer.FlushBufferToLog();
-  return c;
-}
-
-#endif  // !ROCKSDB_LITE
-
 }  // namespace rocksdb

db/compaction_picker.h

@@ -58,7 +58,8 @@ class CompactionPicker {
   virtual Compaction* CompactRange(
       const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
       VersionStorageInfo* vstorage, int input_level, int output_level,
-      uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
+      uint32_t output_path_id, uint32_t max_subcompactions,
+      const InternalKey* begin, const InternalKey* end,
       InternalKey** compaction_end, bool* manual_conflict);
 
   // The maximum allowed output level.  Default value is NumberLevels() - 1.
@@ -88,6 +89,10 @@ class CompactionPicker {
 
   // Takes a list of CompactionInputFiles and returns a (manual) Compaction
   // object.
+  //
+  // Caller must provide a set of input files that has been passed through
+  // `SanitizeCompactionInputFiles` earlier. The lock should not be released
+  // between that call and this one.
   Compaction* CompactFiles(const CompactionOptions& compact_options,
                            const std::vector<CompactionInputFiles>& input_files,
                            int output_level, VersionStorageInfo* vstorage,
@@ -146,7 +151,8 @@ class CompactionPicker {
   // Will return false if it is impossible to apply this compaction.
   bool ExpandInputsToCleanCut(const std::string& cf_name,
                               VersionStorageInfo* vstorage,
-                              CompactionInputFiles* inputs);
+                              CompactionInputFiles* inputs,
+                              InternalKey** next_smallest = nullptr);
 
   // Returns true if any one of the parent files are being compacted
   bool IsRangeInCompaction(VersionStorageInfo* vstorage,
@@ -170,6 +176,15 @@ class CompactionPicker {
                        const CompactionInputFiles& output_level_inputs,
                        std::vector<FileMetaData*>* grandparents);
 
+  void PickFilesMarkedForCompaction(const std::string& cf_name,
+                                    VersionStorageInfo* vstorage,
+                                    int* start_level, int* output_level,
+                                    CompactionInputFiles* start_level_inputs);
+
+  bool GetOverlappingL0Files(VersionStorageInfo* vstorage,
+                             CompactionInputFiles* start_level_inputs,
+                             int output_level, int* parent_index);
+
   // Register this compaction in the set of running compactions
   void RegisterCompaction(Compaction* c);
 
@@ -220,41 +235,6 @@ class LevelCompactionPicker : public CompactionPicker {
 };
 
 #ifndef ROCKSDB_LITE
-class FIFOCompactionPicker : public CompactionPicker {
- public:
-  FIFOCompactionPicker(const ImmutableCFOptions& ioptions,
-                       const InternalKeyComparator* icmp)
-      : CompactionPicker(ioptions, icmp) {}
-
-  virtual Compaction* PickCompaction(const std::string& cf_name,
-                                     const MutableCFOptions& mutable_cf_options,
-                                     VersionStorageInfo* version,
-                                     LogBuffer* log_buffer) override;
-
-  virtual Compaction* CompactRange(
-      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
-      VersionStorageInfo* vstorage, int input_level, int output_level,
-      uint32_t output_path_id, const InternalKey* begin, const InternalKey* end,
-      InternalKey** compaction_end, bool* manual_conflict) override;
-
-  // The maximum allowed output level.  Always returns 0.
-  virtual int MaxOutputLevel() const override { return 0; }
-
-  virtual bool NeedsCompaction(
-      const VersionStorageInfo* vstorage) const override;
-
- private:
-  Compaction* PickTTLCompaction(const std::string& cf_name,
-                                const MutableCFOptions& mutable_cf_options,
-                                VersionStorageInfo* version,
-                                LogBuffer* log_buffer);
-
-  Compaction* PickSizeCompaction(const std::string& cf_name,
-                                 const MutableCFOptions& mutable_cf_options,
-                                 VersionStorageInfo* version,
-                                 LogBuffer* log_buffer);
-};
-
 class NullCompactionPicker : public CompactionPicker {
  public:
   NullCompactionPicker(const ImmutableCFOptions& ioptions,
@@ -263,36 +243,49 @@ class NullCompactionPicker : public CompactionPicker {
   virtual ~NullCompactionPicker() {}
 
   // Always return "nullptr"
-  Compaction* PickCompaction(const std::string& cf_name,
-                             const MutableCFOptions& mutable_cf_options,
-                             VersionStorageInfo* vstorage,
-                             LogBuffer* log_buffer) override {
+  Compaction* PickCompaction(const std::string& /*cf_name*/,
+                             const MutableCFOptions& /*mutable_cf_options*/,
+                             VersionStorageInfo* /*vstorage*/,
+                             LogBuffer* /*log_buffer*/) override {
     return nullptr;
   }
 
   // Always return "nullptr"
-  Compaction* CompactRange(const std::string& cf_name,
-                           const MutableCFOptions& mutable_cf_options,
-                           VersionStorageInfo* vstorage, int input_level,
-                           int output_level, uint32_t output_path_id,
-                           const InternalKey* begin, const InternalKey* end,
-                           InternalKey** compaction_end,
-                           bool* manual_conflict) override {
+  Compaction* CompactRange(const std::string& /*cf_name*/,
+                           const MutableCFOptions& /*mutable_cf_options*/,
+                           VersionStorageInfo* /*vstorage*/,
+                           int /*input_level*/, int /*output_level*/,
+                           uint32_t /*output_path_id*/,
+                           uint32_t /*max_subcompactions*/,
+                           const InternalKey* /*begin*/,
+                           const InternalKey* /*end*/,
+                           InternalKey** /*compaction_end*/,
+                           bool* /*manual_conflict*/) override {
     return nullptr;
   }
 
   // Always returns false.
   virtual bool NeedsCompaction(
-      const VersionStorageInfo* vstorage) const override {
+      const VersionStorageInfo* /*vstorage*/) const override {
     return false;
   }
 };
 #endif  // !ROCKSDB_LITE
 
+bool FindIntraL0Compaction(const std::vector<FileMetaData*>& level_files,
+                           size_t min_files_to_compact,
+                           uint64_t max_compact_bytes_per_del_file,
+                           CompactionInputFiles* comp_inputs);
+
 CompressionType GetCompressionType(const ImmutableCFOptions& ioptions,
                                    const VersionStorageInfo* vstorage,
                                    const MutableCFOptions& mutable_cf_options,
                                    int level, int base_level,
                                    const bool enable_compression = true);
 
+CompressionOptions GetCompressionOptions(const ImmutableCFOptions& ioptions,
+                                         const VersionStorageInfo* vstorage,
+                                         int level,
+                                         const bool enable_compression = true);
+
 }  // namespace rocksdb

db/compaction_picker_fifo.cc

@@ -0,0 +1,235 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file. See the AUTHORS file for names of contributors.
+
+#include "db/compaction_picker_fifo.h"
+#ifndef ROCKSDB_LITE
+
+#ifndef __STDC_FORMAT_MACROS
+#define __STDC_FORMAT_MACROS
+#endif
+
+#include <inttypes.h>
+#include <string>
+#include <vector>
+#include "db/column_family.h"
+#include "util/log_buffer.h"
+#include "util/string_util.h"
+
+namespace rocksdb {
+namespace {
+uint64_t GetTotalFilesSize(const std::vector<FileMetaData*>& files) {
+  uint64_t total_size = 0;
+  for (const auto& f : files) {
+    total_size += f->fd.file_size;
+  }
+  return total_size;
+}
+}  // anonymous namespace
+
+bool FIFOCompactionPicker::NeedsCompaction(
+    const VersionStorageInfo* vstorage) const {
+  const int kLevel0 = 0;
+  return vstorage->CompactionScore(kLevel0) >= 1;
+}
+
+Compaction* FIFOCompactionPicker::PickTTLCompaction(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+  assert(mutable_cf_options.compaction_options_fifo.ttl > 0);
+
+  const int kLevel0 = 0;
+  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
+  uint64_t total_size = GetTotalFilesSize(level_files);
+
+  int64_t _current_time;
+  auto status = ioptions_.env->GetCurrentTime(&_current_time);
+  if (!status.ok()) {
+    ROCKS_LOG_BUFFER(log_buffer,
+                     "[%s] FIFO compaction: Couldn't get current time: %s. "
+                     "Not doing compactions based on TTL. ",
+                     cf_name.c_str(), status.ToString().c_str());
+    return nullptr;
+  }
+  const uint64_t current_time = static_cast<uint64_t>(_current_time);
+
+  std::vector<CompactionInputFiles> inputs;
+  inputs.emplace_back();
+  inputs[0].level = 0;
+
+  // avoid underflow
+  if (current_time > mutable_cf_options.compaction_options_fifo.ttl) {
+    for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
+      auto f = *ritr;
+      if (f->fd.table_reader != nullptr &&
+          f->fd.table_reader->GetTableProperties() != nullptr) {
+        auto creation_time =
+            f->fd.table_reader->GetTableProperties()->creation_time;
+        if (creation_time == 0 ||
+            creation_time >= (current_time -
+                              mutable_cf_options.compaction_options_fifo.ttl)) {
+          break;
+        }
+        total_size -= f->compensated_file_size;
+        inputs[0].files.push_back(f);
+      }
+    }
+  }
+
+  // Return a nullptr and proceed to size-based FIFO compaction if:
+  // 1. there are no files older than ttl OR
+  // 2. there are a few files older than ttl, but deleting them will not bring
+  //    the total size to be less than max_table_files_size threshold.
+  if (inputs[0].files.empty() ||
+      total_size >
+          mutable_cf_options.compaction_options_fifo.max_table_files_size) {
+    return nullptr;
+  }
+
+  for (const auto& f : inputs[0].files) {
+    ROCKS_LOG_BUFFER(log_buffer,
+                     "[%s] FIFO compaction: picking file %" PRIu64
+                     " with creation time %" PRIu64 " for deletion",
+                     cf_name.c_str(), f->fd.GetNumber(),
+                     f->fd.table_reader->GetTableProperties()->creation_time);
+  }
+
+  Compaction* c = new Compaction(
+      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
+      kNoCompression, ioptions_.compression_opts, /* max_subcompactions */ 0,
+      {}, /* is manual */ false, vstorage->CompactionScore(0),
+      /* is deletion compaction */ true, CompactionReason::kFIFOTtl);
+  return c;
+}
+
+Compaction* FIFOCompactionPicker::PickSizeCompaction(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+  const int kLevel0 = 0;
+  const std::vector<FileMetaData*>& level_files = vstorage->LevelFiles(kLevel0);
+  uint64_t total_size = GetTotalFilesSize(level_files);
+
+  if (total_size <=
+          mutable_cf_options.compaction_options_fifo.max_table_files_size ||
+      level_files.size() == 0) {
+    // total size not exceeded
+    if (mutable_cf_options.compaction_options_fifo.allow_compaction &&
+        level_files.size() > 0) {
+      CompactionInputFiles comp_inputs;
+      // try to prevent same files from being compacted multiple times, which
+      // could produce large files that may never TTL-expire. Achieve this by
+      // disallowing compactions with files larger than memtable (inflate its
+      // size by 10% to account for uncompressed L0 files that may have size
+      // slightly greater than memtable size limit).
+      size_t max_compact_bytes_per_del_file =
+          static_cast<size_t>(MultiplyCheckOverflow(
+              static_cast<uint64_t>(mutable_cf_options.write_buffer_size),
+              1.1));
+      if (FindIntraL0Compaction(
+              level_files,
+              mutable_cf_options
+                  .level0_file_num_compaction_trigger /* min_files_to_compact */
+              ,
+              max_compact_bytes_per_del_file, &comp_inputs)) {
+        Compaction* c = new Compaction(
+            vstorage, ioptions_, mutable_cf_options, {comp_inputs}, 0,
+            16 * 1024 * 1024 /* output file size limit */,
+            0 /* max compaction bytes, not applicable */,
+            0 /* output path ID */, mutable_cf_options.compression,
+            ioptions_.compression_opts, 0 /* max_subcompactions */, {},
+            /* is manual */ false, vstorage->CompactionScore(0),
+            /* is deletion compaction */ false,
+            CompactionReason::kFIFOReduceNumFiles);
+        return c;
+      }
+    }
+
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] FIFO compaction: nothing to do. Total size %" PRIu64
+        ", max size %" PRIu64 "\n",
+        cf_name.c_str(), total_size,
+        mutable_cf_options.compaction_options_fifo.max_table_files_size);
+    return nullptr;
+  }
+
+  if (!level0_compactions_in_progress_.empty()) {
+    ROCKS_LOG_BUFFER(
+        log_buffer,
+        "[%s] FIFO compaction: Already executing compaction. No need "
+        "to run parallel compactions since compactions are very fast",
+        cf_name.c_str());
+    return nullptr;
+  }
+
+  std::vector<CompactionInputFiles> inputs;
+  inputs.emplace_back();
+  inputs[0].level = 0;
+
+  for (auto ritr = level_files.rbegin(); ritr != level_files.rend(); ++ritr) {
+    auto f = *ritr;
+    total_size -= f->compensated_file_size;
+    inputs[0].files.push_back(f);
+    char tmp_fsize[16];
+    AppendHumanBytes(f->fd.GetFileSize(), tmp_fsize, sizeof(tmp_fsize));
+    ROCKS_LOG_BUFFER(log_buffer,
+                     "[%s] FIFO compaction: picking file %" PRIu64
+                     " with size %s for deletion",
+                     cf_name.c_str(), f->fd.GetNumber(), tmp_fsize);
+    if (total_size <=
+        mutable_cf_options.compaction_options_fifo.max_table_files_size) {
+      break;
+    }
+  }
+
+  Compaction* c = new Compaction(
+      vstorage, ioptions_, mutable_cf_options, std::move(inputs), 0, 0, 0, 0,
+      kNoCompression, ioptions_.compression_opts, /* max_subcompactions */ 0,
+      {}, /* is manual */ false, vstorage->CompactionScore(0),
+      /* is deletion compaction */ true, CompactionReason::kFIFOMaxSize);
+  return c;
+}
+
+Compaction* FIFOCompactionPicker::PickCompaction(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
+  assert(vstorage->num_levels() == 1);
+
+  Compaction* c = nullptr;
+  if (mutable_cf_options.compaction_options_fifo.ttl > 0) {
+    c = PickTTLCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
+  }
+  if (c == nullptr) {
+    c = PickSizeCompaction(cf_name, mutable_cf_options, vstorage, log_buffer);
+  }
+  RegisterCompaction(c);
+  return c;
+}
+
+Compaction* FIFOCompactionPicker::CompactRange(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, int input_level, int output_level,
+    uint32_t /*output_path_id*/, uint32_t /*max_subcompactions*/,
+    const InternalKey* /*begin*/, const InternalKey* /*end*/,
+    InternalKey** compaction_end, bool* /*manual_conflict*/) {
+#ifdef NDEBUG
+  (void)input_level;
+  (void)output_level;
+#endif
+  assert(input_level == 0);
+  assert(output_level == 0);
+  *compaction_end = nullptr;
+  LogBuffer log_buffer(InfoLogLevel::INFO_LEVEL, ioptions_.info_log);
+  Compaction* c =
+      PickCompaction(cf_name, mutable_cf_options, vstorage, &log_buffer);
+  log_buffer.FlushBufferToLog();
+  return c;
+}
+
+}  // namespace rocksdb
+#endif  // !ROCKSDB_LITE

db/compaction_picker_fifo.h

@@ -0,0 +1,52 @@
+//  Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
+//  This source code is licensed under both the GPLv2 (found in the
+//  COPYING file in the root directory) and Apache 2.0 License
+//  (found in the LICENSE.Apache file in the root directory).
+//
+// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file. See the AUTHORS file for names of contributors.
+
+#pragma once
+#ifndef ROCKSDB_LITE
+
+#include "db/compaction_picker.h"
+
+namespace rocksdb {
+class FIFOCompactionPicker : public CompactionPicker {
+ public:
+  FIFOCompactionPicker(const ImmutableCFOptions& ioptions,
+                       const InternalKeyComparator* icmp)
+      : CompactionPicker(ioptions, icmp) {}
+
+  virtual Compaction* PickCompaction(const std::string& cf_name,
+                                     const MutableCFOptions& mutable_cf_options,
+                                     VersionStorageInfo* version,
+                                     LogBuffer* log_buffer) override;
+
+  virtual Compaction* CompactRange(
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      VersionStorageInfo* vstorage, int input_level, int output_level,
+      uint32_t output_path_id, uint32_t max_subcompactions,
+      const InternalKey* begin, const InternalKey* end,
+      InternalKey** compaction_end, bool* manual_conflict) override;
+
+  // The maximum allowed output level.  Always returns 0.
+  virtual int MaxOutputLevel() const override { return 0; }
+
+  virtual bool NeedsCompaction(
+      const VersionStorageInfo* vstorage) const override;
+
+ private:
+  Compaction* PickTTLCompaction(const std::string& cf_name,
+                                const MutableCFOptions& mutable_cf_options,
+                                VersionStorageInfo* version,
+                                LogBuffer* log_buffer);
+
+  Compaction* PickSizeCompaction(const std::string& cf_name,
+                                 const MutableCFOptions& mutable_cf_options,
+                                 VersionStorageInfo* version,
+                                 LogBuffer* log_buffer);
+};
+}  // namespace rocksdb
+#endif  // !ROCKSDB_LITE

db/compaction_picker_test.cc

@@ -4,10 +4,12 @@
 //  (found in the LICENSE.Apache file in the root directory).
 
 #include "db/compaction_picker.h"
+
 #include <limits>
 #include <string>
 #include <utility>
 #include "db/compaction.h"
+#include "db/compaction_picker_fifo.h"
 #include "db/compaction_picker_universal.h"
 
 #include "util/logging.h"
@@ -20,7 +22,9 @@ namespace rocksdb {
 class CountingLogger : public Logger {
  public:
   using Logger::Logv;
-  virtual void Logv(const char* format, va_list ap) override { log_count++; }
+  virtual void Logv(const char* /*format*/, va_list /*ap*/) override {
+    log_count++;
+  }
   size_t log_count;
 };
 
@@ -55,9 +59,12 @@ class CompactionPickerTest : public testing::Test {
         log_buffer_(InfoLogLevel::INFO_LEVEL, &logger_),
         file_num_(1),
         vstorage_(nullptr) {
+    // ioptions_.compaction_pri = kMinOverlappingRatio has its own set of
+    // tests to cover.
+    ioptions_.compaction_pri = kByCompensatedSize;
     fifo_options_.max_table_files_size = 1;
     mutable_cf_options_.RefreshDerivedOptions(ioptions_);
-    ioptions_.db_paths.emplace_back("dummy",
+    ioptions_.cf_paths.emplace_back("dummy",
                                     std::numeric_limits<uint64_t>::max());
   }
 
@@ -81,16 +88,17 @@ class CompactionPickerTest : public testing::Test {
 
   void Add(int level, uint32_t file_number, const char* smallest,
            const char* largest, uint64_t file_size = 1, uint32_t path_id = 0,
-           SequenceNumber smallest_seq = 100,
-           SequenceNumber largest_seq = 100) {
+           SequenceNumber smallest_seq = 100, SequenceNumber largest_seq = 100,
+           size_t compensated_file_size = 0) {
     assert(level < vstorage_->num_levels());
     FileMetaData* f = new FileMetaData;
     f->fd = FileDescriptor(file_number, path_id, file_size);
     f->smallest = InternalKey(smallest, smallest_seq, kTypeValue);
     f->largest = InternalKey(largest, largest_seq, kTypeValue);
-    f->smallest_seqno = smallest_seq;
-    f->largest_seqno = largest_seq;
-    f->compensated_file_size = file_size;
+    f->fd.smallest_seqno = smallest_seq;
+    f->fd.largest_seqno = largest_seq;
+    f->compensated_file_size =
+        (compensated_file_size != 0) ? compensated_file_size : file_size;
     f->refs = 0;
     vstorage_->AddFile(level, f);
     files_.emplace_back(f);
@@ -175,6 +183,8 @@ TEST_F(CompactionPickerTest, Level1Trigger) {
 }
 
 TEST_F(CompactionPickerTest, Level1Trigger2) {
+  mutable_cf_options_.target_file_size_base = 10000000000;
+  mutable_cf_options_.RefreshDerivedOptions(ioptions_);
   NewVersionStorage(6, kCompactionStyleLevel);
   Add(1, 66U, "150", "200", 1000000001U);
   Add(1, 88U, "201", "300", 1000000000U);
@@ -191,13 +201,14 @@ TEST_F(CompactionPickerTest, Level1Trigger2) {
   ASSERT_EQ(66U, compaction->input(0, 0)->fd.GetNumber());
   ASSERT_EQ(6U, compaction->input(1, 0)->fd.GetNumber());
   ASSERT_EQ(7U, compaction->input(1, 1)->fd.GetNumber());
+  ASSERT_EQ(uint64_t{1073741824}, compaction->OutputFilePreallocationSize());
 }
 
 TEST_F(CompactionPickerTest, LevelMaxScore) {
   NewVersionStorage(6, kCompactionStyleLevel);
   mutable_cf_options_.target_file_size_base = 10000000;
-  mutable_cf_options_.target_file_size_multiplier = 10;
   mutable_cf_options_.max_bytes_for_level_base = 10 * 1024 * 1024;
+  mutable_cf_options_.RefreshDerivedOptions(ioptions_);
   Add(0, 1U, "150", "200", 1000000U);
   // Level 1 score 1.2
   Add(1, 66U, "150", "200", 6000000U);
@@ -218,6 +229,9 @@ TEST_F(CompactionPickerTest, LevelMaxScore) {
   ASSERT_TRUE(compaction.get() != nullptr);
   ASSERT_EQ(1U, compaction->num_input_files(0));
   ASSERT_EQ(7U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(mutable_cf_options_.target_file_size_base +
+                mutable_cf_options_.target_file_size_base / 10,
+            compaction->OutputFilePreallocationSize());
 }
 
 TEST_F(CompactionPickerTest, NeedsCompactionLevel) {
@@ -383,10 +397,10 @@ TEST_F(CompactionPickerTest, NeedsCompactionUniversal) {
   NewVersionStorage(1, kCompactionStyleUniversal);
   UniversalCompactionPicker universal_compaction_picker(
       ioptions_, &icmp_);
+  UpdateVersionStorageInfo();
   // must return false when there's no files.
   ASSERT_EQ(universal_compaction_picker.NeedsCompaction(vstorage_.get()),
             false);
-  UpdateVersionStorageInfo();
 
   // verify the trigger given different number of L0 files.
   for (int i = 1;
@@ -407,6 +421,7 @@ TEST_F(CompactionPickerTest, CompactionUniversalIngestBehindReservedLevel) {
   ioptions_.allow_ingest_behind = true;
   ioptions_.num_levels = 3;
   UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+  UpdateVersionStorageInfo();
   // must return false when there's no files.
   ASSERT_EQ(universal_compaction_picker.NeedsCompaction(vstorage_.get()),
             false);
@@ -437,9 +452,10 @@ TEST_F(CompactionPickerTest, CompactionUniversalIngestBehindReservedLevel) {
 TEST_F(CompactionPickerTest, CannotTrivialMoveUniversal) {
   const uint64_t kFileSize = 100000;
 
-  ioptions_.compaction_options_universal.allow_trivial_move = true;
+  mutable_cf_options_.compaction_options_universal.allow_trivial_move = true;
   NewVersionStorage(1, kCompactionStyleUniversal);
   UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
+  UpdateVersionStorageInfo();
   // must return false when there's no files.
   ASSERT_EQ(universal_compaction_picker.NeedsCompaction(vstorage_.get()),
             false);
@@ -468,7 +484,7 @@ TEST_F(CompactionPickerTest, CannotTrivialMoveUniversal) {
 TEST_F(CompactionPickerTest, AllowsTrivialMoveUniversal) {
   const uint64_t kFileSize = 100000;
 
-  ioptions_.compaction_options_universal.allow_trivial_move = true;
+  mutable_cf_options_.compaction_options_universal.allow_trivial_move = true;
   UniversalCompactionPicker universal_compaction_picker(ioptions_, &icmp_);
 
   NewVersionStorage(3, kCompactionStyleUniversal);
@@ -496,7 +512,7 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
   const uint64_t kMaxSize = kFileSize * kFileCount / 2;
 
   fifo_options_.max_table_files_size = kMaxSize;
-  ioptions_.compaction_options_fifo = fifo_options_;
+  mutable_cf_options_.compaction_options_fifo = fifo_options_;
   FIFOCompactionPicker fifo_compaction_picker(ioptions_, &icmp_);
   UpdateVersionStorageInfo();
   // must return false when there's no files.
@@ -512,7 +528,7 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
         kFileSize, 0, i * 100, i * 100 + 99);
     current_size += kFileSize;
     UpdateVersionStorageInfo();
-    ASSERT_EQ(level_compaction_picker.NeedsCompaction(vstorage_.get()),
+    ASSERT_EQ(fifo_compaction_picker.NeedsCompaction(vstorage_.get()),
               vstorage_->CompactionScore(0) >= 1);
   }
 }
@@ -521,9 +537,10 @@ TEST_F(CompactionPickerTest, NeedsCompactionFIFO) {
 TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
   NewVersionStorage(6, kCompactionStyleLevel);
   ioptions_.compaction_pri = kMinOverlappingRatio;
-  mutable_cf_options_.target_file_size_base = 10000000;
+  mutable_cf_options_.target_file_size_base = 100000000000;
   mutable_cf_options_.target_file_size_multiplier = 10;
   mutable_cf_options_.max_bytes_for_level_base = 10 * 1024 * 1024;
+  mutable_cf_options_.RefreshDerivedOptions(ioptions_);
 
   Add(2, 6U, "150", "179", 50000000U);
   Add(2, 7U, "180", "220", 50000000U);
@@ -543,6 +560,8 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping1) {
   ASSERT_EQ(1U, compaction->num_input_files(0));
   // Pick file 8 because it overlaps with 0 files on level 3.
   ASSERT_EQ(8U, compaction->input(0, 0)->fd.GetNumber());
+  // Compaction input size * 1.1
+  ASSERT_GE(uint64_t{55000000}, compaction->OutputFilePreallocationSize());
 }
 
 TEST_F(CompactionPickerTest, CompactionPriMinOverlapping2) {
@@ -602,6 +621,35 @@ TEST_F(CompactionPickerTest, CompactionPriMinOverlapping3) {
   ASSERT_EQ(8U, compaction->input(0, 0)->fd.GetNumber());
 }
 
+TEST_F(CompactionPickerTest, CompactionPriMinOverlapping4) {
+  NewVersionStorage(6, kCompactionStyleLevel);
+  ioptions_.compaction_pri = kMinOverlappingRatio;
+  mutable_cf_options_.max_bytes_for_level_base = 10000000;
+  mutable_cf_options_.max_bytes_for_level_multiplier = 10;
+
+  // file 7 and 8 over lap with the same file, but file 8 is smaller so
+  // it will be picked.
+  // Overlaps with file 26, 27. And the file is compensated so will be
+  // picked up.
+  Add(2, 6U, "150", "167", 60000000U, 0, 100, 100, 180000000U);
+  Add(2, 7U, "168", "169", 60000000U);  // Overlaps with file 27
+  Add(2, 8U, "201", "300", 61000000U);  // Overlaps with file 28
+
+  Add(3, 26U, "160", "165", 60000000U);
+  // Boosted file size in output level is not considered.
+  Add(3, 27U, "166", "170", 60000000U, 0, 100, 100, 260000000U);
+  Add(3, 28U, "180", "400", 60000000U);
+  Add(3, 29U, "401", "500", 60000000U);
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  // Picking file 8 because overlapping ratio is the biggest.
+  ASSERT_EQ(6U, compaction->input(0, 0)->fd.GetNumber());
+}
+
 // This test exhibits the bug where we don't properly reset parent_index in
 // PickCompaction()
 TEST_F(CompactionPickerTest, ParentIndexResetBug) {
@@ -852,6 +900,80 @@ TEST_F(CompactionPickerTest, OverlappingUserKeys9) {
   ASSERT_EQ(8U, compaction->input(1, 1)->fd.GetNumber());
 }
 
+TEST_F(CompactionPickerTest, OverlappingUserKeys10) {
+  // Locked file encountered when pulling in extra input-level files with same
+  // user keys. Verify we pick the next-best file from the same input level.
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  // file_number 2U is largest and thus first choice. But it overlaps with
+  // file_number 1U which is being compacted. So instead we pick the next-
+  // biggest file, 3U, which is eligible for compaction.
+  Add(1 /* level */, 1U /* file_number */, "100" /* smallest */,
+      "150" /* largest */, 1U /* file_size */);
+  file_map_[1U].first->being_compacted = true;
+  Add(1 /* level */, 2U /* file_number */, "150" /* smallest */,
+      "200" /* largest */, 1000000000U /* file_size */, 0 /* smallest_seq */,
+      0 /* largest_seq */);
+  Add(1 /* level */, 3U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 900000000U /* file_size */);
+  Add(2 /* level */, 4U /* file_number */, "100" /* smallest */,
+      "150" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "151" /* smallest */,
+      "200" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 6U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 1U /* file_size */);
+
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(2U, compaction->num_input_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(1U, compaction->num_input_files(1));
+  ASSERT_EQ(3U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(6U, compaction->input(1, 0)->fd.GetNumber());
+}
+
+TEST_F(CompactionPickerTest, OverlappingUserKeys11) {
+  // Locked file encountered when pulling in extra output-level files with same
+  // user keys. Expected to skip that compaction and pick the next-best choice.
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  // score(L1) = 3.7
+  // score(L2) = 1.85
+  // There is no eligible file in L1 to compact since both candidates pull in
+  // file_number 5U, which overlaps with a file pending compaction (6U). The
+  // first eligible compaction is from L2->L3.
+  Add(1 /* level */, 2U /* file_number */, "151" /* smallest */,
+      "200" /* largest */, 1000000000U /* file_size */);
+  Add(1 /* level */, 3U /* file_number */, "201" /* smallest */,
+      "250" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 4U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 5000000000U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "150" /* smallest */,
+      "201" /* largest */, 1U /* file_size */);
+  Add(2 /* level */, 6U /* file_number */, "201" /* smallest */,
+      "249" /* largest */, 1U /* file_size */, 0 /* smallest_seq */,
+      0 /* largest_seq */);
+  file_map_[6U].first->being_compacted = true;
+  Add(3 /* level */, 7U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 1U /* file_size */);
+
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(2U, compaction->num_input_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(1U, compaction->num_input_files(1));
+  ASSERT_EQ(4U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(7U, compaction->input(1, 0)->fd.GetNumber());
+}
+
 TEST_F(CompactionPickerTest, NotScheduleL1IfL0WithHigherPri1) {
   NewVersionStorage(6, kCompactionStyleLevel);
   mutable_cf_options_.level0_file_num_compaction_trigger = 2;
@@ -1316,6 +1438,49 @@ TEST_F(CompactionPickerTest, IsTrivialMoveOff) {
   ASSERT_FALSE(compaction->IsTrivialMove());
 }
 
+TEST_F(CompactionPickerTest, CacheNextCompactionIndex) {
+  NewVersionStorage(6, kCompactionStyleLevel);
+  mutable_cf_options_.max_compaction_bytes = 100000000000u;
+
+  Add(1 /* level */, 1U /* file_number */, "100" /* smallest */,
+      "149" /* largest */, 1000000000U /* file_size */);
+  file_map_[1U].first->being_compacted = true;
+  Add(1 /* level */, 2U /* file_number */, "150" /* smallest */,
+      "199" /* largest */, 900000000U /* file_size */);
+  Add(1 /* level */, 3U /* file_number */, "200" /* smallest */,
+      "249" /* largest */, 800000000U /* file_size */);
+  Add(1 /* level */, 4U /* file_number */, "250" /* smallest */,
+      "299" /* largest */, 700000000U /* file_size */);
+  Add(2 /* level */, 5U /* file_number */, "150" /* smallest */,
+      "199" /* largest */, 1U /* file_size */);
+  file_map_[5U].first->being_compacted = true;
+
+  UpdateVersionStorageInfo();
+
+  std::unique_ptr<Compaction> compaction(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(1U, compaction->num_input_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(0U, compaction->num_input_files(1));
+  ASSERT_EQ(3U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(2, vstorage_->NextCompactionIndex(1 /* level */));
+
+  compaction.reset(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() != nullptr);
+  ASSERT_EQ(1U, compaction->num_input_levels());
+  ASSERT_EQ(1U, compaction->num_input_files(0));
+  ASSERT_EQ(0U, compaction->num_input_files(1));
+  ASSERT_EQ(4U, compaction->input(0, 0)->fd.GetNumber());
+  ASSERT_EQ(3, vstorage_->NextCompactionIndex(1 /* level */));
+
+  compaction.reset(level_compaction_picker.PickCompaction(
+      cf_name_, mutable_cf_options_, vstorage_.get(), &log_buffer_));
+  ASSERT_TRUE(compaction.get() == nullptr);
+  ASSERT_EQ(4, vstorage_->NextCompactionIndex(1 /* level */));
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {

db/compaction_picker_universal.cc

@@ -35,7 +35,7 @@ namespace {
 // and the index of the file in that level
 
 struct InputFileInfo {
-  InputFileInfo() : f(nullptr) {}
+  InputFileInfo() : f(nullptr), level(0), index(0) {}
 
   FileMetaData* f;
   size_t level;
@@ -97,17 +97,17 @@ void GetSmallestLargestSeqno(const std::vector<FileMetaData*>& files,
                              SequenceNumber* largest_seqno) {
   bool is_first = true;
   for (FileMetaData* f : files) {
-    assert(f->smallest_seqno <= f->largest_seqno);
+    assert(f->fd.smallest_seqno <= f->fd.largest_seqno);
     if (is_first) {
       is_first = false;
-      *smallest_seqno = f->smallest_seqno;
-      *largest_seqno = f->largest_seqno;
+      *smallest_seqno = f->fd.smallest_seqno;
+      *largest_seqno = f->fd.largest_seqno;
     } else {
-      if (f->smallest_seqno < *smallest_seqno) {
-        *smallest_seqno = f->smallest_seqno;
+      if (f->fd.smallest_seqno < *smallest_seqno) {
+        *smallest_seqno = f->fd.smallest_seqno;
       }
-      if (f->largest_seqno > *largest_seqno) {
-        *largest_seqno = f->largest_seqno;
+      if (f->fd.largest_seqno > *largest_seqno) {
+        *largest_seqno = f->fd.largest_seqno;
       }
     }
   }
@@ -162,7 +162,13 @@ bool UniversalCompactionPicker::IsInputFilesNonOverlapping(Compaction* c) {
 bool UniversalCompactionPicker::NeedsCompaction(
     const VersionStorageInfo* vstorage) const {
   const int kLevel0 = 0;
-  return vstorage->CompactionScore(kLevel0) >= 1;
+  if (vstorage->CompactionScore(kLevel0) >= 1) {
+    return true;
+  }
+  if (!vstorage->FilesMarkedForCompaction().empty()) {
+    return true;
+  }
+  return false;
 }
 
 void UniversalCompactionPicker::SortedRun::Dump(char* out_buf,
@@ -204,7 +210,8 @@ void UniversalCompactionPicker::SortedRun::DumpSizeInfo(
 
 std::vector<UniversalCompactionPicker::SortedRun>
 UniversalCompactionPicker::CalculateSortedRuns(
-    const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions) {
+    const VersionStorageInfo& vstorage, const ImmutableCFOptions& /*ioptions*/,
+    const MutableCFOptions& mutable_cf_options) {
   std::vector<UniversalCompactionPicker::SortedRun> ret;
   for (FileMetaData* f : vstorage.LevelFiles(0)) {
     ret.emplace_back(0, f, f->fd.GetFileSize(), f->compensated_file_size,
@@ -218,7 +225,8 @@ UniversalCompactionPicker::CalculateSortedRuns(
     for (FileMetaData* f : vstorage.LevelFiles(level)) {
       total_compensated_size += f->compensated_file_size;
       total_size += f->fd.GetFileSize();
-      if (ioptions.compaction_options_universal.allow_trivial_move == true) {
+      if (mutable_cf_options.compaction_options_universal.allow_trivial_move ==
+          true) {
         if (f->being_compacted) {
           being_compacted = f->being_compacted;
         }
@@ -227,7 +235,8 @@ UniversalCompactionPicker::CalculateSortedRuns(
         // non-zero level, all the files should share the same being_compacted
         // value.
         // This assumption is only valid when
-        // ioptions.compaction_options_universal.allow_trivial_move is false
+        // mutable_cf_options.compaction_options_universal.allow_trivial_move is
+        // false
         assert(is_first || f->being_compacted == being_compacted);
       }
       if (is_first) {
@@ -245,18 +254,18 @@ UniversalCompactionPicker::CalculateSortedRuns(
 
 // Universal style of compaction. Pick files that are contiguous in
 // time-range to compact.
-//
 Compaction* UniversalCompactionPicker::PickCompaction(
     const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
     VersionStorageInfo* vstorage, LogBuffer* log_buffer) {
   const int kLevel0 = 0;
   double score = vstorage->CompactionScore(kLevel0);
   std::vector<SortedRun> sorted_runs =
-      CalculateSortedRuns(*vstorage, ioptions_);
+      CalculateSortedRuns(*vstorage, ioptions_, mutable_cf_options);
 
   if (sorted_runs.size() == 0 ||
-      sorted_runs.size() <
-          (unsigned int)mutable_cf_options.level0_file_num_compaction_trigger) {
+      (vstorage->FilesMarkedForCompaction().empty() &&
+       sorted_runs.size() < (unsigned int)mutable_cf_options
+                                .level0_file_num_compaction_trigger)) {
     ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: nothing to do\n",
                      cf_name.c_str());
     TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
@@ -270,64 +279,81 @@ Compaction* UniversalCompactionPicker::PickCompaction(
       cf_name.c_str(), sorted_runs.size(), vstorage->LevelSummary(&tmp));
 
   // Check for size amplification first.
-  Compaction* c;
-  if ((c = PickCompactionToReduceSizeAmp(cf_name, mutable_cf_options, vstorage,
-                                         score, sorted_runs, log_buffer)) !=
-      nullptr) {
-    ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: compacting for size amp\n",
-                     cf_name.c_str());
-  } else {
-    // Size amplification is within limits. Try reducing read
-    // amplification while maintaining file size ratios.
-    unsigned int ratio = ioptions_.compaction_options_universal.size_ratio;
-
-    if ((c = PickCompactionToReduceSortedRuns(
-             cf_name, mutable_cf_options, vstorage, score, ratio, UINT_MAX,
-             sorted_runs, log_buffer)) != nullptr) {
-      ROCKS_LOG_BUFFER(log_buffer,
-                       "[%s] Universal: compacting for size ratio\n",
+  Compaction* c = nullptr;
+  if (sorted_runs.size() >=
+      static_cast<size_t>(
+          mutable_cf_options.level0_file_num_compaction_trigger)) {
+    if ((c = PickCompactionToReduceSizeAmp(cf_name, mutable_cf_options,
+                                           vstorage, score, sorted_runs,
+                                           log_buffer)) != nullptr) {
+      ROCKS_LOG_BUFFER(log_buffer, "[%s] Universal: compacting for size amp\n",
                        cf_name.c_str());
     } else {
-      // Size amplification and file size ratios are within configured limits.
-      // If max read amplification is exceeding configured limits, then force
-      // compaction without looking at filesize ratios and try to reduce
-      // the number of files to fewer than level0_file_num_compaction_trigger.
-      // This is guaranteed by NeedsCompaction()
-      assert(sorted_runs.size() >=
-             static_cast<size_t>(
-                 mutable_cf_options.level0_file_num_compaction_trigger));
-      // Get the total number of sorted runs that are not being compacted
-      int num_sr_not_compacted = 0;
-      for (size_t i = 0; i < sorted_runs.size(); i++) {
-        if (sorted_runs[i].being_compacted == false) {
-          num_sr_not_compacted++;
-        }
-      }
+      // Size amplification is within limits. Try reducing read
+      // amplification while maintaining file size ratios.
+      unsigned int ratio =
+          mutable_cf_options.compaction_options_universal.size_ratio;
 
-      // The number of sorted runs that are not being compacted is greater than
-      // the maximum allowed number of sorted runs
-      if (num_sr_not_compacted >
-          mutable_cf_options.level0_file_num_compaction_trigger) {
-        unsigned int num_files =
-            num_sr_not_compacted -
-            mutable_cf_options.level0_file_num_compaction_trigger + 1;
-        if ((c = PickCompactionToReduceSortedRuns(
-                 cf_name, mutable_cf_options, vstorage, score, UINT_MAX,
-                 num_files, sorted_runs, log_buffer)) != nullptr) {
-          ROCKS_LOG_BUFFER(log_buffer,
-                           "[%s] Universal: compacting for file num -- %u\n",
-                           cf_name.c_str(), num_files);
+      if ((c = PickCompactionToReduceSortedRuns(
+               cf_name, mutable_cf_options, vstorage, score, ratio, UINT_MAX,
+               sorted_runs, log_buffer)) != nullptr) {
+        ROCKS_LOG_BUFFER(log_buffer,
+                         "[%s] Universal: compacting for size ratio\n",
+                         cf_name.c_str());
+      } else {
+        // Size amplification and file size ratios are within configured limits.
+        // If max read amplification is exceeding configured limits, then force
+        // compaction without looking at filesize ratios and try to reduce
+        // the number of files to fewer than level0_file_num_compaction_trigger.
+        // This is guaranteed by NeedsCompaction()
+        assert(sorted_runs.size() >=
+               static_cast<size_t>(
+                   mutable_cf_options.level0_file_num_compaction_trigger));
+        // Get the total number of sorted runs that are not being compacted
+        int num_sr_not_compacted = 0;
+        for (size_t i = 0; i < sorted_runs.size(); i++) {
+          if (sorted_runs[i].being_compacted == false) {
+            num_sr_not_compacted++;
+          }
+        }
+
+        // The number of sorted runs that are not being compacted is greater
+        // than the maximum allowed number of sorted runs
+        if (num_sr_not_compacted >
+            mutable_cf_options.level0_file_num_compaction_trigger) {
+          unsigned int num_files =
+              num_sr_not_compacted -
+              mutable_cf_options.level0_file_num_compaction_trigger + 1;
+          if ((c = PickCompactionToReduceSortedRuns(
+                   cf_name, mutable_cf_options, vstorage, score, UINT_MAX,
+                   num_files, sorted_runs, log_buffer)) != nullptr) {
+            ROCKS_LOG_BUFFER(log_buffer,
+                             "[%s] Universal: compacting for file num -- %u\n",
+                             cf_name.c_str(), num_files);
+          }
         }
       }
     }
   }
+
+  if (c == nullptr) {
+    if ((c = PickDeleteTriggeredCompaction(cf_name, mutable_cf_options,
+                                           vstorage, score, sorted_runs,
+                                           log_buffer)) != nullptr) {
+      ROCKS_LOG_BUFFER(log_buffer,
+                       "[%s] Universal: delete triggered compaction\n",
+                       cf_name.c_str());
+    }
+  }
+
   if (c == nullptr) {
     TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
                              nullptr);
     return nullptr;
   }
 
-  if (ioptions_.compaction_options_universal.allow_trivial_move == true) {
+  if (mutable_cf_options.compaction_options_universal.allow_trivial_move ==
+      true) {
     c->set_is_trivial_move(IsInputFilesNonOverlapping(c));
   }
 
@@ -339,11 +365,11 @@ Compaction* UniversalCompactionPicker::PickCompaction(
   size_t level_index = 0U;
   if (c->start_level() == 0) {
     for (auto f : *c->inputs(0)) {
-      assert(f->smallest_seqno <= f->largest_seqno);
+      assert(f->fd.smallest_seqno <= f->fd.largest_seqno);
       if (is_first) {
         is_first = false;
       }
-      prev_smallest_seqno = f->smallest_seqno;
+      prev_smallest_seqno = f->fd.smallest_seqno;
     }
     level_index = 1U;
   }
@@ -373,6 +399,7 @@ Compaction* UniversalCompactionPicker::PickCompaction(
               c->inputs(0)->size());
 
   RegisterCompaction(c);
+  vstorage->ComputeCompactionScore(ioptions_, mutable_cf_options);
 
   TEST_SYNC_POINT_CALLBACK("UniversalCompactionPicker::PickCompaction:Return",
                            c);
@@ -380,7 +407,8 @@ Compaction* UniversalCompactionPicker::PickCompaction(
 }
 
 uint32_t UniversalCompactionPicker::GetPathId(
-    const ImmutableCFOptions& ioptions, uint64_t file_size) {
+    const ImmutableCFOptions& ioptions,
+    const MutableCFOptions& mutable_cf_options, uint64_t file_size) {
   // Two conditions need to be satisfied:
   // (1) the target path needs to be able to hold the file's size
   // (2) Total size left in this and previous paths need to be not
@@ -397,12 +425,12 @@ uint32_t UniversalCompactionPicker::GetPathId(
   // that case. We need to improve it.
   uint64_t accumulated_size = 0;
   uint64_t future_size =
-      file_size * (100 - ioptions.compaction_options_universal.size_ratio) /
-      100;
+      file_size *
+      (100 - mutable_cf_options.compaction_options_universal.size_ratio) / 100;
   uint32_t p = 0;
-  assert(!ioptions.db_paths.empty());
-  for (; p < ioptions.db_paths.size() - 1; p++) {
-    uint64_t target_size = ioptions.db_paths[p].target_size;
+  assert(!ioptions.cf_paths.empty());
+  for (; p < ioptions.cf_paths.size() - 1; p++) {
+    uint64_t target_size = ioptions.cf_paths[p].target_size;
     if (target_size > file_size &&
         accumulated_size + (target_size - file_size) > future_size) {
       return p;
@@ -422,9 +450,9 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
     unsigned int max_number_of_files_to_compact,
     const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer) {
   unsigned int min_merge_width =
-      ioptions_.compaction_options_universal.min_merge_width;
+      mutable_cf_options.compaction_options_universal.min_merge_width;
   unsigned int max_merge_width =
-      ioptions_.compaction_options_universal.max_merge_width;
+      mutable_cf_options.compaction_options_universal.max_merge_width;
 
   const SortedRun* sr = nullptr;
   bool done = false;
@@ -491,7 +519,7 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
       if (sz < static_cast<double>(succeeding_sr->size)) {
         break;
       }
-      if (ioptions_.compaction_options_universal.stop_style ==
+      if (mutable_cf_options.compaction_options_universal.stop_style ==
           kCompactionStopStyleSimilarSize) {
         // Similar-size stopping rule: also check the last picked file isn't
         // far larger than the next candidate file.
@@ -534,7 +562,7 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
   // size ratio of compression.
   bool enable_compression = true;
   int ratio_to_compress =
-      ioptions_.compaction_options_universal.compression_size_percent;
+      mutable_cf_options.compaction_options_universal.compression_size_percent;
   if (ratio_to_compress >= 0) {
     uint64_t total_size = 0;
     for (auto& sorted_run : sorted_runs) {
@@ -555,7 +583,8 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
   for (unsigned int i = 0; i < first_index_after; i++) {
     estimated_total_size += sorted_runs[i].size;
   }
-  uint32_t path_id = GetPathId(ioptions_, estimated_total_size);
+  uint32_t path_id =
+      GetPathId(ioptions_, mutable_cf_options, estimated_total_size);
   int start_level = sorted_runs[start_index].level;
   int output_level;
   if (first_index_after == sorted_runs.size()) {
@@ -596,17 +625,21 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSortedRuns(
 
   CompactionReason compaction_reason;
   if (max_number_of_files_to_compact == UINT_MAX) {
-    compaction_reason = CompactionReason::kUniversalSortedRunNum;
-  } else {
     compaction_reason = CompactionReason::kUniversalSizeRatio;
+  } else {
+    compaction_reason = CompactionReason::kUniversalSortedRunNum;
   }
   return new Compaction(
       vstorage, ioptions_, mutable_cf_options, std::move(inputs), output_level,
-      mutable_cf_options.MaxFileSizeForLevel(output_level), LLONG_MAX, path_id,
+      MaxFileSizeForLevel(mutable_cf_options, output_level,
+                          kCompactionStyleUniversal),
+      LLONG_MAX, path_id,
       GetCompressionType(ioptions_, vstorage, mutable_cf_options, start_level,
                          1, enable_compression),
-      /* grandparents */ {}, /* is manual */ false, score,
-      false /* deletion_compaction */, compaction_reason);
+      GetCompressionOptions(ioptions_, vstorage, start_level,
+                            enable_compression),
+      /* max_subcompactions */ 0, /* grandparents */ {}, /* is manual */ false,
+      score, false /* deletion_compaction */, compaction_reason);
 }
 
 // Look at overall size amplification. If size amplification
@@ -620,14 +653,18 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSizeAmp(
     VersionStorageInfo* vstorage, double score,
     const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer) {
   // percentage flexibility while reducing size amplification
-  uint64_t ratio =
-      ioptions_.compaction_options_universal.max_size_amplification_percent;
+  uint64_t ratio = mutable_cf_options.compaction_options_universal
+                       .max_size_amplification_percent;
 
   unsigned int candidate_count = 0;
   uint64_t candidate_size = 0;
   size_t start_index = 0;
   const SortedRun* sr = nullptr;
 
+  if (sorted_runs.back().being_compacted) {
+    return nullptr;
+  }
+
   // Skip files that are already being compacted
   for (size_t loop = 0; loop < sorted_runs.size() - 1; loop++) {
     sr = &sorted_runs[loop];
@@ -699,7 +736,8 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSizeAmp(
   for (size_t loop = start_index; loop < sorted_runs.size(); loop++) {
     estimated_total_size += sorted_runs[loop].size;
   }
-  uint32_t path_id = GetPathId(ioptions_, estimated_total_size);
+  uint32_t path_id =
+      GetPathId(ioptions_, mutable_cf_options, estimated_total_size);
   int start_level = sorted_runs[start_index].level;
 
   std::vector<CompactionInputFiles> inputs(vstorage->num_levels());
@@ -733,15 +771,137 @@ Compaction* UniversalCompactionPicker::PickCompactionToReduceSizeAmp(
   }
 
   return new Compaction(
-      vstorage, ioptions_, mutable_cf_options, std::move(inputs),
-      output_level, mutable_cf_options.MaxFileSizeForLevel(output_level),
+      vstorage, ioptions_, mutable_cf_options, std::move(inputs), output_level,
+      MaxFileSizeForLevel(mutable_cf_options, output_level,
+                          kCompactionStyleUniversal),
       /* max_grandparent_overlap_bytes */ LLONG_MAX, path_id,
-      GetCompressionType(ioptions_, vstorage, mutable_cf_options,
-                         output_level, 1),
-      /* grandparents */ {}, /* is manual */ false, score,
-      false /* deletion_compaction */,
+      GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
+                         1),
+      GetCompressionOptions(ioptions_, vstorage, output_level),
+      /* max_subcompactions */ 0, /* grandparents */ {}, /* is manual */ false,
+      score, false /* deletion_compaction */,
       CompactionReason::kUniversalSizeAmplification);
 }
+
+// Pick files marked for compaction. Typically, files are marked by
+// CompactOnDeleteCollector due to the presence of tombstones.
+Compaction* UniversalCompactionPicker::PickDeleteTriggeredCompaction(
+    const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+    VersionStorageInfo* vstorage, double score,
+    const std::vector<SortedRun>& /*sorted_runs*/, LogBuffer* /*log_buffer*/) {
+  CompactionInputFiles start_level_inputs;
+  int output_level;
+  std::vector<CompactionInputFiles> inputs;
+
+  if (vstorage->num_levels() == 1) {
+    // This is single level universal. Since we're basically trying to reclaim
+    // space by processing files marked for compaction due to high tombstone
+    // density, let's do the same thing as compaction to reduce size amp which
+    // has the same goals.
+    bool compact = false;
+
+    start_level_inputs.level = 0;
+    start_level_inputs.files.clear();
+    output_level = 0;
+    for (FileMetaData* f : vstorage->LevelFiles(0)) {
+      if (f->marked_for_compaction) {
+        compact = true;
+      }
+      if (compact) {
+        start_level_inputs.files.push_back(f);
+      }
+    }
+    if (start_level_inputs.size() <= 1) {
+      // If only the last file in L0 is marked for compaction, ignore it
+      return nullptr;
+    }
+    inputs.push_back(start_level_inputs);
+  } else {
+    int start_level;
+
+    // For multi-level universal, the strategy is to make this look more like
+    // leveled. We pick one of the files marked for compaction and compact with
+    // overlapping files in the adjacent level.
+    PickFilesMarkedForCompaction(cf_name, vstorage, &start_level, &output_level,
+                                 &start_level_inputs);
+    if (start_level_inputs.empty()) {
+      return nullptr;
+    }
+
+    // Pick the first non-empty level after the start_level
+    for (output_level = start_level + 1; output_level < vstorage->num_levels();
+         output_level++) {
+      if (vstorage->NumLevelFiles(output_level) != 0) {
+        break;
+      }
+    }
+
+    // If all higher levels are empty, pick the highest level as output level
+    if (output_level == vstorage->num_levels()) {
+      if (start_level == 0) {
+        output_level = vstorage->num_levels() - 1;
+      } else {
+        // If start level is non-zero and all higher levels are empty, this
+        // compaction will translate into a trivial move. Since the idea is
+        // to reclaim space and trivial move doesn't help with that, we
+        // skip compaction in this case and return nullptr
+        return nullptr;
+      }
+    }
+    if (ioptions_.allow_ingest_behind &&
+        output_level == vstorage->num_levels() - 1) {
+      assert(output_level > 1);
+      output_level--;
+    }
+
+    if (output_level != 0) {
+      if (start_level == 0) {
+        if (!GetOverlappingL0Files(vstorage, &start_level_inputs, output_level,
+                                   nullptr)) {
+          return nullptr;
+        }
+      }
+
+      CompactionInputFiles output_level_inputs;
+      int parent_index = -1;
+
+      output_level_inputs.level = output_level;
+      if (!SetupOtherInputs(cf_name, mutable_cf_options, vstorage,
+                            &start_level_inputs, &output_level_inputs,
+                            &parent_index, -1)) {
+        return nullptr;
+      }
+      inputs.push_back(start_level_inputs);
+      if (!output_level_inputs.empty()) {
+        inputs.push_back(output_level_inputs);
+      }
+      if (FilesRangeOverlapWithCompaction(inputs, output_level)) {
+        return nullptr;
+      }
+    } else {
+      inputs.push_back(start_level_inputs);
+    }
+  }
+
+  uint64_t estimated_total_size = 0;
+  // Use size of the output level as estimated file size
+  for (FileMetaData* f : vstorage->LevelFiles(output_level)) {
+    estimated_total_size += f->fd.GetFileSize();
+  }
+  uint32_t path_id =
+      GetPathId(ioptions_, mutable_cf_options, estimated_total_size);
+  return new Compaction(
+      vstorage, ioptions_, mutable_cf_options, std::move(inputs), output_level,
+      MaxFileSizeForLevel(mutable_cf_options, output_level,
+                          kCompactionStyleUniversal),
+      /* max_grandparent_overlap_bytes */ LLONG_MAX, path_id,
+      GetCompressionType(ioptions_, vstorage, mutable_cf_options, output_level,
+                         1),
+      GetCompressionOptions(ioptions_, vstorage, output_level),
+      /* max_subcompactions */ 0, /* grandparents */ {}, /* is manual */ true,
+      score, false /* deletion_compaction */,
+      CompactionReason::kFilesMarkedForCompaction);
+}
 }  // namespace rocksdb
 
 #endif  // !ROCKSDB_LITE

db/compaction_picker_universal.h

@@ -73,6 +73,11 @@ class UniversalCompactionPicker : public CompactionPicker {
       VersionStorageInfo* vstorage, double score,
       const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer);
 
+  Compaction* PickDeleteTriggeredCompaction(
+      const std::string& cf_name, const MutableCFOptions& mutable_cf_options,
+      VersionStorageInfo* vstorage, double score,
+      const std::vector<SortedRun>& sorted_runs, LogBuffer* log_buffer);
+
   // Used in universal compaction when the enabled_trivial_move
   // option is set. Checks whether there are any overlapping files
   // in the input. Returns true if the input files are non
@@ -80,11 +85,13 @@ class UniversalCompactionPicker : public CompactionPicker {
   bool IsInputFilesNonOverlapping(Compaction* c);
 
   static std::vector<SortedRun> CalculateSortedRuns(
-      const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions);
+      const VersionStorageInfo& vstorage, const ImmutableCFOptions& ioptions,
+      const MutableCFOptions& mutable_cf_options);
 
   // Pick a path ID to place a newly generated file, with its estimated file
   // size.
   static uint32_t GetPathId(const ImmutableCFOptions& ioptions,
+                            const MutableCFOptions& mutable_cf_options,
                             uint64_t file_size);
 };
 }  // namespace rocksdb

db/comparator_db_test.cc

@@ -2,6 +2,7 @@
 //  This source code is licensed under both the GPLv2 (found in the
 //  COPYING file in the root directory) and Apache 2.0 License
 //  (found in the LICENSE.Apache file in the root directory).
+#include <array>
 #include <map>
 #include <string>
 
@@ -188,10 +189,10 @@ class DoubleComparator : public Comparator {
       return -1;
     }
   }
-  virtual void FindShortestSeparator(std::string* start,
-                                     const Slice& limit) const override {}
+  virtual void FindShortestSeparator(std::string* /*start*/,
+                                     const Slice& /*limit*/) const override {}
 
-  virtual void FindShortSuccessor(std::string* key) const override {}
+  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
 };
 
 class HashComparator : public Comparator {
@@ -211,10 +212,10 @@ class HashComparator : public Comparator {
       return -1;
     }
   }
-  virtual void FindShortestSeparator(std::string* start,
-                                     const Slice& limit) const override {}
+  virtual void FindShortestSeparator(std::string* /*start*/,
+                                     const Slice& /*limit*/) const override {}
 
-  virtual void FindShortSuccessor(std::string* key) const override {}
+  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
 };
 
 class TwoStrComparator : public Comparator {
@@ -243,14 +244,16 @@ class TwoStrComparator : public Comparator {
     }
     return a2.compare(b2);
   }
-  virtual void FindShortestSeparator(std::string* start,
-                                     const Slice& limit) const override {}
+  virtual void FindShortestSeparator(std::string* /*start*/,
+                                     const Slice& /*limit*/) const override {}
 
-  virtual void FindShortSuccessor(std::string* key) const override {}
+  virtual void FindShortSuccessor(std::string* /*key*/) const override {}
 };
 }  // namespace
 
-class ComparatorDBTest : public testing::Test {
+class ComparatorDBTest
+    : public testing::Test,
+      virtual public ::testing::WithParamInterface<uint32_t> {
  private:
   std::string dbname_;
   Env* env_;
@@ -261,7 +264,11 @@ class ComparatorDBTest : public testing::Test {
  public:
   ComparatorDBTest() : env_(Env::Default()), db_(nullptr) {
     comparator = BytewiseComparator();
-    dbname_ = test::TmpDir() + "/comparator_db_test";
+    dbname_ = test::PerThreadDBPath("comparator_db_test");
+    BlockBasedTableOptions toptions;
+    toptions.format_version = GetParam();
+    last_options_.table_factory.reset(
+        rocksdb::NewBlockBasedTableFactory(toptions));
     EXPECT_OK(DestroyDB(dbname_, last_options_));
   }
 
@@ -273,8 +280,12 @@ class ComparatorDBTest : public testing::Test {
 
   DB* GetDB() { return db_; }
 
-  void SetOwnedComparator(const Comparator* cmp) {
-    comparator_guard.reset(cmp);
+  void SetOwnedComparator(const Comparator* cmp, bool owner = true) {
+    if (owner) {
+      comparator_guard.reset(cmp);
+    } else {
+      comparator_guard.reset();
+    }
     comparator = cmp;
     last_options_.comparator = cmp;
   }
@@ -303,7 +314,12 @@ class ComparatorDBTest : public testing::Test {
   }
 };
 
-TEST_F(ComparatorDBTest, Bytewise) {
+INSTANTIATE_TEST_CASE_P(FormatDef, ComparatorDBTest,
+                        testing::Values(test::kDefaultFormatVersion));
+INSTANTIATE_TEST_CASE_P(FormatLatest, ComparatorDBTest,
+                        testing::Values(test::kLatestFormatVersion));
+
+TEST_P(ComparatorDBTest, Bytewise) {
   for (int rand_seed = 301; rand_seed < 306; rand_seed++) {
     DestroyAndReopen();
     Random rnd(rand_seed);
@@ -313,7 +329,7 @@ TEST_F(ComparatorDBTest, Bytewise) {
   }
 }
 
-TEST_F(ComparatorDBTest, SimpleSuffixReverseComparator) {
+TEST_P(ComparatorDBTest, SimpleSuffixReverseComparator) {
   SetOwnedComparator(new test::SimpleSuffixReverseComparator());
 
   for (int rnd_seed = 301; rnd_seed < 316; rnd_seed++) {
@@ -339,8 +355,8 @@ TEST_F(ComparatorDBTest, SimpleSuffixReverseComparator) {
   }
 }
 
-TEST_F(ComparatorDBTest, Uint64Comparator) {
-  SetOwnedComparator(test::Uint64Comparator());
+TEST_P(ComparatorDBTest, Uint64Comparator) {
+  SetOwnedComparator(test::Uint64Comparator(), false /* owner */);
 
   for (int rnd_seed = 301; rnd_seed < 316; rnd_seed++) {
     Options* opt = GetOptions();
@@ -363,7 +379,7 @@ TEST_F(ComparatorDBTest, Uint64Comparator) {
   }
 }
 
-TEST_F(ComparatorDBTest, DoubleComparator) {
+TEST_P(ComparatorDBTest, DoubleComparator) {
   SetOwnedComparator(new DoubleComparator());
 
   for (int rnd_seed = 301; rnd_seed < 316; rnd_seed++) {
@@ -388,7 +404,7 @@ TEST_F(ComparatorDBTest, DoubleComparator) {
   }
 }
 
-TEST_F(ComparatorDBTest, HashComparator) {
+TEST_P(ComparatorDBTest, HashComparator) {
   SetOwnedComparator(new HashComparator());
 
   for (int rnd_seed = 301; rnd_seed < 316; rnd_seed++) {
@@ -407,7 +423,7 @@ TEST_F(ComparatorDBTest, HashComparator) {
   }
 }
 
-TEST_F(ComparatorDBTest, TwoStrComparator) {
+TEST_P(ComparatorDBTest, TwoStrComparator) {
   SetOwnedComparator(new TwoStrComparator());
 
   for (int rnd_seed = 301; rnd_seed < 316; rnd_seed++) {
@@ -433,6 +449,209 @@ TEST_F(ComparatorDBTest, TwoStrComparator) {
   }
 }
 
+TEST_P(ComparatorDBTest, IsSameLengthImmediateSuccessor) {
+  {
+    // different length
+    Slice s("abcxy");
+    Slice t("abcxyz");
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    Slice s("abcxyz");
+    Slice t("abcxy");
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    // not last byte different
+    Slice s("abc1xyz");
+    Slice t("abc2xyz");
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    // same string
+    Slice s("abcxyz");
+    Slice t("abcxyz");
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    Slice s("abcxy");
+    Slice t("abcxz");
+    ASSERT_TRUE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    Slice s("abcxz");
+    Slice t("abcxy");
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    const char s_array[] = "\x50\x8a\xac";
+    const char t_array[] = "\x50\x8a\xad";
+    Slice s(s_array);
+    Slice t(t_array);
+    ASSERT_TRUE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    const char s_array[] = "\x50\x8a\xff";
+    const char t_array[] = "\x50\x8b\x00";
+    Slice s(s_array, 3);
+    Slice t(t_array, 3);
+    ASSERT_TRUE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    const char s_array[] = "\x50\x8a\xff\xff";
+    const char t_array[] = "\x50\x8b\x00\x00";
+    Slice s(s_array, 4);
+    Slice t(t_array, 4);
+    ASSERT_TRUE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+  {
+    const char s_array[] = "\x50\x8a\xff\xff";
+    const char t_array[] = "\x50\x8b\x00\x01";
+    Slice s(s_array, 4);
+    Slice t(t_array, 4);
+    ASSERT_FALSE(BytewiseComparator()->IsSameLengthImmediateSuccessor(s, t));
+  }
+}
+
+TEST_P(ComparatorDBTest, FindShortestSeparator) {
+  std::string s1 = "abc1xyz";
+  std::string s2 = "abc3xy";
+
+  BytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_EQ("abc2", s1);
+
+  s1 = "abc5xyztt";
+
+  ReverseBytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_EQ("abc5", s1);
+
+  s1 = "abc3";
+  s2 = "abc2xy";
+  ReverseBytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_EQ("abc3", s1);
+
+  s1 = "abc3xyz";
+  s2 = "abc2xy";
+  ReverseBytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_EQ("abc3", s1);
+
+  s1 = "abc3xyz";
+  s2 = "abc2";
+  ReverseBytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_EQ("abc3", s1);
+
+  std::string old_s1 = s1 = "abc2xy";
+  s2 = "abc2";
+  ReverseBytewiseComparator()->FindShortestSeparator(&s1, s2);
+  ASSERT_TRUE(old_s1 >= s1);
+  ASSERT_TRUE(s1 > s2);
+}
+
+TEST_P(ComparatorDBTest, SeparatorSuccessorRandomizeTest) {
+  // Char list for boundary cases.
+  std::array<unsigned char, 6> char_list{{0, 1, 2, 253, 254, 255}};
+  Random rnd(301);
+
+  for (int attempts = 0; attempts < 1000; attempts++) {
+    uint32_t size1 = rnd.Skewed(4);
+    uint32_t size2;
+
+    if (rnd.OneIn(2)) {
+      // size2 to be random size
+      size2 = rnd.Skewed(4);
+    } else {
+      // size1 is within [-2, +2] of size1
+      int diff = static_cast<int>(rnd.Uniform(5)) - 2;
+      int tmp_size2 = static_cast<int>(size1) + diff;
+      if (tmp_size2 < 0) {
+        tmp_size2 = 0;
+      }
+      size2 = static_cast<uint32_t>(tmp_size2);
+    }
+
+    std::string s1;
+    std::string s2;
+    for (uint32_t i = 0; i < size1; i++) {
+      if (rnd.OneIn(2)) {
+        // Use random byte
+        s1 += static_cast<char>(rnd.Uniform(256));
+      } else {
+        // Use one byte in char_list
+        char c = static_cast<char>(char_list[rnd.Uniform(sizeof(char_list))]);
+        s1 += c;
+      }
+    }
+
+    // First set s2 to be the same as s1, and then modify s2.
+    s2 = s1;
+    s2.resize(size2);
+    // We start from the back of the string
+    if (size2 > 0) {
+      uint32_t pos = size2 - 1;
+      do {
+        if (pos >= size1 || rnd.OneIn(4)) {
+          // For 1/4 chance, use random byte
+          s2[pos] = static_cast<char>(rnd.Uniform(256));
+        } else if (rnd.OneIn(4)) {
+          // In 1/4 chance, stop here.
+          break;
+        } else {
+          // Create a char within [-2, +2] of the matching char of s1.
+          int diff = static_cast<int>(rnd.Uniform(5)) - 2;
+          // char may be signed or unsigned based on platform.
+          int s1_char = static_cast<int>(static_cast<unsigned char>(s1[pos]));
+          int s2_char = s1_char + diff;
+          if (s2_char < 0) {
+            s2_char = 0;
+          }
+          if (s2_char > 255) {
+            s2_char = 255;
+          }
+          s2[pos] = static_cast<char>(s2_char);
+        }
+      } while (pos-- != 0);
+    }
+
+    // Test separators
+    for (int rev = 0; rev < 2; rev++) {
+      if (rev == 1) {
+        // switch s1 and s2
+        std::string t = s1;
+        s1 = s2;
+        s2 = t;
+      }
+      std::string separator = s1;
+      BytewiseComparator()->FindShortestSeparator(&separator, s2);
+      std::string rev_separator = s1;
+      ReverseBytewiseComparator()->FindShortestSeparator(&rev_separator, s2);
+
+      if (s1 == s2) {
+        ASSERT_EQ(s1, separator);
+        ASSERT_EQ(s2, rev_separator);
+      } else if (s1 < s2) {
+        ASSERT_TRUE(s1 <= separator);
+        ASSERT_TRUE(s2 > separator);
+        ASSERT_LE(separator.size(), std::max(s1.size(), s2.size()));
+        ASSERT_EQ(s1, rev_separator);
+      } else {
+        ASSERT_TRUE(s1 >= rev_separator);
+        ASSERT_TRUE(s2 < rev_separator);
+        ASSERT_LE(rev_separator.size(), std::max(s1.size(), s2.size()));
+        ASSERT_EQ(s1, separator);
+      }
+    }
+
+    // Test successors
+    std::string succ = s1;
+    BytewiseComparator()->FindShortSuccessor(&succ);
+    ASSERT_TRUE(succ >= s1);
+
+    succ = s1;
+    ReverseBytewiseComparator()->FindShortSuccessor(&succ);
+    ASSERT_TRUE(succ <= s1);
+  }
+}
+
 }  // namespace rocksdb
 
 int main(int argc, char** argv) {