github-repos/leveldb
1
0
Fork 0
mirror of https://github.com/google/leveldb.git synced 2025-01-27 06:30:07 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into master

Browse Source
This commit is contained in:
rockzheng 2022-01-11 20:40:38 +08:00 committed by GitHub
commit 654381b711
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
151 changed files with 7602 additions and 5849 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
.clang-format Normal file
View File

@ -0,0 +1,18 @@
# Run manually to reformat a file:
# clang-format -i --style=file <file>
# find . -iname '*.cc' -o -iname '*.h' -o -iname '*.h.in' | xargs clang-format -i --style=file
BasedOnStyle: Google
DerivePointerAlignment: false
# Public headers are in a different location in the internal Google repository.
# Order them so that when imported to the authoritative repository they will be
# in correct alphabetical order.
IncludeCategories:
- Regex: '^(<|"(benchmarks|db|helpers)/)'
Priority: 1
- Regex: '^"(leveldb)/'
Priority: 2
- Regex: '^(<|"(issues|port|table|third_party|util)/)'
Priority: 3
- Regex: '.*'
Priority: 4

101
.github/workflows/build.yml vendored Normal file
View File

@ -0,0 +1,101 @@
# Copyright 2021 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.
name: ci
on: [push, pull_request]
permissions:
contents: read
jobs:
build-and-test:
name: >-
CI
${{ matrix.os }}
${{ matrix.compiler }}
${{ matrix.optimized && 'release' || 'debug' }}
runs-on: ${{ matrix.os }}
strategy:
fail-fast: false
matrix:
compiler: [clang, gcc, msvc]
os: [ubuntu-latest, macos-latest, windows-latest]
optimized: [true, false]
exclude:
# MSVC only works on Windows.
- os: ubuntu-latest
compiler: msvc
- os: macos-latest
compiler: msvc
# Not testing with GCC on macOS.
- os: macos-latest
compiler: gcc
# Only testing with MSVC on Windows.
- os: windows-latest
compiler: clang
- os: windows-latest
compiler: gcc
include:
- compiler: clang
CC: clang
CXX: clang++
- compiler: gcc
CC: gcc
CXX: g++
- compiler: msvc
CC:
CXX:
env:
CMAKE_BUILD_DIR: ${{ github.workspace }}/build
CMAKE_BUILD_TYPE: ${{ matrix.optimized && 'RelWithDebInfo' || 'Debug' }}
CC: ${{ matrix.CC }}
CXX: ${{ matrix.CXX }}
BINARY_SUFFIX: ${{ startsWith(matrix.os, 'windows') && '.exe' || '' }}
BINARY_PATH: >-
${{ format(
startsWith(matrix.os, 'windows') && '{0}\build\{1}\' || '{0}/build/',
github.workspace,
matrix.optimized && 'RelWithDebInfo' || 'Debug') }}
steps:
- uses: actions/checkout@v2
with:
submodules: true
- name: Install dependencies on Linux
if: ${{ runner.os == 'Linux' }}
run: |
sudo apt-get update
sudo apt-get install libgoogle-perftools-dev libkyotocabinet-dev \
libsnappy-dev libsqlite3-dev
- name: Generate build config
run: >-
cmake -S "${{ github.workspace }}" -B "${{ env.CMAKE_BUILD_DIR }}"
-DCMAKE_BUILD_TYPE=${{ env.CMAKE_BUILD_TYPE }}
-DCMAKE_INSTALL_PREFIX=${{ runner.temp }}/install_test/
- name: Build
run: >-
cmake --build "${{ env.CMAKE_BUILD_DIR }}"
--config "${{ env.CMAKE_BUILD_TYPE }}"
- name: Run Tests
working-directory: ${{ github.workspace }}/build
run: ctest -C "${{ env.CMAKE_BUILD_TYPE }}" --verbose
- name: Run LevelDB Benchmarks
run: ${{ env.BINARY_PATH }}db_bench${{ env.BINARY_SUFFIX }}
- name: Run SQLite Benchmarks
if: ${{ runner.os != 'Windows' }}
run: ${{ env.BINARY_PATH }}db_bench_sqlite3${{ env.BINARY_SUFFIX }}
- name: Run Kyoto Cabinet Benchmarks
if: ${{ runner.os == 'Linux' && matrix.compiler == 'clang' }}
run: ${{ env.BINARY_PATH }}db_bench_tree_db${{ env.BINARY_SUFFIX }}
- name: Test CMake installation
run: cmake --build "${{ env.CMAKE_BUILD_DIR }}" --target install

17
.gitignore vendored
View File

@ -1,9 +1,8 @@
build_config.mk
*.a
*.o
*.dylib*
*.so
*.so.*
*_test
db_bench
leveldbutil
# Editors.
*.sw*
.vscode
.DS_Store
# Build directory.
build/
out/

6
.gitmodules vendored Normal file
View File

@ -0,0 +1,6 @@
[submodule "third_party/googletest"]
path = third_party/googletest
url = https://github.com/google/googletest.git
[submodule "third_party/benchmark"]
path = third_party/benchmark
url = https://github.com/google/benchmark

69
.travis.yml
View File

@ -1,69 +0,0 @@
# Build matrix / environment variable are explained on:
# http://about.travis-ci.org/docs/user/build-configuration/
# This file can be validated on: http://lint.travis-ci.org/
sudo: false
dist: trusty
language: cpp
compiler:
- gcc
- clang
os:
- linux
- osx
env:
- BUILD_TYPE=Debug
- BUILD_TYPE=RelWithDebInfo
addons:
apt:
# List of whitelisted in travis packages for ubuntu-trusty can be found here:
# https://github.com/travis-ci/apt-package-whitelist/blob/master/ubuntu-trusty
# List of whitelisted in travis apt-sources:
# https://github.com/travis-ci/apt-source-whitelist/blob/master/ubuntu.json
sources:
- ubuntu-toolchain-r-test
- llvm-toolchain-trusty-5.0
packages:
- cmake
- gcc-7
- g++-7
- clang-5.0
- libgoogle-perftools-dev
- libkyotocabinet-dev
- libsnappy-dev
- libsqlite3-dev
install:
# Travis doesn't have a DSL for installing homebrew packages yet. Status tracked
# in https://github.com/travis-ci/travis-ci/issues/5377
# The Travis VM image for Mac already has a link at /usr/local/include/c++,
# causing Homebrew's gcc@7 installation to error out. This was reported to
# Homebrew maintainers at https://github.com/Homebrew/brew/issues/1742 and
# removing the link emerged as a workaround.
- if [ "$TRAVIS_OS_NAME" == "osx" ]; then
brew update;
if [ -L /usr/local/include/c++ ]; then rm /usr/local/include/c++; fi;
brew install gcc@7;
brew install crc32c gperftools kyoto-cabinet snappy sqlite3;
fi
# /usr/bin/gcc is stuck to old versions on both Linux and OSX.
- if [ "$CXX" = "g++" ]; then export CXX="g++-7" CC="gcc-7"; fi
- echo ${CC}
- echo ${CXX}
- ${CXX} --version
- cmake --version
before_script:
- mkdir -p build && cd build
- cmake .. -DCMAKE_BUILD_TYPE=$BUILD_TYPE
- cmake --build .
- cd ..
script:
- cd build ; ctest --verbose ; cd ..
- "if [ -f build/db_bench ] ; then build/db_bench ; fi"
- "if [ -f build/db_bench_sqlite3 ] ; then build/db_bench_sqlite3 ; fi"
- "if [ -f build/db_bench_tree_db ] ; then build/db_bench_tree_db ; fi"

513
CMakeLists.txt
View File

@ -3,25 +3,37 @@
# found in the LICENSE file. See the AUTHORS file for names of contributors.
cmake_minimum_required(VERSION 3.9)
project(leveldb VERSION 1.21.0 LANGUAGES C CXX)
# Keep the version below in sync with the one in db.h
project(leveldb VERSION 1.23.0 LANGUAGES C CXX)
# This project can use C11, but will gracefully decay down to C89.
set(CMAKE_C_STANDARD 11)
set(CMAKE_C_STANDARD_REQUIRED OFF)
set(CMAKE_C_EXTENSIONS OFF)
# C standard can be overridden when this is used as a sub-project.
if(NOT CMAKE_C_STANDARD)
# This project can use C11, but will gracefully decay down to C89.
set(CMAKE_C_STANDARD 11)
set(CMAKE_C_STANDARD_REQUIRED OFF)
set(CMAKE_C_EXTENSIONS OFF)
endif(NOT CMAKE_C_STANDARD)
# This project requires C++11.
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
# C++ standard can be overridden when this is used as a sub-project.
if(NOT CMAKE_CXX_STANDARD)
# This project requires C++11.
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED ON)
set(CMAKE_CXX_EXTENSIONS OFF)
endif(NOT CMAKE_CXX_STANDARD)
if (WIN32)
set(LEVELDB_PLATFORM_NAME LEVELDB_PLATFORM_WINDOWS)
# TODO(cmumford): Make UNICODE configurable for Windows.
add_definitions(-D_UNICODE -DUNICODE)
else (WIN32)
set(LEVELDB_PLATFORM_NAME LEVELDB_PLATFORM_POSIX)
endif (WIN32)
option(LEVELDB_BUILD_TESTS "Build LevelDB's unit tests" ON)
option(LEVELDB_BUILD_BENCHMARKS "Build LevelDB's benchmarks" ON)
option(LEVELDB_INSTALL "Install LevelDB's header and library" ON)
include(TestBigEndian)
test_big_endian(LEVELDB_IS_BIG_ENDIAN)
include(CheckIncludeFile)
check_include_file("unistd.h" HAVE_UNISTD_H)
@ -30,28 +42,50 @@ check_library_exists(crc32c crc32c_value "" HAVE_CRC32C)
check_library_exists(snappy snappy_compress "" HAVE_SNAPPY)
check_library_exists(tcmalloc malloc "" HAVE_TCMALLOC)
include(CheckSymbolExists)
check_symbol_exists(fdatasync "unistd.h" HAVE_FDATASYNC)
include(CheckCXXSymbolExists)
# Using check_cxx_symbol_exists() instead of check_c_symbol_exists() because
# we're including the header from C++, and feature detection should use the same
# compiler language that the project will use later. Principles aside, some
# versions of do not expose fdatasync() in <unistd.h> in standard C mode
# (-std=c11), but do expose the function in standard C++ mode (-std=c++11).
check_cxx_symbol_exists(fdatasync "unistd.h" HAVE_FDATASYNC)
check_cxx_symbol_exists(F_FULLFSYNC "fcntl.h" HAVE_FULLFSYNC)
check_cxx_symbol_exists(O_CLOEXEC "fcntl.h" HAVE_O_CLOEXEC)
include(CheckCXXSourceCompiles)
if(CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
# Disable C++ exceptions.
string(REGEX REPLACE "/EH[a-z]+" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /EHs-c-")
add_definitions(-D_HAS_EXCEPTIONS=0)
# Disable RTTI.
string(REGEX REPLACE "/GR" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /GR-")
else(CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
# Enable strict prototype warnings for C code in clang and gcc.
if(NOT CMAKE_C_FLAGS MATCHES "-Wstrict-prototypes")
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wstrict-prototypes")
endif(NOT CMAKE_C_FLAGS MATCHES "-Wstrict-prototypes")
# Disable C++ exceptions.
string(REGEX REPLACE "-fexceptions" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions")
# Disable RTTI.
string(REGEX REPLACE "-frtti" "" CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-rtti")
endif(CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
# Test whether -Wthread-safety is available. See
# https://clang.llvm.org/docs/ThreadSafetyAnalysis.html
# -Werror is necessary because unknown attributes only generate warnings.
set(OLD_CMAKE_REQUIRED_FLAGS ${CMAKE_REQUIRED_FLAGS})
list(APPEND CMAKE_REQUIRED_FLAGS -Werror -Wthread-safety)
check_cxx_source_compiles("
struct __attribute__((lockable)) Lock {
void Acquire() __attribute__((exclusive_lock_function()));
void Release() __attribute__((unlock_function()));
};
struct ThreadSafeType {
Lock lock_;
int data_ __attribute__((guarded_by(lock_)));
};
int main() { return 0; }
" HAVE_CLANG_THREAD_SAFETY)
set(CMAKE_REQUIRED_FLAGS ${OLD_CMAKE_REQUIRED_FLAGS})
include(CheckCXXCompilerFlag)
check_cxx_compiler_flag(-Wthread-safety HAVE_CLANG_THREAD_SAFETY)
# Used by googletest.
check_cxx_compiler_flag(-Wno-missing-field-initializers
LEVELDB_HAVE_NO_MISSING_FIELD_INITIALIZERS)
include(CheckCXXSourceCompiles)
# Test whether C++17 __has_include is available.
check_cxx_source_compiles("
@ -65,13 +99,13 @@ set(LEVELDB_PUBLIC_INCLUDE_DIR "include/leveldb")
set(LEVELDB_PORT_CONFIG_DIR "include/port")
configure_file(
"${PROJECT_SOURCE_DIR}/port/port_config.h.in"
"port/port_config.h.in"
"${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h"
)
include_directories(
"${PROJECT_BINARY_DIR}/include"
"${PROJECT_SOURCE_DIR}"
"."
)
if(BUILD_SHARED_LIBS)
@ -79,103 +113,82 @@ if(BUILD_SHARED_LIBS)
add_compile_options(-fvisibility=hidden)
endif(BUILD_SHARED_LIBS)
# POSIX code is specified separately so we can leave it out in the future.
add_library(leveldb_port_posix OBJECT "")
target_sources(leveldb_port_posix
PRIVATE
"${PROJECT_SOURCE_DIR}/port/port_posix.cc"
PUBLIC
# The headers below are dependencies for leveldb, but aren't needed by users
# that link to the installed version of leveldb and rely on its public API.
$<BUILD_INTERFACE:${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h>
$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/port/atomic_pointer.h>
$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/port/port_posix.h>
$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/port/port.h>
)
if (NOT HAVE_CXX17_HAS_INCLUDE)
target_compile_definitions(leveldb_port_posix
PRIVATE
LEVELDB_HAS_PORT_CONFIG_H=1
)
endif(NOT HAVE_CXX17_HAS_INCLUDE)
if(BUILD_SHARED_LIBS)
set_property(TARGET leveldb_port_posix PROPERTY POSITION_INDEPENDENT_CODE ON)
endif(BUILD_SHARED_LIBS)
# Must be included before CMAKE_INSTALL_INCLUDEDIR is used.
include(GNUInstallDirs)
add_library(leveldb "")
target_sources(leveldb
PRIVATE
"${PROJECT_SOURCE_DIR}/db/builder.cc"
"${PROJECT_SOURCE_DIR}/db/builder.h"
"${PROJECT_SOURCE_DIR}/db/c.cc"
"${PROJECT_SOURCE_DIR}/db/db_impl.cc"
"${PROJECT_SOURCE_DIR}/db/db_impl.h"
"${PROJECT_SOURCE_DIR}/db/db_iter.cc"
"${PROJECT_SOURCE_DIR}/db/db_iter.h"
"${PROJECT_SOURCE_DIR}/db/dbformat.cc"
"${PROJECT_SOURCE_DIR}/db/dbformat.h"
"${PROJECT_SOURCE_DIR}/db/dumpfile.cc"
"${PROJECT_SOURCE_DIR}/db/filename.cc"
"${PROJECT_SOURCE_DIR}/db/filename.h"
"${PROJECT_SOURCE_DIR}/db/log_format.h"
"${PROJECT_SOURCE_DIR}/db/log_reader.cc"
"${PROJECT_SOURCE_DIR}/db/log_reader.h"
"${PROJECT_SOURCE_DIR}/db/log_writer.cc"
"${PROJECT_SOURCE_DIR}/db/log_writer.h"
"${PROJECT_SOURCE_DIR}/db/memtable.cc"
"${PROJECT_SOURCE_DIR}/db/memtable.h"
"${PROJECT_SOURCE_DIR}/db/repair.cc"
"${PROJECT_SOURCE_DIR}/db/skiplist.h"
"${PROJECT_SOURCE_DIR}/db/snapshot.h"
"${PROJECT_SOURCE_DIR}/db/table_cache.cc"
"${PROJECT_SOURCE_DIR}/db/table_cache.h"
"${PROJECT_SOURCE_DIR}/db/version_edit.cc"
"${PROJECT_SOURCE_DIR}/db/version_edit.h"
"${PROJECT_SOURCE_DIR}/db/version_set.cc"
"${PROJECT_SOURCE_DIR}/db/version_set.h"
"${PROJECT_SOURCE_DIR}/db/write_batch_internal.h"
"${PROJECT_SOURCE_DIR}/db/write_batch.cc"
"${PROJECT_SOURCE_DIR}/port/port.h"
"${PROJECT_SOURCE_DIR}/port/thread_annotations.h"
"${PROJECT_SOURCE_DIR}/table/block_builder.cc"
"${PROJECT_SOURCE_DIR}/table/block_builder.h"
"${PROJECT_SOURCE_DIR}/table/block.cc"
"${PROJECT_SOURCE_DIR}/table/block.h"
"${PROJECT_SOURCE_DIR}/table/filter_block.cc"
"${PROJECT_SOURCE_DIR}/table/filter_block.h"
"${PROJECT_SOURCE_DIR}/table/format.cc"
"${PROJECT_SOURCE_DIR}/table/format.h"
"${PROJECT_SOURCE_DIR}/table/iterator_wrapper.h"
"${PROJECT_SOURCE_DIR}/table/iterator.cc"
"${PROJECT_SOURCE_DIR}/table/merger.cc"
"${PROJECT_SOURCE_DIR}/table/merger.h"
"${PROJECT_SOURCE_DIR}/table/table_builder.cc"
"${PROJECT_SOURCE_DIR}/table/table.cc"
"${PROJECT_SOURCE_DIR}/table/two_level_iterator.cc"
"${PROJECT_SOURCE_DIR}/table/two_level_iterator.h"
"${PROJECT_SOURCE_DIR}/util/arena.cc"
"${PROJECT_SOURCE_DIR}/util/arena.h"
"${PROJECT_SOURCE_DIR}/util/bloom.cc"
"${PROJECT_SOURCE_DIR}/util/cache.cc"
"${PROJECT_SOURCE_DIR}/util/coding.cc"
"${PROJECT_SOURCE_DIR}/util/coding.h"
"${PROJECT_SOURCE_DIR}/util/comparator.cc"
"${PROJECT_SOURCE_DIR}/util/crc32c.cc"
"${PROJECT_SOURCE_DIR}/util/crc32c.h"
"${PROJECT_SOURCE_DIR}/util/env_posix.cc"
"${PROJECT_SOURCE_DIR}/util/env.cc"
"${PROJECT_SOURCE_DIR}/util/filter_policy.cc"
"${PROJECT_SOURCE_DIR}/util/hash.cc"
"${PROJECT_SOURCE_DIR}/util/hash.h"
"${PROJECT_SOURCE_DIR}/util/logging.cc"
"${PROJECT_SOURCE_DIR}/util/logging.h"
"${PROJECT_SOURCE_DIR}/util/mutexlock.h"
"${PROJECT_SOURCE_DIR}/util/options.cc"
"${PROJECT_SOURCE_DIR}/util/posix_logger.h"
"${PROJECT_SOURCE_DIR}/util/random.h"
"${PROJECT_SOURCE_DIR}/util/status.cc"
$<TARGET_OBJECTS:leveldb_port_posix>
"${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h"
"db/builder.cc"
"db/builder.h"
"db/c.cc"
"db/db_impl.cc"
"db/db_impl.h"
"db/db_iter.cc"
"db/db_iter.h"
"db/dbformat.cc"
"db/dbformat.h"
"db/dumpfile.cc"
"db/filename.cc"
"db/filename.h"
"db/log_format.h"
"db/log_reader.cc"
"db/log_reader.h"
"db/log_writer.cc"
"db/log_writer.h"
"db/memtable.cc"
"db/memtable.h"
"db/repair.cc"
"db/skiplist.h"
"db/snapshot.h"
"db/table_cache.cc"
"db/table_cache.h"
"db/version_edit.cc"
"db/version_edit.h"
"db/version_set.cc"
"db/version_set.h"
"db/write_batch_internal.h"
"db/write_batch.cc"
"port/port_stdcxx.h"
"port/port.h"
"port/thread_annotations.h"
"table/block_builder.cc"
"table/block_builder.h"
"table/block.cc"
"table/block.h"
"table/filter_block.cc"
"table/filter_block.h"
"table/format.cc"
"table/format.h"
"table/iterator_wrapper.h"
"table/iterator.cc"
"table/merger.cc"
"table/merger.h"
"table/table_builder.cc"
"table/table.cc"
"table/two_level_iterator.cc"
"table/two_level_iterator.h"
"util/arena.cc"
"util/arena.h"
"util/bloom.cc"
"util/cache.cc"
"util/coding.cc"
"util/coding.h"
"util/comparator.cc"
"util/crc32c.cc"
"util/crc32c.h"
"util/env.cc"
"util/filter_policy.cc"
"util/hash.cc"
"util/hash.h"
"util/logging.cc"
"util/logging.h"
"util/mutexlock.h"
"util/no_destructor.h"
"util/options.cc"
"util/random.h"
"util/status.cc"
# Only CMake 3.3+ supports PUBLIC sources in targets exported by "install".
$<$<VERSION_GREATER:CMAKE_VERSION,3.2>:PUBLIC>
@ -195,23 +208,43 @@ target_sources(leveldb
"${LEVELDB_PUBLIC_INCLUDE_DIR}/table.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/write_batch.h"
)
if (WIN32)
target_sources(leveldb
PRIVATE
"util/env_windows.cc"
"util/windows_logger.h"
)
else (WIN32)
target_sources(leveldb
PRIVATE
"util/env_posix.cc"
"util/posix_logger.h"
)
endif (WIN32)
# MemEnv is not part of the interface and could be pulled to a separate library.
target_sources(leveldb
PRIVATE
"${PROJECT_SOURCE_DIR}/helpers/memenv/memenv.cc"
"${PROJECT_SOURCE_DIR}/helpers/memenv/memenv.h"
"helpers/memenv/memenv.cc"
"helpers/memenv/memenv.h"
)
target_include_directories(leveldb
PUBLIC
$<BUILD_INTERFACE:${PROJECT_SOURCE_DIR}/include>
$<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}>
)
set_target_properties(leveldb
PROPERTIES VERSION ${PROJECT_VERSION} SOVERSION ${PROJECT_VERSION_MAJOR})
target_compile_definitions(leveldb
PRIVATE
# Used by include/export.h when building shared libraries.
LEVELDB_COMPILE_LIBRARY
# Used by port/port.h.
LEVELDB_PLATFORM_POSIX=1
${LEVELDB_PLATFORM_NAME}=1
)
if (NOT HAVE_CXX17_HAS_INCLUDE)
target_compile_definitions(leveldb
@ -234,11 +267,6 @@ if(HAVE_CLANG_THREAD_SAFETY)
-Werror -Wthread-safety)
endif(HAVE_CLANG_THREAD_SAFETY)
# TODO(costan): This is only needed for port_posix.
set(THREADS_PREFER_PTHREAD_FLAG ON)
find_package(Threads REQUIRED)
target_link_libraries(leveldb Threads::Threads)
if(HAVE_CRC32C)
target_link_libraries(leveldb crc32c)
endif(HAVE_CRC32C)
@ -249,14 +277,89 @@ if(HAVE_TCMALLOC)
target_link_libraries(leveldb tcmalloc)
endif(HAVE_TCMALLOC)
# Needed by port_stdcxx.h
find_package(Threads REQUIRED)
target_link_libraries(leveldb Threads::Threads)
add_executable(leveldbutil
"${PROJECT_SOURCE_DIR}/db/leveldbutil.cc"
"db/leveldbutil.cc"
)
target_link_libraries(leveldbutil leveldb)
if(LEVELDB_BUILD_TESTS)
enable_testing()
# Prevent overriding the parent project's compiler/linker settings on Windows.
set(gtest_force_shared_crt ON CACHE BOOL "" FORCE)
set(install_gtest OFF)
set(install_gmock OFF)
set(build_gmock ON)
# This project is tested using GoogleTest.
add_subdirectory("third_party/googletest")
# GoogleTest triggers a missing field initializers warning.
if(LEVELDB_HAVE_NO_MISSING_FIELD_INITIALIZERS)
set_property(TARGET gtest
APPEND PROPERTY COMPILE_OPTIONS -Wno-missing-field-initializers)
set_property(TARGET gmock
APPEND PROPERTY COMPILE_OPTIONS -Wno-missing-field-initializers)
endif(LEVELDB_HAVE_NO_MISSING_FIELD_INITIALIZERS)
add_executable(leveldb_tests "")
target_sources(leveldb_tests
PRIVATE
# "db/fault_injection_test.cc"
# "issues/issue178_test.cc"
# "issues/issue200_test.cc"
# "issues/issue320_test.cc"
"${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h"
# "util/env_test.cc"
"util/status_test.cc"
"util/no_destructor_test.cc"
"util/testutil.cc"
"util/testutil.h"
)
if(NOT BUILD_SHARED_LIBS)
target_sources(leveldb_tests
PRIVATE
"db/autocompact_test.cc"
"db/corruption_test.cc"
"db/db_test.cc"
"db/dbformat_test.cc"
"db/filename_test.cc"
"db/log_test.cc"
"db/recovery_test.cc"
"db/skiplist_test.cc"
"db/version_edit_test.cc"
"db/version_set_test.cc"
"db/write_batch_test.cc"
"helpers/memenv/memenv_test.cc"
"table/filter_block_test.cc"
"table/table_test.cc"
"util/arena_test.cc"
"util/bloom_test.cc"
"util/cache_test.cc"
"util/coding_test.cc"
"util/crc32c_test.cc"
"util/hash_test.cc"
"util/logging_test.cc"
)
endif(NOT BUILD_SHARED_LIBS)
target_link_libraries(leveldb_tests leveldb gmock gtest gtest_main)
target_compile_definitions(leveldb_tests
PRIVATE
${LEVELDB_PLATFORM_NAME}=1
)
if (NOT HAVE_CXX17_HAS_INCLUDE)
target_compile_definitions(leveldb_tests
PRIVATE
LEVELDB_HAS_PORT_CONFIG_H=1
)
endif(NOT HAVE_CXX17_HAS_INCLUDE)
add_test(NAME "leveldb_tests" COMMAND "leveldb_tests")
function(leveldb_test test_file)
get_filename_component(test_target_name "${test_file}" NAME_WE)
@ -264,25 +367,15 @@ if(LEVELDB_BUILD_TESTS)
target_sources("${test_target_name}"
PRIVATE
"${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h"
"${PROJECT_SOURCE_DIR}/util/testharness.cc"
"${PROJECT_SOURCE_DIR}/util/testharness.h"
"${PROJECT_SOURCE_DIR}/util/testutil.cc"
"${PROJECT_SOURCE_DIR}/util/testutil.h"
"util/testutil.cc"
"util/testutil.h"
"${test_file}"
)
if(BUILD_SHARED_LIBS)
# Port functions aren't exposed in the shared library build.
target_sources("${test_target_name}"
PRIVATE
$<TARGET_OBJECTS:leveldb_port_posix>
)
endif(BUILD_SHARED_LIBS)
target_link_libraries("${test_target_name}" leveldb)
target_link_libraries("${test_target_name}" leveldb gmock gtest)
target_compile_definitions("${test_target_name}"
PRIVATE
LEVELDB_PLATFORM_POSIX=1
${LEVELDB_PLATFORM_NAME}=1
)
if (NOT HAVE_CXX17_HAS_INCLUDE)
target_compile_definitions("${test_target_name}"
@ -294,47 +387,25 @@ if(LEVELDB_BUILD_TESTS)
add_test(NAME "${test_target_name}" COMMAND "${test_target_name}")
endfunction(leveldb_test)
leveldb_test("${PROJECT_SOURCE_DIR}/db/c_test.c")
leveldb_test("${PROJECT_SOURCE_DIR}/db/fault_injection_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/issues/issue178_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/issues/issue200_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/env_test.cc")
leveldb_test("db/c_test.c")
if(NOT BUILD_SHARED_LIBS)
leveldb_test("${PROJECT_SOURCE_DIR}/db/autocompact_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/corruption_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/db_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/dbformat_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/filename_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/log_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/recovery_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/skiplist_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/version_edit_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/version_set_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/db/write_batch_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/helpers/memenv/memenv_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/table/filter_block_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/table/table_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/arena_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/bloom_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/cache_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/coding_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/crc32c_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/hash_test.cc")
leveldb_test("${PROJECT_SOURCE_DIR}/util/logging_test.cc")
# TODO(costan): This test also uses
# "${PROJECT_SOURCE_DIR}/util/env_posix_test_helper.h"
leveldb_test("${PROJECT_SOURCE_DIR}/util/env_posix_test.cc")
# "util/env_{posix|windows}_test_helper.h"
if (WIN32)
leveldb_test("util/env_windows_test.cc")
else (WIN32)
leveldb_test("util/env_posix_test.cc")
endif (WIN32)
endif(NOT BUILD_SHARED_LIBS)
endif(LEVELDB_BUILD_TESTS)
if(LEVELDB_BUILD_BENCHMARKS)
# This project uses Google benchmark for benchmarking.
set(BENCHMARK_ENABLE_TESTING OFF CACHE BOOL "" FORCE)
set(BENCHMARK_ENABLE_EXCEPTIONS OFF CACHE BOOL "" FORCE)
add_subdirectory("third_party/benchmark")
function(leveldb_benchmark bench_file)
get_filename_component(bench_target_name "${bench_file}" NAME_WE)
@ -342,27 +413,17 @@ if(LEVELDB_BUILD_BENCHMARKS)
target_sources("${bench_target_name}"
PRIVATE
"${PROJECT_BINARY_DIR}/${LEVELDB_PORT_CONFIG_DIR}/port_config.h"
"${PROJECT_SOURCE_DIR}/util/histogram.cc"
"${PROJECT_SOURCE_DIR}/util/histogram.h"
"${PROJECT_SOURCE_DIR}/util/testharness.cc"
"${PROJECT_SOURCE_DIR}/util/testharness.h"
"${PROJECT_SOURCE_DIR}/util/testutil.cc"
"${PROJECT_SOURCE_DIR}/util/testutil.h"
"util/histogram.cc"
"util/histogram.h"
"util/testutil.cc"
"util/testutil.h"
"${bench_file}"
)
if(BUILD_SHARED_LIBS)
# Port functions aren't exposed in the shared library build.
target_sources("${bench_target_name}"
PRIVATE
$<TARGET_OBJECTS:leveldb_port_posix>
)
endif(BUILD_SHARED_LIBS)
target_link_libraries("${bench_target_name}" leveldb)
target_link_libraries("${bench_target_name}" leveldb gmock gtest benchmark)
target_compile_definitions("${bench_target_name}"
PRIVATE
LEVELDB_PLATFORM_POSIX=1
${LEVELDB_PLATFORM_NAME}=1
)
if (NOT HAVE_CXX17_HAS_INCLUDE)
target_compile_definitions("${bench_target_name}"
@ -373,12 +434,12 @@ if(LEVELDB_BUILD_BENCHMARKS)
endfunction(leveldb_benchmark)
if(NOT BUILD_SHARED_LIBS)
leveldb_benchmark("${PROJECT_SOURCE_DIR}/db/db_bench.cc")
leveldb_benchmark("benchmarks/db_bench.cc")
endif(NOT BUILD_SHARED_LIBS)
check_library_exists(sqlite3 sqlite3_open "" HAVE_SQLITE3)
if(HAVE_SQLITE3)
leveldb_benchmark("${PROJECT_SOURCE_DIR}/doc/bench/db_bench_sqlite3.cc")
leveldb_benchmark("benchmarks/db_bench_sqlite3.cc")
target_link_libraries(db_bench_sqlite3 sqlite3)
endif(HAVE_SQLITE3)
@ -398,13 +459,12 @@ int main() {
" HAVE_KYOTOCABINET)
set(CMAKE_REQUIRED_LIBRARIES ${OLD_CMAKE_REQURED_LIBRARIES})
if(HAVE_KYOTOCABINET)
leveldb_benchmark("${PROJECT_SOURCE_DIR}/doc/bench/db_bench_tree_db.cc")
leveldb_benchmark("benchmarks/db_bench_tree_db.cc")
target_link_libraries(db_bench_tree_db kyotocabinet)
endif(HAVE_KYOTOCABINET)
endif(LEVELDB_BUILD_BENCHMARKS)
if(LEVELDB_INSTALL)
include(GNUInstallDirs)
install(TARGETS leveldb
EXPORT leveldbTargets
RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR}
@ -413,38 +473,43 @@ if(LEVELDB_INSTALL)
)
install(
FILES
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/c.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/cache.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/comparator.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/db.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/dumpfile.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/env.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/export.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/filter_policy.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/iterator.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/options.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/slice.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/status.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/table_builder.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/table.h"
"${PROJECT_SOURCE_DIR}/${LEVELDB_PUBLIC_INCLUDE_DIR}/write_batch.h"
DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/leveldb
"${LEVELDB_PUBLIC_INCLUDE_DIR}/c.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/cache.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/comparator.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/db.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/dumpfile.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/env.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/export.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/filter_policy.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/iterator.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/options.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/slice.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/status.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/table_builder.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/table.h"
"${LEVELDB_PUBLIC_INCLUDE_DIR}/write_batch.h"
DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}/leveldb"
)
include(CMakePackageConfigHelpers)
configure_package_config_file(
"cmake/${PROJECT_NAME}Config.cmake.in"
"${PROJECT_BINARY_DIR}/cmake/${PROJECT_NAME}Config.cmake"
INSTALL_DESTINATION "${CMAKE_INSTALL_LIBDIR}/cmake/${PROJECT_NAME}"
)
write_basic_package_version_file(
"${PROJECT_BINARY_DIR}/leveldbConfigVersion.cmake"
COMPATIBILITY SameMajorVersion
"${PROJECT_BINARY_DIR}/cmake/${PROJECT_NAME}ConfigVersion.cmake"
COMPATIBILITY SameMajorVersion
)
install(
EXPORT leveldbTargets
NAMESPACE leveldb::
DESTINATION "${CMAKE_INSTALL_LIBDIR}/cmake/leveldb"
DESTINATION "${CMAKE_INSTALL_LIBDIR}/cmake/${PROJECT_NAME}"
)
install(
FILES
"${PROJECT_SOURCE_DIR}/cmake/leveldbConfig.cmake"
"${PROJECT_BINARY_DIR}/leveldbConfigVersion.cmake"
DESTINATION "${CMAKE_INSTALL_LIBDIR}/cmake/leveldb"
"${PROJECT_BINARY_DIR}/cmake/${PROJECT_NAME}Config.cmake"
"${PROJECT_BINARY_DIR}/cmake/${PROJECT_NAME}ConfigVersion.cmake"
DESTINATION "${CMAKE_INSTALL_LIBDIR}/cmake/${PROJECT_NAME}"
)
endif(LEVELDB_INSTALL)

2
CONTRIBUTING.md
View File

@ -32,5 +32,5 @@ the CLA.
## Writing Code ##
If your contribution contains code, please make sure that it follows
[the style guide](http://google.github.io/styleguide/cppguide.html).
[the style guide](https://google.github.io/styleguide/cppguide.html).
Otherwise we will have to ask you to make changes, and that's no fun for anyone.

77
README.md
View File

@ -1,6 +1,6 @@
**LevelDB is a fast key-value storage library written at Google that provides an ordered mapping from string keys to string values.**
[![Build Status](https://travis-ci.org/google/leveldb.svg?branch=master)](https://travis-ci.org/google/leveldb)
[![ci](https://github.com/google/leveldb/actions/workflows/build.yml/badge.svg)](https://github.com/google/leveldb/actions/workflows/build.yml)
Authors: Sanjay Ghemawat (sanjay@google.com) and Jeff Dean (jeff@google.com)
@ -13,7 +13,7 @@ Authors: Sanjay Ghemawat (sanjay@google.com) and Jeff Dean (jeff@google.com)
* Multiple changes can be made in one atomic batch.
* Users can create a transient snapshot to get a consistent view of data.
* Forward and backward iteration is supported over the data.
* Data is automatically compressed using the [Snappy compression library](http://google.github.io/snappy/).
* Data is automatically compressed using the [Snappy compression library](https://google.github.io/snappy/).
* External activity (file system operations etc.) is relayed through a virtual interface so users can customize the operating system interactions.
# Documentation
@ -26,10 +26,18 @@ Authors: Sanjay Ghemawat (sanjay@google.com) and Jeff Dean (jeff@google.com)
* Only a single process (possibly multi-threaded) can access a particular database at a time.
* There is no client-server support builtin to the library. An application that needs such support will have to wrap their own server around the library.
# Getting the Source
```bash
git clone --recurse-submodules https://github.com/google/leveldb.git
```
# Building
This project supports [CMake](https://cmake.org/) out of the box.
### Build for POSIX
Quick start:
```bash
@ -37,6 +45,29 @@ mkdir -p build && cd build
cmake -DCMAKE_BUILD_TYPE=Release .. && cmake --build .
```
### Building for Windows
First generate the Visual Studio 2017 project/solution files:
```cmd
mkdir build
cd build
cmake -G "Visual Studio 15" ..
```
The default default will build for x86. For 64-bit run:
```cmd
cmake -G "Visual Studio 15 Win64" ..
```
To compile the Windows solution from the command-line:
```cmd
devenv /build Debug leveldb.sln
```
or open leveldb.sln in Visual Studio and build from within.
Please see the CMake documentation and `CMakeLists.txt` for more advanced usage.
# Contributing to the leveldb Project
@ -48,10 +79,10 @@ will be considered.
Contribution requirements:
1. **POSIX only**. We _generally_ will only accept changes that are both
compiled, and tested on a POSIX platform - usually Linux. Very small
changes will sometimes be accepted, but consider that more of an
exception than the rule.
1. **Tested platforms only**. We _generally_ will only accept changes for
platforms that are compiled and tested. This means POSIX (for Linux and
macOS) or Windows. Very small changes will sometimes be accepted, but
consider that more of an exception than the rule.
2. **Stable API**. We strive very hard to maintain a stable API. Changes that
require changes for projects using leveldb _might_ be rejected without
@ -60,6 +91,14 @@ Contribution requirements:
3. **Tests**: All changes must be accompanied by a new (or changed) test, or
a sufficient explanation as to why a new (or changed) test is not required.
4. **Consistent Style**: This project conforms to the
[Google C++ Style Guide](https://google.github.io/styleguide/cppguide.html).
To ensure your changes are properly formatted please run:
```
clang-format -i --style=file <file>
```
## Submitting a Pull Request
Before any pull request will be accepted the author must first sign a
@ -155,37 +194,37 @@ uncompressed blocks in memory, the read performance improves again:
See [doc/index.md](doc/index.md) for more explanation. See
[doc/impl.md](doc/impl.md) for a brief overview of the implementation.
The public interface is in include/*.h. Callers should not include or
The public interface is in include/leveldb/*.h. Callers should not include or
rely on the details of any other header files in this package. Those
internal APIs may be changed without warning.
Guide to header files:
* **include/db.h**: Main interface to the DB: Start here
* **include/leveldb/db.h**: Main interface to the DB: Start here.
* **include/options.h**: Control over the behavior of an entire database,
* **include/leveldb/options.h**: Control over the behavior of an entire database,
and also control over the behavior of individual reads and writes.
* **include/comparator.h**: Abstraction for user-specified comparison function.
* **include/leveldb/comparator.h**: Abstraction for user-specified comparison function.
If you want just bytewise comparison of keys, you can use the default
comparator, but clients can write their own comparator implementations if they
want custom ordering (e.g. to handle different character encodings, etc.)
want custom ordering (e.g. to handle different character encodings, etc.).
* **include/iterator.h**: Interface for iterating over data. You can get
* **include/leveldb/iterator.h**: Interface for iterating over data. You can get
an iterator from a DB object.
* **include/write_batch.h**: Interface for atomically applying multiple
* **include/leveldb/write_batch.h**: Interface for atomically applying multiple
updates to a database.
* **include/slice.h**: A simple module for maintaining a pointer and a
* **include/leveldb/slice.h**: A simple module for maintaining a pointer and a
length into some other byte array.
* **include/status.h**: Status is returned from many of the public interfaces
* **include/leveldb/status.h**: Status is returned from many of the public interfaces
and is used to report success and various kinds of errors.
* **include/env.h**:
* **include/leveldb/env.h**:
Abstraction of the OS environment. A posix implementation of this interface is
in util/env_posix.cc
in util/env_posix.cc.
* **include/table.h, include/table_builder.h**: Lower-level modules that most
clients probably won't use directly
* **include/leveldb/table.h, include/leveldb/table_builder.h**: Lower-level modules that most
clients probably won't use directly.

422
db/db_bench.cc → benchmarks/db_bench.cc
View File

@ -3,9 +3,13 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include <sys/types.h>
#include <stdio.h>
#include <stdlib.h>
#include <atomic>
#include <cstdio>
#include <cstdlib>
#include "leveldb/cache.h"
#include "leveldb/comparator.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "leveldb/filter_policy.h"
@ -32,9 +36,9 @@
// readmissing -- read N missing keys in random order
// readhot -- read N times in random order from 1% section of DB
// seekrandom -- N random seeks
// seekordered -- N ordered seeks
// open -- cost of opening a DB
// crc32c -- repeated crc32c of 4K of data
// acquireload -- load N*1000 times
// Meta operations:
// compact -- Compact the entire DB
// stats -- Print DB stats
@ -56,9 +60,7 @@ static const char* FLAGS_benchmarks =
"fill100K,"
"crc32c,"
"snappycomp,"
"snappyuncomp,"
"acquireload,"
;
"snappyuncomp,";
// Number of key/values to place in database
static int FLAGS_num = 1000000;
@ -79,6 +81,9 @@ static double FLAGS_compression_ratio = 0.5;
// Print histogram of operation timings
static bool FLAGS_histogram = false;
// Count the number of string comparisons performed
static bool FLAGS_comparisons = false;
// Number of bytes to buffer in memtable before compacting
// (initialized to default value by "main")
static int FLAGS_write_buffer_size = 0;
@ -102,6 +107,9 @@ static int FLAGS_open_files = 0;
// Negative means use default settings.
static int FLAGS_bloom_bits = -1;
// Common key prefix length.
static int FLAGS_key_prefix = 0;
// If true, do not destroy the existing database. If you set this
// flag and also specify a benchmark that wants a fresh database, that
// benchmark will fail.
@ -118,6 +126,33 @@ namespace leveldb {
namespace {
leveldb::Env* g_env = nullptr;
class CountComparator : public Comparator {
public:
CountComparator(const Comparator* wrapped) : wrapped_(wrapped) {}
~CountComparator() override {}
int Compare(const Slice& a, const Slice& b) const override {
count_.fetch_add(1, std::memory_order_relaxed);
return wrapped_->Compare(a, b);
}
const char* Name() const override { return wrapped_->Name(); }
void FindShortestSeparator(std::string* start,
const Slice& limit) const override {
wrapped_->FindShortestSeparator(start, limit);
}
void FindShortSuccessor(std::string* key) const override {
return wrapped_->FindShortSuccessor(key);
}
size_t comparisons() const { return count_.load(std::memory_order_relaxed); }
void reset() { count_.store(0, std::memory_order_relaxed); }
private:
mutable std::atomic<size_t> count_{0};
const Comparator* const wrapped_;
};
// Helper for quickly generating random data.
class RandomGenerator {
private:
@ -150,6 +185,26 @@ class RandomGenerator {
}
};
class KeyBuffer {
public:
KeyBuffer() {
assert(FLAGS_key_prefix < sizeof(buffer_));
memset(buffer_, 'a', FLAGS_key_prefix);
}
KeyBuffer& operator=(KeyBuffer& other) = delete;
KeyBuffer(KeyBuffer& other) = delete;
void Set(int k) {
std::snprintf(buffer_ + FLAGS_key_prefix,
sizeof(buffer_) - FLAGS_key_prefix, "%016d", k);
}
Slice slice() const { return Slice(buffer_, FLAGS_key_prefix + 16); }
private:
char buffer_[1024];
};
#if defined(__linux)
static Slice TrimSpace(Slice s) {
size_t start = 0;
@ -157,7 +212,7 @@ static Slice TrimSpace(Slice s) {
start++;
}
size_t limit = s.size();
while (limit > start && isspace(s[limit-1])) {
while (limit > start && isspace(s[limit - 1])) {
limit--;
}
return Slice(s.data() + start, limit - start);
@ -189,14 +244,12 @@ class Stats {
void Start() {
next_report_ = 100;
last_op_finish_ = start_;
hist_.Clear();
done_ = 0;
bytes_ = 0;
seconds_ = 0;
start_ = g_env->NowMicros();
finish_ = start_;
message_.clear();
start_ = finish_ = last_op_finish_ = g_env->NowMicros();
}
void Merge(const Stats& other) {
@ -216,9 +269,7 @@ class Stats {
seconds_ = (finish_ - start_) * 1e-6;
}
void AddMessage(Slice msg) {
AppendWithSpace(&message_, msg);
}
void AddMessage(Slice msg) { AppendWithSpace(&message_, msg); }
void FinishedSingleOp() {
if (FLAGS_histogram) {
@ -226,29 +277,34 @@ class Stats {
double micros = now - last_op_finish_;
hist_.Add(micros);
if (micros > 20000) {
fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
fflush(stderr);
std::fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
std::fflush(stderr);
}
last_op_finish_ = now;
}
done_++;
if (done_ >= next_report_) {
if (next_report_ < 1000) next_report_ += 100;
else if (next_report_ < 5000) next_report_ += 500;
else if (next_report_ < 10000) next_report_ += 1000;
else if (next_report_ < 50000) next_report_ += 5000;
else if (next_report_ < 100000) next_report_ += 10000;
else if (next_report_ < 500000) next_report_ += 50000;
else next_report_ += 100000;
fprintf(stderr, "... finished %d ops%30s\r", done_, "");
fflush(stderr);
if (next_report_ < 1000)
next_report_ += 100;
else if (next_report_ < 5000)
next_report_ += 500;
else if (next_report_ < 10000)
next_report_ += 1000;
else if (next_report_ < 50000)
next_report_ += 5000;
else if (next_report_ < 100000)
next_report_ += 10000;
else if (next_report_ < 500000)
next_report_ += 50000;
else
next_report_ += 100000;
std::fprintf(stderr, "... finished %d ops%30s\r", done_, "");
std::fflush(stderr);
}
}
void AddBytes(int64_t n) {
bytes_ += n;
}
void AddBytes(int64_t n) { bytes_ += n; }
void Report(const Slice& name) {
// Pretend at least one op was done in case we are running a benchmark
@ -261,21 +317,20 @@ class Stats {
// elapsed times.
double elapsed = (finish_ - start_) * 1e-6;
char rate[100];
snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / elapsed);
std::snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / elapsed);
extra = rate;
}
AppendWithSpace(&extra, message_);
fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(),
seconds_ * 1e6 / done_,
(extra.empty() ? "" : " "),
extra.c_str());
std::fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(), seconds_ * 1e6 / done_,
(extra.empty() ? "" : " "), extra.c_str());
if (FLAGS_histogram) {
fprintf(stdout, "Microseconds per op:\n%s\n", hist_.ToString().c_str());
std::fprintf(stdout, "Microseconds per op:\n%s\n",
hist_.ToString().c_str());
}
fflush(stdout);
std::fflush(stdout);
}
};
@ -296,20 +351,17 @@ struct SharedState {
bool start GUARDED_BY(mu);
SharedState(int total)
: cv(&mu), total(total), num_initialized(0), num_done(0), start(false) { }
: cv(&mu), total(total), num_initialized(0), num_done(0), start(false) {}
};
// Per-thread state for concurrent executions of the same benchmark.
struct ThreadState {
int tid; // 0..n-1 when running in n threads
Random rand; // Has different seeds for different threads
int tid; // 0..n-1 when running in n threads
Random rand; // Has different seeds for different threads
Stats stats;
SharedState* shared;
ThreadState(int index)
: tid(index),
rand(1000 + index) {
}
ThreadState(int index, int seed) : tid(index), rand(seed), shared(nullptr) {}
};
} // namespace
@ -325,55 +377,61 @@ class Benchmark {
WriteOptions write_options_;
int reads_;
int heap_counter_;
CountComparator count_comparator_;
int total_thread_count_;
void PrintHeader() {
const int kKeySize = 16;
const int kKeySize = 16 + FLAGS_key_prefix;
PrintEnvironment();
fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
fprintf(stdout, "Values: %d bytes each (%d bytes after compression)\n",
FLAGS_value_size,
static_cast<int>(FLAGS_value_size * FLAGS_compression_ratio + 0.5));
fprintf(stdout, "Entries: %d\n", num_);
fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_)
/ 1048576.0));
fprintf(stdout, "FileSize: %.1f MB (estimated)\n",
(((kKeySize + FLAGS_value_size * FLAGS_compression_ratio) * num_)
/ 1048576.0));
std::fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
std::fprintf(
stdout, "Values: %d bytes each (%d bytes after compression)\n",
FLAGS_value_size,
static_cast<int>(FLAGS_value_size * FLAGS_compression_ratio + 0.5));
std::fprintf(stdout, "Entries: %d\n", num_);
std::fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_) /
1048576.0));
std::fprintf(
stdout, "FileSize: %.1f MB (estimated)\n",
(((kKeySize + FLAGS_value_size * FLAGS_compression_ratio) * num_) /
1048576.0));
PrintWarnings();
fprintf(stdout, "------------------------------------------------\n");
std::fprintf(stdout, "------------------------------------------------\n");
}
void PrintWarnings() {
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
fprintf(stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n"
);
std::fprintf(
stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n");
#endif
#ifndef NDEBUG
fprintf(stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
std::fprintf(
stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
#endif
// See if snappy is working by attempting to compress a compressible string
const char text[] = "yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy";
std::string compressed;
if (!port::Snappy_Compress(text, sizeof(text), &compressed)) {
fprintf(stdout, "WARNING: Snappy compression is not enabled\n");
std::fprintf(stdout, "WARNING: Snappy compression is not enabled\n");
} else if (compressed.size() >= sizeof(text)) {
fprintf(stdout, "WARNING: Snappy compression is not effective\n");
std::fprintf(stdout, "WARNING: Snappy compression is not effective\n");
}
}
void PrintEnvironment() {
fprintf(stderr, "LevelDB: version %d.%d\n",
kMajorVersion, kMinorVersion);
std::fprintf(stderr, "LevelDB: version %d.%d\n", kMajorVersion,
kMinorVersion);
#if defined(__linux)
time_t now = time(nullptr);
fprintf(stderr, "Date: %s", ctime(&now)); // ctime() adds newline
std::fprintf(stderr, "Date: %s",
ctime(&now)); // ctime() adds newline
FILE* cpuinfo = fopen("/proc/cpuinfo", "r");
FILE* cpuinfo = std::fopen("/proc/cpuinfo", "r");
if (cpuinfo != nullptr) {
char line[1000];
int num_cpus = 0;
@ -393,30 +451,32 @@ class Benchmark {
cache_size = val.ToString();
}
}
fclose(cpuinfo);
fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
std::fclose(cpuinfo);
std::fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
std::fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
}
#endif
}
public:
Benchmark()
: cache_(FLAGS_cache_size >= 0 ? NewLRUCache(FLAGS_cache_size) : nullptr),
filter_policy_(FLAGS_bloom_bits >= 0
? NewBloomFilterPolicy(FLAGS_bloom_bits)
: nullptr),
db_(nullptr),
num_(FLAGS_num),
value_size_(FLAGS_value_size),
entries_per_batch_(1),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
heap_counter_(0) {
: cache_(FLAGS_cache_size >= 0 ? NewLRUCache(FLAGS_cache_size) : nullptr),
filter_policy_(FLAGS_bloom_bits >= 0
? NewBloomFilterPolicy(FLAGS_bloom_bits)
: nullptr),
db_(nullptr),
num_(FLAGS_num),
value_size_(FLAGS_value_size),
entries_per_batch_(1),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
heap_counter_(0),
count_comparator_(BytewiseComparator()),
total_thread_count_(0) {
std::vector<std::string> files;
g_env->GetChildren(FLAGS_db, &files);
for (size_t i = 0; i < files.size(); i++) {
if (Slice(files[i]).starts_with("heap-")) {
g_env->DeleteFile(std::string(FLAGS_db) + "/" + files[i]);
g_env->RemoveFile(std::string(FLAGS_db) + "/" + files[i]);
}
}
if (!FLAGS_use_existing_db) {
@ -494,6 +554,8 @@ class Benchmark {
method = &Benchmark::ReadMissing;
} else if (name == Slice("seekrandom")) {
method = &Benchmark::SeekRandom;
} else if (name == Slice("seekordered")) {
method = &Benchmark::SeekOrdered;
} else if (name == Slice("readhot")) {
method = &Benchmark::ReadHot;
} else if (name == Slice("readrandomsmall")) {
@ -510,8 +572,6 @@ class Benchmark {
method = &Benchmark::Compact;
} else if (name == Slice("crc32c")) {
method = &Benchmark::Crc32c;
} else if (name == Slice("acquireload")) {
method = &Benchmark::AcquireLoad;
} else if (name == Slice("snappycomp")) {
method = &Benchmark::SnappyCompress;
} else if (name == Slice("snappyuncomp")) {
@ -523,15 +583,16 @@ class Benchmark {
} else if (name == Slice("sstables")) {
PrintStats("leveldb.sstables");
} else {
if (name != Slice()) { // No error message for empty name
fprintf(stderr, "unknown benchmark '%s'\n", name.ToString().c_str());
if (!name.empty()) { // No error message for empty name
std::fprintf(stderr, "unknown benchmark '%s'\n",
name.ToString().c_str());
}
}
if (fresh_db) {
if (FLAGS_use_existing_db) {
fprintf(stdout, "%-12s : skipped (--use_existing_db is true)\n",
name.ToString().c_str());
std::fprintf(stdout, "%-12s : skipped (--use_existing_db is true)\n",
name.ToString().c_str());
method = nullptr;
} else {
delete db_;
@ -592,7 +653,11 @@ class Benchmark {
arg[i].bm = this;
arg[i].method = method;
arg[i].shared = &shared;
arg[i].thread = new ThreadState(i);
++total_thread_count_;
// Seed the thread's random state deterministically based upon thread
// creation across all benchmarks. This ensures that the seeds are unique
// but reproducible when rerunning the same set of benchmarks.
arg[i].thread = new ThreadState(i, /*seed=*/1000 + total_thread_count_);
arg[i].thread->shared = &shared;
g_env->StartThread(ThreadBody, &arg[i]);
}
@ -613,6 +678,11 @@ class Benchmark {
arg[0].thread->stats.Merge(arg[i].thread->stats);
}
arg[0].thread->stats.Report(name);
if (FLAGS_comparisons) {
fprintf(stdout, "Comparisons: %zu\n", count_comparator_.comparisons());
count_comparator_.reset();
fflush(stdout);
}
for (int i = 0; i < n; i++) {
delete arg[i].thread;
@ -633,28 +703,12 @@ class Benchmark {
bytes += size;
}
// Print so result is not dead
fprintf(stderr, "... crc=0x%x\r", static_cast<unsigned int>(crc));
std::fprintf(stderr, "... crc=0x%x\r", static_cast<unsigned int>(crc));
thread->stats.AddBytes(bytes);
thread->stats.AddMessage(label);
}
void AcquireLoad(ThreadState* thread) {
int dummy;
port::AtomicPointer ap(&dummy);
int count = 0;
void *ptr = nullptr;
thread->stats.AddMessage("(each op is 1000 loads)");
while (count < 100000) {
for (int i = 0; i < 1000; i++) {
ptr = ap.Acquire_Load();
}
count++;
thread->stats.FinishedSingleOp();
}
if (ptr == nullptr) exit(1); // Disable unused variable warning.
}
void SnappyCompress(ThreadState* thread) {
RandomGenerator gen;
Slice input = gen.Generate(Options().block_size);
@ -673,8 +727,8 @@ class Benchmark {
thread->stats.AddMessage("(snappy failure)");
} else {
char buf[100];
snprintf(buf, sizeof(buf), "(output: %.1f%%)",
(produced * 100.0) / bytes);
std::snprintf(buf, sizeof(buf), "(output: %.1f%%)",
(produced * 100.0) / bytes);
thread->stats.AddMessage(buf);
thread->stats.AddBytes(bytes);
}
@ -688,8 +742,8 @@ class Benchmark {
int64_t bytes = 0;
char* uncompressed = new char[input.size()];
while (ok && bytes < 1024 * 1048576) { // Compress 1G
ok = port::Snappy_Uncompress(compressed.data(), compressed.size(),
uncompressed);
ok = port::Snappy_Uncompress(compressed.data(), compressed.size(),
uncompressed);
bytes += input.size();
thread->stats.FinishedSingleOp();
}
@ -711,13 +765,16 @@ class Benchmark {
options.write_buffer_size = FLAGS_write_buffer_size;
options.max_file_size = FLAGS_max_file_size;
options.block_size = FLAGS_block_size;
if (FLAGS_comparisons) {
options.comparator = &count_comparator_;
}
options.max_open_files = FLAGS_open_files;
options.filter_policy = filter_policy_;
options.reuse_logs = FLAGS_reuse_logs;
Status s = DB::Open(options, FLAGS_db, &db_);
if (!s.ok()) {
fprintf(stderr, "open error: %s\n", s.ToString().c_str());
exit(1);
std::fprintf(stderr, "open error: %s\n", s.ToString().c_str());
std::exit(1);
}
}
@ -729,18 +786,14 @@ class Benchmark {
}
}
void WriteSeq(ThreadState* thread) {
DoWrite(thread, true);
}
void WriteSeq(ThreadState* thread) { DoWrite(thread, true); }
void WriteRandom(ThreadState* thread) {
DoWrite(thread, false);
}
void WriteRandom(ThreadState* thread) { DoWrite(thread, false); }
void DoWrite(ThreadState* thread, bool seq) {
if (num_ != FLAGS_num) {
char msg[100];
snprintf(msg, sizeof(msg), "(%d ops)", num_);
std::snprintf(msg, sizeof(msg), "(%d ops)", num_);
thread->stats.AddMessage(msg);
}
@ -748,20 +801,20 @@ class Benchmark {
WriteBatch batch;
Status s;
int64_t bytes = 0;
KeyBuffer key;
for (int i = 0; i < num_; i += entries_per_batch_) {
batch.Clear();
for (int j = 0; j < entries_per_batch_; j++) {
const int k = seq ? i+j : (thread->rand.Next() % FLAGS_num);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
batch.Put(key, gen.Generate(value_size_));
bytes += value_size_ + strlen(key);
const int k = seq ? i + j : thread->rand.Uniform(FLAGS_num);
key.Set(k);
batch.Put(key.slice(), gen.Generate(value_size_));
bytes += value_size_ + key.slice().size();
thread->stats.FinishedSingleOp();
}
s = db_->Write(write_options_, &batch);
if (!s.ok()) {
fprintf(stderr, "put error: %s\n", s.ToString().c_str());
exit(1);
std::fprintf(stderr, "put error: %s\n", s.ToString().c_str());
std::exit(1);
}
}
thread->stats.AddBytes(bytes);
@ -797,28 +850,29 @@ class Benchmark {
ReadOptions options;
std::string value;
int found = 0;
KeyBuffer key;
for (int i = 0; i < reads_; i++) {
char key[100];
const int k = thread->rand.Next() % FLAGS_num;
snprintf(key, sizeof(key), "%016d", k);
if (db_->Get(options, key, &value).ok()) {
const int k = thread->rand.Uniform(FLAGS_num);
key.Set(k);
if (db_->Get(options, key.slice(), &value).ok()) {
found++;
}
thread->stats.FinishedSingleOp();
}
char msg[100];
snprintf(msg, sizeof(msg), "(%d of %d found)", found, num_);
std::snprintf(msg, sizeof(msg), "(%d of %d found)", found, num_);
thread->stats.AddMessage(msg);
}
void ReadMissing(ThreadState* thread) {
ReadOptions options;
std::string value;
KeyBuffer key;
for (int i = 0; i < reads_; i++) {
char key[100];
const int k = thread->rand.Next() % FLAGS_num;
snprintf(key, sizeof(key), "%016d.", k);
db_->Get(options, key, &value);
const int k = thread->rand.Uniform(FLAGS_num);
key.Set(k);
Slice s = Slice(key.slice().data(), key.slice().size() - 1);
db_->Get(options, s, &value);
thread->stats.FinishedSingleOp();
}
}
@ -827,11 +881,11 @@ class Benchmark {
ReadOptions options;
std::string value;
const int range = (FLAGS_num + 99) / 100;
KeyBuffer key;
for (int i = 0; i < reads_; i++) {
char key[100];
const int k = thread->rand.Next() % range;
snprintf(key, sizeof(key), "%016d", k);
db_->Get(options, key, &value);
const int k = thread->rand.Uniform(range);
key.Set(k);
db_->Get(options, key.slice(), &value);
thread->stats.FinishedSingleOp();
}
}
@ -839,13 +893,13 @@ class Benchmark {
void SeekRandom(ThreadState* thread) {
ReadOptions options;
int found = 0;
KeyBuffer key;
for (int i = 0; i < reads_; i++) {
Iterator* iter = db_->NewIterator(options);
char key[100];
const int k = thread->rand.Next() % FLAGS_num;
snprintf(key, sizeof(key), "%016d", k);
iter->Seek(key);
if (iter->Valid() && iter->key() == key) found++;
const int k = thread->rand.Uniform(FLAGS_num);
key.Set(k);
iter->Seek(key.slice());
if (iter->Valid() && iter->key() == key.slice()) found++;
delete iter;
thread->stats.FinishedSingleOp();
}
@ -854,34 +908,49 @@ class Benchmark {
thread->stats.AddMessage(msg);
}
void SeekOrdered(ThreadState* thread) {
ReadOptions options;
Iterator* iter = db_->NewIterator(options);
int found = 0;
int k = 0;
KeyBuffer key;
for (int i = 0; i < reads_; i++) {
k = (k + (thread->rand.Uniform(100))) % FLAGS_num;
key.Set(k);
iter->Seek(key.slice());
if (iter->Valid() && iter->key() == key.slice()) found++;
thread->stats.FinishedSingleOp();
}
delete iter;
char msg[100];
std::snprintf(msg, sizeof(msg), "(%d of %d found)", found, num_);
thread->stats.AddMessage(msg);
}
void DoDelete(ThreadState* thread, bool seq) {
RandomGenerator gen;
WriteBatch batch;
Status s;
KeyBuffer key;
for (int i = 0; i < num_; i += entries_per_batch_) {
batch.Clear();
for (int j = 0; j < entries_per_batch_; j++) {
const int k = seq ? i+j : (thread->rand.Next() % FLAGS_num);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
batch.Delete(key);
const int k = seq ? i + j : (thread->rand.Uniform(FLAGS_num));
key.Set(k);
batch.Delete(key.slice());
thread->stats.FinishedSingleOp();
}
s = db_->Write(write_options_, &batch);
if (!s.ok()) {
fprintf(stderr, "del error: %s\n", s.ToString().c_str());
exit(1);
std::fprintf(stderr, "del error: %s\n", s.ToString().c_str());
std::exit(1);
}
}
}
void DeleteSeq(ThreadState* thread) {
DoDelete(thread, true);
}
void DeleteSeq(ThreadState* thread) { DoDelete(thread, true); }
void DeleteRandom(ThreadState* thread) {
DoDelete(thread, false);
}
void DeleteRandom(ThreadState* thread) { DoDelete(thread, false); }
void ReadWhileWriting(ThreadState* thread) {
if (thread->tid > 0) {
@ -889,6 +958,7 @@ class Benchmark {
} else {
// Special thread that keeps writing until other threads are done.
RandomGenerator gen;
KeyBuffer key;
while (true) {
{
MutexLock l(&thread->shared->mu);
@ -898,13 +968,13 @@ class Benchmark {
}
}
const int k = thread->rand.Next() % FLAGS_num;
char key[100];
snprintf(key, sizeof(key), "%016d", k);
Status s = db_->Put(write_options_, key, gen.Generate(value_size_));
const int k = thread->rand.Uniform(FLAGS_num);
key.Set(k);
Status s =
db_->Put(write_options_, key.slice(), gen.Generate(value_size_));
if (!s.ok()) {
fprintf(stderr, "put error: %s\n", s.ToString().c_str());
exit(1);
std::fprintf(stderr, "put error: %s\n", s.ToString().c_str());
std::exit(1);
}
}
@ -913,16 +983,14 @@ class Benchmark {
}
}
void Compact(ThreadState* thread) {
db_->CompactRange(nullptr, nullptr);
}
void Compact(ThreadState* thread) { db_->CompactRange(nullptr, nullptr); }
void PrintStats(const char* key) {
std::string stats;
if (!db_->GetProperty(key, &stats)) {
stats = "(failed)";
}
fprintf(stdout, "\n%s\n", stats.c_str());
std::fprintf(stdout, "\n%s\n", stats.c_str());
}
static void WriteToFile(void* arg, const char* buf, int n) {
@ -931,18 +999,19 @@ class Benchmark {
void HeapProfile() {
char fname[100];
snprintf(fname, sizeof(fname), "%s/heap-%04d", FLAGS_db, ++heap_counter_);
std::snprintf(fname, sizeof(fname), "%s/heap-%04d", FLAGS_db,
++heap_counter_);
WritableFile* file;
Status s = g_env->NewWritableFile(fname, &file);
if (!s.ok()) {
fprintf(stderr, "%s\n", s.ToString().c_str());
std::fprintf(stderr, "%s\n", s.ToString().c_str());
return;
}
bool ok = port::GetHeapProfile(WriteToFile, file);
delete file;
if (!ok) {
fprintf(stderr, "heap profiling not supported\n");
g_env->DeleteFile(fname);
std::fprintf(stderr, "heap profiling not supported\n");
g_env->RemoveFile(fname);
}
}
};
@ -967,6 +1036,9 @@ int main(int argc, char** argv) {
} else if (sscanf(argv[i], "--histogram=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_histogram = n;
} else if (sscanf(argv[i], "--comparisons=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_comparisons = n;
} else if (sscanf(argv[i], "--use_existing_db=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_use_existing_db = n;
@ -987,6 +1059,8 @@ int main(int argc, char** argv) {
FLAGS_max_file_size = n;
} else if (sscanf(argv[i], "--block_size=%d%c", &n, &junk) == 1) {
FLAGS_block_size = n;
} else if (sscanf(argv[i], "--key_prefix=%d%c", &n, &junk) == 1) {
FLAGS_key_prefix = n;
} else if (sscanf(argv[i], "--cache_size=%d%c", &n, &junk) == 1) {
FLAGS_cache_size = n;
} else if (sscanf(argv[i], "--bloom_bits=%d%c", &n, &junk) == 1) {
@ -996,8 +1070,8 @@ int main(int argc, char** argv) {
} else if (strncmp(argv[i], "--db=", 5) == 0) {
FLAGS_db = argv[i] + 5;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);
std::fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
std::exit(1);
}
}
@ -1005,9 +1079,9 @@ int main(int argc, char** argv) {
// Choose a location for the test database if none given with --db=<path>
if (FLAGS_db == nullptr) {
leveldb::g_env->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
leveldb::g_env->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
}
leveldb::Benchmark benchmark;

92
benchmarks/db_bench_log.cc Normal file
View File

@ -0,0 +1,92 @@
// Copyright (c) 2019 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 <cinttypes>
#include <cstdio>
#include <string>
#include "gtest/gtest.h"
#include "benchmark/benchmark.h"
#include "db/version_set.h"
#include "leveldb/comparator.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "leveldb/options.h"
#include "port/port.h"
#include "util/mutexlock.h"
#include "util/testutil.h"
namespace leveldb {
namespace {
std::string MakeKey(unsigned int num) {
char buf[30];
std::snprintf(buf, sizeof(buf), "%016u", num);
return std::string(buf);
}
void BM_LogAndApply(benchmark::State& state) {
const int num_base_files = state.range(0);
std::string dbname = testing::TempDir() + "leveldb_test_benchmark";
DestroyDB(dbname, Options());
DB* db = nullptr;
Options opts;
opts.create_if_missing = true;
Status s = DB::Open(opts, dbname, &db);
ASSERT_LEVELDB_OK(s);
ASSERT_TRUE(db != nullptr);
delete db;
db = nullptr;
Env* env = Env::Default();
port::Mutex mu;
MutexLock l(&mu);
InternalKeyComparator cmp(BytewiseComparator());
Options options;
VersionSet vset(dbname, &options, nullptr, &cmp);
bool save_manifest;
ASSERT_LEVELDB_OK(vset.Recover(&save_manifest));
VersionEdit vbase;
uint64_t fnum = 1;
for (int i = 0; i < num_base_files; i++) {
InternalKey start(MakeKey(2 * fnum), 1, kTypeValue);
InternalKey limit(MakeKey(2 * fnum + 1), 1, kTypeDeletion);
vbase.AddFile(2, fnum++, 1 /* file size */, start, limit);
}
ASSERT_LEVELDB_OK(vset.LogAndApply(&vbase, &mu));
uint64_t start_micros = env->NowMicros();
for (auto st : state) {
VersionEdit vedit;
vedit.RemoveFile(2, fnum);
InternalKey start(MakeKey(2 * fnum), 1, kTypeValue);
InternalKey limit(MakeKey(2 * fnum + 1), 1, kTypeDeletion);
vedit.AddFile(2, fnum++, 1 /* file size */, start, limit);
vset.LogAndApply(&vedit, &mu);
}
uint64_t stop_micros = env->NowMicros();
unsigned int us = stop_micros - start_micros;
char buf[16];
std::snprintf(buf, sizeof(buf), "%d", num_base_files);
std::fprintf(stderr,
"BM_LogAndApply/%-6s %8" PRIu64
" iters : %9u us (%7.0f us / iter)\n",
buf, state.iterations(), us, ((float)us) / state.iterations());
}
BENCHMARK(BM_LogAndApply)->Arg(1)->Arg(100)->Arg(10000)->Arg(100000);
} // namespace
} // namespace leveldb
BENCHMARK_MAIN();

237
doc/bench/db_bench_sqlite3.cc → benchmarks/db_bench_sqlite3.cc
View File

@ -2,9 +2,11 @@
// 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 <stdio.h>
#include <stdlib.h>
#include <sqlite3.h>
#include <cstdio>
#include <cstdlib>
#include "util/histogram.h"
#include "util/random.h"
#include "util/testutil.h"
@ -38,8 +40,7 @@ static const char* FLAGS_benchmarks =
"fillrand100K,"
"fillseq100K,"
"readseq,"
"readrand100K,"
;
"readrand100K,";
// Number of key/values to place in database
static int FLAGS_num = 1000000;
@ -69,6 +70,9 @@ static int FLAGS_num_pages = 4096;
// benchmark will fail.
static bool FLAGS_use_existing_db = false;
// If true, the SQLite table has ROWIDs.
static bool FLAGS_use_rowids = false;
// If true, we allow batch writes to occur
static bool FLAGS_transaction = true;
@ -78,33 +82,29 @@ static bool FLAGS_WAL_enabled = true;
// Use the db with the following name.
static const char* FLAGS_db = nullptr;
inline
static void ExecErrorCheck(int status, char *err_msg) {
inline static void ExecErrorCheck(int status, char* err_msg) {
if (status != SQLITE_OK) {
fprintf(stderr, "SQL error: %s\n", err_msg);
std::fprintf(stderr, "SQL error: %s\n", err_msg);
sqlite3_free(err_msg);
exit(1);
std::exit(1);
}
}
inline
static void StepErrorCheck(int status) {
inline static void StepErrorCheck(int status) {
if (status != SQLITE_DONE) {
fprintf(stderr, "SQL step error: status = %d\n", status);
exit(1);
std::fprintf(stderr, "SQL step error: status = %d\n", status);
std::exit(1);
}
}
inline
static void ErrorCheck(int status) {
inline static void ErrorCheck(int status) {
if (status != SQLITE_OK) {
fprintf(stderr, "sqlite3 error: status = %d\n", status);
exit(1);
std::fprintf(stderr, "sqlite3 error: status = %d\n", status);
std::exit(1);
}
}
inline
static void WalCheckpoint(sqlite3* db_) {
inline static void WalCheckpoint(sqlite3* db_) {
// Flush all writes to disk
if (FLAGS_WAL_enabled) {
sqlite3_wal_checkpoint_v2(db_, nullptr, SQLITE_CHECKPOINT_FULL, nullptr,
@ -153,7 +153,7 @@ static Slice TrimSpace(Slice s) {
start++;
}
int limit = s.size();
while (limit > start && isspace(s[limit-1])) {
while (limit > start && isspace(s[limit - 1])) {
limit--;
}
return Slice(s.data() + start, limit - start);
@ -177,41 +177,43 @@ class Benchmark {
// State kept for progress messages
int done_;
int next_report_; // When to report next
int next_report_; // When to report next
void PrintHeader() {
const int kKeySize = 16;
PrintEnvironment();
fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
fprintf(stdout, "Values: %d bytes each\n", FLAGS_value_size);
fprintf(stdout, "Entries: %d\n", num_);
fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_)
/ 1048576.0));
std::fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
std::fprintf(stdout, "Values: %d bytes each\n", FLAGS_value_size);
std::fprintf(stdout, "Entries: %d\n", num_);
std::fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_) /
1048576.0));
PrintWarnings();
fprintf(stdout, "------------------------------------------------\n");
std::fprintf(stdout, "------------------------------------------------\n");
}
void PrintWarnings() {
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
fprintf(stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n"
);
std::fprintf(
stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n");
#endif
#ifndef NDEBUG
fprintf(stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
std::fprintf(
stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
#endif
}
void PrintEnvironment() {
fprintf(stderr, "SQLite: version %s\n", SQLITE_VERSION);
std::fprintf(stderr, "SQLite: version %s\n", SQLITE_VERSION);
#if defined(__linux)
time_t now = time(nullptr);
fprintf(stderr, "Date: %s", ctime(&now)); // ctime() adds newline
std::fprintf(stderr, "Date: %s",
ctime(&now)); // ctime() adds newline
FILE* cpuinfo = fopen("/proc/cpuinfo", "r");
FILE* cpuinfo = std::fopen("/proc/cpuinfo", "r");
if (cpuinfo != nullptr) {
char line[1000];
int num_cpus = 0;
@ -231,9 +233,9 @@ class Benchmark {
cache_size = val.ToString();
}
}
fclose(cpuinfo);
fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
std::fclose(cpuinfo);
std::fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
std::fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
}
#endif
}
@ -254,23 +256,30 @@ class Benchmark {
double micros = (now - last_op_finish_) * 1e6;
hist_.Add(micros);
if (micros > 20000) {
fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
fflush(stderr);
std::fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
std::fflush(stderr);
}
last_op_finish_ = now;
}
done_++;
if (done_ >= next_report_) {
if (next_report_ < 1000) next_report_ += 100;
else if (next_report_ < 5000) next_report_ += 500;
else if (next_report_ < 10000) next_report_ += 1000;
else if (next_report_ < 50000) next_report_ += 5000;
else if (next_report_ < 100000) next_report_ += 10000;
else if (next_report_ < 500000) next_report_ += 50000;
else next_report_ += 100000;
fprintf(stderr, "... finished %d ops%30s\r", done_, "");
fflush(stderr);
if (next_report_ < 1000)
next_report_ += 100;
else if (next_report_ < 5000)
next_report_ += 500;
else if (next_report_ < 10000)
next_report_ += 1000;
else if (next_report_ < 50000)
next_report_ += 5000;
else if (next_report_ < 100000)
next_report_ += 10000;
else if (next_report_ < 500000)
next_report_ += 50000;
else
next_report_ += 100000;
std::fprintf(stderr, "... finished %d ops%30s\r", done_, "");
std::fflush(stderr);
}
}
@ -283,43 +292,36 @@ class Benchmark {
if (bytes_ > 0) {
char rate[100];
snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
std::snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
if (!message_.empty()) {
message_ = std::string(rate) + " " + message_;
message_ = std::string(rate) + " " + message_;
} else {
message_ = rate;
}
}
fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(),
(finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "),
message_.c_str());
std::fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(), (finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "), message_.c_str());
if (FLAGS_histogram) {
fprintf(stdout, "Microseconds per op:\n%s\n", hist_.ToString().c_str());
std::fprintf(stdout, "Microseconds per op:\n%s\n",
hist_.ToString().c_str());
}
fflush(stdout);
std::fflush(stdout);
}
public:
enum Order {
SEQUENTIAL,
RANDOM
};
enum DBState {
FRESH,
EXISTING
};
enum Order { SEQUENTIAL, RANDOM };
enum DBState { FRESH, EXISTING };
Benchmark()
: db_(nullptr),
db_num_(0),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
: db_(nullptr),
db_num_(0),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
std::vector<std::string> files;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
@ -330,7 +332,7 @@ class Benchmark {
std::string file_name(test_dir);
file_name += "/";
file_name += files[i];
Env::Default()->DeleteFile(file_name.c_str());
Env::Default()->RemoveFile(file_name.c_str());
}
}
}
@ -406,7 +408,8 @@ class Benchmark {
} else {
known = false;
if (name != Slice()) { // No error message for empty name
fprintf(stderr, "unknown benchmark '%s'\n", name.ToString().c_str());
std::fprintf(stderr, "unknown benchmark '%s'\n",
name.ToString().c_str());
}
}
if (known) {
@ -426,28 +429,26 @@ class Benchmark {
// Open database
std::string tmp_dir;
Env::Default()->GetTestDirectory(&tmp_dir);
snprintf(file_name, sizeof(file_name),
"%s/dbbench_sqlite3-%d.db",
tmp_dir.c_str(),
db_num_);
std::snprintf(file_name, sizeof(file_name), "%s/dbbench_sqlite3-%d.db",
tmp_dir.c_str(), db_num_);
status = sqlite3_open(file_name, &db_);
if (status) {
fprintf(stderr, "open error: %s\n", sqlite3_errmsg(db_));
exit(1);
std::fprintf(stderr, "open error: %s\n", sqlite3_errmsg(db_));
std::exit(1);
}
// Change SQLite cache size
char cache_size[100];
snprintf(cache_size, sizeof(cache_size), "PRAGMA cache_size = %d",
FLAGS_num_pages);
std::snprintf(cache_size, sizeof(cache_size), "PRAGMA cache_size = %d",
FLAGS_num_pages);
status = sqlite3_exec(db_, cache_size, nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
// FLAGS_page_size is defaulted to 1024
if (FLAGS_page_size != 1024) {
char page_size[100];
snprintf(page_size, sizeof(page_size), "PRAGMA page_size = %d",
FLAGS_page_size);
std::snprintf(page_size, sizeof(page_size), "PRAGMA page_size = %d",
FLAGS_page_size);
status = sqlite3_exec(db_, page_size, nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
}
@ -460,26 +461,27 @@ class Benchmark {
std::string WAL_checkpoint = "PRAGMA wal_autocheckpoint = 4096";
status = sqlite3_exec(db_, WAL_stmt.c_str(), nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
status = sqlite3_exec(db_, WAL_checkpoint.c_str(), nullptr, nullptr,
&err_msg);
status =
sqlite3_exec(db_, WAL_checkpoint.c_str(), nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
}
// Change locking mode to exclusive and create tables/index for database
std::string locking_stmt = "PRAGMA locking_mode = EXCLUSIVE";
std::string create_stmt =
"CREATE TABLE test (key blob, value blob, PRIMARY KEY(key))";
std::string stmt_array[] = { locking_stmt, create_stmt };
"CREATE TABLE test (key blob, value blob, PRIMARY KEY(key))";
if (!FLAGS_use_rowids) create_stmt += " WITHOUT ROWID";
std::string stmt_array[] = {locking_stmt, create_stmt};
int stmt_array_length = sizeof(stmt_array) / sizeof(std::string);
for (int i = 0; i < stmt_array_length; i++) {
status = sqlite3_exec(db_, stmt_array[i].c_str(), nullptr, nullptr,
&err_msg);
status =
sqlite3_exec(db_, stmt_array[i].c_str(), nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
}
}
void Write(bool write_sync, Order order, DBState state,
int num_entries, int value_size, int entries_per_batch) {
void Write(bool write_sync, Order order, DBState state, int num_entries,
int value_size, int entries_per_batch) {
// Create new database if state == FRESH
if (state == FRESH) {
if (FLAGS_use_existing_db) {
@ -494,7 +496,7 @@ class Benchmark {
if (num_entries != num_) {
char msg[100];
snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
std::snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
message_ = msg;
}
@ -507,20 +509,20 @@ class Benchmark {
std::string end_trans_str = "END TRANSACTION;";
// Check for synchronous flag in options
std::string sync_stmt = (write_sync) ? "PRAGMA synchronous = FULL" :
"PRAGMA synchronous = OFF";
std::string sync_stmt =
(write_sync) ? "PRAGMA synchronous = FULL" : "PRAGMA synchronous = OFF";
status = sqlite3_exec(db_, sync_stmt.c_str(), nullptr, nullptr, &err_msg);
ExecErrorCheck(status, err_msg);
// Preparing sqlite3 statements
status = sqlite3_prepare_v2(db_, replace_str.c_str(), -1,
&replace_stmt, nullptr);
status = sqlite3_prepare_v2(db_, replace_str.c_str(), -1, &replace_stmt,
nullptr);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, begin_trans_str.c_str(), -1,
&begin_trans_stmt, nullptr);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1,
&end_trans_stmt, nullptr);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1, &end_trans_stmt,
nullptr);
ErrorCheck(status);
bool transaction = (entries_per_batch > 1);
@ -538,16 +540,16 @@ class Benchmark {
const char* value = gen_.Generate(value_size).data();
// Create values for key-value pair
const int k = (order == SEQUENTIAL) ? i + j :
(rand_.Next() % num_entries);
const int k =
(order == SEQUENTIAL) ? i + j : (rand_.Next() % num_entries);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
std::snprintf(key, sizeof(key), "%016d", k);
// Bind KV values into replace_stmt
status = sqlite3_bind_blob(replace_stmt, 1, key, 16, SQLITE_STATIC);
ErrorCheck(status);
status = sqlite3_bind_blob(replace_stmt, 2, value,
value_size, SQLITE_STATIC);
status = sqlite3_bind_blob(replace_stmt, 2, value, value_size,
SQLITE_STATIC);
ErrorCheck(status);
// Execute replace_stmt
@ -593,8 +595,8 @@ class Benchmark {
status = sqlite3_prepare_v2(db_, begin_trans_str.c_str(), -1,
&begin_trans_stmt, nullptr);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1,
&end_trans_stmt, nullptr);
status = sqlite3_prepare_v2(db_, end_trans_str.c_str(), -1, &end_trans_stmt,
nullptr);
ErrorCheck(status);
status = sqlite3_prepare_v2(db_, read_str.c_str(), -1, &read_stmt, nullptr);
ErrorCheck(status);
@ -614,14 +616,15 @@ class Benchmark {
// Create key value
char key[100];
int k = (order == SEQUENTIAL) ? i + j : (rand_.Next() % reads_);
snprintf(key, sizeof(key), "%016d", k);
std::snprintf(key, sizeof(key), "%016d", k);
// Bind key value into read_stmt
status = sqlite3_bind_blob(read_stmt, 1, key, 16, SQLITE_STATIC);
ErrorCheck(status);
// Execute read statement
while ((status = sqlite3_step(read_stmt)) == SQLITE_ROW) {}
while ((status = sqlite3_step(read_stmt)) == SQLITE_ROW) {
}
StepErrorCheck(status);
// Reset SQLite statement for another use
@ -651,7 +654,7 @@ class Benchmark {
void ReadSequential() {
int status;
sqlite3_stmt *pStmt;
sqlite3_stmt* pStmt;
std::string read_str = "SELECT * FROM test ORDER BY key";
status = sqlite3_prepare_v2(db_, read_str.c_str(), -1, &pStmt, nullptr);
@ -664,7 +667,6 @@ class Benchmark {
status = sqlite3_finalize(pStmt);
ErrorCheck(status);
}
};
} // namespace leveldb
@ -685,6 +687,9 @@ int main(int argc, char** argv) {
} else if (sscanf(argv[i], "--use_existing_db=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_use_existing_db = n;
} else if (sscanf(argv[i], "--use_rowids=%d%c", &n, &junk) == 1 &&
(n == 0 || n == 1)) {
FLAGS_use_rowids = n;
} else if (sscanf(argv[i], "--num=%d%c", &n, &junk) == 1) {
FLAGS_num = n;
} else if (sscanf(argv[i], "--reads=%d%c", &n, &junk) == 1) {
@ -703,16 +708,16 @@ int main(int argc, char** argv) {
} else if (strncmp(argv[i], "--db=", 5) == 0) {
FLAGS_db = argv[i] + 5;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);
std::fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
std::exit(1);
}
}
// Choose a location for the test database if none given with --db=<path>
if (FLAGS_db == nullptr) {
leveldb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
leveldb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
}
leveldb::Benchmark benchmark;

193
doc/bench/db_bench_tree_db.cc → benchmarks/db_bench_tree_db.cc
View File

@ -2,9 +2,11 @@
// 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 <stdio.h>
#include <stdlib.h>
#include <kcpolydb.h>
#include <cstdio>
#include <cstdlib>
#include "util/histogram.h"
#include "util/random.h"
#include "util/testutil.h"
@ -34,8 +36,7 @@ static const char* FLAGS_benchmarks =
"fillrand100K,"
"fillseq100K,"
"readseq100K,"
"readrand100K,"
;
"readrand100K,";
// Number of key/values to place in database
static int FLAGS_num = 1000000;
@ -71,12 +72,10 @@ static bool FLAGS_compression = true;
// Use the db with the following name.
static const char* FLAGS_db = nullptr;
inline
static void DBSynchronize(kyotocabinet::TreeDB* db_)
{
inline static void DBSynchronize(kyotocabinet::TreeDB* db_) {
// Synchronize will flush writes to disk
if (!db_->synchronize()) {
fprintf(stderr, "synchronize error: %s\n", db_->error().name());
std::fprintf(stderr, "synchronize error: %s\n", db_->error().name());
}
}
@ -121,7 +120,7 @@ static Slice TrimSpace(Slice s) {
start++;
}
int limit = s.size();
while (limit > start && isspace(s[limit-1])) {
while (limit > start && isspace(s[limit - 1])) {
limit--;
}
return Slice(s.data() + start, limit - start);
@ -146,47 +145,52 @@ class Benchmark {
// State kept for progress messages
int done_;
int next_report_; // When to report next
int next_report_; // When to report next
void PrintHeader() {
const int kKeySize = 16;
PrintEnvironment();
fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
fprintf(stdout, "Values: %d bytes each (%d bytes after compression)\n",
FLAGS_value_size,
static_cast<int>(FLAGS_value_size * FLAGS_compression_ratio + 0.5));
fprintf(stdout, "Entries: %d\n", num_);
fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_)
/ 1048576.0));
fprintf(stdout, "FileSize: %.1f MB (estimated)\n",
(((kKeySize + FLAGS_value_size * FLAGS_compression_ratio) * num_)
/ 1048576.0));
std::fprintf(stdout, "Keys: %d bytes each\n", kKeySize);
std::fprintf(
stdout, "Values: %d bytes each (%d bytes after compression)\n",
FLAGS_value_size,
static_cast<int>(FLAGS_value_size * FLAGS_compression_ratio + 0.5));
std::fprintf(stdout, "Entries: %d\n", num_);
std::fprintf(stdout, "RawSize: %.1f MB (estimated)\n",
((static_cast<int64_t>(kKeySize + FLAGS_value_size) * num_) /
1048576.0));
std::fprintf(
stdout, "FileSize: %.1f MB (estimated)\n",
(((kKeySize + FLAGS_value_size * FLAGS_compression_ratio) * num_) /
1048576.0));
PrintWarnings();
fprintf(stdout, "------------------------------------------------\n");
std::fprintf(stdout, "------------------------------------------------\n");
}
void PrintWarnings() {
#if defined(__GNUC__) && !defined(__OPTIMIZE__)
fprintf(stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n"
);
std::fprintf(
stdout,
"WARNING: Optimization is disabled: benchmarks unnecessarily slow\n");
#endif
#ifndef NDEBUG
fprintf(stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
std::fprintf(
stdout,
"WARNING: Assertions are enabled; benchmarks unnecessarily slow\n");
#endif
}
void PrintEnvironment() {
fprintf(stderr, "Kyoto Cabinet: version %s, lib ver %d, lib rev %d\n",
kyotocabinet::VERSION, kyotocabinet::LIBVER, kyotocabinet::LIBREV);
std::fprintf(
stderr, "Kyoto Cabinet: version %s, lib ver %d, lib rev %d\n",
kyotocabinet::VERSION, kyotocabinet::LIBVER, kyotocabinet::LIBREV);
#if defined(__linux)
time_t now = time(nullptr);
fprintf(stderr, "Date: %s", ctime(&now)); // ctime() adds newline
std::fprintf(stderr, "Date: %s",
ctime(&now)); // ctime() adds newline
FILE* cpuinfo = fopen("/proc/cpuinfo", "r");
FILE* cpuinfo = std::fopen("/proc/cpuinfo", "r");
if (cpuinfo != nullptr) {
char line[1000];
int num_cpus = 0;
@ -206,9 +210,10 @@ class Benchmark {
cache_size = val.ToString();
}
}
fclose(cpuinfo);
fprintf(stderr, "CPU: %d * %s\n", num_cpus, cpu_type.c_str());
fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
std::fclose(cpuinfo);
std::fprintf(stderr, "CPU: %d * %s\n", num_cpus,
cpu_type.c_str());
std::fprintf(stderr, "CPUCache: %s\n", cache_size.c_str());
}
#endif
}
@ -229,23 +234,30 @@ class Benchmark {
double micros = (now - last_op_finish_) * 1e6;
hist_.Add(micros);
if (micros > 20000) {
fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
fflush(stderr);
std::fprintf(stderr, "long op: %.1f micros%30s\r", micros, "");
std::fflush(stderr);
}
last_op_finish_ = now;
}
done_++;
if (done_ >= next_report_) {
if (next_report_ < 1000) next_report_ += 100;
else if (next_report_ < 5000) next_report_ += 500;
else if (next_report_ < 10000) next_report_ += 1000;
else if (next_report_ < 50000) next_report_ += 5000;
else if (next_report_ < 100000) next_report_ += 10000;
else if (next_report_ < 500000) next_report_ += 50000;
else next_report_ += 100000;
fprintf(stderr, "... finished %d ops%30s\r", done_, "");
fflush(stderr);
if (next_report_ < 1000)
next_report_ += 100;
else if (next_report_ < 5000)
next_report_ += 500;
else if (next_report_ < 10000)
next_report_ += 1000;
else if (next_report_ < 50000)
next_report_ += 5000;
else if (next_report_ < 100000)
next_report_ += 10000;
else if (next_report_ < 500000)
next_report_ += 50000;
else
next_report_ += 100000;
std::fprintf(stderr, "... finished %d ops%30s\r", done_, "");
std::fflush(stderr);
}
}
@ -258,42 +270,35 @@ class Benchmark {
if (bytes_ > 0) {
char rate[100];
snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
std::snprintf(rate, sizeof(rate), "%6.1f MB/s",
(bytes_ / 1048576.0) / (finish - start_));
if (!message_.empty()) {
message_ = std::string(rate) + " " + message_;
message_ = std::string(rate) + " " + message_;
} else {
message_ = rate;
}
}
fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(),
(finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "),
message_.c_str());
std::fprintf(stdout, "%-12s : %11.3f micros/op;%s%s\n",
name.ToString().c_str(), (finish - start_) * 1e6 / done_,
(message_.empty() ? "" : " "), message_.c_str());
if (FLAGS_histogram) {
fprintf(stdout, "Microseconds per op:\n%s\n", hist_.ToString().c_str());
std::fprintf(stdout, "Microseconds per op:\n%s\n",
hist_.ToString().c_str());
}
fflush(stdout);
std::fflush(stdout);
}
public:
enum Order {
SEQUENTIAL,
RANDOM
};
enum DBState {
FRESH,
EXISTING
};
enum Order { SEQUENTIAL, RANDOM };
enum DBState { FRESH, EXISTING };
Benchmark()
: db_(nullptr),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
: db_(nullptr),
num_(FLAGS_num),
reads_(FLAGS_reads < 0 ? FLAGS_num : FLAGS_reads),
bytes_(0),
rand_(301) {
std::vector<std::string> files;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
@ -304,7 +309,7 @@ class Benchmark {
std::string file_name(test_dir);
file_name += "/";
file_name += files[i];
Env::Default()->DeleteFile(file_name.c_str());
Env::Default()->RemoveFile(file_name.c_str());
}
}
}
@ -312,7 +317,7 @@ class Benchmark {
~Benchmark() {
if (!db_->close()) {
fprintf(stderr, "close error: %s\n", db_->error().name());
std::fprintf(stderr, "close error: %s\n", db_->error().name());
}
}
@ -376,7 +381,8 @@ class Benchmark {
} else {
known = false;
if (name != Slice()) { // No error message for empty name
fprintf(stderr, "unknown benchmark '%s'\n", name.ToString().c_str());
std::fprintf(stderr, "unknown benchmark '%s'\n",
name.ToString().c_str());
}
}
if (known) {
@ -386,7 +392,7 @@ class Benchmark {
}
private:
void Open(bool sync) {
void Open(bool sync) {
assert(db_ == nullptr);
// Initialize db_
@ -395,16 +401,14 @@ class Benchmark {
db_num_++;
std::string test_dir;
Env::Default()->GetTestDirectory(&test_dir);
snprintf(file_name, sizeof(file_name),
"%s/dbbench_polyDB-%d.kct",
test_dir.c_str(),
db_num_);
std::snprintf(file_name, sizeof(file_name), "%s/dbbench_polyDB-%d.kct",
test_dir.c_str(), db_num_);
// Create tuning options and open the database
int open_options = kyotocabinet::PolyDB::OWRITER |
kyotocabinet::PolyDB::OCREATE;
int tune_options = kyotocabinet::TreeDB::TSMALL |
kyotocabinet::TreeDB::TLINEAR;
int open_options =
kyotocabinet::PolyDB::OWRITER | kyotocabinet::PolyDB::OCREATE;
int tune_options =
kyotocabinet::TreeDB::TSMALL | kyotocabinet::TreeDB::TLINEAR;
if (FLAGS_compression) {
tune_options |= kyotocabinet::TreeDB::TCOMPRESS;
db_->tune_compressor(&comp_);
@ -412,17 +416,17 @@ class Benchmark {
db_->tune_options(tune_options);
db_->tune_page_cache(FLAGS_cache_size);
db_->tune_page(FLAGS_page_size);
db_->tune_map(256LL<<20);
db_->tune_map(256LL << 20);
if (sync) {
open_options |= kyotocabinet::PolyDB::OAUTOSYNC;
}
if (!db_->open(file_name, open_options)) {
fprintf(stderr, "open error: %s\n", db_->error().name());
std::fprintf(stderr, "open error: %s\n", db_->error().name());
}
}
void Write(bool sync, Order order, DBState state,
int num_entries, int value_size, int entries_per_batch) {
void Write(bool sync, Order order, DBState state, int num_entries,
int value_size, int entries_per_batch) {
// Create new database if state == FRESH
if (state == FRESH) {
if (FLAGS_use_existing_db) {
@ -437,20 +441,19 @@ class Benchmark {
if (num_entries != num_) {
char msg[100];
snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
std::snprintf(msg, sizeof(msg), "(%d ops)", num_entries);
message_ = msg;
}
// Write to database
for (int i = 0; i < num_entries; i++)
{
for (int i = 0; i < num_entries; i++) {
const int k = (order == SEQUENTIAL) ? i : (rand_.Next() % num_entries);
char key[100];
snprintf(key, sizeof(key), "%016d", k);
std::snprintf(key, sizeof(key), "%016d", k);
bytes_ += value_size + strlen(key);
std::string cpp_key = key;
if (!db_->set(cpp_key, gen_.Generate(value_size).ToString())) {
fprintf(stderr, "set error: %s\n", db_->error().name());
std::fprintf(stderr, "set error: %s\n", db_->error().name());
}
FinishedSingleOp();
}
@ -472,7 +475,7 @@ class Benchmark {
for (int i = 0; i < reads_; i++) {
char key[100];
const int k = rand_.Next() % reads_;
snprintf(key, sizeof(key), "%016d", k);
std::snprintf(key, sizeof(key), "%016d", k);
db_->get(key, &value);
FinishedSingleOp();
}
@ -510,16 +513,16 @@ int main(int argc, char** argv) {
} else if (strncmp(argv[i], "--db=", 5) == 0) {
FLAGS_db = argv[i] + 5;
} else {
fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
exit(1);
std::fprintf(stderr, "Invalid flag '%s'\n", argv[i]);
std::exit(1);
}
}
// Choose a location for the test database if none given with --db=<path>
if (FLAGS_db == nullptr) {
leveldb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
leveldb::Env::Default()->GetTestDirectory(&default_db_path);
default_db_path += "/dbbench";
FLAGS_db = default_db_path.c_str();
}
leveldb::Benchmark benchmark;

1
cmake/leveldbConfig.cmake
View File

@ -1 +0,0 @@
include("${CMAKE_CURRENT_LIST_DIR}/leveldbTargets.cmake")

9
cmake/leveldbConfig.cmake.in Normal file
View File

@ -0,0 +1,9 @@
# Copyright 2019 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.
@PACKAGE_INIT@
include("${CMAKE_CURRENT_LIST_DIR}/leveldbTargets.cmake")
check_required_components(leveldb)

54
db/autocompact_test.cc
View File

@ -2,29 +2,24 @@
// 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 "leveldb/db.h"
#include "gtest/gtest.h"
#include "db/db_impl.h"
#include "leveldb/cache.h"
#include "util/testharness.h"
#include "leveldb/db.h"
#include "util/testutil.h"
namespace leveldb {
class AutoCompactTest {
class AutoCompactTest : public testing::Test {
public:
std::string dbname_;
Cache* tiny_cache_;
Options options_;
DB* db_;
AutoCompactTest() {
dbname_ = test::TmpDir() + "/autocompact_test";
dbname_ = testing::TempDir() + "autocompact_test";
tiny_cache_ = NewLRUCache(100);
options_.block_cache = tiny_cache_;
DestroyDB(dbname_, options_);
options_.create_if_missing = true;
options_.compression = kNoCompression;
ASSERT_OK(DB::Open(options_, dbname_, &db_));
EXPECT_LEVELDB_OK(DB::Open(options_, dbname_, &db_));
}
~AutoCompactTest() {
@ -35,7 +30,7 @@ class AutoCompactTest {
std::string Key(int i) {
char buf[100];
snprintf(buf, sizeof(buf), "key%06d", i);
std::snprintf(buf, sizeof(buf), "key%06d", i);
return std::string(buf);
}
@ -47,6 +42,12 @@ class AutoCompactTest {
}
void DoReads(int n);
private:
std::string dbname_;
Cache* tiny_cache_;
Options options_;
DB* db_;
};
static const int kValueSize = 200 * 1024;
@ -61,15 +62,15 @@ void AutoCompactTest::DoReads(int n) {
// Fill database
for (int i = 0; i < kCount; i++) {
ASSERT_OK(db_->Put(WriteOptions(), Key(i), value));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), Key(i), value));
}
ASSERT_OK(dbi->TEST_CompactMemTable());
ASSERT_LEVELDB_OK(dbi->TEST_CompactMemTable());
// Delete everything
for (int i = 0; i < kCount; i++) {
ASSERT_OK(db_->Delete(WriteOptions(), Key(i)));
ASSERT_LEVELDB_OK(db_->Delete(WriteOptions(), Key(i)));
}
ASSERT_OK(dbi->TEST_CompactMemTable());
ASSERT_LEVELDB_OK(dbi->TEST_CompactMemTable());
// Get initial measurement of the space we will be reading.
const int64_t initial_size = Size(Key(0), Key(n));
@ -81,17 +82,16 @@ void AutoCompactTest::DoReads(int n) {
ASSERT_LT(read, 100) << "Taking too long to compact";
Iterator* iter = db_->NewIterator(ReadOptions());
for (iter->SeekToFirst();
iter->Valid() && iter->key().ToString() < limit_key;
iter->Next()) {
iter->Valid() && iter->key().ToString() < limit_key; iter->Next()) {
// Drop data
}
delete iter;
// Wait a little bit to allow any triggered compactions to complete.
Env::Default()->SleepForMicroseconds(1000000);
uint64_t size = Size(Key(0), Key(n));
fprintf(stderr, "iter %3d => %7.3f MB [other %7.3f MB]\n",
read+1, size/1048576.0, Size(Key(n), Key(kCount))/1048576.0);
if (size <= initial_size/10) {
std::fprintf(stderr, "iter %3d => %7.3f MB [other %7.3f MB]\n", read + 1,
size / 1048576.0, Size(Key(n), Key(kCount)) / 1048576.0);
if (size <= initial_size / 10) {
break;
}
}
@ -100,19 +100,11 @@ void AutoCompactTest::DoReads(int n) {
// is pretty much unchanged.
const int64_t final_other_size = Size(Key(n), Key(kCount));
ASSERT_LE(final_other_size, initial_other_size + 1048576);
ASSERT_GE(final_other_size, initial_other_size/5 - 1048576);
ASSERT_GE(final_other_size, initial_other_size / 5 - 1048576);
}
TEST(AutoCompactTest, ReadAll) {
DoReads(kCount);
}
TEST_F(AutoCompactTest, ReadAll) { DoReads(kCount); }
TEST(AutoCompactTest, ReadHalf) {
DoReads(kCount/2);
}
TEST_F(AutoCompactTest, ReadHalf) { DoReads(kCount / 2); }
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

22
db/builder.cc
View File

@ -4,8 +4,8 @@
#include "db/builder.h"
#include "db/filename.h"
#include "db/dbformat.h"
#include "db/filename.h"
#include "db/table_cache.h"
#include "db/version_edit.h"
#include "leveldb/db.h"
@ -14,12 +14,8 @@
namespace leveldb {
Status BuildTable(const std::string& dbname,
Env* env,
const Options& options,
TableCache* table_cache,
Iterator* iter,
FileMetaData* meta) {
Status BuildTable(const std::string& dbname, Env* env, const Options& options,
TableCache* table_cache, Iterator* iter, FileMetaData* meta) {
Status s;
meta->file_size = 0;
iter->SeekToFirst();
@ -34,11 +30,14 @@ Status BuildTable(const std::string& dbname,
TableBuilder* builder = new TableBuilder(options, file);
meta->smallest.DecodeFrom(iter->key());
Slice key;
for (; iter->Valid(); iter->Next()) {
Slice key = iter->key();
meta->largest.DecodeFrom(key);
key = iter->key();
builder->Add(key, iter->value());
}
if (!key.empty()) {
meta->largest.DecodeFrom(key);
}
// Finish and check for builder errors
s = builder->Finish();
@ -60,8 +59,7 @@ Status BuildTable(const std::string& dbname,
if (s.ok()) {
// Verify that the table is usable
Iterator* it = table_cache->NewIterator(ReadOptions(),
meta->number,
Iterator* it = table_cache->NewIterator(ReadOptions(), meta->number,
meta->file_size);
s = it->status();
delete it;
@ -76,7 +74,7 @@ Status BuildTable(const std::string& dbname,
if (s.ok() && meta->file_size > 0) {
// Keep it
} else {
env->DeleteFile(fname);
env->RemoveFile(fname);
}
return s;
}

8
db/builder.h
View File

@ -22,12 +22,8 @@ class VersionEdit;
// *meta will be filled with metadata about the generated table.
// If no data is present in *iter, meta->file_size will be set to
// zero, and no Table file will be produced.
Status BuildTable(const std::string& dbname,
Env* env,
const Options& options,
TableCache* table_cache,
Iterator* iter,
FileMetaData* meta);
Status BuildTable(const std::string& dbname, Env* env, const Options& options,
TableCache* table_cache, Iterator* iter, FileMetaData* meta);
} // namespace leveldb

367
db/c.cc
View File

@ -4,7 +4,11 @@
#include "leveldb/c.h"
#include <stdlib.h>
#include <string.h>
#include <cstdint>
#include <cstdlib>
#include "leveldb/cache.h"
#include "leveldb/comparator.h"
#include "leveldb/db.h"
@ -42,69 +46,72 @@ using leveldb::WriteOptions;
extern "C" {
struct leveldb_t { DB* rep; };
struct leveldb_iterator_t { Iterator* rep; };
struct leveldb_writebatch_t { WriteBatch rep; };
struct leveldb_snapshot_t { const Snapshot* rep; };
struct leveldb_readoptions_t { ReadOptions rep; };
struct leveldb_writeoptions_t { WriteOptions rep; };
struct leveldb_options_t { Options rep; };
struct leveldb_cache_t { Cache* rep; };
struct leveldb_seqfile_t { SequentialFile* rep; };
struct leveldb_randomfile_t { RandomAccessFile* rep; };
struct leveldb_writablefile_t { WritableFile* rep; };
struct leveldb_logger_t { Logger* rep; };
struct leveldb_filelock_t { FileLock* rep; };
struct leveldb_t {
DB* rep;
};
struct leveldb_iterator_t {
Iterator* rep;
};
struct leveldb_writebatch_t {
WriteBatch rep;
};
struct leveldb_snapshot_t {
const Snapshot* rep;
};
struct leveldb_readoptions_t {
ReadOptions rep;
};
struct leveldb_writeoptions_t {
WriteOptions rep;
};
struct leveldb_options_t {
Options rep;
};
struct leveldb_cache_t {
Cache* rep;
};
struct leveldb_seqfile_t {
SequentialFile* rep;
};
struct leveldb_randomfile_t {
RandomAccessFile* rep;
};
struct leveldb_writablefile_t {
WritableFile* rep;
};
struct leveldb_logger_t {
Logger* rep;
};
struct leveldb_filelock_t {
FileLock* rep;
};
struct leveldb_comparator_t : public Comparator {
void* state_;
void (*destructor_)(void*);
int (*compare_)(
void*,
const char* a, size_t alen,
const char* b, size_t blen);
const char* (*name_)(void*);
~leveldb_comparator_t() override { (*destructor_)(state_); }
virtual ~leveldb_comparator_t() {
(*destructor_)(state_);
}
virtual int Compare(const Slice& a, const Slice& b) const {
int Compare(const Slice& a, const Slice& b) const override {
return (*compare_)(state_, a.data(), a.size(), b.data(), b.size());
}
virtual const char* Name() const {
return (*name_)(state_);
}
const char* Name() const override { return (*name_)(state_); }
// No-ops since the C binding does not support key shortening methods.
virtual void FindShortestSeparator(std::string*, const Slice&) const { }
virtual void FindShortSuccessor(std::string* key) const { }
void FindShortestSeparator(std::string*, const Slice&) const override {}
void FindShortSuccessor(std::string* key) const override {}
void* state_;
void (*destructor_)(void*);
int (*compare_)(void*, const char* a, size_t alen, const char* b,
size_t blen);
const char* (*name_)(void*);
};
struct leveldb_filterpolicy_t : public FilterPolicy {
void* state_;
void (*destructor_)(void*);
const char* (*name_)(void*);
char* (*create_)(
void*,
const char* const* key_array, const size_t* key_length_array,
int num_keys,
size_t* filter_length);
unsigned char (*key_match_)(
void*,
const char* key, size_t length,
const char* filter, size_t filter_length);
~leveldb_filterpolicy_t() override { (*destructor_)(state_); }
virtual ~leveldb_filterpolicy_t() {
(*destructor_)(state_);
}
const char* Name() const override { return (*name_)(state_); }
virtual const char* Name() const {
return (*name_)(state_);
}
virtual void CreateFilter(const Slice* keys, int n, std::string* dst) const {
void CreateFilter(const Slice* keys, int n, std::string* dst) const override {
std::vector<const char*> key_pointers(n);
std::vector<size_t> key_sizes(n);
for (int i = 0; i < n; i++) {
@ -114,13 +121,22 @@ struct leveldb_filterpolicy_t : public FilterPolicy {
size_t len;
char* filter = (*create_)(state_, &key_pointers[0], &key_sizes[0], n, &len);
dst->append(filter, len);
free(filter);
std::free(filter);
}
virtual bool KeyMayMatch(const Slice& key, const Slice& filter) const {
return (*key_match_)(state_, key.data(), key.size(),
filter.data(), filter.size());
bool KeyMayMatch(const Slice& key, const Slice& filter) const override {
return (*key_match_)(state_, key.data(), key.size(), filter.data(),
filter.size());
}
void* state_;
void (*destructor_)(void*);
const char* (*name_)(void*);
char* (*create_)(void*, const char* const* key_array,
const size_t* key_length_array, int num_keys,
size_t* filter_length);
uint8_t (*key_match_)(void*, const char* key, size_t length,
const char* filter, size_t filter_length);
};
struct leveldb_env_t {
@ -136,22 +152,21 @@ static bool SaveError(char** errptr, const Status& s) {
*errptr = strdup(s.ToString().c_str());
} else {
// TODO(sanjay): Merge with existing error?
free(*errptr);
std::free(*errptr);
*errptr = strdup(s.ToString().c_str());
}
return true;
}
static char* CopyString(const std::string& str) {
char* result = reinterpret_cast<char*>(malloc(sizeof(char) * str.size()));
memcpy(result, str.data(), sizeof(char) * str.size());
char* result =
reinterpret_cast<char*>(std::malloc(sizeof(char) * str.size()));
std::memcpy(result, str.data(), sizeof(char) * str.size());
return result;
}
leveldb_t* leveldb_open(
const leveldb_options_t* options,
const char* name,
char** errptr) {
leveldb_t* leveldb_open(const leveldb_options_t* options, const char* name,
char** errptr) {
DB* db;
if (SaveError(errptr, DB::Open(options->rep, std::string(name), &db))) {
return nullptr;
@ -166,39 +181,26 @@ void leveldb_close(leveldb_t* db) {
delete db;
}
void leveldb_put(
leveldb_t* db,
const leveldb_writeoptions_t* options,
const char* key, size_t keylen,
const char* val, size_t vallen,
char** errptr) {
void leveldb_put(leveldb_t* db, const leveldb_writeoptions_t* options,
const char* key, size_t keylen, const char* val, size_t vallen,
char** errptr) {
SaveError(errptr,
db->rep->Put(options->rep, Slice(key, keylen), Slice(val, vallen)));
}
void leveldb_delete(
leveldb_t* db,
const leveldb_writeoptions_t* options,
const char* key, size_t keylen,
char** errptr) {
void leveldb_delete(leveldb_t* db, const leveldb_writeoptions_t* options,
const char* key, size_t keylen, char** errptr) {
SaveError(errptr, db->rep->Delete(options->rep, Slice(key, keylen)));
}
void leveldb_write(
leveldb_t* db,
const leveldb_writeoptions_t* options,
leveldb_writebatch_t* batch,
char** errptr) {
void leveldb_write(leveldb_t* db, const leveldb_writeoptions_t* options,
leveldb_writebatch_t* batch, char** errptr) {
SaveError(errptr, db->rep->Write(options->rep, &batch->rep));
}
char* leveldb_get(
leveldb_t* db,
const leveldb_readoptions_t* options,
const char* key, size_t keylen,
size_t* vallen,
char** errptr) {
char* leveldb_get(leveldb_t* db, const leveldb_readoptions_t* options,
const char* key, size_t keylen, size_t* vallen,
char** errptr) {
char* result = nullptr;
std::string tmp;
Status s = db->rep->Get(options->rep, Slice(key, keylen), &tmp);
@ -215,30 +217,25 @@ char* leveldb_get(
}
leveldb_iterator_t* leveldb_create_iterator(
leveldb_t* db,
const leveldb_readoptions_t* options) {
leveldb_t* db, const leveldb_readoptions_t* options) {
leveldb_iterator_t* result = new leveldb_iterator_t;
result->rep = db->rep->NewIterator(options->rep);
return result;
}
const leveldb_snapshot_t* leveldb_create_snapshot(
leveldb_t* db) {
const leveldb_snapshot_t* leveldb_create_snapshot(leveldb_t* db) {
leveldb_snapshot_t* result = new leveldb_snapshot_t;
result->rep = db->rep->GetSnapshot();
return result;
}
void leveldb_release_snapshot(
leveldb_t* db,
const leveldb_snapshot_t* snapshot) {
void leveldb_release_snapshot(leveldb_t* db,
const leveldb_snapshot_t* snapshot) {
db->rep->ReleaseSnapshot(snapshot->rep);
delete snapshot;
}
char* leveldb_property_value(
leveldb_t* db,
const char* propname) {
char* leveldb_property_value(leveldb_t* db, const char* propname) {
std::string tmp;
if (db->rep->GetProperty(Slice(propname), &tmp)) {
// We use strdup() since we expect human readable output.
@ -248,12 +245,12 @@ char* leveldb_property_value(
}
}
void leveldb_approximate_sizes(
leveldb_t* db,
int num_ranges,
const char* const* range_start_key, const size_t* range_start_key_len,
const char* const* range_limit_key, const size_t* range_limit_key_len,
uint64_t* sizes) {
void leveldb_approximate_sizes(leveldb_t* db, int num_ranges,
const char* const* range_start_key,
const size_t* range_start_key_len,
const char* const* range_limit_key,
const size_t* range_limit_key_len,
uint64_t* sizes) {
Range* ranges = new Range[num_ranges];
for (int i = 0; i < num_ranges; i++) {
ranges[i].start = Slice(range_start_key[i], range_start_key_len[i]);
@ -263,10 +260,9 @@ void leveldb_approximate_sizes(
delete[] ranges;
}
void leveldb_compact_range(
leveldb_t* db,
const char* start_key, size_t start_key_len,
const char* limit_key, size_t limit_key_len) {
void leveldb_compact_range(leveldb_t* db, const char* start_key,
size_t start_key_len, const char* limit_key,
size_t limit_key_len) {
Slice a, b;
db->rep->CompactRange(
// Pass null Slice if corresponding "const char*" is null
@ -274,17 +270,13 @@ void leveldb_compact_range(
(limit_key ? (b = Slice(limit_key, limit_key_len), &b) : nullptr));
}
void leveldb_destroy_db(
const leveldb_options_t* options,
const char* name,
char** errptr) {
void leveldb_destroy_db(const leveldb_options_t* options, const char* name,
char** errptr) {
SaveError(errptr, DestroyDB(name, options->rep));
}
void leveldb_repair_db(
const leveldb_options_t* options,
const char* name,
char** errptr) {
void leveldb_repair_db(const leveldb_options_t* options, const char* name,
char** errptr) {
SaveError(errptr, RepairDB(name, options->rep));
}
@ -293,7 +285,7 @@ void leveldb_iter_destroy(leveldb_iterator_t* iter) {
delete iter;
}
unsigned char leveldb_iter_valid(const leveldb_iterator_t* iter) {
uint8_t leveldb_iter_valid(const leveldb_iterator_t* iter) {
return iter->rep->Valid();
}
@ -309,13 +301,9 @@ void leveldb_iter_seek(leveldb_iterator_t* iter, const char* k, size_t klen) {
iter->rep->Seek(Slice(k, klen));
}
void leveldb_iter_next(leveldb_iterator_t* iter) {
iter->rep->Next();
}
void leveldb_iter_next(leveldb_iterator_t* iter) { iter->rep->Next(); }
void leveldb_iter_prev(leveldb_iterator_t* iter) {
iter->rep->Prev();
}
void leveldb_iter_prev(leveldb_iterator_t* iter) { iter->rep->Prev(); }
const char* leveldb_iter_key(const leveldb_iterator_t* iter, size_t* klen) {
Slice s = iter->rep->key();
@ -337,41 +325,34 @@ leveldb_writebatch_t* leveldb_writebatch_create() {
return new leveldb_writebatch_t;
}
void leveldb_writebatch_destroy(leveldb_writebatch_t* b) {
delete b;
}
void leveldb_writebatch_destroy(leveldb_writebatch_t* b) { delete b; }
void leveldb_writebatch_clear(leveldb_writebatch_t* b) {
b->rep.Clear();
}
void leveldb_writebatch_clear(leveldb_writebatch_t* b) { b->rep.Clear(); }
void leveldb_writebatch_put(
leveldb_writebatch_t* b,
const char* key, size_t klen,
const char* val, size_t vlen) {
void leveldb_writebatch_put(leveldb_writebatch_t* b, const char* key,
size_t klen, const char* val, size_t vlen) {
b->rep.Put(Slice(key, klen), Slice(val, vlen));
}
void leveldb_writebatch_delete(
leveldb_writebatch_t* b,
const char* key, size_t klen) {
void leveldb_writebatch_delete(leveldb_writebatch_t* b, const char* key,
size_t klen) {
b->rep.Delete(Slice(key, klen));
}
void leveldb_writebatch_iterate(
leveldb_writebatch_t* b,
void* state,
void (*put)(void*, const char* k, size_t klen, const char* v, size_t vlen),
void (*deleted)(void*, const char* k, size_t klen)) {
void leveldb_writebatch_iterate(const leveldb_writebatch_t* b, void* state,
void (*put)(void*, const char* k, size_t klen,
const char* v, size_t vlen),
void (*deleted)(void*, const char* k,
size_t klen)) {
class H : public WriteBatch::Handler {
public:
void* state_;
void (*put_)(void*, const char* k, size_t klen, const char* v, size_t vlen);
void (*deleted_)(void*, const char* k, size_t klen);
virtual void Put(const Slice& key, const Slice& value) {
void Put(const Slice& key, const Slice& value) override {
(*put_)(state_, key.data(), key.size(), value.data(), value.size());
}
virtual void Delete(const Slice& key) {
void Delete(const Slice& key) override {
(*deleted_)(state_, key.data(), key.size());
}
};
@ -382,38 +363,34 @@ void leveldb_writebatch_iterate(
b->rep.Iterate(&handler);
}
leveldb_options_t* leveldb_options_create() {
return new leveldb_options_t;
void leveldb_writebatch_append(leveldb_writebatch_t* destination,
const leveldb_writebatch_t* source) {
destination->rep.Append(source->rep);
}
void leveldb_options_destroy(leveldb_options_t* options) {
delete options;
}
leveldb_options_t* leveldb_options_create() { return new leveldb_options_t; }
void leveldb_options_set_comparator(
leveldb_options_t* opt,
leveldb_comparator_t* cmp) {
void leveldb_options_destroy(leveldb_options_t* options) { delete options; }
void leveldb_options_set_comparator(leveldb_options_t* opt,
leveldb_comparator_t* cmp) {
opt->rep.comparator = cmp;
}
void leveldb_options_set_filter_policy(
leveldb_options_t* opt,
leveldb_filterpolicy_t* policy) {
void leveldb_options_set_filter_policy(leveldb_options_t* opt,
leveldb_filterpolicy_t* policy) {
opt->rep.filter_policy = policy;
}
void leveldb_options_set_create_if_missing(
leveldb_options_t* opt, unsigned char v) {
void leveldb_options_set_create_if_missing(leveldb_options_t* opt, uint8_t v) {
opt->rep.create_if_missing = v;
}
void leveldb_options_set_error_if_exists(
leveldb_options_t* opt, unsigned char v) {
void leveldb_options_set_error_if_exists(leveldb_options_t* opt, uint8_t v) {
opt->rep.error_if_exists = v;
}
void leveldb_options_set_paranoid_checks(
leveldb_options_t* opt, unsigned char v) {
void leveldb_options_set_paranoid_checks(leveldb_options_t* opt, uint8_t v) {
opt->rep.paranoid_checks = v;
}
@ -454,12 +431,9 @@ void leveldb_options_set_compression(leveldb_options_t* opt, int t) {
}
leveldb_comparator_t* leveldb_comparator_create(
void* state,
void (*destructor)(void*),
int (*compare)(
void*,
const char* a, size_t alen,
const char* b, size_t blen),
void* state, void (*destructor)(void*),
int (*compare)(void*, const char* a, size_t alen, const char* b,
size_t blen),
const char* (*name)(void*)) {
leveldb_comparator_t* result = new leveldb_comparator_t;
result->state_ = state;
@ -469,22 +443,15 @@ leveldb_comparator_t* leveldb_comparator_create(
return result;
}
void leveldb_comparator_destroy(leveldb_comparator_t* cmp) {
delete cmp;
}
void leveldb_comparator_destroy(leveldb_comparator_t* cmp) { delete cmp; }
leveldb_filterpolicy_t* leveldb_filterpolicy_create(
void* state,
void (*destructor)(void*),
char* (*create_filter)(
void*,
const char* const* key_array, const size_t* key_length_array,
int num_keys,
size_t* filter_length),
unsigned char (*key_may_match)(
void*,
const char* key, size_t length,
const char* filter, size_t filter_length),
void* state, void (*destructor)(void*),
char* (*create_filter)(void*, const char* const* key_array,
const size_t* key_length_array, int num_keys,
size_t* filter_length),
uint8_t (*key_may_match)(void*, const char* key, size_t length,
const char* filter, size_t filter_length),
const char* (*name)(void*)) {
leveldb_filterpolicy_t* result = new leveldb_filterpolicy_t;
result->state_ = state;
@ -504,7 +471,8 @@ leveldb_filterpolicy_t* leveldb_filterpolicy_create_bloom(int bits_per_key) {
// they delegate to a NewBloomFilterPolicy() instead of user
// supplied C functions.
struct Wrapper : public leveldb_filterpolicy_t {
const FilterPolicy* rep_;
static void DoNothing(void*) {}
~Wrapper() { delete rep_; }
const char* Name() const { return rep_->Name(); }
void CreateFilter(const Slice* keys, int n, std::string* dst) const {
@ -513,7 +481,8 @@ leveldb_filterpolicy_t* leveldb_filterpolicy_create_bloom(int bits_per_key) {
bool KeyMayMatch(const Slice& key, const Slice& filter) const {
return rep_->KeyMayMatch(key, filter);
}
static void DoNothing(void*) { }
const FilterPolicy* rep_;
};
Wrapper* wrapper = new Wrapper;
wrapper->rep_ = NewBloomFilterPolicy(bits_per_key);
@ -526,24 +495,19 @@ leveldb_readoptions_t* leveldb_readoptions_create() {
return new leveldb_readoptions_t;
}
void leveldb_readoptions_destroy(leveldb_readoptions_t* opt) {
delete opt;
}
void leveldb_readoptions_destroy(leveldb_readoptions_t* opt) { delete opt; }
void leveldb_readoptions_set_verify_checksums(
leveldb_readoptions_t* opt,
unsigned char v) {
void leveldb_readoptions_set_verify_checksums(leveldb_readoptions_t* opt,
uint8_t v) {
opt->rep.verify_checksums = v;
}
void leveldb_readoptions_set_fill_cache(
leveldb_readoptions_t* opt, unsigned char v) {
void leveldb_readoptions_set_fill_cache(leveldb_readoptions_t* opt, uint8_t v) {
opt->rep.fill_cache = v;
}
void leveldb_readoptions_set_snapshot(
leveldb_readoptions_t* opt,
const leveldb_snapshot_t* snap) {
void leveldb_readoptions_set_snapshot(leveldb_readoptions_t* opt,
const leveldb_snapshot_t* snap) {
opt->rep.snapshot = (snap ? snap->rep : nullptr);
}
@ -551,12 +515,9 @@ leveldb_writeoptions_t* leveldb_writeoptions_create() {
return new leveldb_writeoptions_t;
}
void leveldb_writeoptions_destroy(leveldb_writeoptions_t* opt) {
delete opt;
}
void leveldb_writeoptions_destroy(leveldb_writeoptions_t* opt) { delete opt; }
void leveldb_writeoptions_set_sync(
leveldb_writeoptions_t* opt, unsigned char v) {
void leveldb_writeoptions_set_sync(leveldb_writeoptions_t* opt, uint8_t v) {
opt->rep.sync = v;
}
@ -589,22 +550,16 @@ char* leveldb_env_get_test_directory(leveldb_env_t* env) {
return nullptr;
}
char* buffer = static_cast<char*>(malloc(result.size() + 1));
memcpy(buffer, result.data(), result.size());
char* buffer = static_cast<char*>(std::malloc(result.size() + 1));
std::memcpy(buffer, result.data(), result.size());
buffer[result.size()] = '\0';
return buffer;
}
void leveldb_free(void* ptr) {
free(ptr);
}
void leveldb_free(void* ptr) { std::free(ptr); }
int leveldb_major_version() {
return kMajorVersion;
}
int leveldb_major_version() { return kMajorVersion; }
int leveldb_minor_version() {
return kMinorVersion;
}
int leveldb_minor_version() { return kMinorVersion; }
} // end extern "C"

16
db/c_test.c
View File

@ -120,7 +120,7 @@ static const char* CmpName(void* arg) {
}
// Custom filter policy
static unsigned char fake_filter_result = 1;
static uint8_t fake_filter_result = 1;
static void FilterDestroy(void* arg) { }
static const char* FilterName(void* arg) {
return "TestFilter";
@ -135,10 +135,8 @@ static char* FilterCreate(
memcpy(result, "fake", 4);
return result;
}
unsigned char FilterKeyMatch(
void* arg,
const char* key, size_t length,
const char* filter, size_t filter_length) {
uint8_t FilterKeyMatch(void* arg, const char* key, size_t length,
const char* filter, size_t filter_length) {
CheckCondition(filter_length == 4);
CheckCondition(memcmp(filter, "fake", 4) == 0);
return fake_filter_result;
@ -228,12 +226,18 @@ int main(int argc, char** argv) {
leveldb_writebatch_clear(wb);
leveldb_writebatch_put(wb, "bar", 3, "b", 1);
leveldb_writebatch_put(wb, "box", 3, "c", 1);
leveldb_writebatch_delete(wb, "bar", 3);
leveldb_writebatch_t* wb2 = leveldb_writebatch_create();
leveldb_writebatch_delete(wb2, "bar", 3);
leveldb_writebatch_append(wb, wb2);
leveldb_writebatch_destroy(wb2);
leveldb_write(db, woptions, wb, &err);
CheckNoError(err);
CheckGet(db, roptions, "foo", "hello");
CheckGet(db, roptions, "bar", NULL);
CheckGet(db, roptions, "box", "c");
int pos = 0;
leveldb_writebatch_iterate(wb, &pos, CheckPut, CheckDel);
CheckCondition(pos == 3);

154
db/corruption_test.cc
View File

@ -2,53 +2,42 @@
// 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 "leveldb/db.h"
#include <errno.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/types.h>
#include "leveldb/cache.h"
#include "leveldb/env.h"
#include "leveldb/table.h"
#include "leveldb/write_batch.h"
#include "gtest/gtest.h"
#include "db/db_impl.h"
#include "db/filename.h"
#include "db/log_format.h"
#include "db/version_set.h"
#include "leveldb/cache.h"
#include "leveldb/db.h"
#include "leveldb/table.h"
#include "leveldb/write_batch.h"
#include "util/logging.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
static const int kValueSize = 1000;
class CorruptionTest {
class CorruptionTest : public testing::Test {
public:
test::ErrorEnv env_;
std::string dbname_;
Cache* tiny_cache_;
Options options_;
DB* db_;
CorruptionTest() {
tiny_cache_ = NewLRUCache(100);
CorruptionTest()
: db_(nullptr),
dbname_("/memenv/corruption_test"),
tiny_cache_(NewLRUCache(100)) {
options_.env = &env_;
options_.block_cache = tiny_cache_;
dbname_ = test::TmpDir() + "/corruption_test";
DestroyDB(dbname_, options_);
db_ = nullptr;
options_.create_if_missing = true;
Reopen();
options_.create_if_missing = false;
}
~CorruptionTest() {
delete db_;
DestroyDB(dbname_, Options());
delete tiny_cache_;
delete db_;
delete tiny_cache_;
}
Status TryReopen() {
@ -57,21 +46,19 @@ class CorruptionTest {
return DB::Open(options_, dbname_, &db_);
}
void Reopen() {
ASSERT_OK(TryReopen());
}
void Reopen() { ASSERT_LEVELDB_OK(TryReopen()); }
void RepairDB() {
delete db_;
db_ = nullptr;
ASSERT_OK(::leveldb::RepairDB(dbname_, options_));
ASSERT_LEVELDB_OK(::leveldb::RepairDB(dbname_, options_));
}
void Build(int n) {
std::string key_space, value_space;
WriteBatch batch;
for (int i = 0; i < n; i++) {
//if ((i % 100) == 0) fprintf(stderr, "@ %d of %d\n", i, n);
// if ((i % 100) == 0) std::fprintf(stderr, "@ %d of %d\n", i, n);
Slice key = Key(i, &key_space);
batch.Clear();
batch.Put(key, Value(i, &value_space));
@ -81,7 +68,7 @@ class CorruptionTest {
if (i == n - 1) {
options.sync = true;
}
ASSERT_OK(db_->Write(options, &batch));
ASSERT_LEVELDB_OK(db_->Write(options, &batch));
}
}
@ -100,8 +87,7 @@ class CorruptionTest {
// Ignore boundary keys.
continue;
}
if (!ConsumeDecimalNumber(&in, &key) ||
!in.empty() ||
if (!ConsumeDecimalNumber(&in, &key) || !in.empty() ||
key < next_expected) {
bad_keys++;
continue;
@ -116,9 +102,10 @@ class CorruptionTest {
}
delete iter;
fprintf(stderr,
"expected=%d..%d; got=%d; bad_keys=%d; bad_values=%d; missed=%d\n",
min_expected, max_expected, correct, bad_keys, bad_values, missed);
std::fprintf(
stderr,
"expected=%d..%d; got=%d; bad_keys=%d; bad_values=%d; missed=%d\n",
min_expected, max_expected, correct, bad_keys, bad_values, missed);
ASSERT_LE(min_expected, correct);
ASSERT_GE(max_expected, correct);
}
@ -126,14 +113,13 @@ class CorruptionTest {
void Corrupt(FileType filetype, int offset, int bytes_to_corrupt) {
// Pick file to corrupt
std::vector<std::string> filenames;
ASSERT_OK(env_.GetChildren(dbname_, &filenames));
ASSERT_LEVELDB_OK(env_.target()->GetChildren(dbname_, &filenames));
uint64_t number;
FileType type;
std::string fname;
int picked_number = -1;
for (size_t i = 0; i < filenames.size(); i++) {
if (ParseFileName(filenames[i], &number, &type) &&
type == filetype &&
if (ParseFileName(filenames[i], &number, &type) && type == filetype &&
int(number) > picked_number) { // Pick latest file
fname = dbname_ + "/" + filenames[i];
picked_number = number;
@ -141,35 +127,32 @@ class CorruptionTest {
}
ASSERT_TRUE(!fname.empty()) << filetype;
struct stat sbuf;
if (stat(fname.c_str(), &sbuf) != 0) {
const char* msg = strerror(errno);
ASSERT_TRUE(false) << fname << ": " << msg;
}
uint64_t file_size;
ASSERT_LEVELDB_OK(env_.target()->GetFileSize(fname, &file_size));
if (offset < 0) {
// Relative to end of file; make it absolute
if (-offset > sbuf.st_size) {
if (-offset > file_size) {
offset = 0;
} else {
offset = sbuf.st_size + offset;
offset = file_size + offset;
}
}
if (offset > sbuf.st_size) {
offset = sbuf.st_size;
if (offset > file_size) {
offset = file_size;
}
if (offset + bytes_to_corrupt > sbuf.st_size) {
bytes_to_corrupt = sbuf.st_size - offset;
if (offset + bytes_to_corrupt > file_size) {
bytes_to_corrupt = file_size - offset;
}
// Do it
std::string contents;
Status s = ReadFileToString(Env::Default(), fname, &contents);
Status s = ReadFileToString(env_.target(), fname, &contents);
ASSERT_TRUE(s.ok()) << s.ToString();
for (int i = 0; i < bytes_to_corrupt; i++) {
contents[i + offset] ^= 0x80;
}
s = WriteStringToFile(Env::Default(), contents, fname);
s = WriteStringToFile(env_.target(), contents, fname);
ASSERT_TRUE(s.ok()) << s.ToString();
}
@ -187,7 +170,7 @@ class CorruptionTest {
// Return the ith key
Slice Key(int i, std::string* storage) {
char buf[100];
snprintf(buf, sizeof(buf), "%016d", i);
std::snprintf(buf, sizeof(buf), "%016d", i);
storage->assign(buf, strlen(buf));
return Slice(*storage);
}
@ -197,12 +180,20 @@ class CorruptionTest {
Random r(k);
return test::RandomString(&r, kValueSize, storage);
}
test::ErrorEnv env_;
Options options_;
DB* db_;
private:
std::string dbname_;
Cache* tiny_cache_;
};
TEST(CorruptionTest, Recovery) {
TEST_F(CorruptionTest, Recovery) {
Build(100);
Check(100, 100);
Corrupt(kLogFile, 19, 1); // WriteBatch tag for first record
Corrupt(kLogFile, 19, 1); // WriteBatch tag for first record
Corrupt(kLogFile, log::kBlockSize + 1000, 1); // Somewhere in second block
Reopen();
@ -210,13 +201,13 @@ TEST(CorruptionTest, Recovery) {
Check(36, 36);
}
TEST(CorruptionTest, RecoverWriteError) {
TEST_F(CorruptionTest, RecoverWriteError) {
env_.writable_file_error_ = true;
Status s = TryReopen();
ASSERT_TRUE(!s.ok());
}
TEST(CorruptionTest, NewFileErrorDuringWrite) {
TEST_F(CorruptionTest, NewFileErrorDuringWrite) {
// Do enough writing to force minor compaction
env_.writable_file_error_ = true;
const int num = 3 + (Options().write_buffer_size / kValueSize);
@ -233,7 +224,7 @@ TEST(CorruptionTest, NewFileErrorDuringWrite) {
Reopen();
}
TEST(CorruptionTest, TableFile) {
TEST_F(CorruptionTest, TableFile) {
Build(100);
DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
dbi->TEST_CompactMemTable();
@ -244,7 +235,7 @@ TEST(CorruptionTest, TableFile) {
Check(90, 99);
}
TEST(CorruptionTest, TableFileRepair) {
TEST_F(CorruptionTest, TableFileRepair) {
options_.block_size = 2 * kValueSize; // Limit scope of corruption
options_.paranoid_checks = true;
Reopen();
@ -260,7 +251,7 @@ TEST(CorruptionTest, TableFileRepair) {
Check(95, 99);
}
TEST(CorruptionTest, TableFileIndexData) {
TEST_F(CorruptionTest, TableFileIndexData) {
Build(10000); // Enough to build multiple Tables
DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
dbi->TEST_CompactMemTable();
@ -270,36 +261,36 @@ TEST(CorruptionTest, TableFileIndexData) {
Check(5000, 9999);
}
TEST(CorruptionTest, MissingDescriptor) {
TEST_F(CorruptionTest, MissingDescriptor) {
Build(1000);
RepairDB();
Reopen();
Check(1000, 1000);
}
TEST(CorruptionTest, SequenceNumberRecovery) {
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v1"));
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v2"));
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v3"));
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v4"));
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v5"));
TEST_F(CorruptionTest, SequenceNumberRecovery) {
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v1"));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v2"));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v3"));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v4"));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v5"));
RepairDB();
Reopen();
std::string v;
ASSERT_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("v5", v);
// Write something. If sequence number was not recovered properly,
// it will be hidden by an earlier write.
ASSERT_OK(db_->Put(WriteOptions(), "foo", "v6"));
ASSERT_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "v6"));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("v6", v);
Reopen();
ASSERT_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("v6", v);
}
TEST(CorruptionTest, CorruptedDescriptor) {
ASSERT_OK(db_->Put(WriteOptions(), "foo", "hello"));
TEST_F(CorruptionTest, CorruptedDescriptor) {
ASSERT_LEVELDB_OK(db_->Put(WriteOptions(), "foo", "hello"));
DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
dbi->TEST_CompactMemTable();
dbi->TEST_CompactRange(0, nullptr, nullptr);
@ -311,11 +302,11 @@ TEST(CorruptionTest, CorruptedDescriptor) {
RepairDB();
Reopen();
std::string v;
ASSERT_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), "foo", &v));
ASSERT_EQ("hello", v);
}
TEST(CorruptionTest, CompactionInputError) {
TEST_F(CorruptionTest, CompactionInputError) {
Build(10);
DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
dbi->TEST_CompactMemTable();
@ -330,7 +321,7 @@ TEST(CorruptionTest, CompactionInputError) {
Check(10000, 10000);
}
TEST(CorruptionTest, CompactionInputErrorParanoid) {
TEST_F(CorruptionTest, CompactionInputErrorParanoid) {
options_.paranoid_checks = true;
options_.write_buffer_size = 512 << 10;
Reopen();
@ -351,24 +342,21 @@ TEST(CorruptionTest, CompactionInputErrorParanoid) {
ASSERT_TRUE(!s.ok()) << "write did not fail in corrupted paranoid db";
}
TEST(CorruptionTest, UnrelatedKeys) {
TEST_F(CorruptionTest, UnrelatedKeys) {
Build(10);
DBImpl* dbi = reinterpret_cast<DBImpl*>(db_);
dbi->TEST_CompactMemTable();
Corrupt(kTableFile, 100, 1);
std::string tmp1, tmp2;
ASSERT_OK(db_->Put(WriteOptions(), Key(1000, &tmp1), Value(1000, &tmp2)));
ASSERT_LEVELDB_OK(
db_->Put(WriteOptions(), Key(1000, &tmp1), Value(1000, &tmp2)));
std::string v;
ASSERT_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
ASSERT_EQ(Value(1000, &tmp2).ToString(), v);
dbi->TEST_CompactMemTable();
ASSERT_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
ASSERT_LEVELDB_OK(db_->Get(ReadOptions(), Key(1000, &tmp1), &v));
ASSERT_EQ(Value(1000, &tmp2).ToString(), v);
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

346
db/db_impl.cc
View File

@ -4,10 +4,10 @@
#include "db/db_impl.h"
#include <stdint.h>
#include <stdio.h>
#include <algorithm>
#include <atomic>
#include <cstdint>
#include <cstdio>
#include <set>
#include <string>
#include <vector>
@ -41,16 +41,33 @@ const int kNumNonTableCacheFiles = 10;
// Information kept for every waiting writer
struct DBImpl::Writer {
explicit Writer(port::Mutex* mu)
: batch(nullptr), sync(false), done(false), cv(mu) {}
Status status;
WriteBatch* batch;
bool sync;
bool done;
port::CondVar cv;
explicit Writer(port::Mutex* mu) : cv(mu) { }
};
struct DBImpl::CompactionState {
// Files produced by compaction
struct Output {
uint64_t number;
uint64_t file_size;
InternalKey smallest, largest;
};
Output* current_output() { return &outputs[outputs.size() - 1]; }
explicit CompactionState(Compaction* c)
: compaction(c),
smallest_snapshot(0),
outfile(nullptr),
builder(nullptr),
total_bytes(0) {}
Compaction* const compaction;
// Sequence numbers < smallest_snapshot are not significant since we
@ -59,12 +76,6 @@ struct DBImpl::CompactionState {
// we can drop all entries for the same key with sequence numbers < S.
SequenceNumber smallest_snapshot;
// Files produced by compaction
struct Output {
uint64_t number;
uint64_t file_size;
InternalKey smallest, largest;
};
std::vector<Output> outputs;
// State kept for output being generated
@ -72,15 +83,6 @@ struct DBImpl::CompactionState {
TableBuilder* builder;
uint64_t total_bytes;
Output* current_output() { return &outputs[outputs.size()-1]; }
explicit CompactionState(Compaction* c)
: compaction(c),
outfile(nullptr),
builder(nullptr),
total_bytes(0) {
}
};
// Fix user-supplied options to be reasonable
@ -96,10 +98,10 @@ Options SanitizeOptions(const std::string& dbname,
Options result = src;
result.comparator = icmp;
result.filter_policy = (src.filter_policy != nullptr) ? ipolicy : nullptr;
ClipToRange(&result.max_open_files, 64 + kNumNonTableCacheFiles, 50000);
ClipToRange(&result.write_buffer_size, 64<<10, 1<<30);
ClipToRange(&result.max_file_size, 1<<20, 1<<30);
ClipToRange(&result.block_size, 1<<10, 4<<20);
ClipToRange(&result.max_open_files, 64 + kNumNonTableCacheFiles, 50000);
ClipToRange(&result.write_buffer_size, 64 << 10, 1 << 30);
ClipToRange(&result.max_file_size, 1 << 20, 1 << 30);
ClipToRange(&result.block_size, 1 << 10, 4 << 20);
if (result.info_log == nullptr) {
// Open a log file in the same directory as the db
src.env->CreateDir(dbname); // In case it does not exist
@ -132,10 +134,11 @@ DBImpl::DBImpl(const Options& raw_options, const std::string& dbname)
dbname_(dbname),
table_cache_(new TableCache(dbname_, options_, TableCacheSize(options_))),
db_lock_(nullptr),
shutting_down_(nullptr),
shutting_down_(false),
background_work_finished_signal_(&mutex_),
mem_(nullptr),
imm_(nullptr),
has_imm_(false),
logfile_(nullptr),
logfile_number_(0),
log_(nullptr),
@ -144,14 +147,12 @@ DBImpl::DBImpl(const Options& raw_options, const std::string& dbname)
background_compaction_scheduled_(false),
manual_compaction_(nullptr),
versions_(new VersionSet(dbname_, &options_, table_cache_,
&internal_comparator_)) {
has_imm_.Release_Store(nullptr);
}
&internal_comparator_)) {}
DBImpl::~DBImpl() {
// Wait for background work to finish
// Wait for background work to finish.
mutex_.Lock();
shutting_down_.Release_Store(this); // Any non-null value is ok
shutting_down_.store(true, std::memory_order_release);
while (background_compaction_scheduled_) {
background_work_finished_signal_.Wait();
}
@ -195,6 +196,9 @@ Status DBImpl::NewDB() {
std::string record;
new_db.EncodeTo(&record);
s = log.AddRecord(record);
if (s.ok()) {
s = file->Sync();
}
if (s.ok()) {
s = file->Close();
}
@ -204,7 +208,7 @@ Status DBImpl::NewDB() {
// Make "CURRENT" file that points to the new manifest file.
s = SetCurrentFile(env_, dbname_, 1);
} else {
env_->DeleteFile(manifest);
env_->RemoveFile(manifest);
}
return s;
}
@ -218,7 +222,7 @@ void DBImpl::MaybeIgnoreError(Status* s) const {
}
}
void DBImpl::DeleteObsoleteFiles() {
void DBImpl::RemoveObsoleteFiles() {
mutex_.AssertHeld();
if (!bg_error_.ok()) {
@ -235,8 +239,9 @@ void DBImpl::DeleteObsoleteFiles() {
env_->GetChildren(dbname_, &filenames); // Ignoring errors on purpose
uint64_t number;
FileType type;
for (size_t i = 0; i < filenames.size(); i++) {
if (ParseFileName(filenames[i], &number, &type)) {
std::vector<std::string> files_to_delete;
for (std::string& filename : filenames) {
if (ParseFileName(filename, &number, &type)) {
bool keep = true;
switch (type) {
case kLogFile:
@ -264,19 +269,27 @@ void DBImpl::DeleteObsoleteFiles() {
}
if (!keep) {
files_to_delete.push_back(std::move(filename));
if (type == kTableFile) {
table_cache_->Evict(number);
}
Log(options_.info_log, "Delete type=%d #%lld\n",
static_cast<int>(type),
Log(options_.info_log, "Delete type=%d #%lld\n", static_cast<int>(type),
static_cast<unsigned long long>(number));
env_->DeleteFile(dbname_ + "/" + filenames[i]);
}
}
}
// While deleting all files unblock other threads. All files being deleted
// have unique names which will not collide with newly created files and
// are therefore safe to delete while allowing other threads to proceed.
mutex_.Unlock();
for (const std::string& filename : files_to_delete) {
env_->RemoveFile(dbname_ + "/" + filename);
}
mutex_.Lock();
}
Status DBImpl::Recover(VersionEdit* edit, bool *save_manifest) {
Status DBImpl::Recover(VersionEdit* edit, bool* save_manifest) {
mutex_.AssertHeld();
// Ignore error from CreateDir since the creation of the DB is
@ -291,6 +304,8 @@ Status DBImpl::Recover(VersionEdit* edit, bool *save_manifest) {
if (!env_->FileExists(CurrentFileName(dbname_))) {
if (options_.create_if_missing) {
Log(options_.info_log, "Creating DB %s since it was missing.",
dbname_.c_str());
s = NewDB();
if (!s.ok()) {
return s;
@ -301,8 +316,8 @@ Status DBImpl::Recover(VersionEdit* edit, bool *save_manifest) {
}
} else {
if (options_.error_if_exists) {
return Status::InvalidArgument(
dbname_, "exists (error_if_exists is true)");
return Status::InvalidArgument(dbname_,
"exists (error_if_exists is true)");
}
}
@ -340,8 +355,8 @@ Status DBImpl::Recover(VersionEdit* edit, bool *save_manifest) {
}
if (!expected.empty()) {
char buf[50];
snprintf(buf, sizeof(buf), "%d missing files; e.g.",
static_cast<int>(expected.size()));
std::snprintf(buf, sizeof(buf), "%d missing files; e.g.",
static_cast<int>(expected.size()));
return Status::Corruption(buf, TableFileName(dbname_, *(expected.begin())));
}
@ -375,10 +390,10 @@ Status DBImpl::RecoverLogFile(uint64_t log_number, bool last_log,
Logger* info_log;
const char* fname;
Status* status; // null if options_.paranoid_checks==false
virtual void Corruption(size_t bytes, const Status& s) {
void Corruption(size_t bytes, const Status& s) override {
Log(info_log, "%s%s: dropping %d bytes; %s",
(this->status == nullptr ? "(ignoring error) " : ""),
fname, static_cast<int>(bytes), s.ToString().c_str());
(this->status == nullptr ? "(ignoring error) " : ""), fname,
static_cast<int>(bytes), s.ToString().c_str());
if (this->status != nullptr && this->status->ok()) *this->status = s;
}
};
@ -404,10 +419,9 @@ Status DBImpl::RecoverLogFile(uint64_t log_number, bool last_log,
// paranoid_checks==false so that corruptions cause entire commits
// to be skipped instead of propagating bad information (like overly
// large sequence numbers).
log::Reader reader(file, &reporter, true/*checksum*/,
0/*initial_offset*/);
log::Reader reader(file, &reporter, true /*checksum*/, 0 /*initial_offset*/);
Log(options_.info_log, "Recovering log #%llu",
(unsigned long long) log_number);
(unsigned long long)log_number);
// Read all the records and add to a memtable
std::string scratch;
@ -415,11 +429,10 @@ Status DBImpl::RecoverLogFile(uint64_t log_number, bool last_log,
WriteBatch batch;
int compactions = 0;
MemTable* mem = nullptr;
while (reader.ReadRecord(&record, &scratch) &&
status.ok()) {
while (reader.ReadRecord(&record, &scratch) && status.ok()) {
if (record.size() < 12) {
reporter.Corruption(
record.size(), Status::Corruption("log record too small"));
reporter.Corruption(record.size(),
Status::Corruption("log record too small"));
continue;
}
WriteBatchInternal::SetContents(&batch, record);
@ -433,9 +446,8 @@ Status DBImpl::RecoverLogFile(uint64_t log_number, bool last_log,
if (!status.ok()) {
break;
}
const SequenceNumber last_seq =
WriteBatchInternal::Sequence(&batch) +
WriteBatchInternal::Count(&batch) - 1;
const SequenceNumber last_seq = WriteBatchInternal::Sequence(&batch) +
WriteBatchInternal::Count(&batch) - 1;
if (last_seq > *max_sequence) {
*max_sequence = last_seq;
}
@ -499,7 +511,7 @@ Status DBImpl::WriteLevel0Table(MemTable* mem, VersionEdit* edit,
pending_outputs_.insert(meta.number);
Iterator* iter = mem->NewIterator();
Log(options_.info_log, "Level-0 table #%llu: started",
(unsigned long long) meta.number);
(unsigned long long)meta.number);
Status s;
{
@ -509,13 +521,11 @@ Status DBImpl::WriteLevel0Table(MemTable* mem, VersionEdit* edit,
}
Log(options_.info_log, "Level-0 table #%llu: %lld bytes %s",
(unsigned long long) meta.number,
(unsigned long long) meta.file_size,
(unsigned long long)meta.number, (unsigned long long)meta.file_size,
s.ToString().c_str());
delete iter;
pending_outputs_.erase(meta.number);
// Note that if file_size is zero, the file has been deleted and
// should not be added to the manifest.
int level = 0;
@ -525,8 +535,8 @@ Status DBImpl::WriteLevel0Table(MemTable* mem, VersionEdit* edit,
if (base != nullptr) {
level = base->PickLevelForMemTableOutput(min_user_key, max_user_key);
}
edit->AddFile(level, meta.number, meta.file_size,
meta.smallest, meta.largest);
edit->AddFile(level, meta.number, meta.file_size, meta.smallest,
meta.largest);
}
CompactionStats stats;
@ -547,7 +557,7 @@ void DBImpl::CompactMemTable() {
Status s = WriteLevel0Table(imm_, &edit, base);
base->Unref();
if (s.ok() && shutting_down_.Acquire_Load()) {
if (s.ok() && shutting_down_.load(std::memory_order_acquire)) {
s = Status::IOError("Deleting DB during memtable compaction");
}
@ -562,8 +572,8 @@ void DBImpl::CompactMemTable() {
// Commit to the new state
imm_->Unref();
imm_ = nullptr;
has_imm_.Release_Store(nullptr);
DeleteObsoleteFiles();
has_imm_.store(false, std::memory_order_release);
RemoveObsoleteFiles();
} else {
RecordBackgroundError(s);
}
@ -610,7 +620,8 @@ void DBImpl::TEST_CompactRange(int level, const Slice* begin,
}
MutexLock l(&mutex_);
while (!manual.done && !shutting_down_.Acquire_Load() && bg_error_.ok()) {
while (!manual.done && !shutting_down_.load(std::memory_order_acquire) &&
bg_error_.ok()) {
if (manual_compaction_ == nullptr) { // Idle
manual_compaction_ = &manual;
MaybeScheduleCompaction();
@ -652,12 +663,11 @@ void DBImpl::MaybeScheduleCompaction() {
mutex_.AssertHeld();
if (background_compaction_scheduled_) {
// Already scheduled
} else if (shutting_down_.Acquire_Load()) {
} else if (shutting_down_.load(std::memory_order_acquire)) {
// DB is being deleted; no more background compactions
} else if (!bg_error_.ok()) {
// Already got an error; no more changes
} else if (imm_ == nullptr &&
manual_compaction_ == nullptr &&
} else if (imm_ == nullptr && manual_compaction_ == nullptr &&
!versions_->NeedsCompaction()) {
// No work to be done
} else {
@ -673,7 +683,7 @@ void DBImpl::BGWork(void* db) {
void DBImpl::BackgroundCall() {
MutexLock l(&mutex_);
assert(background_compaction_scheduled_);
if (shutting_down_.Acquire_Load()) {
if (shutting_down_.load(std::memory_order_acquire)) {
// No more background work when shutting down.
} else if (!bg_error_.ok()) {
// No more background work after a background error.
@ -709,8 +719,7 @@ void DBImpl::BackgroundCompaction() {
}
Log(options_.info_log,
"Manual compaction at level-%d from %s .. %s; will stop at %s\n",
m->level,
(m->begin ? m->begin->DebugString().c_str() : "(begin)"),
m->level, (m->begin ? m->begin->DebugString().c_str() : "(begin)"),
(m->end ? m->end->DebugString().c_str() : "(end)"),
(m->done ? "(end)" : manual_end.DebugString().c_str()));
} else {
@ -724,20 +733,18 @@ void DBImpl::BackgroundCompaction() {
// Move file to next level
assert(c->num_input_files(0) == 1);
FileMetaData* f = c->input(0, 0);
c->edit()->DeleteFile(c->level(), f->number);
c->edit()->AddFile(c->level() + 1, f->number, f->file_size,
f->smallest, f->largest);
c->edit()->RemoveFile(c->level(), f->number);
c->edit()->AddFile(c->level() + 1, f->number, f->file_size, f->smallest,
f->largest);
status = versions_->LogAndApply(c->edit(), &mutex_);
if (!status.ok()) {
RecordBackgroundError(status);
}
VersionSet::LevelSummaryStorage tmp;
Log(options_.info_log, "Moved #%lld to level-%d %lld bytes %s: %s\n",
static_cast<unsigned long long>(f->number),
c->level() + 1,
static_cast<unsigned long long>(f->number), c->level() + 1,
static_cast<unsigned long long>(f->file_size),
status.ToString().c_str(),
versions_->LevelSummary(&tmp));
status.ToString().c_str(), versions_->LevelSummary(&tmp));
} else {
CompactionState* compact = new CompactionState(c);
status = DoCompactionWork(compact);
@ -746,17 +753,16 @@ void DBImpl::BackgroundCompaction() {
}
CleanupCompaction(compact);
c->ReleaseInputs();
DeleteObsoleteFiles();
RemoveObsoleteFiles();
}
delete c;
if (status.ok()) {
// Done
} else if (shutting_down_.Acquire_Load()) {
} else if (shutting_down_.load(std::memory_order_acquire)) {
// Ignore compaction errors found during shutting down
} else {
Log(options_.info_log,
"Compaction error: %s", status.ToString().c_str());
Log(options_.info_log, "Compaction error: %s", status.ToString().c_str());
}
if (is_manual) {
@ -851,31 +857,25 @@ Status DBImpl::FinishCompactionOutputFile(CompactionState* compact,
if (s.ok() && current_entries > 0) {
// Verify that the table is usable
Iterator* iter = table_cache_->NewIterator(ReadOptions(),
output_number,
current_bytes);
Iterator* iter =
table_cache_->NewIterator(ReadOptions(), output_number, current_bytes);
s = iter->status();
delete iter;
if (s.ok()) {
Log(options_.info_log,
"Generated table #%llu@%d: %lld keys, %lld bytes",
(unsigned long long) output_number,
compact->compaction->level(),
(unsigned long long) current_entries,
(unsigned long long) current_bytes);
Log(options_.info_log, "Generated table #%llu@%d: %lld keys, %lld bytes",
(unsigned long long)output_number, compact->compaction->level(),
(unsigned long long)current_entries,
(unsigned long long)current_bytes);
}
}
return s;
}
Status DBImpl::InstallCompactionResults(CompactionState* compact) {
mutex_.AssertHeld();
Log(options_.info_log, "Compacted %d@%d + %d@%d files => %lld bytes",
compact->compaction->num_input_files(0),
compact->compaction->level(),
compact->compaction->num_input_files(1),
compact->compaction->level() + 1,
Log(options_.info_log, "Compacted %d@%d + %d@%d files => %lld bytes",
compact->compaction->num_input_files(0), compact->compaction->level(),
compact->compaction->num_input_files(1), compact->compaction->level() + 1,
static_cast<long long>(compact->total_bytes));
// Add compaction outputs
@ -883,9 +883,8 @@ Status DBImpl::InstallCompactionResults(CompactionState* compact) {
const int level = compact->compaction->level();
for (size_t i = 0; i < compact->outputs.size(); i++) {
const CompactionState::Output& out = compact->outputs[i];
compact->compaction->edit()->AddFile(
level + 1,
out.number, out.file_size, out.smallest, out.largest);
compact->compaction->edit()->AddFile(level + 1, out.number, out.file_size,
out.smallest, out.largest);
}
return versions_->LogAndApply(compact->compaction->edit(), &mutex_);
}
@ -894,9 +893,8 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
const uint64_t start_micros = env_->NowMicros();
int64_t imm_micros = 0; // Micros spent doing imm_ compactions
Log(options_.info_log, "Compacting %d@%d + %d@%d files",
compact->compaction->num_input_files(0),
compact->compaction->level(),
Log(options_.info_log, "Compacting %d@%d + %d@%d files",
compact->compaction->num_input_files(0), compact->compaction->level(),
compact->compaction->num_input_files(1),
compact->compaction->level() + 1);
@ -906,22 +904,23 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
if (snapshots_.empty()) {
compact->smallest_snapshot = versions_->LastSequence();
} else {
compact->smallest_snapshot = snapshots_.oldest()->number_;
compact->smallest_snapshot = snapshots_.oldest()->sequence_number();
}
Iterator* input = versions_->MakeInputIterator(compact->compaction);
// Release mutex while we're actually doing the compaction work
mutex_.Unlock();
Iterator* input = versions_->MakeInputIterator(compact->compaction);
input->SeekToFirst();
Status status;
ParsedInternalKey ikey;
std::string current_user_key;
bool has_current_user_key = false;
SequenceNumber last_sequence_for_key = kMaxSequenceNumber;
for (; input->Valid() && !shutting_down_.Acquire_Load(); ) {
while (input->Valid() && !shutting_down_.load(std::memory_order_acquire)) {
// Prioritize immutable compaction work
if (has_imm_.NoBarrier_Load() != nullptr) {
if (has_imm_.load(std::memory_order_relaxed)) {
const uint64_t imm_start = env_->NowMicros();
mutex_.Lock();
if (imm_ != nullptr) {
@ -951,8 +950,8 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
last_sequence_for_key = kMaxSequenceNumber;
} else {
if (!has_current_user_key ||
user_comparator()->Compare(ikey.user_key,
Slice(current_user_key)) != 0) {
user_comparator()->Compare(ikey.user_key, Slice(current_user_key)) !=
0) {
// First occurrence of this user key
current_user_key.assign(ikey.user_key.data(), ikey.user_key.size());
has_current_user_key = true;
@ -961,7 +960,7 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
if (last_sequence_for_key <= compact->smallest_snapshot) {
// Hidden by an newer entry for same user key
drop = true; // (A)
drop = true; // (A)
} else if (ikey.type == kTypeDeletion &&
ikey.sequence <= compact->smallest_snapshot &&
compact->compaction->IsBaseLevelForKey(ikey.user_key)) {
@ -1014,7 +1013,7 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
input->Next();
}
if (status.ok() && shutting_down_.Acquire_Load()) {
if (status.ok() && shutting_down_.load(std::memory_order_acquire)) {
status = Status::IOError("Deleting DB during compaction");
}
if (status.ok() && compact->builder != nullptr) {
@ -1047,8 +1046,7 @@ Status DBImpl::DoCompactionWork(CompactionState* compact) {
RecordBackgroundError(status);
}
VersionSet::LevelSummaryStorage tmp;
Log(options_.info_log,
"compacted to: %s", versions_->LevelSummary(&tmp));
Log(options_.info_log, "compacted to: %s", versions_->LevelSummary(&tmp));
return status;
}
@ -1061,7 +1059,7 @@ struct IterState {
MemTable* const imm GUARDED_BY(mu);
IterState(port::Mutex* mutex, MemTable* mem, MemTable* imm, Version* version)
: mu(mutex), version(version), mem(mem), imm(imm) { }
: mu(mutex), version(version), mem(mem), imm(imm) {}
};
static void CleanupIteratorState(void* arg1, void* arg2) {
@ -1114,14 +1112,14 @@ int64_t DBImpl::TEST_MaxNextLevelOverlappingBytes() {
return versions_->MaxNextLevelOverlappingBytes();
}
Status DBImpl::Get(const ReadOptions& options,
const Slice& key,
Status DBImpl::Get(const ReadOptions& options, const Slice& key,
std::string* value) {
Status s;
MutexLock l(&mutex_);
SequenceNumber snapshot;
if (options.snapshot != nullptr) {
snapshot = reinterpret_cast<const SnapshotImpl*>(options.snapshot)->number_;
snapshot =
static_cast<const SnapshotImpl*>(options.snapshot)->sequence_number();
} else {
snapshot = versions_->LastSequence();
}
@ -1165,12 +1163,12 @@ Iterator* DBImpl::NewIterator(const ReadOptions& options) {
SequenceNumber latest_snapshot;
uint32_t seed;
Iterator* iter = NewInternalIterator(options, &latest_snapshot, &seed);
return NewDBIterator(
this, user_comparator(), iter,
(options.snapshot != nullptr
? reinterpret_cast<const SnapshotImpl*>(options.snapshot)->number_
: latest_snapshot),
seed);
return NewDBIterator(this, user_comparator(), iter,
(options.snapshot != nullptr
? static_cast<const SnapshotImpl*>(options.snapshot)
->sequence_number()
: latest_snapshot),
seed);
}
void DBImpl::RecordReadSample(Slice key) {
@ -1185,9 +1183,9 @@ const Snapshot* DBImpl::GetSnapshot() {
return snapshots_.New(versions_->LastSequence());
}
void DBImpl::ReleaseSnapshot(const Snapshot* s) {
void DBImpl::ReleaseSnapshot(const Snapshot* snapshot) {
MutexLock l(&mutex_);
snapshots_.Delete(reinterpret_cast<const SnapshotImpl*>(s));
snapshots_.Delete(static_cast<const SnapshotImpl*>(snapshot));
}
// Convenience methods
@ -1199,9 +1197,9 @@ Status DBImpl::Delete(const WriteOptions& options, const Slice& key) {
return DB::Delete(options, key);
}
Status DBImpl::Write(const WriteOptions& options, WriteBatch* my_batch) {
Status DBImpl::Write(const WriteOptions& options, WriteBatch* updates) {
Writer w(&mutex_);
w.batch = my_batch;
w.batch = updates;
w.sync = options.sync;
w.done = false;
@ -1215,13 +1213,13 @@ Status DBImpl::Write(const WriteOptions& options, WriteBatch* my_batch) {
}
// May temporarily unlock and wait.
Status status = MakeRoomForWrite(my_batch == nullptr);
Status status = MakeRoomForWrite(updates == nullptr);
uint64_t last_sequence = versions_->LastSequence();
Writer* last_writer = &w;
if (status.ok() && my_batch != nullptr) { // nullptr batch is for compactions
WriteBatch* updates = BuildBatchGroup(&last_writer);
WriteBatchInternal::SetSequence(updates, last_sequence + 1);
last_sequence += WriteBatchInternal::Count(updates);
if (status.ok() && updates != nullptr) { // nullptr batch is for compactions
WriteBatch* write_batch = BuildBatchGroup(&last_writer);
WriteBatchInternal::SetSequence(write_batch, last_sequence + 1);
last_sequence += WriteBatchInternal::Count(write_batch);
// Add to log and apply to memtable. We can release the lock
// during this phase since &w is currently responsible for logging
@ -1229,13 +1227,13 @@ Status DBImpl::Write(const WriteOptions& options, WriteBatch* my_batch) {
// into mem_.
{
mutex_.Unlock();
status = log_->AddRecord(WriteBatchInternal::Contents(updates));
status = log_->AddRecord(WriteBatchInternal::Contents(write_batch));
bool log_error = false;
if (status.ok() && options.sync) {
status = logfile_->Sync();
}
if (status.ok()) {
status = WriteBatchInternal::InsertInto(updates, mem_);
status = WriteBatchInternal::InsertInto(write_batch, mem_);
} else {
log_error = true;
}
@ -1247,7 +1245,7 @@ Status DBImpl::Write(const WriteOptions& options, WriteBatch* my_batch) {
RecordBackgroundError(status);
}
}
if (updates == tmp_batch_) tmp_batch_->Clear();
if (write_batch == tmp_batch_) tmp_batch_->Clear();
versions_->SetLastSequence(last_sequence);
}
@ -1286,8 +1284,8 @@ WriteBatch* DBImpl::BuildBatchGroup(Writer** last_writer) {
// original write is small, limit the growth so we do not slow
// down the small write too much.
size_t max_size = 1 << 20;
if (size <= (128<<10)) {
max_size = size + (128<<10);
if (size <= (128 << 10)) {
max_size = size + (128 << 10);
}
*last_writer = first;
@ -1333,9 +1331,8 @@ Status DBImpl::MakeRoomForWrite(bool force) {
// Yield previous error
s = bg_error_;
break;
} else if (
allow_delay &&
versions_->NumLevelFiles(0) >= config::kL0_SlowdownWritesTrigger) {
} else if (allow_delay && versions_->NumLevelFiles(0) >=
config::kL0_SlowdownWritesTrigger) {
// We are getting close to hitting a hard limit on the number of
// L0 files. Rather than delaying a single write by several
// seconds when we hit the hard limit, start delaying each
@ -1376,10 +1373,10 @@ Status DBImpl::MakeRoomForWrite(bool force) {
logfile_number_ = new_log_number;
log_ = new log::Writer(lfile);
imm_ = mem_;
has_imm_.Release_Store(imm_);
has_imm_.store(true, std::memory_order_release);
mem_ = new MemTable(internal_comparator_);
mem_->Ref();
force = false; // Do not force another compaction if have room
force = false; // Do not force another compaction if have room
MaybeScheduleCompaction();
}
}
@ -1403,31 +1400,26 @@ bool DBImpl::GetProperty(const Slice& property, std::string* value) {
return false;
} else {
char buf[100];
snprintf(buf, sizeof(buf), "%d",
versions_->NumLevelFiles(static_cast<int>(level)));
std::snprintf(buf, sizeof(buf), "%d",
versions_->NumLevelFiles(static_cast<int>(level)));
*value = buf;
return true;
}
} else if (in == "stats") {
char buf[200];
snprintf(buf, sizeof(buf),
" Compactions\n"
"Level Files Size(MB) Time(sec) Read(MB) Write(MB)\n"
"--------------------------------------------------\n"
);
std::snprintf(buf, sizeof(buf),
" Compactions\n"
"Level Files Size(MB) Time(sec) Read(MB) Write(MB)\n"
"--------------------------------------------------\n");
value->append(buf);
for (int level = 0; level < config::kNumLevels; level++) {
int files = versions_->NumLevelFiles(level);
if (stats_[level].micros > 0 || files > 0) {
snprintf(
buf, sizeof(buf),
"%3d %8d %8.0f %9.0f %8.0f %9.0f\n",
level,
files,
versions_->NumLevelBytes(level) / 1048576.0,
stats_[level].micros / 1e6,
stats_[level].bytes_read / 1048576.0,
stats_[level].bytes_written / 1048576.0);
std::snprintf(buf, sizeof(buf), "%3d %8d %8.0f %9.0f %8.0f %9.0f\n",
level, files, versions_->NumLevelBytes(level) / 1048576.0,
stats_[level].micros / 1e6,
stats_[level].bytes_read / 1048576.0,
stats_[level].bytes_written / 1048576.0);
value->append(buf);
}
}
@ -1444,8 +1436,8 @@ bool DBImpl::GetProperty(const Slice& property, std::string* value) {
total_usage += imm_->ApproximateMemoryUsage();
}
char buf[50];
snprintf(buf, sizeof(buf), "%llu",
static_cast<unsigned long long>(total_usage));
std::snprintf(buf, sizeof(buf), "%llu",
static_cast<unsigned long long>(total_usage));
value->append(buf);
return true;
}
@ -1453,16 +1445,11 @@ bool DBImpl::GetProperty(const Slice& property, std::string* value) {
return false;
}
void DBImpl::GetApproximateSizes(
const Range* range, int n,
uint64_t* sizes) {
void DBImpl::GetApproximateSizes(const Range* range, int n, uint64_t* sizes) {
// TODO(opt): better implementation
Version* v;
{
MutexLock l(&mutex_);
versions_->current()->Ref();
v = versions_->current();
}
MutexLock l(&mutex_);
Version* v = versions_->current();
v->Ref();
for (int i = 0; i < n; i++) {
// Convert user_key into a corresponding internal key.
@ -1473,10 +1460,7 @@ void DBImpl::GetApproximateSizes(
sizes[i] = (limit >= start ? limit - start : 0);
}
{
MutexLock l(&mutex_);
v->Unref();
}
v->Unref();
}
// Default implementations of convenience methods that subclasses of DB
@ -1493,10 +1477,9 @@ Status DB::Delete(const WriteOptions& opt, const Slice& key) {
return Write(opt, &batch);
}
DB::~DB() { }
DB::~DB() = default;
Status DB::Open(const Options& options, const std::string& dbname,
DB** dbptr) {
Status DB::Open(const Options& options, const std::string& dbname, DB** dbptr) {
*dbptr = nullptr;
DBImpl* impl = new DBImpl(options, dbname);
@ -1526,7 +1509,7 @@ Status DB::Open(const Options& options, const std::string& dbname,
s = impl->versions_->LogAndApply(&edit, &impl->mutex_);
}
if (s.ok()) {
impl->DeleteObsoleteFiles();
impl->RemoveObsoleteFiles();
impl->MaybeScheduleCompaction();
}
impl->mutex_.Unlock();
@ -1539,8 +1522,7 @@ Status DB::Open(const Options& options, const std::string& dbname,
return s;
}
Snapshot::~Snapshot() {
}
Snapshot::~Snapshot() = default;
Status DestroyDB(const std::string& dbname, const Options& options) {
Env* env = options.env;
@ -1560,15 +1542,15 @@ Status DestroyDB(const std::string& dbname, const Options& options) {
for (size_t i = 0; i < filenames.size(); i++) {
if (ParseFileName(filenames[i], &number, &type) &&
type != kDBLockFile) { // Lock file will be deleted at end
Status del = env->DeleteFile(dbname + "/" + filenames[i]);
Status del = env->RemoveFile(dbname + "/" + filenames[i]);
if (result.ok() && !del.ok()) {
result = del;
}
}
}
env->UnlockFile(lock); // Ignore error since state is already gone
env->DeleteFile(lockname);
env->DeleteDir(dbname); // Ignore error in case dir contains other files
env->RemoveFile(lockname);
env->RemoveDir(dbname); // Ignore error in case dir contains other files
}
return result;
}

101
db/db_impl.h
View File

@ -5,8 +5,11 @@
#ifndef STORAGE_LEVELDB_DB_DB_IMPL_H_
#define STORAGE_LEVELDB_DB_DB_IMPL_H_
#include <atomic>
#include <deque>
#include <set>
#include <string>
#include "db/dbformat.h"
#include "db/log_writer.h"
#include "db/snapshot.h"
@ -26,21 +29,25 @@ class VersionSet;
class DBImpl : public DB {
public:
DBImpl(const Options& options, const std::string& dbname);
virtual ~DBImpl();
DBImpl(const DBImpl&) = delete;
DBImpl& operator=(const DBImpl&) = delete;
~DBImpl() override;
// Implementations of the DB interface
virtual Status Put(const WriteOptions&, const Slice& key, const Slice& value);
virtual Status Delete(const WriteOptions&, const Slice& key);
virtual Status Write(const WriteOptions& options, WriteBatch* updates);
virtual Status Get(const ReadOptions& options,
const Slice& key,
std::string* value);
virtual Iterator* NewIterator(const ReadOptions&);
virtual const Snapshot* GetSnapshot();
virtual void ReleaseSnapshot(const Snapshot* snapshot);
virtual bool GetProperty(const Slice& property, std::string* value);
virtual void GetApproximateSizes(const Range* range, int n, uint64_t* sizes);
virtual void CompactRange(const Slice* begin, const Slice* end);
Status Put(const WriteOptions&, const Slice& key,
const Slice& value) override;
Status Delete(const WriteOptions&, const Slice& key) override;
Status Write(const WriteOptions& options, WriteBatch* updates) override;
Status Get(const ReadOptions& options, const Slice& key,
std::string* value) override;
Iterator* NewIterator(const ReadOptions&) override;
const Snapshot* GetSnapshot() override;
void ReleaseSnapshot(const Snapshot* snapshot) override;
bool GetProperty(const Slice& property, std::string* value) override;
void GetApproximateSizes(const Range* range, int n, uint64_t* sizes) override;
void CompactRange(const Slice* begin, const Slice* end) override;
// Extra methods (for testing) that are not in the public DB interface
@ -69,6 +76,31 @@ class DBImpl : public DB {
struct CompactionState;
struct Writer;
// Information for a manual compaction
struct ManualCompaction {
int level;
bool done;
const InternalKey* begin; // null means beginning of key range
const InternalKey* end; // null means end of key range
InternalKey tmp_storage; // Used to keep track of compaction progress
};
// Per level compaction stats. stats_[level] stores the stats for
// compactions that produced data for the specified "level".
struct CompactionStats {
CompactionStats() : micros(0), bytes_read(0), bytes_written(0) {}
void Add(const CompactionStats& c) {
this->micros += c.micros;
this->bytes_read += c.bytes_read;
this->bytes_written += c.bytes_written;
}
int64_t micros;
int64_t bytes_read;
int64_t bytes_written;
};
Iterator* NewInternalIterator(const ReadOptions&,
SequenceNumber* latest_snapshot,
uint32_t* seed);
@ -84,7 +116,7 @@ class DBImpl : public DB {
void MaybeIgnoreError(Status* s) const;
// Delete any unneeded files and stale in-memory entries.
void DeleteObsoleteFiles() EXCLUSIVE_LOCKS_REQUIRED(mutex_);
void RemoveObsoleteFiles() EXCLUSIVE_LOCKS_REQUIRED(mutex_);
// Compact the in-memory write buffer to disk. Switches to a new
// log-file/memtable and writes a new descriptor iff successful.
@ -119,6 +151,10 @@ class DBImpl : public DB {
Status InstallCompactionResults(CompactionState* compact)
EXCLUSIVE_LOCKS_REQUIRED(mutex_);
const Comparator* user_comparator() const {
return internal_comparator_.user_comparator();
}
// Constant after construction
Env* const env_;
const InternalKeyComparator internal_comparator_;
@ -136,11 +172,11 @@ class DBImpl : public DB {
// State below is protected by mutex_
port::Mutex mutex_;
port::AtomicPointer shutting_down_;
std::atomic<bool> shutting_down_;
port::CondVar background_work_finished_signal_ GUARDED_BY(mutex_);
MemTable* mem_;
MemTable* imm_ GUARDED_BY(mutex_); // Memtable being compacted
port::AtomicPointer has_imm_; // So bg thread can detect non-null imm_
std::atomic<bool> has_imm_; // So bg thread can detect non-null imm_
WritableFile* logfile_;
uint64_t logfile_number_ GUARDED_BY(mutex_);
log::Writer* log_;
@ -159,45 +195,14 @@ class DBImpl : public DB {
// Has a background compaction been scheduled or is running?
bool background_compaction_scheduled_ GUARDED_BY(mutex_);
// Information for a manual compaction
struct ManualCompaction {
int level;
bool done;
const InternalKey* begin; // null means beginning of key range
const InternalKey* end; // null means end of key range
InternalKey tmp_storage; // Used to keep track of compaction progress
};
ManualCompaction* manual_compaction_ GUARDED_BY(mutex_);
VersionSet* const versions_;
VersionSet* const versions_ GUARDED_BY(mutex_);
// Have we encountered a background error in paranoid mode?
Status bg_error_ GUARDED_BY(mutex_);
// Per level compaction stats. stats_[level] stores the stats for
// compactions that produced data for the specified "level".
struct CompactionStats {
int64_t micros;
int64_t bytes_read;
int64_t bytes_written;
CompactionStats() : micros(0), bytes_read(0), bytes_written(0) { }
void Add(const CompactionStats& c) {
this->micros += c.micros;
this->bytes_read += c.bytes_read;
this->bytes_written += c.bytes_written;
}
};
CompactionStats stats_[config::kNumLevels] GUARDED_BY(mutex_);
// No copying allowed
DBImpl(const DBImpl&);
void operator=(const DBImpl&);
const Comparator* user_comparator() const {
return internal_comparator_.user_comparator();
}
};
// Sanitize db options. The caller should delete result.info_log if

97
db/db_iter.cc
View File

@ -4,9 +4,9 @@
#include "db/db_iter.h"
#include "db/filename.h"
#include "db/db_impl.h"
#include "db/dbformat.h"
#include "db/filename.h"
#include "leveldb/env.h"
#include "leveldb/iterator.h"
#include "port/port.h"
@ -21,9 +21,9 @@ static void DumpInternalIter(Iterator* iter) {
for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
ParsedInternalKey k;
if (!ParseInternalKey(iter->key(), &k)) {
fprintf(stderr, "Corrupt '%s'\n", EscapeString(iter->key()).c_str());
std::fprintf(stderr, "Corrupt '%s'\n", EscapeString(iter->key()).c_str());
} else {
fprintf(stderr, "@ '%s'\n", k.DebugString().c_str());
std::fprintf(stderr, "@ '%s'\n", k.DebugString().c_str());
}
}
}
@ -36,17 +36,14 @@ namespace {
// combines multiple entries for the same userkey found in the DB
// representation into a single entry while accounting for sequence
// numbers, deletion markers, overwrites, etc.
class DBIter: public Iterator {
class DBIter : public Iterator {
public:
// Which direction is the iterator currently moving?
// (1) When moving forward, the internal iterator is positioned at
// the exact entry that yields this->key(), this->value()
// (2) When moving backwards, the internal iterator is positioned
// just before all entries whose user key == this->key().
enum Direction {
kForward,
kReverse
};
enum Direction { kForward, kReverse };
DBIter(DBImpl* db, const Comparator* cmp, Iterator* iter, SequenceNumber s,
uint32_t seed)
@ -57,21 +54,22 @@ class DBIter: public Iterator {
direction_(kForward),
valid_(false),
rnd_(seed),
bytes_counter_(RandomPeriod()) {
}
virtual ~DBIter() {
delete iter_;
}
virtual bool Valid() const { return valid_; }
virtual Slice key() const {
bytes_until_read_sampling_(RandomCompactionPeriod()) {}
DBIter(const DBIter&) = delete;
DBIter& operator=(const DBIter&) = delete;
~DBIter() override { delete iter_; }
bool Valid() const override { return valid_; }
Slice key() const override {
assert(valid_);
return (direction_ == kForward) ? ExtractUserKey(iter_->key()) : saved_key_;
}
virtual Slice value() const {
Slice value() const override {
assert(valid_);
return (direction_ == kForward) ? iter_->value() : saved_value_;
}
virtual Status status() const {
Status status() const override {
if (status_.ok()) {
return iter_->status();
} else {
@ -79,11 +77,11 @@ class DBIter: public Iterator {
}
}
virtual void Next();
virtual void Prev();
virtual void Seek(const Slice& target);
virtual void SeekToFirst();
virtual void SeekToLast();
void Next() override;
void Prev() override;
void Seek(const Slice& target) override;
void SeekToFirst() override;
void SeekToLast() override;
private:
void FindNextUserEntry(bool skipping, std::string* skip);
@ -103,38 +101,35 @@ class DBIter: public Iterator {
}
}
// Pick next gap with average value of config::kReadBytesPeriod.
ssize_t RandomPeriod() {
return rnd_.Uniform(2*config::kReadBytesPeriod);
// Picks the number of bytes that can be read until a compaction is scheduled.
size_t RandomCompactionPeriod() {
return rnd_.Uniform(2 * config::kReadBytesPeriod);
}
DBImpl* db_;
const Comparator* const user_comparator_;
Iterator* const iter_;
SequenceNumber const sequence_;
Status status_;
std::string saved_key_; // == current key when direction_==kReverse
std::string saved_value_; // == current raw value when direction_==kReverse
std::string saved_key_; // == current key when direction_==kReverse
std::string saved_value_; // == current raw value when direction_==kReverse
Direction direction_;
bool valid_;
Random rnd_;
ssize_t bytes_counter_;
// No copying allowed
DBIter(const DBIter&);
void operator=(const DBIter&);
size_t bytes_until_read_sampling_;
};
inline bool DBIter::ParseKey(ParsedInternalKey* ikey) {
Slice k = iter_->key();
ssize_t n = k.size() + iter_->value().size();
bytes_counter_ -= n;
while (bytes_counter_ < 0) {
bytes_counter_ += RandomPeriod();
size_t bytes_read = k.size() + iter_->value().size();
while (bytes_until_read_sampling_ < bytes_read) {
bytes_until_read_sampling_ += RandomCompactionPeriod();
db_->RecordReadSample(k);
}
assert(bytes_until_read_sampling_ >= bytes_read);
bytes_until_read_sampling_ -= bytes_read;
if (!ParseInternalKey(k, ikey)) {
status_ = Status::Corruption("corrupted internal key in DBIter");
return false;
@ -165,6 +160,15 @@ void DBIter::Next() {
} else {
// Store in saved_key_ the current key so we skip it below.
SaveKey(ExtractUserKey(iter_->key()), &saved_key_);
// iter_ is pointing to current key. We can now safely move to the next to
// avoid checking current key.
iter_->Next();
if (!iter_->Valid()) {
valid_ = false;
saved_key_.clear();
return;
}
}
FindNextUserEntry(true, &saved_key_);
@ -218,8 +222,8 @@ void DBIter::Prev() {
ClearSavedValue();
return;
}
if (user_comparator_->Compare(ExtractUserKey(iter_->key()),
saved_key_) < 0) {
if (user_comparator_->Compare(ExtractUserKey(iter_->key()), saved_key_) <
0) {
break;
}
}
@ -275,8 +279,8 @@ void DBIter::Seek(const Slice& target) {
direction_ = kForward;
ClearSavedValue();
saved_key_.clear();
AppendInternalKey(
&saved_key_, ParsedInternalKey(target, sequence_, kValueTypeForSeek));
AppendInternalKey(&saved_key_,
ParsedInternalKey(target, sequence_, kValueTypeForSeek));
iter_->Seek(saved_key_);
if (iter_->Valid()) {
FindNextUserEntry(false, &saved_key_ /* temporary storage */);
@ -305,12 +309,9 @@ void DBIter::SeekToLast() {
} // anonymous namespace
Iterator* NewDBIterator(
DBImpl* db,
const Comparator* user_key_comparator,
Iterator* internal_iter,
SequenceNumber sequence,
uint32_t seed) {
Iterator* NewDBIterator(DBImpl* db, const Comparator* user_key_comparator,
Iterator* internal_iter, SequenceNumber sequence,
uint32_t seed) {
return new DBIter(db, user_key_comparator, internal_iter, sequence, seed);
}

11
db/db_iter.h
View File

@ -5,9 +5,10 @@
#ifndef STORAGE_LEVELDB_DB_DB_ITER_H_
#define STORAGE_LEVELDB_DB_DB_ITER_H_
#include <stdint.h>
#include "leveldb/db.h"
#include <cstdint>
#include "db/dbformat.h"
#include "leveldb/db.h"
namespace leveldb {
@ -16,10 +17,8 @@ class DBImpl;
// Return a new iterator that converts internal keys (yielded by
// "*internal_iter") that were live at the specified "sequence" number
// into appropriate user keys.
Iterator* NewDBIterator(DBImpl* db,
const Comparator* user_key_comparator,
Iterator* internal_iter,
SequenceNumber sequence,
Iterator* NewDBIterator(DBImpl* db, const Comparator* user_key_comparator,
Iterator* internal_iter, SequenceNumber sequence,
uint32_t seed);
} // namespace leveldb

1042
db/db_test.cc
View File

File diff suppressed because it is too large Load Diff

44
db/dbformat.cc
View File

@ -2,8 +2,11 @@
// 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 <stdio.h>
#include "db/dbformat.h"
#include <cstdio>
#include <sstream>
#include "port/port.h"
#include "util/coding.h"
@ -21,26 +24,20 @@ void AppendInternalKey(std::string* result, const ParsedInternalKey& key) {
}
std::string ParsedInternalKey::DebugString() const {
char buf[50];
snprintf(buf, sizeof(buf), "' @ %llu : %d",
(unsigned long long) sequence,
int(type));
std::string result = "'";
result += EscapeString(user_key.ToString());
result += buf;
return result;
std::ostringstream ss;
ss << '\'' << EscapeString(user_key.ToString()) << "' @ " << sequence << " : "
<< static_cast<int>(type);
return ss.str();
}
std::string InternalKey::DebugString() const {
std::string result;
ParsedInternalKey parsed;
if (ParseInternalKey(rep_, &parsed)) {
result = parsed.DebugString();
} else {
result = "(bad)";
result.append(EscapeString(rep_));
return parsed.DebugString();
}
return result;
std::ostringstream ss;
ss << "(bad)" << EscapeString(rep_);
return ss.str();
}
const char* InternalKeyComparator::Name() const {
@ -65,9 +62,8 @@ int InternalKeyComparator::Compare(const Slice& akey, const Slice& bkey) const {
return r;
}
void InternalKeyComparator::FindShortestSeparator(
std::string* start,
const Slice& limit) const {
void InternalKeyComparator::FindShortestSeparator(std::string* start,
const Slice& limit) const {
// Attempt to shorten the user portion of the key
Slice user_start = ExtractUserKey(*start);
Slice user_limit = ExtractUserKey(limit);
@ -77,7 +73,8 @@ void InternalKeyComparator::FindShortestSeparator(
user_comparator_->Compare(user_start, tmp) < 0) {
// User key has become shorter physically, but larger logically.
// Tack on the earliest possible number to the shortened user key.
PutFixed64(&tmp, PackSequenceAndType(kMaxSequenceNumber,kValueTypeForSeek));
PutFixed64(&tmp,
PackSequenceAndType(kMaxSequenceNumber, kValueTypeForSeek));
assert(this->Compare(*start, tmp) < 0);
assert(this->Compare(tmp, limit) < 0);
start->swap(tmp);
@ -92,15 +89,14 @@ void InternalKeyComparator::FindShortSuccessor(std::string* key) const {
user_comparator_->Compare(user_key, tmp) < 0) {
// User key has become shorter physically, but larger logically.
// Tack on the earliest possible number to the shortened user key.
PutFixed64(&tmp, PackSequenceAndType(kMaxSequenceNumber,kValueTypeForSeek));
PutFixed64(&tmp,
PackSequenceAndType(kMaxSequenceNumber, kValueTypeForSeek));
assert(this->Compare(*key, tmp) < 0);
key->swap(tmp);
}
}
const char* InternalFilterPolicy::Name() const {
return user_policy_->Name();
}
const char* InternalFilterPolicy::Name() const { return user_policy_->Name(); }
void InternalFilterPolicy::CreateFilter(const Slice* keys, int n,
std::string* dst) const {
@ -130,7 +126,7 @@ LookupKey::LookupKey(const Slice& user_key, SequenceNumber s) {
start_ = dst;
dst = EncodeVarint32(dst, usize + 8);
kstart_ = dst;
memcpy(dst, user_key.data(), usize);
std::memcpy(dst, user_key.data(), usize);
dst += usize;
EncodeFixed64(dst, PackSequenceAndType(s, kValueTypeForSeek));
dst += 8;

66
db/dbformat.h
View File

@ -5,7 +5,10 @@
#ifndef STORAGE_LEVELDB_DB_DBFORMAT_H_
#define STORAGE_LEVELDB_DB_DBFORMAT_H_
#include <stdio.h>
#include <cstddef>
#include <cstdint>
#include <string>
#include "leveldb/comparator.h"
#include "leveldb/db.h"
#include "leveldb/filter_policy.h"
@ -48,10 +51,7 @@ class InternalKey;
// Value types encoded as the last component of internal keys.
// DO NOT CHANGE THESE ENUM VALUES: they are embedded in the on-disk
// data structures.
enum ValueType {
kTypeDeletion = 0x0,
kTypeValue = 0x1
};
enum ValueType { kTypeDeletion = 0x0, kTypeValue = 0x1 };
// kValueTypeForSeek defines the ValueType that should be passed when
// constructing a ParsedInternalKey object for seeking to a particular
// sequence number (since we sort sequence numbers in decreasing order
@ -64,17 +64,16 @@ typedef uint64_t SequenceNumber;
// We leave eight bits empty at the bottom so a type and sequence#
// can be packed together into 64-bits.
static const SequenceNumber kMaxSequenceNumber =
((0x1ull << 56) - 1);
static const SequenceNumber kMaxSequenceNumber = ((0x1ull << 56) - 1);
struct ParsedInternalKey {
Slice user_key;
SequenceNumber sequence;
ValueType type;
ParsedInternalKey() { } // Intentionally left uninitialized (for speed)
ParsedInternalKey() {} // Intentionally left uninitialized (for speed)
ParsedInternalKey(const Slice& u, const SequenceNumber& seq, ValueType t)
: user_key(u), sequence(seq), type(t) { }
: user_key(u), sequence(seq), type(t) {}
std::string DebugString() const;
};
@ -103,14 +102,14 @@ inline Slice ExtractUserKey(const Slice& internal_key) {
class InternalKeyComparator : public Comparator {
private:
const Comparator* user_comparator_;
public:
explicit InternalKeyComparator(const Comparator* c) : user_comparator_(c) { }
virtual const char* Name() const;
virtual int Compare(const Slice& a, const Slice& b) const;
virtual void FindShortestSeparator(
std::string* start,
const Slice& limit) const;
virtual void FindShortSuccessor(std::string* key) const;
explicit InternalKeyComparator(const Comparator* c) : user_comparator_(c) {}
const char* Name() const override;
int Compare(const Slice& a, const Slice& b) const override;
void FindShortestSeparator(std::string* start,
const Slice& limit) const override;
void FindShortSuccessor(std::string* key) const override;
const Comparator* user_comparator() const { return user_comparator_; }
@ -121,11 +120,12 @@ class InternalKeyComparator : public Comparator {
class InternalFilterPolicy : public FilterPolicy {
private:
const FilterPolicy* const user_policy_;
public:
explicit InternalFilterPolicy(const FilterPolicy* p) : user_policy_(p) { }
virtual const char* Name() const;
virtual void CreateFilter(const Slice* keys, int n, std::string* dst) const;
virtual bool KeyMayMatch(const Slice& key, const Slice& filter) const;
explicit InternalFilterPolicy(const FilterPolicy* p) : user_policy_(p) {}
const char* Name() const override;
void CreateFilter(const Slice* keys, int n, std::string* dst) const override;
bool KeyMayMatch(const Slice& key, const Slice& filter) const override;
};
// Modules in this directory should keep internal keys wrapped inside
@ -134,13 +134,18 @@ class InternalFilterPolicy : public FilterPolicy {
class InternalKey {
private:
std::string rep_;
public:
InternalKey() { } // Leave rep_ as empty to indicate it is invalid
InternalKey() {} // Leave rep_ as empty to indicate it is invalid
InternalKey(const Slice& user_key, SequenceNumber s, ValueType t) {
AppendInternalKey(&rep_, ParsedInternalKey(user_key, s, t));
}
void DecodeFrom(const Slice& s) { rep_.assign(s.data(), s.size()); }
bool DecodeFrom(const Slice& s) {
rep_.assign(s.data(), s.size());
return !rep_.empty();
}
Slice Encode() const {
assert(!rep_.empty());
return rep_;
@ -158,8 +163,8 @@ class InternalKey {
std::string DebugString() const;
};
inline int InternalKeyComparator::Compare(
const InternalKey& a, const InternalKey& b) const {
inline int InternalKeyComparator::Compare(const InternalKey& a,
const InternalKey& b) const {
return Compare(a.Encode(), b.Encode());
}
@ -168,11 +173,11 @@ inline bool ParseInternalKey(const Slice& internal_key,
const size_t n = internal_key.size();
if (n < 8) return false;
uint64_t num = DecodeFixed64(internal_key.data() + n - 8);
unsigned char c = num & 0xff;
uint8_t c = num & 0xff;
result->sequence = num >> 8;
result->type = static_cast<ValueType>(c);
result->user_key = Slice(internal_key.data(), n - 8);
return (c <= static_cast<unsigned char>(kTypeValue));
return (c <= static_cast<uint8_t>(kTypeValue));
}
// A helper class useful for DBImpl::Get()
@ -182,6 +187,9 @@ class LookupKey {
// the specified sequence number.
LookupKey(const Slice& user_key, SequenceNumber sequence);
LookupKey(const LookupKey&) = delete;
LookupKey& operator=(const LookupKey&) = delete;
~LookupKey();
// Return a key suitable for lookup in a MemTable.
@ -204,11 +212,7 @@ class LookupKey {
const char* start_;
const char* kstart_;
const char* end_;
char space_[200]; // Avoid allocation for short keys
// No copying allowed
LookupKey(const LookupKey&);
void operator=(const LookupKey&);
char space_[200]; // Avoid allocation for short keys
};
inline LookupKey::~LookupKey() {

96
db/dbformat_test.cc
View File

@ -3,13 +3,13 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/dbformat.h"
#include "gtest/gtest.h"
#include "util/logging.h"
#include "util/testharness.h"
namespace leveldb {
static std::string IKey(const std::string& user_key,
uint64_t seq,
static std::string IKey(const std::string& user_key, uint64_t seq,
ValueType vt) {
std::string encoded;
AppendInternalKey(&encoded, ParsedInternalKey(user_key, seq, vt));
@ -28,9 +28,7 @@ static std::string ShortSuccessor(const std::string& s) {
return result;
}
static void TestKey(const std::string& key,
uint64_t seq,
ValueType vt) {
static void TestKey(const std::string& key, uint64_t seq, ValueType vt) {
std::string encoded = IKey(key, seq, vt);
Slice in(encoded);
@ -44,16 +42,20 @@ static void TestKey(const std::string& key,
ASSERT_TRUE(!ParseInternalKey(Slice("bar"), &decoded));
}
class FormatTest { };
TEST(FormatTest, InternalKey_EncodeDecode) {
const char* keys[] = { "", "k", "hello", "longggggggggggggggggggggg" };
const uint64_t seq[] = {
1, 2, 3,
(1ull << 8) - 1, 1ull << 8, (1ull << 8) + 1,
(1ull << 16) - 1, 1ull << 16, (1ull << 16) + 1,
(1ull << 32) - 1, 1ull << 32, (1ull << 32) + 1
};
const char* keys[] = {"", "k", "hello", "longggggggggggggggggggggg"};
const uint64_t seq[] = {1,
2,
3,
(1ull << 8) - 1,
1ull << 8,
(1ull << 8) + 1,
(1ull << 16) - 1,
1ull << 16,
(1ull << 16) + 1,
(1ull << 32) - 1,
1ull << 32,
(1ull << 32) + 1};
for (int k = 0; k < sizeof(keys) / sizeof(keys[0]); k++) {
for (int s = 0; s < sizeof(seq) / sizeof(seq[0]); s++) {
TestKey(keys[k], seq[s], kTypeValue);
@ -62,40 +64,44 @@ TEST(FormatTest, InternalKey_EncodeDecode) {
}
}
TEST(FormatTest, InternalKey_DecodeFromEmpty) {
InternalKey internal_key;
ASSERT_TRUE(!internal_key.DecodeFrom(""));
}
TEST(FormatTest, InternalKeyShortSeparator) {
// When user keys are same
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("foo", 99, kTypeValue)));
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("foo", 101, kTypeValue)));
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("foo", 100, kTypeValue)));
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("foo", 100, kTypeDeletion)));
Shorten(IKey("foo", 100, kTypeValue), IKey("foo", 99, kTypeValue)));
ASSERT_EQ(
IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue), IKey("foo", 101, kTypeValue)));
ASSERT_EQ(
IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue), IKey("foo", 100, kTypeValue)));
ASSERT_EQ(
IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue), IKey("foo", 100, kTypeDeletion)));
// When user keys are misordered
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("bar", 99, kTypeValue)));
Shorten(IKey("foo", 100, kTypeValue), IKey("bar", 99, kTypeValue)));
// When user keys are different, but correctly ordered
ASSERT_EQ(IKey("g", kMaxSequenceNumber, kValueTypeForSeek),
Shorten(IKey("foo", 100, kTypeValue),
IKey("hello", 200, kTypeValue)));
ASSERT_EQ(
IKey("g", kMaxSequenceNumber, kValueTypeForSeek),
Shorten(IKey("foo", 100, kTypeValue), IKey("hello", 200, kTypeValue)));
// When start user key is prefix of limit user key
ASSERT_EQ(IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue),
IKey("foobar", 200, kTypeValue)));
ASSERT_EQ(
IKey("foo", 100, kTypeValue),
Shorten(IKey("foo", 100, kTypeValue), IKey("foobar", 200, kTypeValue)));
// When limit user key is prefix of start user key
ASSERT_EQ(IKey("foobar", 100, kTypeValue),
Shorten(IKey("foobar", 100, kTypeValue),
IKey("foo", 200, kTypeValue)));
ASSERT_EQ(
IKey("foobar", 100, kTypeValue),
Shorten(IKey("foobar", 100, kTypeValue), IKey("foo", 200, kTypeValue)));
}
TEST(FormatTest, InternalKeyShortestSuccessor) {
@ -105,8 +111,18 @@ TEST(FormatTest, InternalKeyShortestSuccessor) {
ShortSuccessor(IKey("\xff\xff", 100, kTypeValue)));
}
} // namespace leveldb
TEST(FormatTest, ParsedInternalKeyDebugString) {
ParsedInternalKey key("The \"key\" in 'single quotes'", 42, kTypeValue);
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
ASSERT_EQ("'The \"key\" in 'single quotes'' @ 42 : 1", key.DebugString());
}
TEST(FormatTest, InternalKeyDebugString) {
InternalKey key("The \"key\" in 'single quotes'", 42, kTypeValue);
ASSERT_EQ("'The \"key\" in 'single quotes'' @ 42 : 1", key.DebugString());
InternalKey invalid_key;
ASSERT_EQ("(bad)", invalid_key.DebugString());
}
} // namespace leveldb

24
db/dumpfile.cc
View File

@ -4,7 +4,7 @@
#include "leveldb/dumpfile.h"
#include <stdio.h>
#include <cstdio>
#include "db/dbformat.h"
#include "db/filename.h"
@ -38,8 +38,7 @@ bool GuessType(const std::string& fname, FileType* type) {
// Notified when log reader encounters corruption.
class CorruptionReporter : public log::Reader::Reporter {
public:
WritableFile* dst_;
virtual void Corruption(size_t bytes, const Status& status) {
void Corruption(size_t bytes, const Status& status) override {
std::string r = "corruption: ";
AppendNumberTo(&r, bytes);
r += " bytes; ";
@ -47,6 +46,8 @@ class CorruptionReporter : public log::Reader::Reporter {
r.push_back('\n');
dst_->Append(r);
}
WritableFile* dst_;
};
// Print contents of a log file. (*func)() is called on every record.
@ -73,8 +74,7 @@ Status PrintLogContents(Env* env, const std::string& fname,
// Called on every item found in a WriteBatch.
class WriteBatchItemPrinter : public WriteBatch::Handler {
public:
WritableFile* dst_;
virtual void Put(const Slice& key, const Slice& value) {
void Put(const Slice& key, const Slice& value) override {
std::string r = " put '";
AppendEscapedStringTo(&r, key);
r += "' '";
@ -82,14 +82,15 @@ class WriteBatchItemPrinter : public WriteBatch::Handler {
r += "'\n";
dst_->Append(r);
}
virtual void Delete(const Slice& key) {
void Delete(const Slice& key) override {
std::string r = " del '";
AppendEscapedStringTo(&r, key);
r += "'\n";
dst_->Append(r);
}
};
WritableFile* dst_;
};
// Called on every log record (each one of which is a WriteBatch)
// found in a kLogFile.
@ -216,9 +217,12 @@ Status DumpFile(Env* env, const std::string& fname, WritableFile* dst) {
return Status::InvalidArgument(fname + ": unknown file type");
}
switch (ftype) {
case kLogFile: return DumpLog(env, fname, dst);
case kDescriptorFile: return DumpDescriptor(env, fname, dst);
case kTableFile: return DumpTable(env, fname, dst);
case kLogFile:
return DumpLog(env, fname, dst);
case kDescriptorFile:
return DumpDescriptor(env, fname, dst);
case kTableFile:
return DumpTable(env, fname, dst);
default:
break;
}

142
db/fault_injection_test.cc
View File

@ -9,12 +9,13 @@
#include <map>
#include <set>
#include "leveldb/db.h"
#include "gtest/gtest.h"
#include "db/db_impl.h"
#include "db/filename.h"
#include "db/log_format.h"
#include "db/version_set.h"
#include "leveldb/cache.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "leveldb/table.h"
#include "leveldb/write_batch.h"
@ -22,7 +23,6 @@
#include "port/thread_annotations.h"
#include "util/logging.h"
#include "util/mutexlock.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
@ -56,8 +56,7 @@ Status Truncate(const std::string& filename, uint64_t length) {
SequentialFile* orig_file;
Status s = env->NewSequentialFile(filename, &orig_file);
if (!s.ok())
return s;
if (!s.ok()) return s;
char* scratch = new char[length];
leveldb::Slice result;
@ -73,7 +72,7 @@ Status Truncate(const std::string& filename, uint64_t length) {
if (s.ok()) {
s = env->RenameFile(tmp_name, filename);
} else {
env->DeleteFile(tmp_name);
env->RemoveFile(tmp_name);
}
}
}
@ -85,15 +84,15 @@ Status Truncate(const std::string& filename, uint64_t length) {
struct FileState {
std::string filename_;
ssize_t pos_;
ssize_t pos_at_last_sync_;
ssize_t pos_at_last_flush_;
int64_t pos_;
int64_t pos_at_last_sync_;
int64_t pos_at_last_flush_;
FileState(const std::string& filename)
: filename_(filename),
pos_(-1),
pos_at_last_sync_(-1),
pos_at_last_flush_(-1) { }
pos_at_last_flush_(-1) {}
FileState() : pos_(-1), pos_at_last_sync_(-1), pos_at_last_flush_(-1) {}
@ -108,14 +107,13 @@ struct FileState {
// is written to or sync'ed.
class TestWritableFile : public WritableFile {
public:
TestWritableFile(const FileState& state,
WritableFile* f,
TestWritableFile(const FileState& state, WritableFile* f,
FaultInjectionTestEnv* env);
virtual ~TestWritableFile();
virtual Status Append(const Slice& data);
virtual Status Close();
virtual Status Flush();
virtual Status Sync();
~TestWritableFile() override;
Status Append(const Slice& data) override;
Status Close() override;
Status Flush() override;
Status Sync() override;
private:
FileState state_;
@ -130,17 +128,17 @@ class FaultInjectionTestEnv : public EnvWrapper {
public:
FaultInjectionTestEnv()
: EnvWrapper(Env::Default()), filesystem_active_(true) {}
virtual ~FaultInjectionTestEnv() { }
virtual Status NewWritableFile(const std::string& fname,
WritableFile** result);
virtual Status NewAppendableFile(const std::string& fname,
WritableFile** result);
virtual Status DeleteFile(const std::string& f);
virtual Status RenameFile(const std::string& s, const std::string& t);
~FaultInjectionTestEnv() override = default;
Status NewWritableFile(const std::string& fname,
WritableFile** result) override;
Status NewAppendableFile(const std::string& fname,
WritableFile** result) override;
Status RemoveFile(const std::string& f) override;
Status RenameFile(const std::string& s, const std::string& t) override;
void WritableFileClosed(const FileState& state);
Status DropUnsyncedFileData();
Status DeleteFilesCreatedAfterLastDirSync();
Status RemoveFilesCreatedAfterLastDirSync();
void DirWasSynced();
bool IsFileCreatedSinceLastDirSync(const std::string& filename);
void ResetState();
@ -165,13 +163,9 @@ class FaultInjectionTestEnv : public EnvWrapper {
bool filesystem_active_ GUARDED_BY(mutex_); // Record flushes, syncs, writes
};
TestWritableFile::TestWritableFile(const FileState& state,
WritableFile* f,
TestWritableFile::TestWritableFile(const FileState& state, WritableFile* f,
FaultInjectionTestEnv* env)
: state_(state),
target_(f),
writable_file_opened_(true),
env_(env) {
: state_(state), target_(f), writable_file_opened_(true), env_(env) {
assert(f != nullptr);
}
@ -274,10 +268,11 @@ Status FaultInjectionTestEnv::NewAppendableFile(const std::string& fname,
Status FaultInjectionTestEnv::DropUnsyncedFileData() {
Status s;
MutexLock l(&mutex_);
for (std::map<std::string, FileState>::const_iterator it =
db_file_state_.begin();
s.ok() && it != db_file_state_.end(); ++it) {
const FileState& state = it->second;
for (const auto& kvp : db_file_state_) {
if (!s.ok()) {
break;
}
const FileState& state = kvp.second;
if (!state.IsFullySynced()) {
s = state.DropUnsyncedData();
}
@ -303,9 +298,9 @@ void FaultInjectionTestEnv::UntrackFile(const std::string& f) {
new_files_since_last_dir_sync_.erase(f);
}
Status FaultInjectionTestEnv::DeleteFile(const std::string& f) {
Status s = EnvWrapper::DeleteFile(f);
ASSERT_OK(s);
Status FaultInjectionTestEnv::RemoveFile(const std::string& f) {
Status s = EnvWrapper::RemoveFile(f);
EXPECT_LEVELDB_OK(s);
if (s.ok()) {
UntrackFile(f);
}
@ -340,18 +335,20 @@ void FaultInjectionTestEnv::ResetState() {
SetFilesystemActive(true);
}
Status FaultInjectionTestEnv::DeleteFilesCreatedAfterLastDirSync() {
// Because DeleteFile access this container make a copy to avoid deadlock
Status FaultInjectionTestEnv::RemoveFilesCreatedAfterLastDirSync() {
// Because RemoveFile access this container make a copy to avoid deadlock
mutex_.Lock();
std::set<std::string> new_files(new_files_since_last_dir_sync_.begin(),
new_files_since_last_dir_sync_.end());
mutex_.Unlock();
Status s;
std::set<std::string>::const_iterator it;
for (it = new_files.begin(); s.ok() && it != new_files.end(); ++it) {
s = DeleteFile(*it);
Status status;
for (const auto& new_file : new_files) {
Status remove_status = RemoveFile(new_file);
if (!remove_status.ok() && status.ok()) {
status = std::move(remove_status);
}
}
return s;
return status;
}
void FaultInjectionTestEnv::WritableFileClosed(const FileState& state) {
@ -360,11 +357,11 @@ void FaultInjectionTestEnv::WritableFileClosed(const FileState& state) {
}
Status FileState::DropUnsyncedData() const {
ssize_t sync_pos = pos_at_last_sync_ == -1 ? 0 : pos_at_last_sync_;
int64_t sync_pos = pos_at_last_sync_ == -1 ? 0 : pos_at_last_sync_;
return Truncate(filename_, sync_pos);
}
class FaultInjectionTest {
class FaultInjectionTest : public testing::Test {
public:
enum ExpectedVerifResult { VAL_EXPECT_NO_ERROR, VAL_EXPECT_ERROR };
enum ResetMethod { RESET_DROP_UNSYNCED_DATA, RESET_DELETE_UNSYNCED_FILES };
@ -379,7 +376,7 @@ class FaultInjectionTest {
: env_(new FaultInjectionTestEnv),
tiny_cache_(NewLRUCache(100)),
db_(nullptr) {
dbname_ = test::TmpDir() + "/fault_test";
dbname_ = testing::TempDir() + "fault_test";
DestroyDB(dbname_, Options()); // Destroy any db from earlier run
options_.reuse_logs = true;
options_.env = env_;
@ -395,9 +392,7 @@ class FaultInjectionTest {
delete env_;
}
void ReuseLogs(bool reuse) {
options_.reuse_logs = reuse;
}
void ReuseLogs(bool reuse) { options_.reuse_logs = reuse; }
void Build(int start_idx, int num_vals) {
std::string key_space, value_space;
@ -407,7 +402,7 @@ class FaultInjectionTest {
batch.Clear();
batch.Put(key, Value(i, &value_space));
WriteOptions options;
ASSERT_OK(db_->Write(options, &batch));
ASSERT_LEVELDB_OK(db_->Write(options, &batch));
}
}
@ -429,10 +424,10 @@ class FaultInjectionTest {
s = ReadValue(i, &val);
if (expected == VAL_EXPECT_NO_ERROR) {
if (s.ok()) {
ASSERT_EQ(value_space, val);
EXPECT_EQ(value_space, val);
}
} else if (s.ok()) {
fprintf(stderr, "Expected an error at %d, but was OK\n", i);
std::fprintf(stderr, "Expected an error at %d, but was OK\n", i);
s = Status::IOError(dbname_, "Expected value error:");
} else {
s = Status::OK(); // An expected error
@ -444,7 +439,7 @@ class FaultInjectionTest {
// Return the ith key
Slice Key(int i, std::string* storage) const {
char buf[100];
snprintf(buf, sizeof(buf), "%016d", i);
std::snprintf(buf, sizeof(buf), "%016d", i);
storage->assign(buf, strlen(buf));
return Slice(*storage);
}
@ -469,9 +464,8 @@ class FaultInjectionTest {
void DeleteAllData() {
Iterator* iter = db_->NewIterator(ReadOptions());
WriteOptions options;
for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
ASSERT_OK(db_->Delete(WriteOptions(), iter->key()));
ASSERT_LEVELDB_OK(db_->Delete(WriteOptions(), iter->key()));
}
delete iter;
@ -480,10 +474,10 @@ class FaultInjectionTest {
void ResetDBState(ResetMethod reset_method) {
switch (reset_method) {
case RESET_DROP_UNSYNCED_DATA:
ASSERT_OK(env_->DropUnsyncedFileData());
ASSERT_LEVELDB_OK(env_->DropUnsyncedFileData());
break;
case RESET_DELETE_UNSYNCED_FILES:
ASSERT_OK(env_->DeleteFilesCreatedAfterLastDirSync());
ASSERT_LEVELDB_OK(env_->RemoveFilesCreatedAfterLastDirSync());
break;
default:
assert(false);
@ -498,35 +492,34 @@ class FaultInjectionTest {
}
void PartialCompactTestReopenWithFault(ResetMethod reset_method,
int num_pre_sync,
int num_post_sync) {
int num_pre_sync, int num_post_sync) {
env_->SetFilesystemActive(false);
CloseDB();
ResetDBState(reset_method);
ASSERT_OK(OpenDB());
ASSERT_OK(Verify(0, num_pre_sync, FaultInjectionTest::VAL_EXPECT_NO_ERROR));
ASSERT_OK(Verify(num_pre_sync, num_post_sync, FaultInjectionTest::VAL_EXPECT_ERROR));
ASSERT_LEVELDB_OK(OpenDB());
ASSERT_LEVELDB_OK(
Verify(0, num_pre_sync, FaultInjectionTest::VAL_EXPECT_NO_ERROR));
ASSERT_LEVELDB_OK(Verify(num_pre_sync, num_post_sync,
FaultInjectionTest::VAL_EXPECT_ERROR));
}
void NoWriteTestPreFault() {
}
void NoWriteTestPreFault() {}
void NoWriteTestReopenWithFault(ResetMethod reset_method) {
CloseDB();
ResetDBState(reset_method);
ASSERT_OK(OpenDB());
ASSERT_LEVELDB_OK(OpenDB());
}
void DoTest() {
Random rnd(0);
ASSERT_OK(OpenDB());
ASSERT_LEVELDB_OK(OpenDB());
for (size_t idx = 0; idx < kNumIterations; idx++) {
int num_pre_sync = rnd.Uniform(kMaxNumValues);
int num_post_sync = rnd.Uniform(kMaxNumValues);
PartialCompactTestPreFault(num_pre_sync, num_post_sync);
PartialCompactTestReopenWithFault(RESET_DROP_UNSYNCED_DATA,
num_pre_sync,
PartialCompactTestReopenWithFault(RESET_DROP_UNSYNCED_DATA, num_pre_sync,
num_post_sync);
NoWriteTestPreFault();
@ -536,8 +529,7 @@ class FaultInjectionTest {
// No new files created so we expect all values since no files will be
// dropped.
PartialCompactTestReopenWithFault(RESET_DELETE_UNSYNCED_FILES,
num_pre_sync + num_post_sync,
0);
num_pre_sync + num_post_sync, 0);
NoWriteTestPreFault();
NoWriteTestReopenWithFault(RESET_DELETE_UNSYNCED_FILES);
@ -545,18 +537,14 @@ class FaultInjectionTest {
}
};
TEST(FaultInjectionTest, FaultTestNoLogReuse) {
TEST_F(FaultInjectionTest, FaultTestNoLogReuse) {
ReuseLogs(false);
DoTest();
}
TEST(FaultInjectionTest, FaultTestWithLogReuse) {
TEST_F(FaultInjectionTest, FaultTestWithLogReuse) {
ReuseLogs(true);
DoTest();
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

25
db/filename.cc
View File

@ -2,9 +2,11 @@
// 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 <ctype.h>
#include <stdio.h>
#include "db/filename.h"
#include <cassert>
#include <cstdio>
#include "db/dbformat.h"
#include "leveldb/env.h"
#include "util/logging.h"
@ -18,9 +20,8 @@ Status WriteStringToFileSync(Env* env, const Slice& data,
static std::string MakeFileName(const std::string& dbname, uint64_t number,
const char* suffix) {
char buf[100];
snprintf(buf, sizeof(buf), "/%06llu.%s",
static_cast<unsigned long long>(number),
suffix);
std::snprintf(buf, sizeof(buf), "/%06llu.%s",
static_cast<unsigned long long>(number), suffix);
return dbname + buf;
}
@ -42,8 +43,8 @@ std::string SSTTableFileName(const std::string& dbname, uint64_t number) {
std::string DescriptorFileName(const std::string& dbname, uint64_t number) {
assert(number > 0);
char buf[100];
snprintf(buf, sizeof(buf), "/MANIFEST-%06llu",
static_cast<unsigned long long>(number));
std::snprintf(buf, sizeof(buf), "/MANIFEST-%06llu",
static_cast<unsigned long long>(number));
return dbname + buf;
}
@ -51,9 +52,7 @@ std::string CurrentFileName(const std::string& dbname) {
return dbname + "/CURRENT";
}
std::string LockFileName(const std::string& dbname) {
return dbname + "/LOCK";
}
std::string LockFileName(const std::string& dbname) { return dbname + "/LOCK"; }
std::string TempFileName(const std::string& dbname, uint64_t number) {
assert(number > 0);
@ -69,7 +68,6 @@ std::string OldInfoLogFileName(const std::string& dbname) {
return dbname + "/LOG.old";
}
// Owned filenames have the form:
// dbname/CURRENT
// dbname/LOCK
@ -77,8 +75,7 @@ std::string OldInfoLogFileName(const std::string& dbname) {
// dbname/LOG.old
// dbname/MANIFEST-[0-9]+
// dbname/[0-9]+.(log|sst|ldb)
bool ParseFileName(const std::string& filename,
uint64_t* number,
bool ParseFileName(const std::string& filename, uint64_t* number,
FileType* type) {
Slice rest(filename);
if (rest == "CURRENT") {
@ -136,7 +133,7 @@ Status SetCurrentFile(Env* env, const std::string& dbname,
s = env->RenameFile(tmp, CurrentFileName(dbname));
}
if (!s.ok()) {
env->DeleteFile(tmp);
env->RemoveFile(tmp);
}
return s;
}

6
db/filename.h
View File

@ -7,8 +7,9 @@
#ifndef STORAGE_LEVELDB_DB_FILENAME_H_
#define STORAGE_LEVELDB_DB_FILENAME_H_
#include <stdint.h>
#include <cstdint>
#include <string>
#include "leveldb/slice.h"
#include "leveldb/status.h"
#include "port/port.h"
@ -69,8 +70,7 @@ std::string OldInfoLogFileName(const std::string& dbname);
// If filename is a leveldb file, store the type of the file in *type.
// The number encoded in the filename is stored in *number. If the
// filename was successfully parsed, returns true. Else return false.
bool ParseFileName(const std::string& filename,
uint64_t* number,
bool ParseFileName(const std::string& filename, uint64_t* number,
FileType* type);
// Make the CURRENT file point to the descriptor file with the

76
db/filename_test.cc
View File

@ -4,15 +4,13 @@
#include "db/filename.h"
#include "gtest/gtest.h"
#include "db/dbformat.h"
#include "port/port.h"
#include "util/logging.h"
#include "util/testharness.h"
namespace leveldb {
class FileNameTest { };
TEST(FileNameTest, Parse) {
Slice db;
FileType type;
@ -24,17 +22,17 @@ TEST(FileNameTest, Parse) {
uint64_t number;
FileType type;
} cases[] = {
{ "100.log", 100, kLogFile },
{ "0.log", 0, kLogFile },
{ "0.sst", 0, kTableFile },
{ "0.ldb", 0, kTableFile },
{ "CURRENT", 0, kCurrentFile },
{ "LOCK", 0, kDBLockFile },
{ "MANIFEST-2", 2, kDescriptorFile },
{ "MANIFEST-7", 7, kDescriptorFile },
{ "LOG", 0, kInfoLogFile },
{ "LOG.old", 0, kInfoLogFile },
{ "18446744073709551615.log", 18446744073709551615ull, kLogFile },
{"100.log", 100, kLogFile},
{"0.log", 0, kLogFile},
{"0.sst", 0, kTableFile},
{"0.ldb", 0, kTableFile},
{"CURRENT", 0, kCurrentFile},
{"LOCK", 0, kDBLockFile},
{"MANIFEST-2", 2, kDescriptorFile},
{"MANIFEST-7", 7, kDescriptorFile},
{"LOG", 0, kInfoLogFile},
{"LOG.old", 0, kInfoLogFile},
{"18446744073709551615.log", 18446744073709551615ull, kLogFile},
};
for (int i = 0; i < sizeof(cases) / sizeof(cases[0]); i++) {
std::string f = cases[i].fname;
@ -44,30 +42,28 @@ TEST(FileNameTest, Parse) {
}
// Errors
static const char* errors[] = {
"",
"foo",
"foo-dx-100.log",
".log",
"",
"manifest",
"CURREN",
"CURRENTX",
"MANIFES",
"MANIFEST",
"MANIFEST-",
"XMANIFEST-3",
"MANIFEST-3x",
"LOC",
"LOCKx",
"LO",
"LOGx",
"18446744073709551616.log",
"184467440737095516150.log",
"100",
"100.",
"100.lop"
};
static const char* errors[] = {"",
"foo",
"foo-dx-100.log",
".log",
"",
"manifest",
"CURREN",
"CURRENTX",
"MANIFES",
"MANIFEST",
"MANIFEST-",
"XMANIFEST-3",
"MANIFEST-3x",
"LOC",
"LOCKx",
"LO",
"LOGx",
"18446744073709551616.log",
"184467440737095516150.log",
"100",
"100.",
"100.lop"};
for (int i = 0; i < sizeof(errors) / sizeof(errors[0]); i++) {
std::string f = errors[i];
ASSERT_TRUE(!ParseFileName(f, &number, &type)) << f;
@ -129,7 +125,3 @@ TEST(FileNameTest, Construction) {
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

20
db/leveldbutil.cc
View File

@ -2,7 +2,8 @@
// 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 <stdio.h>
#include <cstdio>
#include "leveldb/dumpfile.h"
#include "leveldb/env.h"
#include "leveldb/status.h"
@ -12,13 +13,13 @@ namespace {
class StdoutPrinter : public WritableFile {
public:
virtual Status Append(const Slice& data) {
Status Append(const Slice& data) override {
fwrite(data.data(), 1, data.size(), stdout);
return Status::OK();
}
virtual Status Close() { return Status::OK(); }
virtual Status Flush() { return Status::OK(); }
virtual Status Sync() { return Status::OK(); }
Status Close() override { return Status::OK(); }
Status Flush() override { return Status::OK(); }
Status Sync() override { return Status::OK(); }
};
bool HandleDumpCommand(Env* env, char** files, int num) {
@ -27,7 +28,7 @@ bool HandleDumpCommand(Env* env, char** files, int num) {
for (int i = 0; i < num; i++) {
Status s = DumpFile(env, files[i], &printer);
if (!s.ok()) {
fprintf(stderr, "%s\n", s.ToString().c_str());
std::fprintf(stderr, "%s\n", s.ToString().c_str());
ok = false;
}
}
@ -38,11 +39,10 @@ bool HandleDumpCommand(Env* env, char** files, int num) {
} // namespace leveldb
static void Usage() {
fprintf(
std::fprintf(
stderr,
"Usage: leveldbutil command...\n"
" dump files... -- dump contents of specified files\n"
);
" dump files... -- dump contents of specified files\n");
}
int main(int argc, char** argv) {
@ -54,7 +54,7 @@ int main(int argc, char** argv) {
} else {
std::string command = argv[1];
if (command == "dump") {
ok = leveldb::HandleDumpCommand(env, argv+2, argc-2);
ok = leveldb::HandleDumpCommand(env, argv + 2, argc - 2);
} else {
Usage();
ok = false;

19
db/log_reader.cc
View File

@ -4,7 +4,8 @@
#include "db/log_reader.h"
#include <stdio.h>
#include <cstdio>
#include "leveldb/env.h"
#include "util/coding.h"
#include "util/crc32c.h"
@ -12,8 +13,7 @@
namespace leveldb {
namespace log {
Reader::Reporter::~Reporter() {
}
Reader::Reporter::~Reporter() = default;
Reader::Reader(SequentialFile* file, Reporter* reporter, bool checksum,
uint64_t initial_offset)
@ -26,12 +26,9 @@ Reader::Reader(SequentialFile* file, Reporter* reporter, bool checksum,
last_record_offset_(0),
end_of_buffer_offset_(0),
initial_offset_(initial_offset),
resyncing_(initial_offset > 0) {
}
resyncing_(initial_offset > 0) {}
Reader::~Reader() {
delete[] backing_store_;
}
Reader::~Reader() { delete[] backing_store_; }
bool Reader::SkipToInitialBlock() {
const size_t offset_in_block = initial_offset_ % kBlockSize;
@ -163,7 +160,7 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch) {
default: {
char buf[40];
snprintf(buf, sizeof(buf), "unknown record type %u", record_type);
std::snprintf(buf, sizeof(buf), "unknown record type %u", record_type);
ReportCorruption(
(fragment.size() + (in_fragmented_record ? scratch->size() : 0)),
buf);
@ -176,9 +173,7 @@ bool Reader::ReadRecord(Slice* record, std::string* scratch) {
return false;
}
uint64_t Reader::LastRecordOffset() {
return last_record_offset_;
}
uint64_t Reader::LastRecordOffset() { return last_record_offset_; }
void Reader::ReportCorruption(uint64_t bytes, const char* reason) {
ReportDrop(bytes, Status::Corruption(reason));

49
db/log_reader.h
View File

@ -5,7 +5,7 @@
#ifndef STORAGE_LEVELDB_DB_LOG_READER_H_
#define STORAGE_LEVELDB_DB_LOG_READER_H_
#include <stdint.h>
#include <cstdint>
#include "db/log_format.h"
#include "leveldb/slice.h"
@ -24,7 +24,7 @@ class Reader {
public:
virtual ~Reporter();
// Some corruption was detected. "size" is the approximate number
// Some corruption was detected. "bytes" is the approximate number
// of bytes dropped due to the corruption.
virtual void Corruption(size_t bytes, const Status& status) = 0;
};
@ -43,6 +43,9 @@ class Reader {
Reader(SequentialFile* file, Reporter* reporter, bool checksum,
uint64_t initial_offset);
Reader(const Reader&) = delete;
Reader& operator=(const Reader&) = delete;
~Reader();
// Read the next record into *record. Returns true if read
@ -58,26 +61,6 @@ class Reader {
uint64_t LastRecordOffset();
private:
SequentialFile* const file_;
Reporter* const reporter_;
bool const checksum_;
char* const backing_store_;
Slice buffer_;
bool eof_; // Last Read() indicated EOF by returning < kBlockSize
// Offset of the last record returned by ReadRecord.
uint64_t last_record_offset_;
// Offset of the first location past the end of buffer_.
uint64_t end_of_buffer_offset_;
// Offset at which to start looking for the first record to return
uint64_t const initial_offset_;
// True if we are resynchronizing after a seek (initial_offset_ > 0). In
// particular, a run of kMiddleType and kLastType records can be silently
// skipped in this mode
bool resyncing_;
// Extend record types with the following special values
enum {
kEof = kMaxRecordType + 1,
@ -102,9 +85,25 @@ class Reader {
void ReportCorruption(uint64_t bytes, const char* reason);
void ReportDrop(uint64_t bytes, const Status& reason);
// No copying allowed
Reader(const Reader&);
void operator=(const Reader&);
SequentialFile* const file_;
Reporter* const reporter_;
bool const checksum_;
char* const backing_store_;
Slice buffer_;
bool eof_; // Last Read() indicated EOF by returning < kBlockSize
// Offset of the last record returned by ReadRecord.
uint64_t last_record_offset_;
// Offset of the first location past the end of buffer_.
uint64_t end_of_buffer_offset_;
// Offset at which to start looking for the first record to return
uint64_t const initial_offset_;
// True if we are resynchronizing after a seek (initial_offset_ > 0). In
// particular, a run of kMiddleType and kLastType records can be silently
// skipped in this mode
bool resyncing_;
};
} // namespace log

349
db/log_test.cc
View File

@ -2,13 +2,13 @@
// 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 "gtest/gtest.h"
#include "db/log_reader.h"
#include "db/log_writer.h"
#include "leveldb/env.h"
#include "util/coding.h"
#include "util/crc32c.h"
#include "util/random.h"
#include "util/testharness.h"
namespace leveldb {
namespace log {
@ -27,7 +27,7 @@ static std::string BigString(const std::string& partial_string, size_t n) {
// Construct a string from a number
static std::string NumberString(int n) {
char buf[50];
snprintf(buf, sizeof(buf), "%d.", n);
std::snprintf(buf, sizeof(buf), "%d.", n);
return std::string(buf);
}
@ -36,88 +36,13 @@ static std::string RandomSkewedString(int i, Random* rnd) {
return BigString(NumberString(i), rnd->Skewed(17));
}
class LogTest {
private:
class StringDest : public WritableFile {
public:
std::string contents_;
virtual Status Close() { return Status::OK(); }
virtual Status Flush() { return Status::OK(); }
virtual Status Sync() { return Status::OK(); }
virtual Status Append(const Slice& slice) {
contents_.append(slice.data(), slice.size());
return Status::OK();
}
};
class StringSource : public SequentialFile {
public:
Slice contents_;
bool force_error_;
bool returned_partial_;
StringSource() : force_error_(false), returned_partial_(false) { }
virtual Status Read(size_t n, Slice* result, char* scratch) {
ASSERT_TRUE(!returned_partial_) << "must not Read() after eof/error";
if (force_error_) {
force_error_ = false;
returned_partial_ = true;
return Status::Corruption("read error");
}
if (contents_.size() < n) {
n = contents_.size();
returned_partial_ = true;
}
*result = Slice(contents_.data(), n);
contents_.remove_prefix(n);
return Status::OK();
}
virtual Status Skip(uint64_t n) {
if (n > contents_.size()) {
contents_.clear();
return Status::NotFound("in-memory file skipped past end");
}
contents_.remove_prefix(n);
return Status::OK();
}
};
class ReportCollector : public Reader::Reporter {
public:
size_t dropped_bytes_;
std::string message_;
ReportCollector() : dropped_bytes_(0) { }
virtual void Corruption(size_t bytes, const Status& status) {
dropped_bytes_ += bytes;
message_.append(status.ToString());
}
};
StringDest dest_;
StringSource source_;
ReportCollector report_;
bool reading_;
Writer* writer_;
Reader* reader_;
// Record metadata for testing initial offset functionality
static size_t initial_offset_record_sizes_[];
static uint64_t initial_offset_last_record_offsets_[];
static int num_initial_offset_records_;
class LogTest : public testing::Test {
public:
LogTest() : reading_(false),
writer_(new Writer(&dest_)),
reader_(new Reader(&source_, &report_, true/*checksum*/,
0/*initial_offset*/)) {
}
LogTest()
: reading_(false),
writer_(new Writer(&dest_)),
reader_(new Reader(&source_, &report_, true /*checksum*/,
0 /*initial_offset*/)) {}
~LogTest() {
delete writer_;
@ -134,9 +59,7 @@ class LogTest {
writer_->AddRecord(Slice(msg));
}
size_t WrittenBytes() const {
return dest_.contents_.size();
}
size_t WrittenBytes() const { return dest_.contents_.size(); }
std::string Read() {
if (!reading_) {
@ -166,22 +89,16 @@ class LogTest {
void FixChecksum(int header_offset, int len) {
// Compute crc of type/len/data
uint32_t crc = crc32c::Value(&dest_.contents_[header_offset+6], 1 + len);
uint32_t crc = crc32c::Value(&dest_.contents_[header_offset + 6], 1 + len);
crc = crc32c::Mask(crc);
EncodeFixed32(&dest_.contents_[header_offset], crc);
}
void ForceError() {
source_.force_error_ = true;
}
void ForceError() { source_.force_error_ = true; }
size_t DroppedBytes() const {
return report_.dropped_bytes_;
}
size_t DroppedBytes() const { return report_.dropped_bytes_; }
std::string ReportMessage() const {
return report_.message_;
}
std::string ReportMessage() const { return report_.message_; }
// Returns OK iff recorded error message contains "msg"
std::string MatchError(const std::string& msg) const {
@ -202,14 +119,14 @@ class LogTest {
void StartReadingAt(uint64_t initial_offset) {
delete reader_;
reader_ = new Reader(&source_, &report_, true/*checksum*/, initial_offset);
reader_ = new Reader(&source_, &report_, true /*checksum*/, initial_offset);
}
void CheckOffsetPastEndReturnsNoRecords(uint64_t offset_past_end) {
WriteInitialOffsetLog();
reading_ = true;
source_.contents_ = Slice(dest_.contents_);
Reader* offset_reader = new Reader(&source_, &report_, true/*checksum*/,
Reader* offset_reader = new Reader(&source_, &report_, true /*checksum*/,
WrittenBytes() + offset_past_end);
Slice record;
std::string scratch;
@ -222,8 +139,8 @@ class LogTest {
WriteInitialOffsetLog();
reading_ = true;
source_.contents_ = Slice(dest_.contents_);
Reader* offset_reader = new Reader(&source_, &report_, true/*checksum*/,
initial_offset);
Reader* offset_reader =
new Reader(&source_, &report_, true /*checksum*/, initial_offset);
// Read all records from expected_record_offset through the last one.
ASSERT_LT(expected_record_offset, num_initial_offset_records_);
@ -240,38 +157,110 @@ class LogTest {
}
delete offset_reader;
}
private:
class StringDest : public WritableFile {
public:
Status Close() override { return Status::OK(); }
Status Flush() override { return Status::OK(); }
Status Sync() override { return Status::OK(); }
Status Append(const Slice& slice) override {
contents_.append(slice.data(), slice.size());
return Status::OK();
}
std::string contents_;
};
class StringSource : public SequentialFile {
public:
StringSource() : force_error_(false), returned_partial_(false) {}
Status Read(size_t n, Slice* result, char* scratch) override {
EXPECT_TRUE(!returned_partial_) << "must not Read() after eof/error";
if (force_error_) {
force_error_ = false;
returned_partial_ = true;
return Status::Corruption("read error");
}
if (contents_.size() < n) {
n = contents_.size();
returned_partial_ = true;
}
*result = Slice(contents_.data(), n);
contents_.remove_prefix(n);
return Status::OK();
}
Status Skip(uint64_t n) override {
if (n > contents_.size()) {
contents_.clear();
return Status::NotFound("in-memory file skipped past end");
}
contents_.remove_prefix(n);
return Status::OK();
}
Slice contents_;
bool force_error_;
bool returned_partial_;
};
class ReportCollector : public Reader::Reporter {
public:
ReportCollector() : dropped_bytes_(0) {}
void Corruption(size_t bytes, const Status& status) override {
dropped_bytes_ += bytes;
message_.append(status.ToString());
}
size_t dropped_bytes_;
std::string message_;
};
// Record metadata for testing initial offset functionality
static size_t initial_offset_record_sizes_[];
static uint64_t initial_offset_last_record_offsets_[];
static int num_initial_offset_records_;
StringDest dest_;
StringSource source_;
ReportCollector report_;
bool reading_;
Writer* writer_;
Reader* reader_;
};
size_t LogTest::initial_offset_record_sizes_[] =
{10000, // Two sizable records in first block
10000,
2 * log::kBlockSize - 1000, // Span three blocks
1,
13716, // Consume all but two bytes of block 3.
log::kBlockSize - kHeaderSize, // Consume the entirety of block 4.
};
size_t LogTest::initial_offset_record_sizes_[] = {
10000, // Two sizable records in first block
10000,
2 * log::kBlockSize - 1000, // Span three blocks
1,
13716, // Consume all but two bytes of block 3.
log::kBlockSize - kHeaderSize, // Consume the entirety of block 4.
};
uint64_t LogTest::initial_offset_last_record_offsets_[] =
{0,
kHeaderSize + 10000,
2 * (kHeaderSize + 10000),
2 * (kHeaderSize + 10000) +
(2 * log::kBlockSize - 1000) + 3 * kHeaderSize,
2 * (kHeaderSize + 10000) +
(2 * log::kBlockSize - 1000) + 3 * kHeaderSize
+ kHeaderSize + 1,
3 * log::kBlockSize,
};
uint64_t LogTest::initial_offset_last_record_offsets_[] = {
0,
kHeaderSize + 10000,
2 * (kHeaderSize + 10000),
2 * (kHeaderSize + 10000) + (2 * log::kBlockSize - 1000) + 3 * kHeaderSize,
2 * (kHeaderSize + 10000) + (2 * log::kBlockSize - 1000) + 3 * kHeaderSize +
kHeaderSize + 1,
3 * log::kBlockSize,
};
// LogTest::initial_offset_last_record_offsets_ must be defined before this.
int LogTest::num_initial_offset_records_ =
sizeof(LogTest::initial_offset_last_record_offsets_)/sizeof(uint64_t);
sizeof(LogTest::initial_offset_last_record_offsets_) / sizeof(uint64_t);
TEST(LogTest, Empty) {
ASSERT_EQ("EOF", Read());
}
TEST_F(LogTest, Empty) { ASSERT_EQ("EOF", Read()); }
TEST(LogTest, ReadWrite) {
TEST_F(LogTest, ReadWrite) {
Write("foo");
Write("bar");
Write("");
@ -284,7 +273,7 @@ TEST(LogTest, ReadWrite) {
ASSERT_EQ("EOF", Read()); // Make sure reads at eof work
}
TEST(LogTest, ManyBlocks) {
TEST_F(LogTest, ManyBlocks) {
for (int i = 0; i < 100000; i++) {
Write(NumberString(i));
}
@ -294,7 +283,7 @@ TEST(LogTest, ManyBlocks) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, Fragmentation) {
TEST_F(LogTest, Fragmentation) {
Write("small");
Write(BigString("medium", 50000));
Write(BigString("large", 100000));
@ -304,9 +293,9 @@ TEST(LogTest, Fragmentation) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, MarginalTrailer) {
TEST_F(LogTest, MarginalTrailer) {
// Make a trailer that is exactly the same length as an empty record.
const int n = kBlockSize - 2*kHeaderSize;
const int n = kBlockSize - 2 * kHeaderSize;
Write(BigString("foo", n));
ASSERT_EQ(kBlockSize - kHeaderSize, WrittenBytes());
Write("");
@ -317,9 +306,9 @@ TEST(LogTest, MarginalTrailer) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, MarginalTrailer2) {
TEST_F(LogTest, MarginalTrailer2) {
// Make a trailer that is exactly the same length as an empty record.
const int n = kBlockSize - 2*kHeaderSize;
const int n = kBlockSize - 2 * kHeaderSize;
Write(BigString("foo", n));
ASSERT_EQ(kBlockSize - kHeaderSize, WrittenBytes());
Write("bar");
@ -330,8 +319,8 @@ TEST(LogTest, MarginalTrailer2) {
ASSERT_EQ("", ReportMessage());
}
TEST(LogTest, ShortTrailer) {
const int n = kBlockSize - 2*kHeaderSize + 4;
TEST_F(LogTest, ShortTrailer) {
const int n = kBlockSize - 2 * kHeaderSize + 4;
Write(BigString("foo", n));
ASSERT_EQ(kBlockSize - kHeaderSize + 4, WrittenBytes());
Write("");
@ -342,15 +331,15 @@ TEST(LogTest, ShortTrailer) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, AlignedEof) {
const int n = kBlockSize - 2*kHeaderSize + 4;
TEST_F(LogTest, AlignedEof) {
const int n = kBlockSize - 2 * kHeaderSize + 4;
Write(BigString("foo", n));
ASSERT_EQ(kBlockSize - kHeaderSize + 4, WrittenBytes());
ASSERT_EQ(BigString("foo", n), Read());
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, OpenForAppend) {
TEST_F(LogTest, OpenForAppend) {
Write("hello");
ReopenForAppend();
Write("world");
@ -359,7 +348,7 @@ TEST(LogTest, OpenForAppend) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, RandomRead) {
TEST_F(LogTest, RandomRead) {
const int N = 500;
Random write_rnd(301);
for (int i = 0; i < N; i++) {
@ -374,7 +363,7 @@ TEST(LogTest, RandomRead) {
// Tests of all the error paths in log_reader.cc follow:
TEST(LogTest, ReadError) {
TEST_F(LogTest, ReadError) {
Write("foo");
ForceError();
ASSERT_EQ("EOF", Read());
@ -382,7 +371,7 @@ TEST(LogTest, ReadError) {
ASSERT_EQ("OK", MatchError("read error"));
}
TEST(LogTest, BadRecordType) {
TEST_F(LogTest, BadRecordType) {
Write("foo");
// Type is stored in header[6]
IncrementByte(6, 100);
@ -392,16 +381,16 @@ TEST(LogTest, BadRecordType) {
ASSERT_EQ("OK", MatchError("unknown record type"));
}
TEST(LogTest, TruncatedTrailingRecordIsIgnored) {
TEST_F(LogTest, TruncatedTrailingRecordIsIgnored) {
Write("foo");
ShrinkSize(4); // Drop all payload as well as a header byte
ShrinkSize(4); // Drop all payload as well as a header byte
ASSERT_EQ("EOF", Read());
// Truncated last record is ignored, not treated as an error.
ASSERT_EQ(0, DroppedBytes());
ASSERT_EQ("", ReportMessage());
}
TEST(LogTest, BadLength) {
TEST_F(LogTest, BadLength) {
const int kPayloadSize = kBlockSize - kHeaderSize;
Write(BigString("bar", kPayloadSize));
Write("foo");
@ -412,7 +401,7 @@ TEST(LogTest, BadLength) {
ASSERT_EQ("OK", MatchError("bad record length"));
}
TEST(LogTest, BadLengthAtEndIsIgnored) {
TEST_F(LogTest, BadLengthAtEndIsIgnored) {
Write("foo");
ShrinkSize(1);
ASSERT_EQ("EOF", Read());
@ -420,7 +409,7 @@ TEST(LogTest, BadLengthAtEndIsIgnored) {
ASSERT_EQ("", ReportMessage());
}
TEST(LogTest, ChecksumMismatch) {
TEST_F(LogTest, ChecksumMismatch) {
Write("foo");
IncrementByte(0, 10);
ASSERT_EQ("EOF", Read());
@ -428,7 +417,7 @@ TEST(LogTest, ChecksumMismatch) {
ASSERT_EQ("OK", MatchError("checksum mismatch"));
}
TEST(LogTest, UnexpectedMiddleType) {
TEST_F(LogTest, UnexpectedMiddleType) {
Write("foo");
SetByte(6, kMiddleType);
FixChecksum(0, 3);
@ -437,7 +426,7 @@ TEST(LogTest, UnexpectedMiddleType) {
ASSERT_EQ("OK", MatchError("missing start"));
}
TEST(LogTest, UnexpectedLastType) {
TEST_F(LogTest, UnexpectedLastType) {
Write("foo");
SetByte(6, kLastType);
FixChecksum(0, 3);
@ -446,7 +435,7 @@ TEST(LogTest, UnexpectedLastType) {
ASSERT_EQ("OK", MatchError("missing start"));
}
TEST(LogTest, UnexpectedFullType) {
TEST_F(LogTest, UnexpectedFullType) {
Write("foo");
Write("bar");
SetByte(6, kFirstType);
@ -457,7 +446,7 @@ TEST(LogTest, UnexpectedFullType) {
ASSERT_EQ("OK", MatchError("partial record without end"));
}
TEST(LogTest, UnexpectedFirstType) {
TEST_F(LogTest, UnexpectedFirstType) {
Write("foo");
Write(BigString("bar", 100000));
SetByte(6, kFirstType);
@ -468,7 +457,7 @@ TEST(LogTest, UnexpectedFirstType) {
ASSERT_EQ("OK", MatchError("partial record without end"));
}
TEST(LogTest, MissingLastIsIgnored) {
TEST_F(LogTest, MissingLastIsIgnored) {
Write(BigString("bar", kBlockSize));
// Remove the LAST block, including header.
ShrinkSize(14);
@ -477,7 +466,7 @@ TEST(LogTest, MissingLastIsIgnored) {
ASSERT_EQ(0, DroppedBytes());
}
TEST(LogTest, PartialLastIsIgnored) {
TEST_F(LogTest, PartialLastIsIgnored) {
Write(BigString("bar", kBlockSize));
// Cause a bad record length in the LAST block.
ShrinkSize(1);
@ -486,13 +475,13 @@ TEST(LogTest, PartialLastIsIgnored) {
ASSERT_EQ(0, DroppedBytes());
}
TEST(LogTest, SkipIntoMultiRecord) {
TEST_F(LogTest, SkipIntoMultiRecord) {
// Consider a fragmented record:
// first(R1), middle(R1), last(R1), first(R2)
// If initial_offset points to a record after first(R1) but before first(R2)
// incomplete fragment errors are not actual errors, and must be suppressed
// until a new first or full record is encountered.
Write(BigString("foo", 3*kBlockSize));
Write(BigString("foo", 3 * kBlockSize));
Write("correct");
StartReadingAt(kBlockSize);
@ -502,7 +491,7 @@ TEST(LogTest, SkipIntoMultiRecord) {
ASSERT_EQ("EOF", Read());
}
TEST(LogTest, ErrorJoinsRecords) {
TEST_F(LogTest, ErrorJoinsRecords) {
// Consider two fragmented records:
// first(R1) last(R1) first(R2) last(R2)
// where the middle two fragments disappear. We do not want
@ -514,78 +503,56 @@ TEST(LogTest, ErrorJoinsRecords) {
Write("correct");
// Wipe the middle block
for (int offset = kBlockSize; offset < 2*kBlockSize; offset++) {
for (int offset = kBlockSize; offset < 2 * kBlockSize; offset++) {
SetByte(offset, 'x');
}
ASSERT_EQ("correct", Read());
ASSERT_EQ("EOF", Read());
const size_t dropped = DroppedBytes();
ASSERT_LE(dropped, 2*kBlockSize + 100);
ASSERT_GE(dropped, 2*kBlockSize);
ASSERT_LE(dropped, 2 * kBlockSize + 100);
ASSERT_GE(dropped, 2 * kBlockSize);
}
TEST(LogTest, ReadStart) {
CheckInitialOffsetRecord(0, 0);
}
TEST_F(LogTest, ReadStart) { CheckInitialOffsetRecord(0, 0); }
TEST(LogTest, ReadSecondOneOff) {
CheckInitialOffsetRecord(1, 1);
}
TEST_F(LogTest, ReadSecondOneOff) { CheckInitialOffsetRecord(1, 1); }
TEST(LogTest, ReadSecondTenThousand) {
CheckInitialOffsetRecord(10000, 1);
}
TEST_F(LogTest, ReadSecondTenThousand) { CheckInitialOffsetRecord(10000, 1); }
TEST(LogTest, ReadSecondStart) {
CheckInitialOffsetRecord(10007, 1);
}
TEST_F(LogTest, ReadSecondStart) { CheckInitialOffsetRecord(10007, 1); }
TEST(LogTest, ReadThirdOneOff) {
CheckInitialOffsetRecord(10008, 2);
}
TEST_F(LogTest, ReadThirdOneOff) { CheckInitialOffsetRecord(10008, 2); }
TEST(LogTest, ReadThirdStart) {
CheckInitialOffsetRecord(20014, 2);
}
TEST_F(LogTest, ReadThirdStart) { CheckInitialOffsetRecord(20014, 2); }
TEST(LogTest, ReadFourthOneOff) {
CheckInitialOffsetRecord(20015, 3);
}
TEST_F(LogTest, ReadFourthOneOff) { CheckInitialOffsetRecord(20015, 3); }
TEST(LogTest, ReadFourthFirstBlockTrailer) {
TEST_F(LogTest, ReadFourthFirstBlockTrailer) {
CheckInitialOffsetRecord(log::kBlockSize - 4, 3);
}
TEST(LogTest, ReadFourthMiddleBlock) {
TEST_F(LogTest, ReadFourthMiddleBlock) {
CheckInitialOffsetRecord(log::kBlockSize + 1, 3);
}
TEST(LogTest, ReadFourthLastBlock) {
TEST_F(LogTest, ReadFourthLastBlock) {
CheckInitialOffsetRecord(2 * log::kBlockSize + 1, 3);
}
TEST(LogTest, ReadFourthStart) {
TEST_F(LogTest, ReadFourthStart) {
CheckInitialOffsetRecord(
2 * (kHeaderSize + 1000) + (2 * log::kBlockSize - 1000) + 3 * kHeaderSize,
3);
}
TEST(LogTest, ReadInitialOffsetIntoBlockPadding) {
TEST_F(LogTest, ReadInitialOffsetIntoBlockPadding) {
CheckInitialOffsetRecord(3 * log::kBlockSize - 3, 5);
}
TEST(LogTest, ReadEnd) {
CheckOffsetPastEndReturnsNoRecords(0);
}
TEST_F(LogTest, ReadEnd) { CheckOffsetPastEndReturnsNoRecords(0); }
TEST(LogTest, ReadPastEnd) {
CheckOffsetPastEndReturnsNoRecords(5);
}
TEST_F(LogTest, ReadPastEnd) { CheckOffsetPastEndReturnsNoRecords(5); }
} // namespace log
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

31
db/log_writer.cc
View File

@ -4,7 +4,8 @@
#include "db/log_writer.h"
#include <stdint.h>
#include <cstdint>
#include "leveldb/env.h"
#include "util/coding.h"
#include "util/crc32c.h"
@ -19,9 +20,7 @@ static void InitTypeCrc(uint32_t* type_crc) {
}
}
Writer::Writer(WritableFile* dest)
: dest_(dest),
block_offset_(0) {
Writer::Writer(WritableFile* dest) : dest_(dest), block_offset_(0) {
InitTypeCrc(type_crc_);
}
@ -30,8 +29,7 @@ Writer::Writer(WritableFile* dest, uint64_t dest_length)
InitTypeCrc(type_crc_);
}
Writer::~Writer() {
}
Writer::~Writer() = default;
Status Writer::AddRecord(const Slice& slice) {
const char* ptr = slice.data();
@ -49,7 +47,7 @@ Status Writer::AddRecord(const Slice& slice) {
// Switch to a new block
if (leftover > 0) {
// Fill the trailer (literal below relies on kHeaderSize being 7)
assert(kHeaderSize == 7);
static_assert(kHeaderSize == 7, "");
dest_->Append(Slice("\x00\x00\x00\x00\x00\x00", leftover));
}
block_offset_ = 0;
@ -81,30 +79,31 @@ Status Writer::AddRecord(const Slice& slice) {
return s;
}
Status Writer::EmitPhysicalRecord(RecordType t, const char* ptr, size_t n) {
assert(n <= 0xffff); // Must fit in two bytes
assert(block_offset_ + kHeaderSize + n <= kBlockSize);
Status Writer::EmitPhysicalRecord(RecordType t, const char* ptr,
size_t length) {
assert(length <= 0xffff); // Must fit in two bytes
assert(block_offset_ + kHeaderSize + length <= kBlockSize);
// Format the header
char buf[kHeaderSize];
buf[4] = static_cast<char>(n & 0xff);
buf[5] = static_cast<char>(n >> 8);
buf[4] = static_cast<char>(length & 0xff);
buf[5] = static_cast<char>(length >> 8);
buf[6] = static_cast<char>(t);
// Compute the crc of the record type and the payload.
uint32_t crc = crc32c::Extend(type_crc_[t], ptr, n);
crc = crc32c::Mask(crc); // Adjust for storage
uint32_t crc = crc32c::Extend(type_crc_[t], ptr, length);
crc = crc32c::Mask(crc); // Adjust for storage
EncodeFixed32(buf, crc);
// Write the header and the payload
Status s = dest_->Append(Slice(buf, kHeaderSize));
if (s.ok()) {
s = dest_->Append(Slice(ptr, n));
s = dest_->Append(Slice(ptr, length));
if (s.ok()) {
s = dest_->Flush();
}
}
block_offset_ += kHeaderSize + n;
block_offset_ += kHeaderSize + length;
return s;
}

16
db/log_writer.h
View File

@ -5,7 +5,8 @@
#ifndef STORAGE_LEVELDB_DB_LOG_WRITER_H_
#define STORAGE_LEVELDB_DB_LOG_WRITER_H_
#include <stdint.h>
#include <cstdint>
#include "db/log_format.h"
#include "leveldb/slice.h"
#include "leveldb/status.h"
@ -28,24 +29,23 @@ class Writer {
// "*dest" must remain live while this Writer is in use.
Writer(WritableFile* dest, uint64_t dest_length);
Writer(const Writer&) = delete;
Writer& operator=(const Writer&) = delete;
~Writer();
Status AddRecord(const Slice& slice);
private:
Status EmitPhysicalRecord(RecordType type, const char* ptr, size_t length);
WritableFile* dest_;
int block_offset_; // Current offset in block
int block_offset_; // Current offset in block
// crc32c values for all supported record types. These are
// pre-computed to reduce the overhead of computing the crc of the
// record type stored in the header.
uint32_t type_crc_[kMaxRecordType + 1];
Status EmitPhysicalRecord(RecordType type, const char* ptr, size_t length);
// No copying allowed
Writer(const Writer&);
void operator=(const Writer&);
};
} // namespace log

73
db/memtable.cc
View File

@ -18,20 +18,15 @@ static Slice GetLengthPrefixedSlice(const char* data) {
return Slice(p, len);
}
MemTable::MemTable(const InternalKeyComparator& cmp)
: comparator_(cmp),
refs_(0),
table_(comparator_, &arena_) {
}
MemTable::MemTable(const InternalKeyComparator& comparator)
: comparator_(comparator), refs_(0), table_(comparator_, &arena_) {}
MemTable::~MemTable() {
assert(refs_ == 0);
}
MemTable::~MemTable() { assert(refs_ == 0); }
size_t MemTable::ApproximateMemoryUsage() { return arena_.MemoryUsage(); }
int MemTable::KeyComparator::operator()(const char* aptr, const char* bptr)
const {
int MemTable::KeyComparator::operator()(const char* aptr,
const char* bptr) const {
// Internal keys are encoded as length-prefixed strings.
Slice a = GetLengthPrefixedSlice(aptr);
Slice b = GetLengthPrefixedSlice(bptr);
@ -48,60 +43,59 @@ static const char* EncodeKey(std::string* scratch, const Slice& target) {
return scratch->data();
}
class MemTableIterator: public Iterator {
class MemTableIterator : public Iterator {
public:
explicit MemTableIterator(MemTable::Table* table) : iter_(table) { }
explicit MemTableIterator(MemTable::Table* table) : iter_(table) {}
virtual bool Valid() const { return iter_.Valid(); }
virtual void Seek(const Slice& k) { iter_.Seek(EncodeKey(&tmp_, k)); }
virtual void SeekToFirst() { iter_.SeekToFirst(); }
virtual void SeekToLast() { iter_.SeekToLast(); }
virtual void Next() { iter_.Next(); }
virtual void Prev() { iter_.Prev(); }
virtual Slice key() const { return GetLengthPrefixedSlice(iter_.key()); }
virtual Slice value() const {
MemTableIterator(const MemTableIterator&) = delete;
MemTableIterator& operator=(const MemTableIterator&) = delete;
~MemTableIterator() override = default;
bool Valid() const override { return iter_.Valid(); }
void Seek(const Slice& k) override { iter_.Seek(EncodeKey(&tmp_, k)); }
void SeekToFirst() override { iter_.SeekToFirst(); }
void SeekToLast() override { iter_.SeekToLast(); }
void Next() override { iter_.Next(); }
void Prev() override { iter_.Prev(); }
Slice key() const override { return GetLengthPrefixedSlice(iter_.key()); }
Slice value() const override {
Slice key_slice = GetLengthPrefixedSlice(iter_.key());
return GetLengthPrefixedSlice(key_slice.data() + key_slice.size());
}
virtual Status status() const { return Status::OK(); }
Status status() const override { return Status::OK(); }
private:
MemTable::Table::Iterator iter_;
std::string tmp_; // For passing to EncodeKey
// No copying allowed
MemTableIterator(const MemTableIterator&);
void operator=(const MemTableIterator&);
std::string tmp_; // For passing to EncodeKey
};
Iterator* MemTable::NewIterator() {
return new MemTableIterator(&table_);
}
Iterator* MemTable::NewIterator() { return new MemTableIterator(&table_); }
void MemTable::Add(SequenceNumber s, ValueType type,
const Slice& key,
void MemTable::Add(SequenceNumber s, ValueType type, const Slice& key,
const Slice& value) {
// Format of an entry is concatenation of:
// key_size : varint32 of internal_key.size()
// key bytes : char[internal_key.size()]
// tag : uint64((sequence << 8) | type)
// value_size : varint32 of value.size()
// value bytes : char[value.size()]
size_t key_size = key.size();
size_t val_size = value.size();
size_t internal_key_size = key_size + 8;
const size_t encoded_len =
VarintLength(internal_key_size) + internal_key_size +
VarintLength(val_size) + val_size;
const size_t encoded_len = VarintLength(internal_key_size) +
internal_key_size + VarintLength(val_size) +
val_size;
char* buf = arena_.Allocate(encoded_len);
char* p = EncodeVarint32(buf, internal_key_size);
memcpy(p, key.data(), key_size);
std::memcpy(p, key.data(), key_size);
p += key_size;
EncodeFixed64(p, (s << 8) | type);
p += 8;
p = EncodeVarint32(p, val_size);
memcpy(p, value.data(), val_size);
assert((p + val_size) - buf == encoded_len);
std::memcpy(p, value.data(), val_size);
assert(p + val_size == buf + encoded_len);
table_.Insert(buf);
}
@ -121,10 +115,9 @@ bool MemTable::Get(const LookupKey& key, std::string* value, Status* s) {
// all entries with overly large sequence numbers.
const char* entry = iter.key();
uint32_t key_length;
const char* key_ptr = GetVarint32Ptr(entry, entry+5, &key_length);
const char* key_ptr = GetVarint32Ptr(entry, entry + 5, &key_length);
if (comparator_.comparator.user_comparator()->Compare(
Slice(key_ptr, key_length - 8),
key.user_key()) == 0) {
Slice(key_ptr, key_length - 8), key.user_key()) == 0) {
// Correct user key
const uint64_t tag = DecodeFixed64(key_ptr + key_length - 8);
switch (static_cast<ValueType>(tag & 0xff)) {

28
db/memtable.h
View File

@ -6,9 +6,10 @@
#define STORAGE_LEVELDB_DB_MEMTABLE_H_
#include <string>
#include "leveldb/db.h"
#include "db/dbformat.h"
#include "db/skiplist.h"
#include "leveldb/db.h"
#include "util/arena.h"
namespace leveldb {
@ -22,6 +23,9 @@ class MemTable {
// is zero and the caller must call Ref() at least once.
explicit MemTable(const InternalKeyComparator& comparator);
MemTable(const MemTable&) = delete;
MemTable& operator=(const MemTable&) = delete;
// Increase reference count.
void Ref() { ++refs_; }
@ -49,8 +53,7 @@ class MemTable {
// Add an entry into memtable that maps key to value at the
// specified sequence number and with the specified type.
// Typically value will be empty if type==kTypeDeletion.
void Add(SequenceNumber seq, ValueType type,
const Slice& key,
void Add(SequenceNumber seq, ValueType type, const Slice& key,
const Slice& value);
// If memtable contains a value for key, store it in *value and return true.
@ -60,26 +63,23 @@ class MemTable {
bool Get(const LookupKey& key, std::string* value, Status* s);
private:
~MemTable(); // Private since only Unref() should be used to delete it
struct KeyComparator {
const InternalKeyComparator comparator;
explicit KeyComparator(const InternalKeyComparator& c) : comparator(c) { }
int operator()(const char* a, const char* b) const;
};
friend class MemTableIterator;
friend class MemTableBackwardIterator;
struct KeyComparator {
const InternalKeyComparator comparator;
explicit KeyComparator(const InternalKeyComparator& c) : comparator(c) {}
int operator()(const char* a, const char* b) const;
};
typedef SkipList<const char*, KeyComparator> Table;
~MemTable(); // Private since only Unref() should be used to delete it
KeyComparator comparator_;
int refs_;
Arena arena_;
Table table_;
// No copying allowed
MemTable(const MemTable&);
void operator=(const MemTable&);
};
} // namespace leveldb

123
db/recovery_test.cc
View File

@ -2,6 +2,7 @@
// 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 "gtest/gtest.h"
#include "db/db_impl.h"
#include "db/filename.h"
#include "db/version_set.h"
@ -10,15 +11,14 @@
#include "leveldb/env.h"
#include "leveldb/write_batch.h"
#include "util/logging.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
class RecoveryTest {
class RecoveryTest : public testing::Test {
public:
RecoveryTest() : env_(Env::Default()), db_(nullptr) {
dbname_ = test::TmpDir() + "/recovery_test";
dbname_ = testing::TempDir() + "recovery_test";
DestroyDB(dbname_, Options());
Open();
}
@ -63,7 +63,7 @@ class RecoveryTest {
}
void Open(Options* options = nullptr) {
ASSERT_OK(OpenWithStatus(options));
ASSERT_LEVELDB_OK(OpenWithStatus(options));
ASSERT_EQ(1, NumLogs());
}
@ -84,37 +84,37 @@ class RecoveryTest {
std::string ManifestFileName() {
std::string current;
ASSERT_OK(ReadFileToString(env_, CurrentFileName(dbname_), &current));
EXPECT_LEVELDB_OK(
ReadFileToString(env_, CurrentFileName(dbname_), &current));
size_t len = current.size();
if (len > 0 && current[len-1] == '\n') {
if (len > 0 && current[len - 1] == '\n') {
current.resize(len - 1);
}
return dbname_ + "/" + current;
}
std::string LogName(uint64_t number) {
return LogFileName(dbname_, number);
}
std::string LogName(uint64_t number) { return LogFileName(dbname_, number); }
size_t DeleteLogFiles() {
size_t RemoveLogFiles() {
// Linux allows unlinking open files, but Windows does not.
// Closing the db allows for file deletion.
Close();
std::vector<uint64_t> logs = GetFiles(kLogFile);
for (size_t i = 0; i < logs.size(); i++) {
ASSERT_OK(env_->DeleteFile(LogName(logs[i]))) << LogName(logs[i]);
EXPECT_LEVELDB_OK(env_->RemoveFile(LogName(logs[i]))) << LogName(logs[i]);
}
return logs.size();
}
void DeleteManifestFile() {
ASSERT_OK(env_->DeleteFile(ManifestFileName()));
void RemoveManifestFile() {
ASSERT_LEVELDB_OK(env_->RemoveFile(ManifestFileName()));
}
uint64_t FirstLogFile() {
return GetFiles(kLogFile)[0];
}
uint64_t FirstLogFile() { return GetFiles(kLogFile)[0]; }
std::vector<uint64_t> GetFiles(FileType t) {
std::vector<std::string> filenames;
ASSERT_OK(env_->GetChildren(dbname_, &filenames));
EXPECT_LEVELDB_OK(env_->GetChildren(dbname_, &filenames));
std::vector<uint64_t> result;
for (size_t i = 0; i < filenames.size(); i++) {
uint64_t number;
@ -126,35 +126,29 @@ class RecoveryTest {
return result;
}
int NumLogs() {
return GetFiles(kLogFile).size();
}
int NumLogs() { return GetFiles(kLogFile).size(); }
int NumTables() {
return GetFiles(kTableFile).size();
}
int NumTables() { return GetFiles(kTableFile).size(); }
uint64_t FileSize(const std::string& fname) {
uint64_t result;
ASSERT_OK(env_->GetFileSize(fname, &result)) << fname;
EXPECT_LEVELDB_OK(env_->GetFileSize(fname, &result)) << fname;
return result;
}
void CompactMemTable() {
dbfull()->TEST_CompactMemTable();
}
void CompactMemTable() { dbfull()->TEST_CompactMemTable(); }
// Directly construct a log file that sets key to val.
void MakeLogFile(uint64_t lognum, SequenceNumber seq, Slice key, Slice val) {
std::string fname = LogFileName(dbname_, lognum);
WritableFile* file;
ASSERT_OK(env_->NewWritableFile(fname, &file));
ASSERT_LEVELDB_OK(env_->NewWritableFile(fname, &file));
log::Writer writer(file);
WriteBatch batch;
batch.Put(key, val);
WriteBatchInternal::SetSequence(&batch, seq);
ASSERT_OK(writer.AddRecord(WriteBatchInternal::Contents(&batch)));
ASSERT_OK(file->Flush());
ASSERT_LEVELDB_OK(writer.AddRecord(WriteBatchInternal::Contents(&batch)));
ASSERT_LEVELDB_OK(file->Flush());
delete file;
}
@ -164,12 +158,13 @@ class RecoveryTest {
DB* db_;
};
TEST(RecoveryTest, ManifestReused) {
TEST_F(RecoveryTest, ManifestReused) {
if (!CanAppend()) {
fprintf(stderr, "skipping test because env does not support appending\n");
std::fprintf(stderr,
"skipping test because env does not support appending\n");
return;
}
ASSERT_OK(Put("foo", "bar"));
ASSERT_LEVELDB_OK(Put("foo", "bar"));
Close();
std::string old_manifest = ManifestFileName();
Open();
@ -180,12 +175,13 @@ TEST(RecoveryTest, ManifestReused) {
ASSERT_EQ("bar", Get("foo"));
}
TEST(RecoveryTest, LargeManifestCompacted) {
TEST_F(RecoveryTest, LargeManifestCompacted) {
if (!CanAppend()) {
fprintf(stderr, "skipping test because env does not support appending\n");
std::fprintf(stderr,
"skipping test because env does not support appending\n");
return;
}
ASSERT_OK(Put("foo", "bar"));
ASSERT_LEVELDB_OK(Put("foo", "bar"));
Close();
std::string old_manifest = ManifestFileName();
@ -193,10 +189,10 @@ TEST(RecoveryTest, LargeManifestCompacted) {
{
uint64_t len = FileSize(old_manifest);
WritableFile* file;
ASSERT_OK(env()->NewAppendableFile(old_manifest, &file));
std::string zeroes(3*1048576 - static_cast<size_t>(len), 0);
ASSERT_OK(file->Append(zeroes));
ASSERT_OK(file->Flush());
ASSERT_LEVELDB_OK(env()->NewAppendableFile(old_manifest, &file));
std::string zeroes(3 * 1048576 - static_cast<size_t>(len), 0);
ASSERT_LEVELDB_OK(file->Append(zeroes));
ASSERT_LEVELDB_OK(file->Flush());
delete file;
}
@ -211,22 +207,23 @@ TEST(RecoveryTest, LargeManifestCompacted) {
ASSERT_EQ("bar", Get("foo"));
}
TEST(RecoveryTest, NoLogFiles) {
ASSERT_OK(Put("foo", "bar"));
ASSERT_EQ(1, DeleteLogFiles());
TEST_F(RecoveryTest, NoLogFiles) {
ASSERT_LEVELDB_OK(Put("foo", "bar"));
ASSERT_EQ(1, RemoveLogFiles());
Open();
ASSERT_EQ("NOT_FOUND", Get("foo"));
Open();
ASSERT_EQ("NOT_FOUND", Get("foo"));
}
TEST(RecoveryTest, LogFileReuse) {
TEST_F(RecoveryTest, LogFileReuse) {
if (!CanAppend()) {
fprintf(stderr, "skipping test because env does not support appending\n");
std::fprintf(stderr,
"skipping test because env does not support appending\n");
return;
}
for (int i = 0; i < 2; i++) {
ASSERT_OK(Put("foo", "bar"));
ASSERT_LEVELDB_OK(Put("foo", "bar"));
if (i == 0) {
// Compact to ensure current log is empty
CompactMemTable();
@ -250,13 +247,13 @@ TEST(RecoveryTest, LogFileReuse) {
}
}
TEST(RecoveryTest, MultipleMemTables) {
TEST_F(RecoveryTest, MultipleMemTables) {
// Make a large log.
const int kNum = 1000;
for (int i = 0; i < kNum; i++) {
char buf[100];
snprintf(buf, sizeof(buf), "%050d", i);
ASSERT_OK(Put(buf, buf));
std::snprintf(buf, sizeof(buf), "%050d", i);
ASSERT_LEVELDB_OK(Put(buf, buf));
}
ASSERT_EQ(0, NumTables());
Close();
@ -267,35 +264,35 @@ TEST(RecoveryTest, MultipleMemTables) {
// Force creation of multiple memtables by reducing the write buffer size.
Options opt;
opt.reuse_logs = true;
opt.write_buffer_size = (kNum*100) / 2;
opt.write_buffer_size = (kNum * 100) / 2;
Open(&opt);
ASSERT_LE(2, NumTables());
ASSERT_EQ(1, NumLogs());
ASSERT_NE(old_log_file, FirstLogFile()) << "must not reuse log";
for (int i = 0; i < kNum; i++) {
char buf[100];
snprintf(buf, sizeof(buf), "%050d", i);
std::snprintf(buf, sizeof(buf), "%050d", i);
ASSERT_EQ(buf, Get(buf));
}
}
TEST(RecoveryTest, MultipleLogFiles) {
ASSERT_OK(Put("foo", "bar"));
TEST_F(RecoveryTest, MultipleLogFiles) {
ASSERT_LEVELDB_OK(Put("foo", "bar"));
Close();
ASSERT_EQ(1, NumLogs());
// Make a bunch of uncompacted log files.
uint64_t old_log = FirstLogFile();
MakeLogFile(old_log+1, 1000, "hello", "world");
MakeLogFile(old_log+2, 1001, "hi", "there");
MakeLogFile(old_log+3, 1002, "foo", "bar2");
MakeLogFile(old_log + 1, 1000, "hello", "world");
MakeLogFile(old_log + 2, 1001, "hi", "there");
MakeLogFile(old_log + 3, 1002, "foo", "bar2");
// Recover and check that all log files were processed.
Open();
ASSERT_LE(1, NumTables());
ASSERT_EQ(1, NumLogs());
uint64_t new_log = FirstLogFile();
ASSERT_LE(old_log+3, new_log);
ASSERT_LE(old_log + 3, new_log);
ASSERT_EQ("bar2", Get("foo"));
ASSERT_EQ("world", Get("hello"));
ASSERT_EQ("there", Get("hi"));
@ -313,7 +310,7 @@ TEST(RecoveryTest, MultipleLogFiles) {
// Check that introducing an older log file does not cause it to be re-read.
Close();
MakeLogFile(old_log+1, 2000, "hello", "stale write");
MakeLogFile(old_log + 1, 2000, "hello", "stale write");
Open();
ASSERT_LE(1, NumTables());
ASSERT_EQ(1, NumLogs());
@ -325,17 +322,13 @@ TEST(RecoveryTest, MultipleLogFiles) {
ASSERT_EQ("there", Get("hi"));
}
TEST(RecoveryTest, ManifestMissing) {
ASSERT_OK(Put("foo", "bar"));
TEST_F(RecoveryTest, ManifestMissing) {
ASSERT_LEVELDB_OK(Put("foo", "bar"));
Close();
DeleteManifestFile();
RemoveManifestFile();
Status status = OpenWithStatus();
ASSERT_TRUE(status.IsCorruption());
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

88
db/repair.cc
View File

@ -84,9 +84,7 @@ class Repairer {
"recovered %d files; %llu bytes. "
"Some data may have been lost. "
"****",
dbname_.c_str(),
static_cast<int>(tables_.size()),
bytes);
dbname_.c_str(), static_cast<int>(tables_.size()), bytes);
}
return status;
}
@ -97,22 +95,6 @@ class Repairer {
SequenceNumber max_sequence;
};
std::string const dbname_;
Env* const env_;
InternalKeyComparator const icmp_;
InternalFilterPolicy const ipolicy_;
Options const options_;
bool owns_info_log_;
bool owns_cache_;
TableCache* table_cache_;
VersionEdit edit_;
std::vector<std::string> manifests_;
std::vector<uint64_t> table_numbers_;
std::vector<uint64_t> logs_;
std::vector<TableInfo> tables_;
uint64_t next_file_number_;
Status FindFiles() {
std::vector<std::string> filenames;
Status status = env_->GetChildren(dbname_, &filenames);
@ -152,8 +134,7 @@ class Repairer {
Status status = ConvertLogToTable(logs_[i]);
if (!status.ok()) {
Log(options_.info_log, "Log #%llu: ignoring conversion error: %s",
(unsigned long long) logs_[i],
status.ToString().c_str());
(unsigned long long)logs_[i], status.ToString().c_str());
}
ArchiveFile(logname);
}
@ -164,11 +145,10 @@ class Repairer {
Env* env;
Logger* info_log;
uint64_t lognum;
virtual void Corruption(size_t bytes, const Status& s) {
void Corruption(size_t bytes, const Status& s) override {
// We print error messages for corruption, but continue repairing.
Log(info_log, "Log #%llu: dropping %d bytes; %s",
(unsigned long long) lognum,
static_cast<int>(bytes),
(unsigned long long)lognum, static_cast<int>(bytes),
s.ToString().c_str());
}
};
@ -190,8 +170,8 @@ class Repairer {
// corruptions cause entire commits to be skipped instead of
// propagating bad information (like overly large sequence
// numbers).
log::Reader reader(lfile, &reporter, false/*do not checksum*/,
0/*initial_offset*/);
log::Reader reader(lfile, &reporter, false /*do not checksum*/,
0 /*initial_offset*/);
// Read all the records and add to a memtable
std::string scratch;
@ -202,8 +182,8 @@ class Repairer {
int counter = 0;
while (reader.ReadRecord(&record, &scratch)) {
if (record.size() < 12) {
reporter.Corruption(
record.size(), Status::Corruption("log record too small"));
reporter.Corruption(record.size(),
Status::Corruption("log record too small"));
continue;
}
WriteBatchInternal::SetContents(&batch, record);
@ -212,8 +192,7 @@ class Repairer {
counter += WriteBatchInternal::Count(&batch);
} else {
Log(options_.info_log, "Log #%llu: ignoring %s",
(unsigned long long) log,
status.ToString().c_str());
(unsigned long long)log, status.ToString().c_str());
status = Status::OK(); // Keep going with rest of file
}
}
@ -234,9 +213,7 @@ class Repairer {
}
}
Log(options_.info_log, "Log #%llu: %d ops saved to Table #%llu %s",
(unsigned long long) log,
counter,
(unsigned long long) meta.number,
(unsigned long long)log, counter, (unsigned long long)meta.number,
status.ToString().c_str());
return status;
}
@ -272,8 +249,7 @@ class Repairer {
ArchiveFile(TableFileName(dbname_, number));
ArchiveFile(SSTTableFileName(dbname_, number));
Log(options_.info_log, "Table #%llu: dropped: %s",
(unsigned long long) t.meta.number,
status.ToString().c_str());
(unsigned long long)t.meta.number, status.ToString().c_str());
return;
}
@ -287,8 +263,7 @@ class Repairer {
Slice key = iter->key();
if (!ParseInternalKey(key, &parsed)) {
Log(options_.info_log, "Table #%llu: unparsable key %s",
(unsigned long long) t.meta.number,
EscapeString(key).c_str());
(unsigned long long)t.meta.number, EscapeString(key).c_str());
continue;
}
@ -307,9 +282,7 @@ class Repairer {
}
delete iter;
Log(options_.info_log, "Table #%llu: %d entries %s",
(unsigned long long) t.meta.number,
counter,
status.ToString().c_str());
(unsigned long long)t.meta.number, counter, status.ToString().c_str());
if (status.ok()) {
tables_.push_back(t);
@ -363,12 +336,12 @@ class Repairer {
s = env_->RenameFile(copy, orig);
if (s.ok()) {
Log(options_.info_log, "Table #%llu: %d entries repaired",
(unsigned long long) t.meta.number, counter);
(unsigned long long)t.meta.number, counter);
tables_.push_back(t);
}
}
if (!s.ok()) {
env_->DeleteFile(copy);
env_->RemoveFile(copy);
}
}
@ -395,11 +368,12 @@ class Repairer {
for (size_t i = 0; i < tables_.size(); i++) {
// TODO(opt): separate out into multiple levels
const TableInfo& t = tables_[i];
edit_.AddFile(0, t.meta.number, t.meta.file_size,
t.meta.smallest, t.meta.largest);
edit_.AddFile(0, t.meta.number, t.meta.file_size, t.meta.smallest,
t.meta.largest);
}
//fprintf(stderr, "NewDescriptor:\n%s\n", edit_.DebugString().c_str());
// std::fprintf(stderr,
// "NewDescriptor:\n%s\n", edit_.DebugString().c_str());
{
log::Writer log(file);
std::string record;
@ -413,7 +387,7 @@ class Repairer {
file = nullptr;
if (!status.ok()) {
env_->DeleteFile(tmp);
env_->RemoveFile(tmp);
} else {
// Discard older manifests
for (size_t i = 0; i < manifests_.size(); i++) {
@ -425,7 +399,7 @@ class Repairer {
if (status.ok()) {
status = SetCurrentFile(env_, dbname_, 1);
} else {
env_->DeleteFile(tmp);
env_->RemoveFile(tmp);
}
}
return status;
@ -447,9 +421,25 @@ class Repairer {
new_file.append("/");
new_file.append((slash == nullptr) ? fname.c_str() : slash + 1);
Status s = env_->RenameFile(fname, new_file);
Log(options_.info_log, "Archiving %s: %s\n",
fname.c_str(), s.ToString().c_str());
Log(options_.info_log, "Archiving %s: %s\n", fname.c_str(),
s.ToString().c_str());
}
const std::string dbname_;
Env* const env_;
InternalKeyComparator const icmp_;
InternalFilterPolicy const ipolicy_;
const Options options_;
bool owns_info_log_;
bool owns_cache_;
TableCache* table_cache_;
VersionEdit edit_;
std::vector<std::string> manifests_;
std::vector<uint64_t> table_numbers_;
std::vector<uint64_t> logs_;
std::vector<TableInfo> tables_;
uint64_t next_file_number_;
};
} // namespace

150
db/skiplist.h
View File

@ -27,17 +27,16 @@
//
// ... prev vs. next pointer ordering ...
#include <assert.h>
#include <stdlib.h>
#include "port/port.h"
#include <atomic>
#include <cassert>
#include <cstdlib>
#include "util/arena.h"
#include "util/random.h"
namespace leveldb {
class Arena;
template<typename Key, class Comparator>
template <typename Key, class Comparator>
class SkipList {
private:
struct Node;
@ -48,6 +47,9 @@ class SkipList {
// must remain allocated for the lifetime of the skiplist object.
explicit SkipList(Comparator cmp, Arena* arena);
SkipList(const SkipList&) = delete;
SkipList& operator=(const SkipList&) = delete;
// Insert key into the list.
// REQUIRES: nothing that compares equal to key is currently in the list.
void Insert(const Key& key);
@ -97,24 +99,10 @@ class SkipList {
private:
enum { kMaxHeight = 12 };
// Immutable after construction
Comparator const compare_;
Arena* const arena_; // Arena used for allocations of nodes
Node* const head_;
// Modified only by Insert(). Read racily by readers, but stale
// values are ok.
port::AtomicPointer max_height_; // Height of the entire list
inline int GetMaxHeight() const {
return static_cast<int>(
reinterpret_cast<intptr_t>(max_height_.NoBarrier_Load()));
return max_height_.load(std::memory_order_relaxed);
}
// Read/written only by Insert().
Random rnd_;
Node* NewNode(const Key& key, int height);
int RandomHeight();
bool Equal(const Key& a, const Key& b) const { return (compare_(a, b) == 0); }
@ -137,15 +125,24 @@ class SkipList {
// Return head_ if list is empty.
Node* FindLast() const;
// No copying allowed
SkipList(const SkipList&);
void operator=(const SkipList&);
// Immutable after construction
Comparator const compare_;
Arena* const arena_; // Arena used for allocations of nodes
Node* const head_;
// Modified only by Insert(). Read racily by readers, but stale
// values are ok.
std::atomic<int> max_height_; // Height of the entire list
// Read/written only by Insert().
Random rnd_;
};
// Implementation details follow
template<typename Key, class Comparator>
struct SkipList<Key,Comparator>::Node {
explicit Node(const Key& k) : key(k) { }
template <typename Key, class Comparator>
struct SkipList<Key, Comparator>::Node {
explicit Node(const Key& k) : key(k) {}
Key const key;
@ -155,63 +152,63 @@ struct SkipList<Key,Comparator>::Node {
assert(n >= 0);
// Use an 'acquire load' so that we observe a fully initialized
// version of the returned Node.
return reinterpret_cast<Node*>(next_[n].Acquire_Load());
return next_[n].load(std::memory_order_acquire);
}
void SetNext(int n, Node* x) {
assert(n >= 0);
// Use a 'release store' so that anybody who reads through this
// pointer observes a fully initialized version of the inserted node.
next_[n].Release_Store(x);
next_[n].store(x, std::memory_order_release);
}
// No-barrier variants that can be safely used in a few locations.
Node* NoBarrier_Next(int n) {
assert(n >= 0);
return reinterpret_cast<Node*>(next_[n].NoBarrier_Load());
return next_[n].load(std::memory_order_relaxed);
}
void NoBarrier_SetNext(int n, Node* x) {
assert(n >= 0);
next_[n].NoBarrier_Store(x);
next_[n].store(x, std::memory_order_relaxed);
}
private:
// Array of length equal to the node height. next_[0] is lowest level link.
port::AtomicPointer next_[1];
std::atomic<Node*> next_[1];
};
template<typename Key, class Comparator>
typename SkipList<Key,Comparator>::Node*
SkipList<Key,Comparator>::NewNode(const Key& key, int height) {
char* mem = arena_->AllocateAligned(
sizeof(Node) + sizeof(port::AtomicPointer) * (height - 1));
return new (mem) Node(key);
template <typename Key, class Comparator>
typename SkipList<Key, Comparator>::Node* SkipList<Key, Comparator>::NewNode(
const Key& key, int height) {
char* const node_memory = arena_->AllocateAligned(
sizeof(Node) + sizeof(std::atomic<Node*>) * (height - 1));
return new (node_memory) Node(key);
}
template<typename Key, class Comparator>
inline SkipList<Key,Comparator>::Iterator::Iterator(const SkipList* list) {
template <typename Key, class Comparator>
inline SkipList<Key, Comparator>::Iterator::Iterator(const SkipList* list) {
list_ = list;
node_ = nullptr;
}
template<typename Key, class Comparator>
inline bool SkipList<Key,Comparator>::Iterator::Valid() const {
template <typename Key, class Comparator>
inline bool SkipList<Key, Comparator>::Iterator::Valid() const {
return node_ != nullptr;
}
template<typename Key, class Comparator>
inline const Key& SkipList<Key,Comparator>::Iterator::key() const {
template <typename Key, class Comparator>
inline const Key& SkipList<Key, Comparator>::Iterator::key() const {
assert(Valid());
return node_->key;
}
template<typename Key, class Comparator>
inline void SkipList<Key,Comparator>::Iterator::Next() {
template <typename Key, class Comparator>
inline void SkipList<Key, Comparator>::Iterator::Next() {
assert(Valid());
node_ = node_->Next(0);
}
template<typename Key, class Comparator>
inline void SkipList<Key,Comparator>::Iterator::Prev() {
template <typename Key, class Comparator>
inline void SkipList<Key, Comparator>::Iterator::Prev() {
// Instead of using explicit "prev" links, we just search for the
// last node that falls before key.
assert(Valid());
@ -221,30 +218,30 @@ inline void SkipList<Key,Comparator>::Iterator::Prev() {
}
}
template<typename Key, class Comparator>
inline void SkipList<Key,Comparator>::Iterator::Seek(const Key& target) {
template <typename Key, class Comparator>
inline void SkipList<Key, Comparator>::Iterator::Seek(const Key& target) {
node_ = list_->FindGreaterOrEqual(target, nullptr);
}
template<typename Key, class Comparator>
inline void SkipList<Key,Comparator>::Iterator::SeekToFirst() {
template <typename Key, class Comparator>
inline void SkipList<Key, Comparator>::Iterator::SeekToFirst() {
node_ = list_->head_->Next(0);
}
template<typename Key, class Comparator>
inline void SkipList<Key,Comparator>::Iterator::SeekToLast() {
template <typename Key, class Comparator>
inline void SkipList<Key, Comparator>::Iterator::SeekToLast() {
node_ = list_->FindLast();
if (node_ == list_->head_) {
node_ = nullptr;
}
}
template<typename Key, class Comparator>
int SkipList<Key,Comparator>::RandomHeight() {
template <typename Key, class Comparator>
int SkipList<Key, Comparator>::RandomHeight() {
// Increase height with probability 1 in kBranching
static const unsigned int kBranching = 4;
int height = 1;
while (height < kMaxHeight && ((rnd_.Next() % kBranching) == 0)) {
while (height < kMaxHeight && rnd_.OneIn(kBranching)) {
height++;
}
assert(height > 0);
@ -252,15 +249,16 @@ int SkipList<Key,Comparator>::RandomHeight() {
return height;
}
template<typename Key, class Comparator>
bool SkipList<Key,Comparator>::KeyIsAfterNode(const Key& key, Node* n) const {
template <typename Key, class Comparator>
bool SkipList<Key, Comparator>::KeyIsAfterNode(const Key& key, Node* n) const {
// null n is considered infinite
return (n != nullptr) && (compare_(n->key, key) < 0);
}
template<typename Key, class Comparator>
typename SkipList<Key,Comparator>::Node* SkipList<Key,Comparator>::FindGreaterOrEqual(const Key& key, Node** prev)
const {
template <typename Key, class Comparator>
typename SkipList<Key, Comparator>::Node*
SkipList<Key, Comparator>::FindGreaterOrEqual(const Key& key,
Node** prev) const {
Node* x = head_;
int level = GetMaxHeight() - 1;
while (true) {
@ -280,9 +278,9 @@ typename SkipList<Key,Comparator>::Node* SkipList<Key,Comparator>::FindGreaterOr
}
}
template<typename Key, class Comparator>
typename SkipList<Key,Comparator>::Node*
SkipList<Key,Comparator>::FindLessThan(const Key& key) const {
template <typename Key, class Comparator>
typename SkipList<Key, Comparator>::Node*
SkipList<Key, Comparator>::FindLessThan(const Key& key) const {
Node* x = head_;
int level = GetMaxHeight() - 1;
while (true) {
@ -301,8 +299,8 @@ SkipList<Key,Comparator>::FindLessThan(const Key& key) const {
}
}
template<typename Key, class Comparator>
typename SkipList<Key,Comparator>::Node* SkipList<Key,Comparator>::FindLast()
template <typename Key, class Comparator>
typename SkipList<Key, Comparator>::Node* SkipList<Key, Comparator>::FindLast()
const {
Node* x = head_;
int level = GetMaxHeight() - 1;
@ -321,20 +319,20 @@ typename SkipList<Key,Comparator>::Node* SkipList<Key,Comparator>::FindLast()
}
}
template<typename Key, class Comparator>
SkipList<Key,Comparator>::SkipList(Comparator cmp, Arena* arena)
template <typename Key, class Comparator>
SkipList<Key, Comparator>::SkipList(Comparator cmp, Arena* arena)
: compare_(cmp),
arena_(arena),
head_(NewNode(0 /* any key will do */, kMaxHeight)),
max_height_(reinterpret_cast<void*>(1)),
max_height_(1),
rnd_(0xdeadbeef) {
for (int i = 0; i < kMaxHeight; i++) {
head_->SetNext(i, nullptr);
}
}
template<typename Key, class Comparator>
void SkipList<Key,Comparator>::Insert(const Key& key) {
template <typename Key, class Comparator>
void SkipList<Key, Comparator>::Insert(const Key& key) {
// TODO(opt): We can use a barrier-free variant of FindGreaterOrEqual()
// here since Insert() is externally synchronized.
Node* prev[kMaxHeight];
@ -348,8 +346,6 @@ void SkipList<Key,Comparator>::Insert(const Key& key) {
for (int i = GetMaxHeight(); i < height; i++) {
prev[i] = head_;
}
//fprintf(stderr, "Change height from %d to %d\n", max_height_, height);
// It is ok to mutate max_height_ without any synchronization
// with concurrent readers. A concurrent reader that observes
// the new value of max_height_ will see either the old value of
@ -357,7 +353,7 @@ void SkipList<Key,Comparator>::Insert(const Key& key) {
// the loop below. In the former case the reader will
// immediately drop to the next level since nullptr sorts after all
// keys. In the latter case the reader will use the new node.
max_height_.NoBarrier_Store(reinterpret_cast<void*>(height));
max_height_.store(height, std::memory_order_relaxed);
}
x = NewNode(key, height);
@ -369,8 +365,8 @@ void SkipList<Key,Comparator>::Insert(const Key& key) {
}
}
template<typename Key, class Comparator>
bool SkipList<Key,Comparator>::Contains(const Key& key) const {
template <typename Key, class Comparator>
bool SkipList<Key, Comparator>::Contains(const Key& key) const {
Node* x = FindGreaterOrEqual(key, nullptr);
if (x != nullptr && Equal(key, x->key)) {
return true;

66
db/skiplist_test.cc
View File

@ -3,14 +3,18 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/skiplist.h"
#include <atomic>
#include <set>
#include "gtest/gtest.h"
#include "leveldb/env.h"
#include "port/port.h"
#include "port/thread_annotations.h"
#include "util/arena.h"
#include "util/hash.h"
#include "util/random.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
@ -28,8 +32,6 @@ struct Comparator {
}
};
class SkipTest { };
TEST(SkipTest, Empty) {
Arena arena;
Comparator cmp;
@ -114,8 +116,7 @@ TEST(SkipTest, InsertAndLookup) {
// Compare against model iterator
for (std::set<Key>::reverse_iterator model_iter = keys.rbegin();
model_iter != keys.rend();
++model_iter) {
model_iter != keys.rend(); ++model_iter) {
ASSERT_TRUE(iter.Valid());
ASSERT_EQ(*model_iter, iter.key());
iter.Prev();
@ -128,7 +129,7 @@ TEST(SkipTest, InsertAndLookup) {
// concurrent readers (with no synchronization other than when a
// reader's iterator is created), the reader always observes all the
// data that was present in the skip list when the iterator was
// constructor. Because insertions are happening concurrently, we may
// constructed. Because insertions are happening concurrently, we may
// also observe new values that were inserted since the iterator was
// constructed, but we should never miss any values that were present
// at iterator construction time.
@ -150,19 +151,19 @@ TEST(SkipTest, InsertAndLookup) {
// been concurrently added since the iterator started.
class ConcurrentTest {
private:
static const uint32_t K = 4;
static constexpr uint32_t K = 4;
static uint64_t key(Key key) { return (key >> 40); }
static uint64_t gen(Key key) { return (key >> 8) & 0xffffffffu; }
static uint64_t hash(Key key) { return key & 0xff; }
static uint64_t HashNumbers(uint64_t k, uint64_t g) {
uint64_t data[2] = { k, g };
uint64_t data[2] = {k, g};
return Hash(reinterpret_cast<char*>(data), sizeof(data), 0);
}
static Key MakeKey(uint64_t k, uint64_t g) {
assert(sizeof(Key) == sizeof(uint64_t));
static_assert(sizeof(Key) == sizeof(uint64_t), "");
assert(k <= K); // We sometimes pass K to seek to the end of the skiplist
assert(g <= 0xffffffffu);
return ((k << 40) | (g << 8) | (HashNumbers(k, g) & 0xff));
@ -188,13 +189,11 @@ class ConcurrentTest {
// Per-key generation
struct State {
port::AtomicPointer generation[K];
void Set(int k, intptr_t v) {
generation[k].Release_Store(reinterpret_cast<void*>(v));
}
intptr_t Get(int k) {
return reinterpret_cast<intptr_t>(generation[k].Acquire_Load());
std::atomic<int> generation[K];
void Set(int k, int v) {
generation[k].store(v, std::memory_order_release);
}
int Get(int k) { return generation[k].load(std::memory_order_acquire); }
State() {
for (int k = 0; k < K; k++) {
@ -213,7 +212,7 @@ class ConcurrentTest {
SkipList<Key, Comparator> list_;
public:
ConcurrentTest() : list_(Comparator(), &arena_) { }
ConcurrentTest() : list_(Comparator(), &arena_) {}
// REQUIRES: External synchronization
void WriteStep(Random* rnd) {
@ -252,11 +251,9 @@ class ConcurrentTest {
// Note that generation 0 is never inserted, so it is ok if
// <*,0,*> is missing.
ASSERT_TRUE((gen(pos) == 0) ||
(gen(pos) > static_cast<Key>(initial_state.Get(key(pos))))
) << "key: " << key(pos)
<< "; gen: " << gen(pos)
<< "; initgen: "
<< initial_state.Get(key(pos));
(gen(pos) > static_cast<Key>(initial_state.Get(key(pos)))))
<< "key: " << key(pos) << "; gen: " << gen(pos)
<< "; initgen: " << initial_state.Get(key(pos));
// Advance to next key in the valid key space
if (key(pos) < key(current)) {
@ -283,7 +280,9 @@ class ConcurrentTest {
}
}
};
const uint32_t ConcurrentTest::K;
// Needed when building in C++11 mode.
constexpr uint32_t ConcurrentTest::K;
// Simple test that does single-threaded testing of the ConcurrentTest
// scaffolding.
@ -300,19 +299,12 @@ class TestState {
public:
ConcurrentTest t_;
int seed_;
port::AtomicPointer quit_flag_;
std::atomic<bool> quit_flag_;
enum ReaderState {
STARTING,
RUNNING,
DONE
};
enum ReaderState { STARTING, RUNNING, DONE };
explicit TestState(int s)
: seed_(s),
quit_flag_(nullptr),
state_(STARTING),
state_cv_(&mu_) {}
: seed_(s), quit_flag_(false), state_(STARTING), state_cv_(&mu_) {}
void Wait(ReaderState s) LOCKS_EXCLUDED(mu_) {
mu_.Lock();
@ -340,7 +332,7 @@ static void ConcurrentReader(void* arg) {
Random rnd(state->seed_);
int64_t reads = 0;
state->Change(TestState::RUNNING);
while (!state->quit_flag_.Acquire_Load()) {
while (!state->quit_flag_.load(std::memory_order_acquire)) {
state->t_.ReadStep(&rnd);
++reads;
}
@ -354,7 +346,7 @@ static void RunConcurrent(int run) {
const int kSize = 1000;
for (int i = 0; i < N; i++) {
if ((i % 100) == 0) {
fprintf(stderr, "Run %d of %d\n", i, N);
std::fprintf(stderr, "Run %d of %d\n", i, N);
}
TestState state(seed + 1);
Env::Default()->Schedule(ConcurrentReader, &state);
@ -362,7 +354,7 @@ static void RunConcurrent(int run) {
for (int i = 0; i < kSize; i++) {
state.t_.WriteStep(&rnd);
}
state.quit_flag_.Release_Store(&state); // Any non-null arg will do
state.quit_flag_.store(true, std::memory_order_release);
state.Wait(TestState::DONE);
}
}
@ -374,7 +366,3 @@ TEST(SkipTest, Concurrent4) { RunConcurrent(4); }
TEST(SkipTest, Concurrent5) { RunConcurrent(5); }
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

78
db/snapshot.h
View File

@ -16,50 +16,78 @@ class SnapshotList;
// Each SnapshotImpl corresponds to a particular sequence number.
class SnapshotImpl : public Snapshot {
public:
SequenceNumber number_; // const after creation
SnapshotImpl(SequenceNumber sequence_number)
: sequence_number_(sequence_number) {}
SequenceNumber sequence_number() const { return sequence_number_; }
private:
friend class SnapshotList;
// SnapshotImpl is kept in a doubly-linked circular list
// SnapshotImpl is kept in a doubly-linked circular list. The SnapshotList
// implementation operates on the next/previous fields directly.
SnapshotImpl* prev_;
SnapshotImpl* next_;
SnapshotList* list_; // just for sanity checks
const SequenceNumber sequence_number_;
#if !defined(NDEBUG)
SnapshotList* list_ = nullptr;
#endif // !defined(NDEBUG)
};
class SnapshotList {
public:
SnapshotList() {
list_.prev_ = &list_;
list_.next_ = &list_;
SnapshotList() : head_(0) {
head_.prev_ = &head_;
head_.next_ = &head_;
}
bool empty() const { return list_.next_ == &list_; }
SnapshotImpl* oldest() const { assert(!empty()); return list_.next_; }
SnapshotImpl* newest() const { assert(!empty()); return list_.prev_; }
const SnapshotImpl* New(SequenceNumber seq) {
SnapshotImpl* s = new SnapshotImpl;
s->number_ = seq;
s->list_ = this;
s->next_ = &list_;
s->prev_ = list_.prev_;
s->prev_->next_ = s;
s->next_->prev_ = s;
return s;
bool empty() const { return head_.next_ == &head_; }
SnapshotImpl* oldest() const {
assert(!empty());
return head_.next_;
}
SnapshotImpl* newest() const {
assert(!empty());
return head_.prev_;
}
void Delete(const SnapshotImpl* s) {
assert(s->list_ == this);
s->prev_->next_ = s->next_;
s->next_->prev_ = s->prev_;
delete s;
// Creates a SnapshotImpl and appends it to the end of the list.
SnapshotImpl* New(SequenceNumber sequence_number) {
assert(empty() || newest()->sequence_number_ <= sequence_number);
SnapshotImpl* snapshot = new SnapshotImpl(sequence_number);
#if !defined(NDEBUG)
snapshot->list_ = this;
#endif // !defined(NDEBUG)
snapshot->next_ = &head_;
snapshot->prev_ = head_.prev_;
snapshot->prev_->next_ = snapshot;
snapshot->next_->prev_ = snapshot;
return snapshot;
}
// Removes a SnapshotImpl from this list.
//
// The snapshot must have been created by calling New() on this list.
//
// The snapshot pointer should not be const, because its memory is
// deallocated. However, that would force us to change DB::ReleaseSnapshot(),
// which is in the API, and currently takes a const Snapshot.
void Delete(const SnapshotImpl* snapshot) {
#if !defined(NDEBUG)
assert(snapshot->list_ == this);
#endif // !defined(NDEBUG)
snapshot->prev_->next_ = snapshot->next_;
snapshot->next_->prev_ = snapshot->prev_;
delete snapshot;
}
private:
// Dummy head of doubly-linked list of snapshots
SnapshotImpl list_;
SnapshotImpl head_;
};
} // namespace leveldb

25
db/table_cache.cc
View File

@ -29,18 +29,14 @@ static void UnrefEntry(void* arg1, void* arg2) {
cache->Release(h);
}
TableCache::TableCache(const std::string& dbname,
const Options& options,
TableCache::TableCache(const std::string& dbname, const Options& options,
int entries)
: env_(options.env),
dbname_(dbname),
options_(options),
cache_(NewLRUCache(entries)) {
}
cache_(NewLRUCache(entries)) {}
TableCache::~TableCache() {
delete cache_;
}
TableCache::~TableCache() { delete cache_; }
Status TableCache::FindTable(uint64_t file_number, uint64_t file_size,
Cache::Handle** handle) {
@ -80,8 +76,7 @@ Status TableCache::FindTable(uint64_t file_number, uint64_t file_size,
}
Iterator* TableCache::NewIterator(const ReadOptions& options,
uint64_t file_number,
uint64_t file_size,
uint64_t file_number, uint64_t file_size,
Table** tableptr) {
if (tableptr != nullptr) {
*tableptr = nullptr;
@ -102,17 +97,15 @@ Iterator* TableCache::NewIterator(const ReadOptions& options,
return result;
}
Status TableCache::Get(const ReadOptions& options,
uint64_t file_number,
uint64_t file_size,
const Slice& k,
void* arg,
void (*saver)(void*, const Slice&, const Slice&)) {
Status TableCache::Get(const ReadOptions& options, uint64_t file_number,
uint64_t file_size, const Slice& k, void* arg,
void (*handle_result)(void*, const Slice&,
const Slice&)) {
Cache::Handle* handle = nullptr;
Status s = FindTable(file_number, file_size, &handle);
if (s.ok()) {
Table* t = reinterpret_cast<TableAndFile*>(cache_->Value(handle))->table;
s = t->InternalGet(options, k, arg, saver);
s = t->InternalGet(options, k, arg, handle_result);
cache_->Release(handle);
}
return s;

24
db/table_cache.h
View File

@ -7,8 +7,9 @@
#ifndef STORAGE_LEVELDB_DB_TABLE_CACHE_H_
#define STORAGE_LEVELDB_DB_TABLE_CACHE_H_
#include <cstdint>
#include <string>
#include <stdint.h>
#include "db/dbformat.h"
#include "leveldb/cache.h"
#include "leveldb/table.h"
@ -21,6 +22,10 @@ class Env;
class TableCache {
public:
TableCache(const std::string& dbname, const Options& options, int entries);
TableCache(const TableCache&) = delete;
TableCache& operator=(const TableCache&) = delete;
~TableCache();
// Return an iterator for the specified file number (the corresponding
@ -30,30 +35,25 @@ class TableCache {
// underlies the returned iterator. The returned "*tableptr" object is owned
// by the cache and should not be deleted, and is valid for as long as the
// returned iterator is live.
Iterator* NewIterator(const ReadOptions& options,
uint64_t file_number,
uint64_t file_size,
Table** tableptr = nullptr);
Iterator* NewIterator(const ReadOptions& options, uint64_t file_number,
uint64_t file_size, Table** tableptr = nullptr);
// If a seek to internal key "k" in specified file finds an entry,
// call (*handle_result)(arg, found_key, found_value).
Status Get(const ReadOptions& options,
uint64_t file_number,
uint64_t file_size,
const Slice& k,
void* arg,
Status Get(const ReadOptions& options, uint64_t file_number,
uint64_t file_size, const Slice& k, void* arg,
void (*handle_result)(void*, const Slice&, const Slice&));
// Evict any entry for the specified file number
void Evict(uint64_t file_number);
private:
Status FindTable(uint64_t file_number, uint64_t file_size, Cache::Handle**);
Env* const env_;
const std::string dbname_;
const Options& options_;
Cache* cache_;
Status FindTable(uint64_t file_number, uint64_t file_size, Cache::Handle**);
};
} // namespace leveldb

50
db/version_edit.cc
View File

@ -12,15 +12,15 @@ namespace leveldb {
// Tag numbers for serialized VersionEdit. These numbers are written to
// disk and should not be changed.
enum Tag {
kComparator = 1,
kLogNumber = 2,
kNextFileNumber = 3,
kLastSequence = 4,
kCompactPointer = 5,
kDeletedFile = 6,
kNewFile = 7,
kComparator = 1,
kLogNumber = 2,
kNextFileNumber = 3,
kLastSequence = 4,
kCompactPointer = 5,
kDeletedFile = 6,
kNewFile = 7,
// 8 was used for large value refs
kPrevLogNumber = 9
kPrevLogNumber = 9
};
void VersionEdit::Clear() {
@ -34,6 +34,7 @@ void VersionEdit::Clear() {
has_prev_log_number_ = false;
has_next_file_number_ = false;
has_last_sequence_ = false;
compact_pointers_.clear();
deleted_files_.clear();
new_files_.clear();
}
@ -66,12 +67,10 @@ void VersionEdit::EncodeTo(std::string* dst) const {
PutLengthPrefixedSlice(dst, compact_pointers_[i].second.Encode());
}
for (DeletedFileSet::const_iterator iter = deleted_files_.begin();
iter != deleted_files_.end();
++iter) {
for (const auto& deleted_file_kvp : deleted_files_) {
PutVarint32(dst, kDeletedFile);
PutVarint32(dst, iter->first); // level
PutVarint64(dst, iter->second); // file number
PutVarint32(dst, deleted_file_kvp.first); // level
PutVarint64(dst, deleted_file_kvp.second); // file number
}
for (size_t i = 0; i < new_files_.size(); i++) {
@ -88,8 +87,7 @@ void VersionEdit::EncodeTo(std::string* dst) const {
static bool GetInternalKey(Slice* input, InternalKey* dst) {
Slice str;
if (GetLengthPrefixedSlice(input, &str)) {
dst->DecodeFrom(str);
return true;
return dst->DecodeFrom(str);
} else {
return false;
}
@ -97,8 +95,7 @@ static bool GetInternalKey(Slice* input, InternalKey* dst) {
static bool GetLevel(Slice* input, int* level) {
uint32_t v;
if (GetVarint32(input, &v) &&
v < config::kNumLevels) {
if (GetVarint32(input, &v) && v < config::kNumLevels) {
*level = v;
return true;
} else {
@ -163,8 +160,7 @@ Status VersionEdit::DecodeFrom(const Slice& src) {
break;
case kCompactPointer:
if (GetLevel(&input, &level) &&
GetInternalKey(&input, &key)) {
if (GetLevel(&input, &level) && GetInternalKey(&input, &key)) {
compact_pointers_.push_back(std::make_pair(level, key));
} else {
msg = "compaction pointer";
@ -172,8 +168,7 @@ Status VersionEdit::DecodeFrom(const Slice& src) {
break;
case kDeletedFile:
if (GetLevel(&input, &level) &&
GetVarint64(&input, &number)) {
if (GetLevel(&input, &level) && GetVarint64(&input, &number)) {
deleted_files_.insert(std::make_pair(level, number));
} else {
msg = "deleted file";
@ -181,8 +176,7 @@ Status VersionEdit::DecodeFrom(const Slice& src) {
break;
case kNewFile:
if (GetLevel(&input, &level) &&
GetVarint64(&input, &f.number) &&
if (GetLevel(&input, &level) && GetVarint64(&input, &f.number) &&
GetVarint64(&input, &f.file_size) &&
GetInternalKey(&input, &f.smallest) &&
GetInternalKey(&input, &f.largest)) {
@ -238,13 +232,11 @@ std::string VersionEdit::DebugString() const {
r.append(" ");
r.append(compact_pointers_[i].second.DebugString());
}
for (DeletedFileSet::const_iterator iter = deleted_files_.begin();
iter != deleted_files_.end();
++iter) {
r.append("\n DeleteFile: ");
AppendNumberTo(&r, iter->first);
for (const auto& deleted_files_kvp : deleted_files_) {
r.append("\n RemoveFile: ");
AppendNumberTo(&r, deleted_files_kvp.first);
r.append(" ");
AppendNumberTo(&r, iter->second);
AppendNumberTo(&r, deleted_files_kvp.second);
}
for (size_t i = 0; i < new_files_.size(); i++) {
const FileMetaData& f = new_files_[i].second;

31
db/version_edit.h
View File

@ -8,6 +8,7 @@
#include <set>
#include <utility>
#include <vector>
#include "db/dbformat.h"
namespace leveldb {
@ -15,20 +16,20 @@ namespace leveldb {
class VersionSet;
struct FileMetaData {
int refs;
int allowed_seeks; // Seeks allowed until compaction
uint64_t number;
uint64_t file_size; // File size in bytes
InternalKey smallest; // Smallest internal key served by table
InternalKey largest; // Largest internal key served by table
FileMetaData() : refs(0), allowed_seeks(1 << 30), file_size(0) {}
FileMetaData() : refs(0), allowed_seeks(1 << 30), file_size(0) { }
int refs;
int allowed_seeks; // Seeks allowed until compaction
uint64_t number;
uint64_t file_size; // File size in bytes
InternalKey smallest; // Smallest internal key served by table
InternalKey largest; // Largest internal key served by table
};
class VersionEdit {
public:
VersionEdit() { Clear(); }
~VersionEdit() { }
~VersionEdit() = default;
void Clear();
@ -59,10 +60,8 @@ class VersionEdit {
// Add the specified file at the specified number.
// REQUIRES: This version has not been saved (see VersionSet::SaveTo)
// REQUIRES: "smallest" and "largest" are smallest and largest keys in file
void AddFile(int level, uint64_t file,
uint64_t file_size,
const InternalKey& smallest,
const InternalKey& largest) {
void AddFile(int level, uint64_t file, uint64_t file_size,
const InternalKey& smallest, const InternalKey& largest) {
FileMetaData f;
f.number = file;
f.file_size = file_size;
@ -72,7 +71,7 @@ class VersionEdit {
}
// Delete the specified "file" from the specified "level".
void DeleteFile(int level, uint64_t file) {
void RemoveFile(int level, uint64_t file) {
deleted_files_.insert(std::make_pair(level, file));
}
@ -84,7 +83,7 @@ class VersionEdit {
private:
friend class VersionSet;
typedef std::set< std::pair<int, uint64_t> > DeletedFileSet;
typedef std::set<std::pair<int, uint64_t>> DeletedFileSet;
std::string comparator_;
uint64_t log_number_;
@ -97,9 +96,9 @@ class VersionEdit {
bool has_next_file_number_;
bool has_last_sequence_;
std::vector< std::pair<int, InternalKey> > compact_pointers_;
std::vector<std::pair<int, InternalKey>> compact_pointers_;
DeletedFileSet deleted_files_;
std::vector< std::pair<int, FileMetaData> > new_files_;
std::vector<std::pair<int, FileMetaData>> new_files_;
};
} // namespace leveldb

11
db/version_edit_test.cc
View File

@ -3,7 +3,8 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/version_edit.h"
#include "util/testharness.h"
#include "gtest/gtest.h"
namespace leveldb {
@ -17,8 +18,6 @@ static void TestEncodeDecode(const VersionEdit& edit) {
ASSERT_EQ(encoded, encoded2);
}
class VersionEditTest { };
TEST(VersionEditTest, EncodeDecode) {
static const uint64_t kBig = 1ull << 50;
@ -28,7 +27,7 @@ TEST(VersionEditTest, EncodeDecode) {
edit.AddFile(3, kBig + 300 + i, kBig + 400 + i,
InternalKey("foo", kBig + 500 + i, kTypeValue),
InternalKey("zoo", kBig + 600 + i, kTypeDeletion));
edit.DeleteFile(4, kBig + 700 + i);
edit.RemoveFile(4, kBig + 700 + i);
edit.SetCompactPointer(i, InternalKey("x", kBig + 900 + i, kTypeValue));
}
@ -40,7 +39,3 @@ TEST(VersionEditTest, EncodeDecode) {
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

472
db/version_set.cc
View File

@ -5,7 +5,8 @@
#include "db/version_set.h"
#include <algorithm>
#include <stdio.h>
#include <cstdio>
#include "db/filename.h"
#include "db/log_reader.h"
#include "db/log_writer.h"
@ -20,7 +21,7 @@
namespace leveldb {
static int TargetFileSize(const Options* options) {
static size_t TargetFileSize(const Options* options) {
return options->max_file_size;
}
@ -84,8 +85,7 @@ Version::~Version() {
}
int FindFile(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& files,
const Slice& key) {
const std::vector<FileMetaData*>& files, const Slice& key) {
uint32_t left = 0;
uint32_t right = files.size();
while (left < right) {
@ -104,26 +104,25 @@ int FindFile(const InternalKeyComparator& icmp,
return right;
}
static bool AfterFile(const Comparator* ucmp,
const Slice* user_key, const FileMetaData* f) {
static bool AfterFile(const Comparator* ucmp, const Slice* user_key,
const FileMetaData* f) {
// null user_key occurs before all keys and is therefore never after *f
return (user_key != nullptr &&
ucmp->Compare(*user_key, f->largest.user_key()) > 0);
}
static bool BeforeFile(const Comparator* ucmp,
const Slice* user_key, const FileMetaData* f) {
static bool BeforeFile(const Comparator* ucmp, const Slice* user_key,
const FileMetaData* f) {
// null user_key occurs after all keys and is therefore never before *f
return (user_key != nullptr &&
ucmp->Compare(*user_key, f->smallest.user_key()) < 0);
}
bool SomeFileOverlapsRange(
const InternalKeyComparator& icmp,
bool disjoint_sorted_files,
const std::vector<FileMetaData*>& files,
const Slice* smallest_user_key,
const Slice* largest_user_key) {
bool SomeFileOverlapsRange(const InternalKeyComparator& icmp,
bool disjoint_sorted_files,
const std::vector<FileMetaData*>& files,
const Slice* smallest_user_key,
const Slice* largest_user_key) {
const Comparator* ucmp = icmp.user_comparator();
if (!disjoint_sorted_files) {
// Need to check against all files
@ -143,8 +142,9 @@ bool SomeFileOverlapsRange(
uint32_t index = 0;
if (smallest_user_key != nullptr) {
// Find the earliest possible internal key for smallest_user_key
InternalKey small(*smallest_user_key, kMaxSequenceNumber,kValueTypeForSeek);
index = FindFile(icmp, files, small.Encode());
InternalKey small_key(*smallest_user_key, kMaxSequenceNumber,
kValueTypeForSeek);
index = FindFile(icmp, files, small_key.Encode());
}
if (index >= files.size()) {
@ -164,25 +164,21 @@ class Version::LevelFileNumIterator : public Iterator {
public:
LevelFileNumIterator(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>* flist)
: icmp_(icmp),
flist_(flist),
index_(flist->size()) { // Marks as invalid
: icmp_(icmp), flist_(flist), index_(flist->size()) { // Marks as invalid
}
virtual bool Valid() const {
return index_ < flist_->size();
}
virtual void Seek(const Slice& target) {
bool Valid() const override { return index_ < flist_->size(); }
void Seek(const Slice& target) override {
index_ = FindFile(icmp_, *flist_, target);
}
virtual void SeekToFirst() { index_ = 0; }
virtual void SeekToLast() {
void SeekToFirst() override { index_ = 0; }
void SeekToLast() override {
index_ = flist_->empty() ? 0 : flist_->size() - 1;
}
virtual void Next() {
void Next() override {
assert(Valid());
index_++;
}
virtual void Prev() {
void Prev() override {
assert(Valid());
if (index_ == 0) {
index_ = flist_->size(); // Marks as invalid
@ -190,17 +186,18 @@ class Version::LevelFileNumIterator : public Iterator {
index_--;
}
}
Slice key() const {
Slice key() const override {
assert(Valid());
return (*flist_)[index_]->largest.Encode();
}
Slice value() const {
Slice value() const override {
assert(Valid());
EncodeFixed64(value_buf_, (*flist_)[index_]->number);
EncodeFixed64(value_buf_+8, (*flist_)[index_]->file_size);
EncodeFixed64(value_buf_ + 8, (*flist_)[index_]->file_size);
return Slice(value_buf_, sizeof(value_buf_));
}
virtual Status status() const { return Status::OK(); }
Status status() const override { return Status::OK(); }
private:
const InternalKeyComparator icmp_;
const std::vector<FileMetaData*>* const flist_;
@ -210,16 +207,14 @@ class Version::LevelFileNumIterator : public Iterator {
mutable char value_buf_[16];
};
static Iterator* GetFileIterator(void* arg,
const ReadOptions& options,
static Iterator* GetFileIterator(void* arg, const ReadOptions& options,
const Slice& file_value) {
TableCache* cache = reinterpret_cast<TableCache*>(arg);
if (file_value.size() != 16) {
return NewErrorIterator(
Status::Corruption("FileReader invoked with unexpected value"));
} else {
return cache->NewIterator(options,
DecodeFixed64(file_value.data()),
return cache->NewIterator(options, DecodeFixed64(file_value.data()),
DecodeFixed64(file_value.data() + 8));
}
}
@ -227,17 +222,16 @@ static Iterator* GetFileIterator(void* arg,
Iterator* Version::NewConcatenatingIterator(const ReadOptions& options,
int level) const {
return NewTwoLevelIterator(
new LevelFileNumIterator(vset_->icmp_, &files_[level]),
&GetFileIterator, vset_->table_cache_, options);
new LevelFileNumIterator(vset_->icmp_, &files_[level]), &GetFileIterator,
vset_->table_cache_, options);
}
void Version::AddIterators(const ReadOptions& options,
std::vector<Iterator*>* iters) {
// Merge all level zero files together since they may overlap
for (size_t i = 0; i < files_[0].size(); i++) {
iters->push_back(
vset_->table_cache_->NewIterator(
options, files_[0][i]->number, files_[0][i]->file_size));
iters->push_back(vset_->table_cache_->NewIterator(
options, files_[0][i]->number, files_[0][i]->file_size));
}
// For levels > 0, we can use a concatenating iterator that sequentially
@ -264,7 +258,7 @@ struct Saver {
Slice user_key;
std::string* value;
};
}
} // namespace
static void SaveValue(void* arg, const Slice& ikey, const Slice& v) {
Saver* s = reinterpret_cast<Saver*>(arg);
ParsedInternalKey parsed_key;
@ -284,10 +278,8 @@ static bool NewestFirst(FileMetaData* a, FileMetaData* b) {
return a->number > b->number;
}
void Version::ForEachOverlapping(Slice user_key, Slice internal_key,
void* arg,
void Version::ForEachOverlapping(Slice user_key, Slice internal_key, void* arg,
bool (*func)(void*, int, FileMetaData*)) {
// TODO(sanjay): Change Version::Get() to use this function.
const Comparator* ucmp = vset_->icmp_.user_comparator();
// Search level-0 in order from newest to oldest.
@ -329,103 +321,82 @@ void Version::ForEachOverlapping(Slice user_key, Slice internal_key,
}
}
Status Version::Get(const ReadOptions& options,
const LookupKey& k,
std::string* value,
GetStats* stats) {
Slice ikey = k.internal_key();
Slice user_key = k.user_key();
const Comparator* ucmp = vset_->icmp_.user_comparator();
Status s;
Status Version::Get(const ReadOptions& options, const LookupKey& k,
std::string* value, GetStats* stats) {
stats->seek_file = nullptr;
stats->seek_file_level = -1;
FileMetaData* last_file_read = nullptr;
int last_file_read_level = -1;
// We can search level-by-level since entries never hop across
// levels. Therefore we are guaranteed that if we find data
// in an smaller level, later levels are irrelevant.
std::vector<FileMetaData*> tmp;
FileMetaData* tmp2;
for (int level = 0; level < config::kNumLevels; level++) {
size_t num_files = files_[level].size();
if (num_files == 0) continue;
struct State {
Saver saver;
GetStats* stats;
const ReadOptions* options;
Slice ikey;
FileMetaData* last_file_read;
int last_file_read_level;
// Get the list of files to search in this level
FileMetaData* const* files = &files_[level][0];
if (level == 0) {
// Level-0 files may overlap each other. Find all files that
// overlap user_key and process them in order from newest to oldest.
tmp.reserve(num_files);
for (uint32_t i = 0; i < num_files; i++) {
FileMetaData* f = files[i];
if (ucmp->Compare(user_key, f->smallest.user_key()) >= 0 &&
ucmp->Compare(user_key, f->largest.user_key()) <= 0) {
tmp.push_back(f);
}
}
if (tmp.empty()) continue;
VersionSet* vset;
Status s;
bool found;
std::sort(tmp.begin(), tmp.end(), NewestFirst);
files = &tmp[0];
num_files = tmp.size();
} else {
// Binary search to find earliest index whose largest key >= ikey.
uint32_t index = FindFile(vset_->icmp_, files_[level], ikey);
if (index >= num_files) {
files = nullptr;
num_files = 0;
} else {
tmp2 = files[index];
if (ucmp->Compare(user_key, tmp2->smallest.user_key()) < 0) {
// All of "tmp2" is past any data for user_key
files = nullptr;
num_files = 0;
} else {
files = &tmp2;
num_files = 1;
}
}
}
static bool Match(void* arg, int level, FileMetaData* f) {
State* state = reinterpret_cast<State*>(arg);
for (uint32_t i = 0; i < num_files; ++i) {
if (last_file_read != nullptr && stats->seek_file == nullptr) {
if (state->stats->seek_file == nullptr &&
state->last_file_read != nullptr) {
// We have had more than one seek for this read. Charge the 1st file.
stats->seek_file = last_file_read;
stats->seek_file_level = last_file_read_level;
state->stats->seek_file = state->last_file_read;
state->stats->seek_file_level = state->last_file_read_level;
}
FileMetaData* f = files[i];
last_file_read = f;
last_file_read_level = level;
state->last_file_read = f;
state->last_file_read_level = level;
Saver saver;
saver.state = kNotFound;
saver.ucmp = ucmp;
saver.user_key = user_key;
saver.value = value;
s = vset_->table_cache_->Get(options, f->number, f->file_size,
ikey, &saver, SaveValue);
if (!s.ok()) {
return s;
state->s = state->vset->table_cache_->Get(*state->options, f->number,
f->file_size, state->ikey,
&state->saver, SaveValue);
if (!state->s.ok()) {
state->found = true;
return false;
}
switch (saver.state) {
switch (state->saver.state) {
case kNotFound:
break; // Keep searching in other files
return true; // Keep searching in other files
case kFound:
return s;
state->found = true;
return false;
case kDeleted:
s = Status::NotFound(Slice()); // Use empty error message for speed
return s;
return false;
case kCorrupt:
s = Status::Corruption("corrupted key for ", user_key);
return s;
state->s =
Status::Corruption("corrupted key for ", state->saver.user_key);
state->found = true;
return false;
}
}
}
return Status::NotFound(Slice()); // Use an empty error message for speed
// Not reached. Added to avoid false compilation warnings of
// "control reaches end of non-void function".
return false;
}
};
State state;
state.found = false;
state.stats = stats;
state.last_file_read = nullptr;
state.last_file_read_level = -1;
state.options = &options;
state.ikey = k.internal_key();
state.vset = vset_;
state.saver.state = kNotFound;
state.saver.ucmp = vset_->icmp_.user_comparator();
state.saver.user_key = k.user_key();
state.saver.value = value;
ForEachOverlapping(state.saver.user_key, state.ikey, &state, &State::Match);
return state.found ? state.s : Status::NotFound(Slice());
}
bool Version::UpdateStats(const GetStats& stats) {
@ -479,9 +450,7 @@ bool Version::RecordReadSample(Slice internal_key) {
return false;
}
void Version::Ref() {
++refs_;
}
void Version::Ref() { ++refs_; }
void Version::Unref() {
assert(this != &vset_->dummy_versions_);
@ -492,16 +461,14 @@ void Version::Unref() {
}
}
bool Version::OverlapInLevel(int level,
const Slice* smallest_user_key,
bool Version::OverlapInLevel(int level, const Slice* smallest_user_key,
const Slice* largest_user_key) {
return SomeFileOverlapsRange(vset_->icmp_, (level > 0), files_[level],
smallest_user_key, largest_user_key);
}
int Version::PickLevelForMemTableOutput(
const Slice& smallest_user_key,
const Slice& largest_user_key) {
int Version::PickLevelForMemTableOutput(const Slice& smallest_user_key,
const Slice& largest_user_key) {
int level = 0;
if (!OverlapInLevel(0, &smallest_user_key, &largest_user_key)) {
// Push to next level if there is no overlap in next level,
@ -528,11 +495,9 @@ int Version::PickLevelForMemTableOutput(
}
// Store in "*inputs" all files in "level" that overlap [begin,end]
void Version::GetOverlappingInputs(
int level,
const InternalKey* begin,
const InternalKey* end,
std::vector<FileMetaData*>* inputs) {
void Version::GetOverlappingInputs(int level, const InternalKey* begin,
const InternalKey* end,
std::vector<FileMetaData*>* inputs) {
assert(level >= 0);
assert(level < config::kNumLevels);
inputs->clear();
@ -544,7 +509,7 @@ void Version::GetOverlappingInputs(
user_end = end->user_key();
}
const Comparator* user_cmp = vset_->icmp_.user_comparator();
for (size_t i = 0; i < files_[level].size(); ) {
for (size_t i = 0; i < files_[level].size();) {
FileMetaData* f = files_[level][i++];
const Slice file_start = f->smallest.user_key();
const Slice file_limit = f->largest.user_key();
@ -561,8 +526,8 @@ void Version::GetOverlappingInputs(
user_begin = file_start;
inputs->clear();
i = 0;
} else if (end != nullptr && user_cmp->Compare(file_limit,
user_end) > 0) {
} else if (end != nullptr &&
user_cmp->Compare(file_limit, user_end) > 0) {
user_end = file_limit;
inputs->clear();
i = 0;
@ -630,9 +595,7 @@ class VersionSet::Builder {
public:
// Initialize a builder with the files from *base and other info from *vset
Builder(VersionSet* vset, Version* base)
: vset_(vset),
base_(base) {
Builder(VersionSet* vset, Version* base) : vset_(vset), base_(base) {
base_->Ref();
BySmallestKey cmp;
cmp.internal_comparator = &vset_->icmp_;
@ -646,8 +609,8 @@ class VersionSet::Builder {
const FileSet* added = levels_[level].added_files;
std::vector<FileMetaData*> to_unref;
to_unref.reserve(added->size());
for (FileSet::const_iterator it = added->begin();
it != added->end(); ++it) {
for (FileSet::const_iterator it = added->begin(); it != added->end();
++it) {
to_unref.push_back(*it);
}
delete added;
@ -663,7 +626,7 @@ class VersionSet::Builder {
}
// Apply all of the edits in *edit to the current state.
void Apply(VersionEdit* edit) {
void Apply(const VersionEdit* edit) {
// Update compaction pointers
for (size_t i = 0; i < edit->compact_pointers_.size(); i++) {
const int level = edit->compact_pointers_[i].first;
@ -672,12 +635,9 @@ class VersionSet::Builder {
}
// Delete files
const VersionEdit::DeletedFileSet& del = edit->deleted_files_;
for (VersionEdit::DeletedFileSet::const_iterator iter = del.begin();
iter != del.end();
++iter) {
const int level = iter->first;
const uint64_t number = iter->second;
for (const auto& deleted_file_set_kvp : edit->deleted_files_) {
const int level = deleted_file_set_kvp.first;
const uint64_t number = deleted_file_set_kvp.second;
levels_[level].deleted_files.insert(number);
}
@ -700,7 +660,7 @@ class VersionSet::Builder {
// same as the compaction of 40KB of data. We are a little
// conservative and allow approximately one seek for every 16KB
// of data before triggering a compaction.
f->allowed_seeks = (f->file_size / 16384);
f->allowed_seeks = static_cast<int>((f->file_size / 16384U));
if (f->allowed_seeks < 100) f->allowed_seeks = 100;
levels_[level].deleted_files.erase(f->number);
@ -718,20 +678,17 @@ class VersionSet::Builder {
const std::vector<FileMetaData*>& base_files = base_->files_[level];
std::vector<FileMetaData*>::const_iterator base_iter = base_files.begin();
std::vector<FileMetaData*>::const_iterator base_end = base_files.end();
const FileSet* added = levels_[level].added_files;
v->files_[level].reserve(base_files.size() + added->size());
for (FileSet::const_iterator added_iter = added->begin();
added_iter != added->end();
++added_iter) {
const FileSet* added_files = levels_[level].added_files;
v->files_[level].reserve(base_files.size() + added_files->size());
for (const auto& added_file : *added_files) {
// Add all smaller files listed in base_
for (std::vector<FileMetaData*>::const_iterator bpos
= std::upper_bound(base_iter, base_end, *added_iter, cmp);
base_iter != bpos;
++base_iter) {
for (std::vector<FileMetaData*>::const_iterator bpos =
std::upper_bound(base_iter, base_end, added_file, cmp);
base_iter != bpos; ++base_iter) {
MaybeAddFile(v, level, *base_iter);
}
MaybeAddFile(v, level, *added_iter);
MaybeAddFile(v, level, added_file);
}
// Add remaining base files
@ -743,13 +700,13 @@ class VersionSet::Builder {
// Make sure there is no overlap in levels > 0
if (level > 0) {
for (uint32_t i = 1; i < v->files_[level].size(); i++) {
const InternalKey& prev_end = v->files_[level][i-1]->largest;
const InternalKey& prev_end = v->files_[level][i - 1]->largest;
const InternalKey& this_begin = v->files_[level][i]->smallest;
if (vset_->icmp_.Compare(prev_end, this_begin) >= 0) {
fprintf(stderr, "overlapping ranges in same level %s vs. %s\n",
prev_end.DebugString().c_str(),
this_begin.DebugString().c_str());
abort();
std::fprintf(stderr, "overlapping ranges in same level %s vs. %s\n",
prev_end.DebugString().c_str(),
this_begin.DebugString().c_str());
std::abort();
}
}
}
@ -764,7 +721,7 @@ class VersionSet::Builder {
std::vector<FileMetaData*>* files = &v->files_[level];
if (level > 0 && !files->empty()) {
// Must not overlap
assert(vset_->icmp_.Compare((*files)[files->size()-1]->largest,
assert(vset_->icmp_.Compare((*files)[files->size() - 1]->largest,
f->smallest) < 0);
}
f->refs++;
@ -773,8 +730,7 @@ class VersionSet::Builder {
}
};
VersionSet::VersionSet(const std::string& dbname,
const Options* options,
VersionSet::VersionSet(const std::string& dbname, const Options* options,
TableCache* table_cache,
const InternalKeyComparator* cmp)
: env_(options->env),
@ -850,7 +806,6 @@ Status VersionSet::LogAndApply(VersionEdit* edit, port::Mutex* mu) {
// first call to LogAndApply (when opening the database).
assert(descriptor_file_ == nullptr);
new_manifest_file = DescriptorFileName(dbname_, manifest_file_number_);
edit->SetNextFile(next_file_number_);
s = env_->NewWritableFile(new_manifest_file, &descriptor_file_);
if (s.ok()) {
descriptor_log_ = new log::Writer(descriptor_file_);
@ -896,17 +851,17 @@ Status VersionSet::LogAndApply(VersionEdit* edit, port::Mutex* mu) {
delete descriptor_file_;
descriptor_log_ = nullptr;
descriptor_file_ = nullptr;
env_->DeleteFile(new_manifest_file);
env_->RemoveFile(new_manifest_file);
}
}
return s;
}
Status VersionSet::Recover(bool *save_manifest) {
Status VersionSet::Recover(bool* save_manifest) {
struct LogReporter : public log::Reader::Reporter {
Status* status;
virtual void Corruption(size_t bytes, const Status& s) {
void Corruption(size_t bytes, const Status& s) override {
if (this->status->ok()) *this->status = s;
}
};
@ -917,7 +872,7 @@ Status VersionSet::Recover(bool *save_manifest) {
if (!s.ok()) {
return s;
}
if (current.empty() || current[current.size()-1] != '\n') {
if (current.empty() || current[current.size() - 1] != '\n') {
return Status::Corruption("CURRENT file does not end with newline");
}
current.resize(current.size() - 1);
@ -927,8 +882,8 @@ Status VersionSet::Recover(bool *save_manifest) {
s = env_->NewSequentialFile(dscname, &file);
if (!s.ok()) {
if (s.IsNotFound()) {
return Status::Corruption(
"CURRENT points to a non-existent file", s.ToString());
return Status::Corruption("CURRENT points to a non-existent file",
s.ToString());
}
return s;
}
@ -942,14 +897,17 @@ Status VersionSet::Recover(bool *save_manifest) {
uint64_t log_number = 0;
uint64_t prev_log_number = 0;
Builder builder(this, current_);
int read_records = 0;
{
LogReporter reporter;
reporter.status = &s;
log::Reader reader(file, &reporter, true/*checksum*/, 0/*initial_offset*/);
log::Reader reader(file, &reporter, true /*checksum*/,
0 /*initial_offset*/);
Slice record;
std::string scratch;
while (reader.ReadRecord(&record, &scratch) && s.ok()) {
++read_records;
VersionEdit edit;
s = edit.DecodeFrom(record);
if (s.ok()) {
@ -1024,6 +982,10 @@ Status VersionSet::Recover(bool *save_manifest) {
} else {
*save_manifest = true;
}
} else {
std::string error = s.ToString();
Log(options_->info_log, "Error recovering version set with %d records: %s",
read_records, error.c_str());
}
return s;
@ -1071,7 +1033,7 @@ void VersionSet::Finalize(Version* v) {
int best_level = -1;
double best_score = -1;
for (int level = 0; level < config::kNumLevels-1; level++) {
for (int level = 0; level < config::kNumLevels - 1; level++) {
double score;
if (level == 0) {
// We treat level-0 specially by bounding the number of files
@ -1086,7 +1048,7 @@ void VersionSet::Finalize(Version* v) {
// setting, or very high compression ratios, or lots of
// overwrites/deletions).
score = v->files_[level].size() /
static_cast<double>(config::kL0_CompactionTrigger);
static_cast<double>(config::kL0_CompactionTrigger);
} else {
// Compute the ratio of current size to size limit.
const uint64_t level_bytes = TotalFileSize(v->files_[level]);
@ -1142,16 +1104,13 @@ int VersionSet::NumLevelFiles(int level) const {
const char* VersionSet::LevelSummary(LevelSummaryStorage* scratch) const {
// Update code if kNumLevels changes
assert(config::kNumLevels == 7);
snprintf(scratch->buffer, sizeof(scratch->buffer),
"files[ %d %d %d %d %d %d %d ]",
int(current_->files_[0].size()),
int(current_->files_[1].size()),
int(current_->files_[2].size()),
int(current_->files_[3].size()),
int(current_->files_[4].size()),
int(current_->files_[5].size()),
int(current_->files_[6].size()));
static_assert(config::kNumLevels == 7, "");
std::snprintf(
scratch->buffer, sizeof(scratch->buffer), "files[ %d %d %d %d %d %d %d ]",
int(current_->files_[0].size()), int(current_->files_[1].size()),
int(current_->files_[2].size()), int(current_->files_[3].size()),
int(current_->files_[4].size()), int(current_->files_[5].size()),
int(current_->files_[6].size()));
return scratch->buffer;
}
@ -1188,8 +1147,7 @@ uint64_t VersionSet::ApproximateOffsetOf(Version* v, const InternalKey& ikey) {
}
void VersionSet::AddLiveFiles(std::set<uint64_t>* live) {
for (Version* v = dummy_versions_.next_;
v != &dummy_versions_;
for (Version* v = dummy_versions_.next_; v != &dummy_versions_;
v = v->next_) {
for (int level = 0; level < config::kNumLevels; level++) {
const std::vector<FileMetaData*>& files = v->files_[level];
@ -1212,7 +1170,7 @@ int64_t VersionSet::MaxNextLevelOverlappingBytes() {
for (int level = 1; level < config::kNumLevels - 1; level++) {
for (size_t i = 0; i < current_->files_[level].size(); i++) {
const FileMetaData* f = current_->files_[level][i];
current_->GetOverlappingInputs(level+1, &f->smallest, &f->largest,
current_->GetOverlappingInputs(level + 1, &f->smallest, &f->largest,
&overlaps);
const int64_t sum = TotalFileSize(overlaps);
if (sum > result) {
@ -1227,8 +1185,7 @@ int64_t VersionSet::MaxNextLevelOverlappingBytes() {
// *smallest, *largest.
// REQUIRES: inputs is not empty
void VersionSet::GetRange(const std::vector<FileMetaData*>& inputs,
InternalKey* smallest,
InternalKey* largest) {
InternalKey* smallest, InternalKey* largest) {
assert(!inputs.empty());
smallest->Clear();
largest->Clear();
@ -1253,8 +1210,7 @@ void VersionSet::GetRange(const std::vector<FileMetaData*>& inputs,
// REQUIRES: inputs is not empty
void VersionSet::GetRange2(const std::vector<FileMetaData*>& inputs1,
const std::vector<FileMetaData*>& inputs2,
InternalKey* smallest,
InternalKey* largest) {
InternalKey* smallest, InternalKey* largest) {
std::vector<FileMetaData*> all = inputs1;
all.insert(all.end(), inputs2.begin(), inputs2.end());
GetRange(all, smallest, largest);
@ -1276,8 +1232,8 @@ Iterator* VersionSet::MakeInputIterator(Compaction* c) {
if (c->level() + which == 0) {
const std::vector<FileMetaData*>& files = c->inputs_[which];
for (size_t i = 0; i < files.size(); i++) {
list[num++] = table_cache_->NewIterator(
options, files[i]->number, files[i]->file_size);
list[num++] = table_cache_->NewIterator(options, files[i]->number,
files[i]->file_size);
}
} else {
// Create concatenating iterator for the files from this level
@ -1304,7 +1260,7 @@ Compaction* VersionSet::PickCompaction() {
if (size_compaction) {
level = current_->compaction_level_;
assert(level >= 0);
assert(level+1 < config::kNumLevels);
assert(level + 1 < config::kNumLevels);
c = new Compaction(options_, level);
// Pick the first file that comes after compact_pointer_[level]
@ -1347,12 +1303,95 @@ Compaction* VersionSet::PickCompaction() {
return c;
}
// Finds the largest key in a vector of files. Returns true if files is not
// empty.
bool FindLargestKey(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& files,
InternalKey* largest_key) {
if (files.empty()) {
return false;
}
*largest_key = files[0]->largest;
for (size_t i = 1; i < files.size(); ++i) {
FileMetaData* f = files[i];
if (icmp.Compare(f->largest, *largest_key) > 0) {
*largest_key = f->largest;
}
}
return true;
}
// Finds minimum file b2=(l2, u2) in level file for which l2 > u1 and
// user_key(l2) = user_key(u1)
FileMetaData* FindSmallestBoundaryFile(
const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& level_files,
const InternalKey& largest_key) {
const Comparator* user_cmp = icmp.user_comparator();
FileMetaData* smallest_boundary_file = nullptr;
for (size_t i = 0; i < level_files.size(); ++i) {
FileMetaData* f = level_files[i];
if (icmp.Compare(f->smallest, largest_key) > 0 &&
user_cmp->Compare(f->smallest.user_key(), largest_key.user_key()) ==
0) {
if (smallest_boundary_file == nullptr ||
icmp.Compare(f->smallest, smallest_boundary_file->smallest) < 0) {
smallest_boundary_file = f;
}
}
}
return smallest_boundary_file;
}
// Extracts the largest file b1 from |compaction_files| and then searches for a
// b2 in |level_files| for which user_key(u1) = user_key(l2). If it finds such a
// file b2 (known as a boundary file) it adds it to |compaction_files| and then
// searches again using this new upper bound.
//
// If there are two blocks, b1=(l1, u1) and b2=(l2, u2) and
// user_key(u1) = user_key(l2), and if we compact b1 but not b2 then a
// subsequent get operation will yield an incorrect result because it will
// return the record from b2 in level i rather than from b1 because it searches
// level by level for records matching the supplied user key.
//
// parameters:
// in level_files: List of files to search for boundary files.
// in/out compaction_files: List of files to extend by adding boundary files.
void AddBoundaryInputs(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& level_files,
std::vector<FileMetaData*>* compaction_files) {
InternalKey largest_key;
// Quick return if compaction_files is empty.
if (!FindLargestKey(icmp, *compaction_files, &largest_key)) {
return;
}
bool continue_searching = true;
while (continue_searching) {
FileMetaData* smallest_boundary_file =
FindSmallestBoundaryFile(icmp, level_files, largest_key);
// If a boundary file was found advance largest_key, otherwise we're done.
if (smallest_boundary_file != NULL) {
compaction_files->push_back(smallest_boundary_file);
largest_key = smallest_boundary_file->largest;
} else {
continue_searching = false;
}
}
}
void VersionSet::SetupOtherInputs(Compaction* c) {
const int level = c->level();
InternalKey smallest, largest;
AddBoundaryInputs(icmp_, current_->files_[level], &c->inputs_[0]);
GetRange(c->inputs_[0], &smallest, &largest);
current_->GetOverlappingInputs(level+1, &smallest, &largest, &c->inputs_[1]);
current_->GetOverlappingInputs(level + 1, &smallest, &largest,
&c->inputs_[1]);
AddBoundaryInputs(icmp_, current_->files_[level + 1], &c->inputs_[1]);
// Get entire range covered by compaction
InternalKey all_start, all_limit;
@ -1363,6 +1402,7 @@ void VersionSet::SetupOtherInputs(Compaction* c) {
if (!c->inputs_[1].empty()) {
std::vector<FileMetaData*> expanded0;
current_->GetOverlappingInputs(level, &all_start, &all_limit, &expanded0);
AddBoundaryInputs(icmp_, current_->files_[level], &expanded0);
const int64_t inputs0_size = TotalFileSize(c->inputs_[0]);
const int64_t inputs1_size = TotalFileSize(c->inputs_[1]);
const int64_t expanded0_size = TotalFileSize(expanded0);
@ -1372,18 +1412,15 @@ void VersionSet::SetupOtherInputs(Compaction* c) {
InternalKey new_start, new_limit;
GetRange(expanded0, &new_start, &new_limit);
std::vector<FileMetaData*> expanded1;
current_->GetOverlappingInputs(level+1, &new_start, &new_limit,
current_->GetOverlappingInputs(level + 1, &new_start, &new_limit,
&expanded1);
AddBoundaryInputs(icmp_, current_->files_[level + 1], &expanded1);
if (expanded1.size() == c->inputs_[1].size()) {
Log(options_->info_log,
"Expanding@%d %d+%d (%ld+%ld bytes) to %d+%d (%ld+%ld bytes)\n",
level,
int(c->inputs_[0].size()),
int(c->inputs_[1].size()),
long(inputs0_size), long(inputs1_size),
int(expanded0.size()),
int(expanded1.size()),
long(expanded0_size), long(inputs1_size));
level, int(c->inputs_[0].size()), int(c->inputs_[1].size()),
long(inputs0_size), long(inputs1_size), int(expanded0.size()),
int(expanded1.size()), long(expanded0_size), long(inputs1_size));
smallest = new_start;
largest = new_limit;
c->inputs_[0] = expanded0;
@ -1408,10 +1445,8 @@ void VersionSet::SetupOtherInputs(Compaction* c) {
c->edit_.SetCompactPointer(level, largest);
}
Compaction* VersionSet::CompactRange(
int level,
const InternalKey* begin,
const InternalKey* end) {
Compaction* VersionSet::CompactRange(int level, const InternalKey* begin,
const InternalKey* end) {
std::vector<FileMetaData*> inputs;
current_->GetOverlappingInputs(level, begin, end, &inputs);
if (inputs.empty()) {
@ -1474,7 +1509,7 @@ bool Compaction::IsTrivialMove() const {
void Compaction::AddInputDeletions(VersionEdit* edit) {
for (int which = 0; which < 2; which++) {
for (size_t i = 0; i < inputs_[which].size(); i++) {
edit->DeleteFile(level_ + which, inputs_[which][i]->number);
edit->RemoveFile(level_ + which, inputs_[which][i]->number);
}
}
}
@ -1484,7 +1519,7 @@ bool Compaction::IsBaseLevelForKey(const Slice& user_key) {
const Comparator* user_cmp = input_version_->vset_->icmp_.user_comparator();
for (int lvl = level_ + 2; lvl < config::kNumLevels; lvl++) {
const std::vector<FileMetaData*>& files = input_version_->files_[lvl];
for (; level_ptrs_[lvl] < files.size(); ) {
while (level_ptrs_[lvl] < files.size()) {
FileMetaData* f = files[level_ptrs_[lvl]];
if (user_cmp->Compare(user_key, f->largest.user_key()) <= 0) {
// We've advanced far enough
@ -1505,8 +1540,9 @@ bool Compaction::ShouldStopBefore(const Slice& internal_key) {
// Scan to find earliest grandparent file that contains key.
const InternalKeyComparator* icmp = &vset->icmp_;
while (grandparent_index_ < grandparents_.size() &&
icmp->Compare(internal_key,
grandparents_[grandparent_index_]->largest.Encode()) > 0) {
icmp->Compare(internal_key,
grandparents_[grandparent_index_]->largest.Encode()) >
0) {
if (seen_key_) {
overlapped_bytes_ += grandparents_[grandparent_index_]->file_size;
}

96
db/version_set.h
View File

@ -18,6 +18,7 @@
#include <map>
#include <set>
#include <vector>
#include "db/dbformat.h"
#include "db/version_edit.h"
#include "port/port.h"
@ -25,7 +26,9 @@
namespace leveldb {
namespace log { class Writer; }
namespace log {
class Writer;
}
class Compaction;
class Iterator;
@ -40,8 +43,7 @@ class WritableFile;
// Return files.size() if there is no such file.
// REQUIRES: "files" contains a sorted list of non-overlapping files.
int FindFile(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& files,
const Slice& key);
const std::vector<FileMetaData*>& files, const Slice& key);
// Returns true iff some file in "files" overlaps the user key range
// [*smallest,*largest].
@ -57,6 +59,11 @@ bool SomeFileOverlapsRange(const InternalKeyComparator& icmp,
class Version {
public:
struct GetStats {
FileMetaData* seek_file;
int seek_file_level;
};
// Append to *iters a sequence of iterators that will
// yield the contents of this Version when merged together.
// REQUIRES: This version has been saved (see VersionSet::SaveTo)
@ -65,10 +72,6 @@ class Version {
// Lookup the value for key. If found, store it in *val and
// return OK. Else return a non-OK status. Fills *stats.
// REQUIRES: lock is not held
struct GetStats {
FileMetaData* seek_file;
int seek_file_level;
};
Status Get(const ReadOptions&, const LookupKey& key, std::string* val,
GetStats* stats);
@ -90,16 +93,15 @@ class Version {
void GetOverlappingInputs(
int level,
const InternalKey* begin, // nullptr means before all keys
const InternalKey* end, // nullptr means after all keys
const InternalKey* begin, // nullptr means before all keys
const InternalKey* end, // nullptr means after all keys
std::vector<FileMetaData*>* inputs);
// Returns true iff some file in the specified level overlaps
// some part of [*smallest_user_key,*largest_user_key].
// smallest_user_key==nullptr represents a key smaller than all the DB's keys.
// largest_user_key==nullptr represents a key largest than all the DB's keys.
bool OverlapInLevel(int level,
const Slice* smallest_user_key,
bool OverlapInLevel(int level, const Slice* smallest_user_key,
const Slice* largest_user_key);
// Return the level at which we should place a new memtable compaction
@ -117,6 +119,22 @@ class Version {
friend class VersionSet;
class LevelFileNumIterator;
explicit Version(VersionSet* vset)
: vset_(vset),
next_(this),
prev_(this),
refs_(0),
file_to_compact_(nullptr),
file_to_compact_level_(-1),
compaction_score_(-1),
compaction_level_(-1) {}
Version(const Version&) = delete;
Version& operator=(const Version&) = delete;
~Version();
Iterator* NewConcatenatingIterator(const ReadOptions&, int level) const;
// Call func(arg, level, f) for every file that overlaps user_key in
@ -124,14 +142,13 @@ class Version {
// false, makes no more calls.
//
// REQUIRES: user portion of internal_key == user_key.
void ForEachOverlapping(Slice user_key, Slice internal_key,
void* arg,
void ForEachOverlapping(Slice user_key, Slice internal_key, void* arg,
bool (*func)(void*, int, FileMetaData*));
VersionSet* vset_; // VersionSet to which this Version belongs
Version* next_; // Next version in linked list
Version* prev_; // Previous version in linked list
int refs_; // Number of live refs to this version
VersionSet* vset_; // VersionSet to which this Version belongs
Version* next_; // Next version in linked list
Version* prev_; // Previous version in linked list
int refs_; // Number of live refs to this version
// List of files per level
std::vector<FileMetaData*> files_[config::kNumLevels];
@ -145,28 +162,15 @@ class Version {
// are initialized by Finalize().
double compaction_score_;
int compaction_level_;
explicit Version(VersionSet* vset)
: vset_(vset), next_(this), prev_(this), refs_(0),
file_to_compact_(nullptr),
file_to_compact_level_(-1),
compaction_score_(-1),
compaction_level_(-1) {
}
~Version();
// No copying allowed
Version(const Version&);
void operator=(const Version&);
};
class VersionSet {
public:
VersionSet(const std::string& dbname,
const Options* options,
TableCache* table_cache,
const InternalKeyComparator*);
VersionSet(const std::string& dbname, const Options* options,
TableCache* table_cache, const InternalKeyComparator*);
VersionSet(const VersionSet&) = delete;
VersionSet& operator=(const VersionSet&) = delete;
~VersionSet();
// Apply *edit to the current version to form a new descriptor that
@ -178,7 +182,7 @@ class VersionSet {
EXCLUSIVE_LOCKS_REQUIRED(mu);
// Recover the last saved descriptor from persistent storage.
Status Recover(bool *save_manifest);
Status Recover(bool* save_manifest);
// Return the current version.
Version* current() const { return current_; }
@ -233,10 +237,8 @@ class VersionSet {
// the specified level. Returns nullptr if there is nothing in that
// level that overlaps the specified range. Caller should delete
// the result.
Compaction* CompactRange(
int level,
const InternalKey* begin,
const InternalKey* end);
Compaction* CompactRange(int level, const InternalKey* begin,
const InternalKey* end);
// Return the maximum overlapping data (in bytes) at next level for any
// file at a level >= 1.
@ -277,14 +279,12 @@ class VersionSet {
void Finalize(Version* v);
void GetRange(const std::vector<FileMetaData*>& inputs,
InternalKey* smallest,
void GetRange(const std::vector<FileMetaData*>& inputs, InternalKey* smallest,
InternalKey* largest);
void GetRange2(const std::vector<FileMetaData*>& inputs1,
const std::vector<FileMetaData*>& inputs2,
InternalKey* smallest,
InternalKey* largest);
InternalKey* smallest, InternalKey* largest);
void SetupOtherInputs(Compaction* c);
@ -313,10 +313,6 @@ class VersionSet {
// Per-level key at which the next compaction at that level should start.
// Either an empty string, or a valid InternalKey.
std::string compact_pointer_[config::kNumLevels];
// No copying allowed
VersionSet(const VersionSet&);
void operator=(const VersionSet&);
};
// A Compaction encapsulates information about a compaction.
@ -373,9 +369,9 @@ class Compaction {
VersionEdit edit_;
// Each compaction reads inputs from "level_" and "level_+1"
std::vector<FileMetaData*> inputs_[2]; // The two sets of inputs
std::vector<FileMetaData*> inputs_[2]; // The two sets of inputs
// State used to check for number of of overlapping grandparent files
// State used to check for number of overlapping grandparent files
// (parent == level_ + 1, grandparent == level_ + 2)
std::vector<FileMetaData*> grandparents_;
size_t grandparent_index_; // Index in grandparent_starts_

220
db/version_set_test.cc
View File

@ -3,18 +3,16 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
#include "db/version_set.h"
#include "gtest/gtest.h"
#include "util/logging.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
class FindFileTest {
class FindFileTest : public testing::Test {
public:
std::vector<FileMetaData*> files_;
bool disjoint_sorted_files_;
FindFileTest() : disjoint_sorted_files_(true) { }
FindFileTest() : disjoint_sorted_files_(true) {}
~FindFileTest() {
for (int i = 0; i < files_.size(); i++) {
@ -46,17 +44,22 @@ class FindFileTest {
(smallest != nullptr ? &s : nullptr),
(largest != nullptr ? &l : nullptr));
}
bool disjoint_sorted_files_;
private:
std::vector<FileMetaData*> files_;
};
TEST(FindFileTest, Empty) {
TEST_F(FindFileTest, Empty) {
ASSERT_EQ(0, Find("foo"));
ASSERT_TRUE(! Overlaps("a", "z"));
ASSERT_TRUE(! Overlaps(nullptr, "z"));
ASSERT_TRUE(! Overlaps("a", nullptr));
ASSERT_TRUE(! Overlaps(nullptr, nullptr));
ASSERT_TRUE(!Overlaps("a", "z"));
ASSERT_TRUE(!Overlaps(nullptr, "z"));
ASSERT_TRUE(!Overlaps("a", nullptr));
ASSERT_TRUE(!Overlaps(nullptr, nullptr));
}
TEST(FindFileTest, Single) {
TEST_F(FindFileTest, Single) {
Add("p", "q");
ASSERT_EQ(0, Find("a"));
ASSERT_EQ(0, Find("p"));
@ -65,8 +68,8 @@ TEST(FindFileTest, Single) {
ASSERT_EQ(1, Find("q1"));
ASSERT_EQ(1, Find("z"));
ASSERT_TRUE(! Overlaps("a", "b"));
ASSERT_TRUE(! Overlaps("z1", "z2"));
ASSERT_TRUE(!Overlaps("a", "b"));
ASSERT_TRUE(!Overlaps("z1", "z2"));
ASSERT_TRUE(Overlaps("a", "p"));
ASSERT_TRUE(Overlaps("a", "q"));
ASSERT_TRUE(Overlaps("a", "z"));
@ -78,16 +81,15 @@ TEST(FindFileTest, Single) {
ASSERT_TRUE(Overlaps("q", "q"));
ASSERT_TRUE(Overlaps("q", "q1"));
ASSERT_TRUE(! Overlaps(nullptr, "j"));
ASSERT_TRUE(! Overlaps("r", nullptr));
ASSERT_TRUE(!Overlaps(nullptr, "j"));
ASSERT_TRUE(!Overlaps("r", nullptr));
ASSERT_TRUE(Overlaps(nullptr, "p"));
ASSERT_TRUE(Overlaps(nullptr, "p1"));
ASSERT_TRUE(Overlaps("q", nullptr));
ASSERT_TRUE(Overlaps(nullptr, nullptr));
}
TEST(FindFileTest, Multiple) {
TEST_F(FindFileTest, Multiple) {
Add("150", "200");
Add("200", "250");
Add("300", "350");
@ -110,10 +112,10 @@ TEST(FindFileTest, Multiple) {
ASSERT_EQ(3, Find("450"));
ASSERT_EQ(4, Find("451"));
ASSERT_TRUE(! Overlaps("100", "149"));
ASSERT_TRUE(! Overlaps("251", "299"));
ASSERT_TRUE(! Overlaps("451", "500"));
ASSERT_TRUE(! Overlaps("351", "399"));
ASSERT_TRUE(!Overlaps("100", "149"));
ASSERT_TRUE(!Overlaps("251", "299"));
ASSERT_TRUE(!Overlaps("451", "500"));
ASSERT_TRUE(!Overlaps("351", "399"));
ASSERT_TRUE(Overlaps("100", "150"));
ASSERT_TRUE(Overlaps("100", "200"));
@ -125,13 +127,13 @@ TEST(FindFileTest, Multiple) {
ASSERT_TRUE(Overlaps("450", "500"));
}
TEST(FindFileTest, MultipleNullBoundaries) {
TEST_F(FindFileTest, MultipleNullBoundaries) {
Add("150", "200");
Add("200", "250");
Add("300", "350");
Add("400", "450");
ASSERT_TRUE(! Overlaps(nullptr, "149"));
ASSERT_TRUE(! Overlaps("451", nullptr));
ASSERT_TRUE(!Overlaps(nullptr, "149"));
ASSERT_TRUE(!Overlaps("451", nullptr));
ASSERT_TRUE(Overlaps(nullptr, nullptr));
ASSERT_TRUE(Overlaps(nullptr, "150"));
ASSERT_TRUE(Overlaps(nullptr, "199"));
@ -145,21 +147,21 @@ TEST(FindFileTest, MultipleNullBoundaries) {
ASSERT_TRUE(Overlaps("450", nullptr));
}
TEST(FindFileTest, OverlapSequenceChecks) {
TEST_F(FindFileTest, OverlapSequenceChecks) {
Add("200", "200", 5000, 3000);
ASSERT_TRUE(! Overlaps("199", "199"));
ASSERT_TRUE(! Overlaps("201", "300"));
ASSERT_TRUE(!Overlaps("199", "199"));
ASSERT_TRUE(!Overlaps("201", "300"));
ASSERT_TRUE(Overlaps("200", "200"));
ASSERT_TRUE(Overlaps("190", "200"));
ASSERT_TRUE(Overlaps("200", "210"));
}
TEST(FindFileTest, OverlappingFiles) {
TEST_F(FindFileTest, OverlappingFiles) {
Add("150", "600");
Add("400", "500");
disjoint_sorted_files_ = false;
ASSERT_TRUE(! Overlaps("100", "149"));
ASSERT_TRUE(! Overlaps("601", "700"));
ASSERT_TRUE(!Overlaps("100", "149"));
ASSERT_TRUE(!Overlaps("601", "700"));
ASSERT_TRUE(Overlaps("100", "150"));
ASSERT_TRUE(Overlaps("100", "200"));
ASSERT_TRUE(Overlaps("100", "300"));
@ -172,8 +174,158 @@ TEST(FindFileTest, OverlappingFiles) {
ASSERT_TRUE(Overlaps("600", "700"));
}
} // namespace leveldb
void AddBoundaryInputs(const InternalKeyComparator& icmp,
const std::vector<FileMetaData*>& level_files,
std::vector<FileMetaData*>* compaction_files);
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
class AddBoundaryInputsTest : public testing::Test {
public:
std::vector<FileMetaData*> level_files_;
std::vector<FileMetaData*> compaction_files_;
std::vector<FileMetaData*> all_files_;
InternalKeyComparator icmp_;
AddBoundaryInputsTest() : icmp_(BytewiseComparator()) {}
~AddBoundaryInputsTest() {
for (size_t i = 0; i < all_files_.size(); ++i) {
delete all_files_[i];
}
all_files_.clear();
}
FileMetaData* CreateFileMetaData(uint64_t number, InternalKey smallest,
InternalKey largest) {
FileMetaData* f = new FileMetaData();
f->number = number;
f->smallest = smallest;
f->largest = largest;
all_files_.push_back(f);
return f;
}
};
TEST_F(AddBoundaryInputsTest, TestEmptyFileSets) {
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_TRUE(compaction_files_.empty());
ASSERT_TRUE(level_files_.empty());
}
TEST_F(AddBoundaryInputsTest, TestEmptyLevelFiles) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 2, kTypeValue),
InternalKey(InternalKey("100", 1, kTypeValue)));
compaction_files_.push_back(f1);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_EQ(1, compaction_files_.size());
ASSERT_EQ(f1, compaction_files_[0]);
ASSERT_TRUE(level_files_.empty());
}
TEST_F(AddBoundaryInputsTest, TestEmptyCompactionFiles) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 2, kTypeValue),
InternalKey(InternalKey("100", 1, kTypeValue)));
level_files_.push_back(f1);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_TRUE(compaction_files_.empty());
ASSERT_EQ(1, level_files_.size());
ASSERT_EQ(f1, level_files_[0]);
}
TEST_F(AddBoundaryInputsTest, TestNoBoundaryFiles) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 2, kTypeValue),
InternalKey(InternalKey("100", 1, kTypeValue)));
FileMetaData* f2 =
CreateFileMetaData(1, InternalKey("200", 2, kTypeValue),
InternalKey(InternalKey("200", 1, kTypeValue)));
FileMetaData* f3 =
CreateFileMetaData(1, InternalKey("300", 2, kTypeValue),
InternalKey(InternalKey("300", 1, kTypeValue)));
level_files_.push_back(f3);
level_files_.push_back(f2);
level_files_.push_back(f1);
compaction_files_.push_back(f2);
compaction_files_.push_back(f3);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_EQ(2, compaction_files_.size());
}
TEST_F(AddBoundaryInputsTest, TestOneBoundaryFiles) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 3, kTypeValue),
InternalKey(InternalKey("100", 2, kTypeValue)));
FileMetaData* f2 =
CreateFileMetaData(1, InternalKey("100", 1, kTypeValue),
InternalKey(InternalKey("200", 3, kTypeValue)));
FileMetaData* f3 =
CreateFileMetaData(1, InternalKey("300", 2, kTypeValue),
InternalKey(InternalKey("300", 1, kTypeValue)));
level_files_.push_back(f3);
level_files_.push_back(f2);
level_files_.push_back(f1);
compaction_files_.push_back(f1);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_EQ(2, compaction_files_.size());
ASSERT_EQ(f1, compaction_files_[0]);
ASSERT_EQ(f2, compaction_files_[1]);
}
TEST_F(AddBoundaryInputsTest, TestTwoBoundaryFiles) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 6, kTypeValue),
InternalKey(InternalKey("100", 5, kTypeValue)));
FileMetaData* f2 =
CreateFileMetaData(1, InternalKey("100", 2, kTypeValue),
InternalKey(InternalKey("300", 1, kTypeValue)));
FileMetaData* f3 =
CreateFileMetaData(1, InternalKey("100", 4, kTypeValue),
InternalKey(InternalKey("100", 3, kTypeValue)));
level_files_.push_back(f2);
level_files_.push_back(f3);
level_files_.push_back(f1);
compaction_files_.push_back(f1);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_EQ(3, compaction_files_.size());
ASSERT_EQ(f1, compaction_files_[0]);
ASSERT_EQ(f3, compaction_files_[1]);
ASSERT_EQ(f2, compaction_files_[2]);
}
TEST_F(AddBoundaryInputsTest, TestDisjoinFilePointers) {
FileMetaData* f1 =
CreateFileMetaData(1, InternalKey("100", 6, kTypeValue),
InternalKey(InternalKey("100", 5, kTypeValue)));
FileMetaData* f2 =
CreateFileMetaData(1, InternalKey("100", 6, kTypeValue),
InternalKey(InternalKey("100", 5, kTypeValue)));
FileMetaData* f3 =
CreateFileMetaData(1, InternalKey("100", 2, kTypeValue),
InternalKey(InternalKey("300", 1, kTypeValue)));
FileMetaData* f4 =
CreateFileMetaData(1, InternalKey("100", 4, kTypeValue),
InternalKey(InternalKey("100", 3, kTypeValue)));
level_files_.push_back(f2);
level_files_.push_back(f3);
level_files_.push_back(f4);
compaction_files_.push_back(f1);
AddBoundaryInputs(icmp_, level_files_, &compaction_files_);
ASSERT_EQ(3, compaction_files_.size());
ASSERT_EQ(f1, compaction_files_[0]);
ASSERT_EQ(f4, compaction_files_[1]);
ASSERT_EQ(f3, compaction_files_[2]);
}
} // namespace leveldb

25
db/write_batch.cc
View File

@ -15,10 +15,10 @@
#include "leveldb/write_batch.h"
#include "leveldb/db.h"
#include "db/dbformat.h"
#include "db/memtable.h"
#include "db/write_batch_internal.h"
#include "leveldb/db.h"
#include "util/coding.h"
namespace leveldb {
@ -26,22 +26,18 @@ namespace leveldb {
// WriteBatch header has an 8-byte sequence number followed by a 4-byte count.
static const size_t kHeader = 12;
WriteBatch::WriteBatch() {
Clear();
}
WriteBatch::WriteBatch() { Clear(); }
WriteBatch::~WriteBatch() { }
WriteBatch::~WriteBatch() = default;
WriteBatch::Handler::~Handler() { }
WriteBatch::Handler::~Handler() = default;
void WriteBatch::Clear() {
rep_.clear();
rep_.resize(kHeader);
}
size_t WriteBatch::ApproximateSize() {
return rep_.size();
}
size_t WriteBatch::ApproximateSize() const { return rep_.size(); }
Status WriteBatch::Iterate(Handler* handler) const {
Slice input(rep_);
@ -112,25 +108,28 @@ void WriteBatch::Delete(const Slice& key) {
PutLengthPrefixedSlice(&rep_, key);
}
void WriteBatch::Append(const WriteBatch& source) {
WriteBatchInternal::Append(this, &source);
}
namespace {
class MemTableInserter : public WriteBatch::Handler {
public:
SequenceNumber sequence_;
MemTable* mem_;
virtual void Put(const Slice& key, const Slice& value) {
void Put(const Slice& key, const Slice& value) override {
mem_->Add(sequence_, kTypeValue, key, value);
sequence_++;
}
virtual void Delete(const Slice& key) {
void Delete(const Slice& key) override {
mem_->Add(sequence_, kTypeDeletion, key, Slice());
sequence_++;
}
};
} // namespace
Status WriteBatchInternal::InsertInto(const WriteBatch* b,
MemTable* memtable) {
Status WriteBatchInternal::InsertInto(const WriteBatch* b, MemTable* memtable) {
MemTableInserter inserter;
inserter.sequence_ = WriteBatchInternal::Sequence(b);
inserter.mem_ = memtable;

9
db/write_batch_internal.h
View File

@ -29,13 +29,9 @@ class WriteBatchInternal {
// this batch.
static void SetSequence(WriteBatch* batch, SequenceNumber seq);
static Slice Contents(const WriteBatch* batch) {
return Slice(batch->rep_);
}
static Slice Contents(const WriteBatch* batch) { return Slice(batch->rep_); }
static size_t ByteSize(const WriteBatch* batch) {
return batch->rep_.size();
}
static size_t ByteSize(const WriteBatch* batch) { return batch->rep_.size(); }
static void SetContents(WriteBatch* batch, const Slice& contents);
@ -46,5 +42,4 @@ class WriteBatchInternal {
} // namespace leveldb
#endif // STORAGE_LEVELDB_DB_WRITE_BATCH_INTERNAL_H_

63
db/write_batch_test.cc
View File

@ -2,13 +2,12 @@
// 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 "leveldb/db.h"
#include "gtest/gtest.h"
#include "db/memtable.h"
#include "db/write_batch_internal.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "util/logging.h"
#include "util/testharness.h"
namespace leveldb {
@ -22,7 +21,7 @@ static std::string PrintContents(WriteBatch* b) {
Iterator* iter = mem->NewIterator();
for (iter->SeekToFirst(); iter->Valid(); iter->Next()) {
ParsedInternalKey ikey;
ASSERT_TRUE(ParseInternalKey(iter->key(), &ikey));
EXPECT_TRUE(ParseInternalKey(iter->key(), &ikey));
switch (ikey.type) {
case kTypeValue:
state.append("Put(");
@ -52,8 +51,6 @@ static std::string PrintContents(WriteBatch* b) {
return state;
}
class WriteBatchTest { };
TEST(WriteBatchTest, Empty) {
WriteBatch batch;
ASSERT_EQ("", PrintContents(&batch));
@ -68,10 +65,11 @@ TEST(WriteBatchTest, Multiple) {
WriteBatchInternal::SetSequence(&batch, 100);
ASSERT_EQ(100, WriteBatchInternal::Sequence(&batch));
ASSERT_EQ(3, WriteBatchInternal::Count(&batch));
ASSERT_EQ("Put(baz, boo)@102"
"Delete(box)@101"
"Put(foo, bar)@100",
PrintContents(&batch));
ASSERT_EQ(
"Put(baz, boo)@102"
"Delete(box)@101"
"Put(foo, bar)@100",
PrintContents(&batch));
}
TEST(WriteBatchTest, Corruption) {
@ -81,36 +79,37 @@ TEST(WriteBatchTest, Corruption) {
WriteBatchInternal::SetSequence(&batch, 200);
Slice contents = WriteBatchInternal::Contents(&batch);
WriteBatchInternal::SetContents(&batch,
Slice(contents.data(),contents.size()-1));
ASSERT_EQ("Put(foo, bar)@200"
"ParseError()",
PrintContents(&batch));
Slice(contents.data(), contents.size() - 1));
ASSERT_EQ(
"Put(foo, bar)@200"
"ParseError()",
PrintContents(&batch));
}
TEST(WriteBatchTest, Append) {
WriteBatch b1, b2;
WriteBatchInternal::SetSequence(&b1, 200);
WriteBatchInternal::SetSequence(&b2, 300);
WriteBatchInternal::Append(&b1, &b2);
ASSERT_EQ("",
PrintContents(&b1));
b1.Append(b2);
ASSERT_EQ("", PrintContents(&b1));
b2.Put("a", "va");
WriteBatchInternal::Append(&b1, &b2);
ASSERT_EQ("Put(a, va)@200",
PrintContents(&b1));
b1.Append(b2);
ASSERT_EQ("Put(a, va)@200", PrintContents(&b1));
b2.Clear();
b2.Put("b", "vb");
WriteBatchInternal::Append(&b1, &b2);
ASSERT_EQ("Put(a, va)@200"
"Put(b, vb)@201",
PrintContents(&b1));
b1.Append(b2);
ASSERT_EQ(
"Put(a, va)@200"
"Put(b, vb)@201",
PrintContents(&b1));
b2.Delete("foo");
WriteBatchInternal::Append(&b1, &b2);
ASSERT_EQ("Put(a, va)@200"
"Put(b, vb)@202"
"Put(b, vb)@201"
"Delete(foo)@203",
PrintContents(&b1));
b1.Append(b2);
ASSERT_EQ(
"Put(a, va)@200"
"Put(b, vb)@202"
"Put(b, vb)@201"
"Delete(foo)@203",
PrintContents(&b1));
}
TEST(WriteBatchTest, ApproximateSize) {
@ -131,7 +130,3 @@ TEST(WriteBatchTest, ApproximateSize) {
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

16
doc/benchmark.html
View File

@ -83,23 +83,23 @@ div.bsql {
<p>Google, July 2011</p>
<hr>
<p>In order to test LevelDB's performance, we benchmark it against other well-established database implementations. We compare LevelDB (revision 39) against <a href="http://www.sqlite.org/">SQLite3</a> (version 3.7.6.3) and <a href="http://fallabs.com/kyotocabinet/spex.html">Kyoto Cabinet's</a> (version 1.2.67) TreeDB (a B+Tree based key-value store). We would like to acknowledge Scott Hess and Mikio Hirabayashi for their suggestions and contributions to the SQLite3 and Kyoto Cabinet benchmarks, respectively.</p>
<p>In order to test LevelDB's performance, we benchmark it against other well-established database implementations. We compare LevelDB (revision 39) against <a href="https://www.sqlite.org/">SQLite3</a> (version 3.7.6.3) and <a href="https://dbmx.net/kyotocabinet/spex.html">Kyoto Cabinet's</a> (version 1.2.67) TreeDB (a B+Tree based key-value store). We would like to acknowledge Scott Hess and Mikio Hirabayashi for their suggestions and contributions to the SQLite3 and Kyoto Cabinet benchmarks, respectively.</p>
<p>Benchmarks were all performed on a six-core Intel(R) Xeon(R) CPU X5650 @ 2.67GHz, with 12288 KB of total L3 cache and 12 GB of DDR3 RAM at 1333 MHz. (Note that LevelDB uses at most two CPUs since the benchmarks are single threaded: one to run the benchmark, and one for background compactions.) We ran the benchmarks on two machines (with identical processors), one with an Ext3 file system and one with an Ext4 file system. The machine with the Ext3 file system has a SATA Hitachi HDS721050CLA362 hard drive. The machine with the Ext4 file system has a SATA Samsung HD502HJ hard drive. Both hard drives spin at 7200 RPM and have hard drive write-caching enabled (using `hdparm -W 1 [device]`). The numbers reported below are the median of three measurements.</p>
<h4>Benchmark Source Code</h4>
<p>We wrote benchmark tools for SQLite and Kyoto TreeDB based on LevelDB's <span class="code">db_bench</span>. The code for each of the benchmarks resides here:</p>
<ul>
<li> <b>LevelDB:</b> <a href="http://code.google.com/p/leveldb/source/browse/trunk/db/db_bench.cc">db/db_bench.cc</a>.</li>
<li> <b>SQLite:</b> <a href="http://code.google.com/p/leveldb/source/browse/#svn%2Ftrunk%2Fdoc%2Fbench%2Fdb_bench_sqlite3.cc">doc/bench/db_bench_sqlite3.cc</a>.</li>
<li> <b>Kyoto TreeDB:</b> <a href="http://code.google.com/p/leveldb/source/browse/#svn%2Ftrunk%2Fdoc%2Fbench%2Fdb_bench_tree_db.cc">doc/bench/db_bench_tree_db.cc</a>.</li>
<li> <b>LevelDB:</b> <a href="https://github.com/google/leveldb/blob/master/benchmarks/db_bench.cc">benchmarks/db_bench.cc</a>.</li>
<li> <b>SQLite:</b> <a href="https://github.com/google/leveldb/blob/master/benchmarks/db_bench_sqlite3.cc">benchmarks/db_bench_sqlite3.cc</a>.</li>
<li> <b>Kyoto TreeDB:</b> <a href="https://github.com/google/leveldb/blob/master/benchmarks/db_bench_tree_db.cc">benchmarks/db_bench_tree_db.cc</a>.</li>
</ul>
<h4>Custom Build Specifications</h4>
<ul>
<li>LevelDB: LevelDB was compiled with the <a href="http://code.google.com/p/google-perftools">tcmalloc</a> library and the <a href="http://code.google.com/p/snappy/">Snappy</a> compression library (revision 33). Assertions were disabled.</li>
<li>TreeDB: TreeDB was compiled using the <a href="http://www.oberhumer.com/opensource/lzo/">LZO</a> compression library (version 2.03). Furthermore, we enabled the TSMALL and TLINEAR options when opening the database in order to reduce the footprint of each record.</li>
<li>SQLite: We tuned SQLite's performance, by setting its locking mode to exclusive. We also enabled SQLite's <a href="http://www.sqlite.org/draft/wal.html">write-ahead logging</a>.</li>
<li>LevelDB: LevelDB was compiled with the <a href="https://github.com/gperftools/gperftools">tcmalloc</a> library and the <a href="https://github.com/google/snappy">Snappy</a> compression library (revision 33). Assertions were disabled.</li>
<li>TreeDB: TreeDB was compiled using the <a href="https://www.oberhumer.com/opensource/lzo/">LZO</a> compression library (version 2.03). Furthermore, we enabled the TSMALL and TLINEAR options when opening the database in order to reduce the footprint of each record.</li>
<li>SQLite: We tuned SQLite's performance, by setting its locking mode to exclusive. We also enabled SQLite's <a href="https://www.sqlite.org/draft/wal.html">write-ahead logging</a>.</li>
</ul>
<h2>1. Baseline Performance</h2>
@ -451,7 +451,7 @@ performance may very well be better with compression if it allows more
of the working set to fit in memory.</p>
<h2>Note about Ext4 Filesystems</h2>
<p>The preceding numbers are for an ext3 file system. Synchronous writes are much slower under <a href="http://en.wikipedia.org/wiki/Ext4">ext4</a> (LevelDB drops to ~31 writes / second and TreeDB drops to ~5 writes / second; SQLite3's synchronous writes do not noticeably drop) due to ext4's different handling of <span class="code">fsync</span> / <span class="code">msync</span> calls. Even LevelDB's asynchronous write performance drops somewhat since it spreads its storage across multiple files and issues <span class="code">fsync</span> calls when switching to a new file.</p>
<p>The preceding numbers are for an ext3 file system. Synchronous writes are much slower under <a href="https://en.wikipedia.org/wiki/Ext4">ext4</a> (LevelDB drops to ~31 writes / second and TreeDB drops to ~5 writes / second; SQLite3's synchronous writes do not noticeably drop) due to ext4's different handling of <span class="code">fsync</span> / <span class="code">msync</span> calls. Even LevelDB's asynchronous write performance drops somewhat since it spreads its storage across multiple files and issues <span class="code">fsync</span> calls when switching to a new file.</p>
<h2>Acknowledgements</h2>
<p>Jeff Dean and Sanjay Ghemawat wrote LevelDB. Kevin Tseng wrote and compiled these benchmarks. Mikio Hirabayashi, Scott Hess, and Gabor Cselle provided help and advice.</p>

18
doc/impl.md
View File

@ -1,7 +1,7 @@
## Files
The implementation of leveldb is similar in spirit to the representation of a
single [Bigtable tablet (section 5.3)](http://research.google.com/archive/bigtable.html).
single [Bigtable tablet (section 5.3)](https://research.google/pubs/pub27898/).
However the organization of the files that make up the representation is
somewhat different and is explained below.
@ -64,13 +64,15 @@ Other files used for miscellaneous purposes may also be present (LOCK, *.dbtmp).
## Level 0
When the log file grows above a certain size (1MB by default):
Create a brand new memtable and log file and direct future updates here
When the log file grows above a certain size (4MB by default):
Create a brand new memtable and log file and direct future updates here.
In the background:
Write the contents of the previous memtable to an sstable
Discard the memtable
Delete the old log file and the old memtable
Add the new sstable to the young (level-0) level.
1. Write the contents of the previous memtable to an sstable.
2. Discard the memtable.
3. Delete the old log file and the old memtable.
4. Add the new sstable to the young (level-0) level.
## Compactions
@ -164,7 +166,7 @@ So maybe even the sharding is not necessary on modern filesystems?
## Garbage collection of files
`DeleteObsoleteFiles()` is called at the end of every compaction and at the end
`RemoveObsoleteFiles()` is called at the end of every compaction and at the end
of recovery. It finds the names of all files in the database. It deletes all log
files that are not the current log file. It deletes all table files that are not
referenced from some level and are not the output of an active compaction.

23
doc/index.md
View File

@ -307,7 +307,7 @@ version numbers found in the keys to decide how to interpret them.
## Performance
Performance can be tuned by changing the default values of the types defined in
`include/leveldb/options.h`.
`include/options.h`.
### Block size
@ -338,19 +338,19 @@ options.compression = leveldb::kNoCompression;
### Cache
The contents of the database are stored in a set of files in the filesystem and
each file stores a sequence of compressed blocks. If options.cache is non-NULL,
it is used to cache frequently used uncompressed block contents.
each file stores a sequence of compressed blocks. If options.block_cache is
non-NULL, it is used to cache frequently used uncompressed block contents.
```c++
#include "leveldb/cache.h"
leveldb::Options options;
options.cache = leveldb::NewLRUCache(100 * 1048576); // 100MB cache
options.block_cache = leveldb::NewLRUCache(100 * 1048576); // 100MB cache
leveldb::DB* db;
leveldb::DB::Open(options, name, &db);
... use the db ...
delete db
delete options.cache;
delete options.block_cache;
```
Note that the cache holds uncompressed data, and therefore it should be sized
@ -369,6 +369,7 @@ leveldb::Iterator* it = db->NewIterator(options);
for (it->SeekToFirst(); it->Valid(); it->Next()) {
...
}
delete it;
```
### Key Layout
@ -424,21 +425,21 @@ spaces. For example:
```c++
class CustomFilterPolicy : public leveldb::FilterPolicy {
private:
FilterPolicy* builtin_policy_;
leveldb::FilterPolicy* builtin_policy_;
public:
CustomFilterPolicy() : builtin_policy_(NewBloomFilterPolicy(10)) {}
CustomFilterPolicy() : builtin_policy_(leveldb::NewBloomFilterPolicy(10)) {}
~CustomFilterPolicy() { delete builtin_policy_; }
const char* Name() const { return "IgnoreTrailingSpacesFilter"; }
void CreateFilter(const Slice* keys, int n, std::string* dst) const {
void CreateFilter(const leveldb::Slice* keys, int n, std::string* dst) const {
// Use builtin bloom filter code after removing trailing spaces
std::vector<Slice> trimmed(n);
std::vector<leveldb::Slice> trimmed(n);
for (int i = 0; i < n; i++) {
trimmed[i] = RemoveTrailingSpaces(keys[i]);
}
return builtin_policy_->CreateFilter(&trimmed[i], n, dst);
builtin_policy_->CreateFilter(trimmed.data(), n, dst);
}
};
```
@ -478,7 +479,7 @@ leveldb::Range ranges[2];
ranges[0] = leveldb::Range("a", "c");
ranges[1] = leveldb::Range("x", "z");
uint64_t sizes[2];
leveldb::Status s = db->GetApproximateSizes(ranges, 2, sizes);
db->GetApproximateSizes(ranges, 2, sizes);
```
The preceding call will set `sizes[0]` to the approximate number of bytes of

183
helpers/memenv/memenv.cc
View File

@ -4,8 +4,7 @@
#include "helpers/memenv/memenv.h"
#include <string.h>
#include <cstring>
#include <limits>
#include <map>
#include <string>
@ -27,6 +26,10 @@ class FileState {
// and the caller must call Ref() at least once.
FileState() : refs_(0), size_(0) {}
// No copying allowed.
FileState(const FileState&) = delete;
FileState& operator=(const FileState&) = delete;
// Increase the reference count.
void Ref() {
MutexLock lock(&refs_mutex_);
@ -51,9 +54,22 @@ class FileState {
}
}
uint64_t Size() const { return size_; }
uint64_t Size() const {
MutexLock lock(&blocks_mutex_);
return size_;
}
void Truncate() {
MutexLock lock(&blocks_mutex_);
for (char*& block : blocks_) {
delete[] block;
}
blocks_.clear();
size_ = 0;
}
Status Read(uint64_t offset, size_t n, Slice* result, char* scratch) const {
MutexLock lock(&blocks_mutex_);
if (offset > size_) {
return Status::IOError("Offset greater than file size.");
}
@ -69,13 +85,6 @@ class FileState {
assert(offset / kBlockSize <= std::numeric_limits<size_t>::max());
size_t block = static_cast<size_t>(offset / kBlockSize);
size_t block_offset = offset % kBlockSize;
if (n <= kBlockSize - block_offset) {
// The requested bytes are all in the first block.
*result = Slice(blocks_[block] + block_offset, n);
return Status::OK();
}
size_t bytes_to_copy = n;
char* dst = scratch;
@ -84,7 +93,7 @@ class FileState {
if (avail > bytes_to_copy) {
avail = bytes_to_copy;
}
memcpy(dst, blocks_[block] + block_offset, avail);
std::memcpy(dst, blocks_[block] + block_offset, avail);
bytes_to_copy -= avail;
dst += avail;
@ -100,6 +109,7 @@ class FileState {
const char* src = data.data();
size_t src_len = data.size();
MutexLock lock(&blocks_mutex_);
while (src_len > 0) {
size_t avail;
size_t offset = size_ % kBlockSize;
@ -116,7 +126,7 @@ class FileState {
if (avail > src_len) {
avail = src_len;
}
memcpy(blocks_.back() + offset, src, avail);
std::memcpy(blocks_.back() + offset, src, avail);
src_len -= avail;
src += avail;
size_ += avail;
@ -126,28 +136,17 @@ class FileState {
}
private:
// Private since only Unref() should be used to delete it.
~FileState() {
for (std::vector<char*>::iterator i = blocks_.begin(); i != blocks_.end();
++i) {
delete [] *i;
}
}
enum { kBlockSize = 8 * 1024 };
// No copying allowed.
FileState(const FileState&);
void operator=(const FileState&);
// Private since only Unref() should be used to delete it.
~FileState() { Truncate(); }
port::Mutex refs_mutex_;
int refs_ GUARDED_BY(refs_mutex_);
// The following fields are not protected by any mutex. They are only mutable
// while the file is being written, and concurrent access is not allowed
// to writable files.
std::vector<char*> blocks_;
uint64_t size_;
enum { kBlockSize = 8 * 1024 };
mutable port::Mutex blocks_mutex_;
std::vector<char*> blocks_ GUARDED_BY(blocks_mutex_);
uint64_t size_ GUARDED_BY(blocks_mutex_);
};
class SequentialFileImpl : public SequentialFile {
@ -156,11 +155,9 @@ class SequentialFileImpl : public SequentialFile {
file_->Ref();
}
~SequentialFileImpl() {
file_->Unref();
}
~SequentialFileImpl() override { file_->Unref(); }
virtual Status Read(size_t n, Slice* result, char* scratch) {
Status Read(size_t n, Slice* result, char* scratch) override {
Status s = file_->Read(pos_, n, result, scratch);
if (s.ok()) {
pos_ += result->size();
@ -168,7 +165,7 @@ class SequentialFileImpl : public SequentialFile {
return s;
}
virtual Status Skip(uint64_t n) {
Status Skip(uint64_t n) override {
if (pos_ > file_->Size()) {
return Status::IOError("pos_ > file_->Size()");
}
@ -187,16 +184,12 @@ class SequentialFileImpl : public SequentialFile {
class RandomAccessFileImpl : public RandomAccessFile {
public:
explicit RandomAccessFileImpl(FileState* file) : file_(file) {
file_->Ref();
}
explicit RandomAccessFileImpl(FileState* file) : file_(file) { file_->Ref(); }
~RandomAccessFileImpl() {
file_->Unref();
}
~RandomAccessFileImpl() override { file_->Unref(); }
virtual Status Read(uint64_t offset, size_t n, Slice* result,
char* scratch) const {
Status Read(uint64_t offset, size_t n, Slice* result,
char* scratch) const override {
return file_->Read(offset, n, result, scratch);
}
@ -206,21 +199,15 @@ class RandomAccessFileImpl : public RandomAccessFile {
class WritableFileImpl : public WritableFile {
public:
WritableFileImpl(FileState* file) : file_(file) {
file_->Ref();
}
WritableFileImpl(FileState* file) : file_(file) { file_->Ref(); }
~WritableFileImpl() {
file_->Unref();
}
~WritableFileImpl() override { file_->Unref(); }
virtual Status Append(const Slice& data) {
return file_->Append(data);
}
Status Append(const Slice& data) override { return file_->Append(data); }
virtual Status Close() { return Status::OK(); }
virtual Status Flush() { return Status::OK(); }
virtual Status Sync() { return Status::OK(); }
Status Close() override { return Status::OK(); }
Status Flush() override { return Status::OK(); }
Status Sync() override { return Status::OK(); }
private:
FileState* file_;
@ -228,22 +215,22 @@ class WritableFileImpl : public WritableFile {
class NoOpLogger : public Logger {
public:
virtual void Logv(const char* format, va_list ap) { }
void Logv(const char* format, std::va_list ap) override {}
};
class InMemoryEnv : public EnvWrapper {
public:
explicit InMemoryEnv(Env* base_env) : EnvWrapper(base_env) { }
explicit InMemoryEnv(Env* base_env) : EnvWrapper(base_env) {}
virtual ~InMemoryEnv() {
for (FileSystem::iterator i = file_map_.begin(); i != file_map_.end(); ++i){
i->second->Unref();
~InMemoryEnv() override {
for (const auto& kvp : file_map_) {
kvp.second->Unref();
}
}
// Partial implementation of the Env interface.
virtual Status NewSequentialFile(const std::string& fname,
SequentialFile** result) {
Status NewSequentialFile(const std::string& fname,
SequentialFile** result) override {
MutexLock lock(&mutex_);
if (file_map_.find(fname) == file_map_.end()) {
*result = nullptr;
@ -254,8 +241,8 @@ class InMemoryEnv : public EnvWrapper {
return Status::OK();
}
virtual Status NewRandomAccessFile(const std::string& fname,
RandomAccessFile** result) {
Status NewRandomAccessFile(const std::string& fname,
RandomAccessFile** result) override {
MutexLock lock(&mutex_);
if (file_map_.find(fname) == file_map_.end()) {
*result = nullptr;
@ -266,23 +253,28 @@ class InMemoryEnv : public EnvWrapper {
return Status::OK();
}
virtual Status NewWritableFile(const std::string& fname,
WritableFile** result) {
Status NewWritableFile(const std::string& fname,
WritableFile** result) override {
MutexLock lock(&mutex_);
if (file_map_.find(fname) != file_map_.end()) {
DeleteFileInternal(fname);
}
FileSystem::iterator it = file_map_.find(fname);
FileState* file = new FileState();
file->Ref();
file_map_[fname] = file;
FileState* file;
if (it == file_map_.end()) {
// File is not currently open.
file = new FileState();
file->Ref();
file_map_[fname] = file;
} else {
file = it->second;
file->Truncate();
}
*result = new WritableFileImpl(file);
return Status::OK();
}
virtual Status NewAppendableFile(const std::string& fname,
WritableFile** result) {
Status NewAppendableFile(const std::string& fname,
WritableFile** result) override {
MutexLock lock(&mutex_);
FileState** sptr = &file_map_[fname];
FileState* file = *sptr;
@ -294,18 +286,18 @@ class InMemoryEnv : public EnvWrapper {
return Status::OK();
}
virtual bool FileExists(const std::string& fname) {
bool FileExists(const std::string& fname) override {
MutexLock lock(&mutex_);
return file_map_.find(fname) != file_map_.end();
}
virtual Status GetChildren(const std::string& dir,
std::vector<std::string>* result) {
Status GetChildren(const std::string& dir,
std::vector<std::string>* result) override {
MutexLock lock(&mutex_);
result->clear();
for (FileSystem::iterator i = file_map_.begin(); i != file_map_.end(); ++i){
const std::string& filename = i->first;
for (const auto& kvp : file_map_) {
const std::string& filename = kvp.first;
if (filename.size() >= dir.size() + 1 && filename[dir.size()] == '/' &&
Slice(filename).starts_with(Slice(dir))) {
@ -316,7 +308,7 @@ class InMemoryEnv : public EnvWrapper {
return Status::OK();
}
void DeleteFileInternal(const std::string& fname)
void RemoveFileInternal(const std::string& fname)
EXCLUSIVE_LOCKS_REQUIRED(mutex_) {
if (file_map_.find(fname) == file_map_.end()) {
return;
@ -326,25 +318,21 @@ class InMemoryEnv : public EnvWrapper {
file_map_.erase(fname);
}
virtual Status DeleteFile(const std::string& fname) {
Status RemoveFile(const std::string& fname) override {
MutexLock lock(&mutex_);
if (file_map_.find(fname) == file_map_.end()) {
return Status::IOError(fname, "File not found");
}
DeleteFileInternal(fname);
RemoveFileInternal(fname);
return Status::OK();
}
virtual Status CreateDir(const std::string& dirname) {
return Status::OK();
}
Status CreateDir(const std::string& dirname) override { return Status::OK(); }
virtual Status DeleteDir(const std::string& dirname) {
return Status::OK();
}
Status RemoveDir(const std::string& dirname) override { return Status::OK(); }
virtual Status GetFileSize(const std::string& fname, uint64_t* file_size) {
Status GetFileSize(const std::string& fname, uint64_t* file_size) override {
MutexLock lock(&mutex_);
if (file_map_.find(fname) == file_map_.end()) {
return Status::IOError(fname, "File not found");
@ -354,35 +342,35 @@ class InMemoryEnv : public EnvWrapper {
return Status::OK();
}
virtual Status RenameFile(const std::string& src,
const std::string& target) {
Status RenameFile(const std::string& src,
const std::string& target) override {
MutexLock lock(&mutex_);
if (file_map_.find(src) == file_map_.end()) {
return Status::IOError(src, "File not found");
}
DeleteFileInternal(target);
RemoveFileInternal(target);
file_map_[target] = file_map_[src];
file_map_.erase(src);
return Status::OK();
}
virtual Status LockFile(const std::string& fname, FileLock** lock) {
Status LockFile(const std::string& fname, FileLock** lock) override {
*lock = new FileLock;
return Status::OK();
}
virtual Status UnlockFile(FileLock* lock) {
Status UnlockFile(FileLock* lock) override {
delete lock;
return Status::OK();
}
virtual Status GetTestDirectory(std::string* path) {
Status GetTestDirectory(std::string* path) override {
*path = "/test";
return Status::OK();
}
virtual Status NewLogger(const std::string& fname, Logger** result) {
Status NewLogger(const std::string& fname, Logger** result) override {
*result = new NoOpLogger;
return Status::OK();
}
@ -390,14 +378,13 @@ class InMemoryEnv : public EnvWrapper {
private:
// Map from filenames to FileState objects, representing a simple file system.
typedef std::map<std::string, FileState*> FileSystem;
port::Mutex mutex_;
FileSystem file_map_ GUARDED_BY(mutex_);
};
} // namespace
Env* NewMemEnv(Env* base_env) {
return new InMemoryEnv(base_env);
}
Env* NewMemEnv(Env* base_env) { return new InMemoryEnv(base_env); }
} // namespace leveldb

166
helpers/memenv/memenv_test.cc
View File

@ -4,76 +4,74 @@
#include "helpers/memenv/memenv.h"
#include "db/db_impl.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "util/testharness.h"
#include <string>
#include <vector>
#include "gtest/gtest.h"
#include "db/db_impl.h"
#include "leveldb/db.h"
#include "leveldb/env.h"
#include "util/testutil.h"
namespace leveldb {
class MemEnvTest {
class MemEnvTest : public testing::Test {
public:
Env* env_;
MemEnvTest() : env_(NewMemEnv(Env::Default())) {}
~MemEnvTest() { delete env_; }
MemEnvTest()
: env_(NewMemEnv(Env::Default())) {
}
~MemEnvTest() {
delete env_;
}
Env* env_;
};
TEST(MemEnvTest, Basics) {
TEST_F(MemEnvTest, Basics) {
uint64_t file_size;
WritableFile* writable_file;
std::vector<std::string> children;
ASSERT_OK(env_->CreateDir("/dir"));
ASSERT_LEVELDB_OK(env_->CreateDir("/dir"));
// Check that the directory is empty.
ASSERT_TRUE(!env_->FileExists("/dir/non_existent"));
ASSERT_TRUE(!env_->GetFileSize("/dir/non_existent", &file_size).ok());
ASSERT_OK(env_->GetChildren("/dir", &children));
ASSERT_LEVELDB_OK(env_->GetChildren("/dir", &children));
ASSERT_EQ(0, children.size());
// Create a file.
ASSERT_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_LEVELDB_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_LEVELDB_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_EQ(0, file_size);
delete writable_file;
// Check that the file exists.
ASSERT_TRUE(env_->FileExists("/dir/f"));
ASSERT_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_LEVELDB_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_EQ(0, file_size);
ASSERT_OK(env_->GetChildren("/dir", &children));
ASSERT_LEVELDB_OK(env_->GetChildren("/dir", &children));
ASSERT_EQ(1, children.size());
ASSERT_EQ("f", children[0]);
// Write to the file.
ASSERT_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_OK(writable_file->Append("abc"));
ASSERT_LEVELDB_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_LEVELDB_OK(writable_file->Append("abc"));
delete writable_file;
// Check that append works.
ASSERT_OK(env_->NewAppendableFile("/dir/f", &writable_file));
ASSERT_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_LEVELDB_OK(env_->NewAppendableFile("/dir/f", &writable_file));
ASSERT_LEVELDB_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_EQ(3, file_size);
ASSERT_OK(writable_file->Append("hello"));
ASSERT_LEVELDB_OK(writable_file->Append("hello"));
delete writable_file;
// Check for expected size.
ASSERT_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_LEVELDB_OK(env_->GetFileSize("/dir/f", &file_size));
ASSERT_EQ(8, file_size);
// Check that renaming works.
ASSERT_TRUE(!env_->RenameFile("/dir/non_existent", "/dir/g").ok());
ASSERT_OK(env_->RenameFile("/dir/f", "/dir/g"));
ASSERT_LEVELDB_OK(env_->RenameFile("/dir/f", "/dir/g"));
ASSERT_TRUE(!env_->FileExists("/dir/f"));
ASSERT_TRUE(env_->FileExists("/dir/g"));
ASSERT_OK(env_->GetFileSize("/dir/g", &file_size));
ASSERT_LEVELDB_OK(env_->GetFileSize("/dir/g", &file_size));
ASSERT_EQ(8, file_size);
// Check that opening non-existent file fails.
@ -85,49 +83,50 @@ TEST(MemEnvTest, Basics) {
ASSERT_TRUE(!rand_file);
// Check that deleting works.
ASSERT_TRUE(!env_->DeleteFile("/dir/non_existent").ok());
ASSERT_OK(env_->DeleteFile("/dir/g"));
ASSERT_TRUE(!env_->RemoveFile("/dir/non_existent").ok());
ASSERT_LEVELDB_OK(env_->RemoveFile("/dir/g"));
ASSERT_TRUE(!env_->FileExists("/dir/g"));
ASSERT_OK(env_->GetChildren("/dir", &children));
ASSERT_LEVELDB_OK(env_->GetChildren("/dir", &children));
ASSERT_EQ(0, children.size());
ASSERT_OK(env_->DeleteDir("/dir"));
ASSERT_LEVELDB_OK(env_->RemoveDir("/dir"));
}
TEST(MemEnvTest, ReadWrite) {
TEST_F(MemEnvTest, ReadWrite) {
WritableFile* writable_file;
SequentialFile* seq_file;
RandomAccessFile* rand_file;
Slice result;
char scratch[100];
ASSERT_OK(env_->CreateDir("/dir"));
ASSERT_LEVELDB_OK(env_->CreateDir("/dir"));
ASSERT_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_OK(writable_file->Append("hello "));
ASSERT_OK(writable_file->Append("world"));
ASSERT_LEVELDB_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_LEVELDB_OK(writable_file->Append("hello "));
ASSERT_LEVELDB_OK(writable_file->Append("world"));
delete writable_file;
// Read sequentially.
ASSERT_OK(env_->NewSequentialFile("/dir/f", &seq_file));
ASSERT_OK(seq_file->Read(5, &result, scratch)); // Read "hello".
ASSERT_LEVELDB_OK(env_->NewSequentialFile("/dir/f", &seq_file));
ASSERT_LEVELDB_OK(seq_file->Read(5, &result, scratch)); // Read "hello".
ASSERT_EQ(0, result.compare("hello"));
ASSERT_OK(seq_file->Skip(1));
ASSERT_OK(seq_file->Read(1000, &result, scratch)); // Read "world".
ASSERT_LEVELDB_OK(seq_file->Skip(1));
ASSERT_LEVELDB_OK(seq_file->Read(1000, &result, scratch)); // Read "world".
ASSERT_EQ(0, result.compare("world"));
ASSERT_OK(seq_file->Read(1000, &result, scratch)); // Try reading past EOF.
ASSERT_LEVELDB_OK(
seq_file->Read(1000, &result, scratch)); // Try reading past EOF.
ASSERT_EQ(0, result.size());
ASSERT_OK(seq_file->Skip(100)); // Try to skip past end of file.
ASSERT_OK(seq_file->Read(1000, &result, scratch));
ASSERT_LEVELDB_OK(seq_file->Skip(100)); // Try to skip past end of file.
ASSERT_LEVELDB_OK(seq_file->Read(1000, &result, scratch));
ASSERT_EQ(0, result.size());
delete seq_file;
// Random reads.
ASSERT_OK(env_->NewRandomAccessFile("/dir/f", &rand_file));
ASSERT_OK(rand_file->Read(6, 5, &result, scratch)); // Read "world".
ASSERT_LEVELDB_OK(env_->NewRandomAccessFile("/dir/f", &rand_file));
ASSERT_LEVELDB_OK(rand_file->Read(6, 5, &result, scratch)); // Read "world".
ASSERT_EQ(0, result.compare("world"));
ASSERT_OK(rand_file->Read(0, 5, &result, scratch)); // Read "hello".
ASSERT_LEVELDB_OK(rand_file->Read(0, 5, &result, scratch)); // Read "hello".
ASSERT_EQ(0, result.compare("hello"));
ASSERT_OK(rand_file->Read(10, 100, &result, scratch)); // Read "d".
ASSERT_LEVELDB_OK(rand_file->Read(10, 100, &result, scratch)); // Read "d".
ASSERT_EQ(0, result.compare("d"));
// Too high offset.
@ -135,30 +134,30 @@ TEST(MemEnvTest, ReadWrite) {
delete rand_file;
}
TEST(MemEnvTest, Locks) {
TEST_F(MemEnvTest, Locks) {
FileLock* lock;
// These are no-ops, but we test they return success.
ASSERT_OK(env_->LockFile("some file", &lock));
ASSERT_OK(env_->UnlockFile(lock));
ASSERT_LEVELDB_OK(env_->LockFile("some file", &lock));
ASSERT_LEVELDB_OK(env_->UnlockFile(lock));
}
TEST(MemEnvTest, Misc) {
TEST_F(MemEnvTest, Misc) {
std::string test_dir;
ASSERT_OK(env_->GetTestDirectory(&test_dir));
ASSERT_LEVELDB_OK(env_->GetTestDirectory(&test_dir));
ASSERT_TRUE(!test_dir.empty());
WritableFile* writable_file;
ASSERT_OK(env_->NewWritableFile("/a/b", &writable_file));
ASSERT_LEVELDB_OK(env_->NewWritableFile("/a/b", &writable_file));
// These are no-ops, but we test they return success.
ASSERT_OK(writable_file->Sync());
ASSERT_OK(writable_file->Flush());
ASSERT_OK(writable_file->Close());
ASSERT_LEVELDB_OK(writable_file->Sync());
ASSERT_LEVELDB_OK(writable_file->Flush());
ASSERT_LEVELDB_OK(writable_file->Close());
delete writable_file;
}
TEST(MemEnvTest, LargeWrite) {
TEST_F(MemEnvTest, LargeWrite) {
const size_t kWriteSize = 300 * 1024;
char* scratch = new char[kWriteSize * 2];
@ -168,30 +167,53 @@ TEST(MemEnvTest, LargeWrite) {
}
WritableFile* writable_file;
ASSERT_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_OK(writable_file->Append("foo"));
ASSERT_OK(writable_file->Append(write_data));
ASSERT_LEVELDB_OK(env_->NewWritableFile("/dir/f", &writable_file));
ASSERT_LEVELDB_OK(writable_file->Append("foo"));
ASSERT_LEVELDB_OK(writable_file->Append(write_data));
delete writable_file;
SequentialFile* seq_file;
Slice result;
ASSERT_OK(env_->NewSequentialFile("/dir/f", &seq_file));
ASSERT_OK(seq_file->Read(3, &result, scratch)); // Read "foo".
ASSERT_LEVELDB_OK(env_->NewSequentialFile("/dir/f", &seq_file));
ASSERT_LEVELDB_OK(seq_file->Read(3, &result, scratch)); // Read "foo".
ASSERT_EQ(0, result.compare("foo"));
size_t read = 0;
std::string read_data;
while (read < kWriteSize) {
ASSERT_OK(seq_file->Read(kWriteSize - read, &result, scratch));
ASSERT_LEVELDB_OK(seq_file->Read(kWriteSize - read, &result, scratch));
read_data.append(result.data(), result.size());
read += result.size();
}
ASSERT_TRUE(write_data == read_data);
delete seq_file;
delete [] scratch;
delete[] scratch;
}
TEST(MemEnvTest, DBTest) {
TEST_F(MemEnvTest, OverwriteOpenFile) {
const char kWrite1Data[] = "Write #1 data";
const size_t kFileDataLen = sizeof(kWrite1Data) - 1;
const std::string kTestFileName = testing::TempDir() + "leveldb-TestFile.dat";
ASSERT_LEVELDB_OK(WriteStringToFile(env_, kWrite1Data, kTestFileName));
RandomAccessFile* rand_file;
ASSERT_LEVELDB_OK(env_->NewRandomAccessFile(kTestFileName, &rand_file));
const char kWrite2Data[] = "Write #2 data";
ASSERT_LEVELDB_OK(WriteStringToFile(env_, kWrite2Data, kTestFileName));
// Verify that overwriting an open file will result in the new file data
// being read from files opened before the write.
Slice result;
char scratch[kFileDataLen];
ASSERT_LEVELDB_OK(rand_file->Read(0, kFileDataLen, &result, scratch));
ASSERT_EQ(0, result.compare(kWrite2Data));
delete rand_file;
}
TEST_F(MemEnvTest, DBTest) {
Options options;
options.create_if_missing = true;
options.env = env_;
@ -200,14 +222,14 @@ TEST(MemEnvTest, DBTest) {
const Slice keys[] = {Slice("aaa"), Slice("bbb"), Slice("ccc")};
const Slice vals[] = {Slice("foo"), Slice("bar"), Slice("baz")};
ASSERT_OK(DB::Open(options, "/dir/db", &db));
ASSERT_LEVELDB_OK(DB::Open(options, "/dir/db", &db));
for (size_t i = 0; i < 3; ++i) {
ASSERT_OK(db->Put(WriteOptions(), keys[i], vals[i]));
ASSERT_LEVELDB_OK(db->Put(WriteOptions(), keys[i], vals[i]));
}
for (size_t i = 0; i < 3; ++i) {
std::string res;
ASSERT_OK(db->Get(ReadOptions(), keys[i], &res));
ASSERT_LEVELDB_OK(db->Get(ReadOptions(), keys[i], &res));
ASSERT_TRUE(res == vals[i]);
}
@ -223,11 +245,11 @@ TEST(MemEnvTest, DBTest) {
delete iterator;
DBImpl* dbi = reinterpret_cast<DBImpl*>(db);
ASSERT_OK(dbi->TEST_CompactMemTable());
ASSERT_LEVELDB_OK(dbi->TEST_CompactMemTable());
for (size_t i = 0; i < 3; ++i) {
std::string res;
ASSERT_OK(db->Get(ReadOptions(), keys[i], &res));
ASSERT_LEVELDB_OK(db->Get(ReadOptions(), keys[i], &res));
ASSERT_TRUE(res == vals[i]);
}
@ -235,7 +257,3 @@ TEST(MemEnvTest, DBTest) {
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

90
include/leveldb/c.h
View File

@ -32,7 +32,7 @@
On failure, leveldb frees the old value of *errptr and
set *errptr to a malloc()ed error message.
(4) Bools have the type unsigned char (0 == false; rest == true)
(4) Bools have the type uint8_t (0 == false; rest == true)
(5) All of the pointer arguments must be non-NULL.
*/
@ -40,33 +40,34 @@
#ifndef STORAGE_LEVELDB_INCLUDE_C_H_
#define STORAGE_LEVELDB_INCLUDE_C_H_
#include <stdarg.h>
#include <stddef.h>
#include <stdint.h>
#include "leveldb/export.h"
#ifdef __cplusplus
extern "C" {
#endif
#include <stdarg.h>
#include <stddef.h>
#include <stdint.h>
#include "leveldb/export.h"
/* Exported types */
typedef struct leveldb_t leveldb_t;
typedef struct leveldb_cache_t leveldb_cache_t;
typedef struct leveldb_comparator_t leveldb_comparator_t;
typedef struct leveldb_env_t leveldb_env_t;
typedef struct leveldb_filelock_t leveldb_filelock_t;
typedef struct leveldb_filterpolicy_t leveldb_filterpolicy_t;
typedef struct leveldb_iterator_t leveldb_iterator_t;
typedef struct leveldb_logger_t leveldb_logger_t;
typedef struct leveldb_options_t leveldb_options_t;
typedef struct leveldb_randomfile_t leveldb_randomfile_t;
typedef struct leveldb_readoptions_t leveldb_readoptions_t;
typedef struct leveldb_seqfile_t leveldb_seqfile_t;
typedef struct leveldb_snapshot_t leveldb_snapshot_t;
typedef struct leveldb_writablefile_t leveldb_writablefile_t;
typedef struct leveldb_writebatch_t leveldb_writebatch_t;
typedef struct leveldb_writeoptions_t leveldb_writeoptions_t;
typedef struct leveldb_t leveldb_t;
typedef struct leveldb_cache_t leveldb_cache_t;
typedef struct leveldb_comparator_t leveldb_comparator_t;
typedef struct leveldb_env_t leveldb_env_t;
typedef struct leveldb_filelock_t leveldb_filelock_t;
typedef struct leveldb_filterpolicy_t leveldb_filterpolicy_t;
typedef struct leveldb_iterator_t leveldb_iterator_t;
typedef struct leveldb_logger_t leveldb_logger_t;
typedef struct leveldb_options_t leveldb_options_t;
typedef struct leveldb_randomfile_t leveldb_randomfile_t;
typedef struct leveldb_readoptions_t leveldb_readoptions_t;
typedef struct leveldb_seqfile_t leveldb_seqfile_t;
typedef struct leveldb_snapshot_t leveldb_snapshot_t;
typedef struct leveldb_writablefile_t leveldb_writablefile_t;
typedef struct leveldb_writebatch_t leveldb_writebatch_t;
typedef struct leveldb_writeoptions_t leveldb_writeoptions_t;
/* DB operations */
@ -130,7 +131,7 @@ LEVELDB_EXPORT void leveldb_repair_db(const leveldb_options_t* options,
/* Iterator */
LEVELDB_EXPORT void leveldb_iter_destroy(leveldb_iterator_t*);
LEVELDB_EXPORT unsigned char leveldb_iter_valid(const leveldb_iterator_t*);
LEVELDB_EXPORT uint8_t leveldb_iter_valid(const leveldb_iterator_t*);
LEVELDB_EXPORT void leveldb_iter_seek_to_first(leveldb_iterator_t*);
LEVELDB_EXPORT void leveldb_iter_seek_to_last(leveldb_iterator_t*);
LEVELDB_EXPORT void leveldb_iter_seek(leveldb_iterator_t*, const char* k,
@ -146,7 +147,7 @@ LEVELDB_EXPORT void leveldb_iter_get_error(const leveldb_iterator_t*,
/* Write batch */
LEVELDB_EXPORT leveldb_writebatch_t* leveldb_writebatch_create();
LEVELDB_EXPORT leveldb_writebatch_t* leveldb_writebatch_create(void);
LEVELDB_EXPORT void leveldb_writebatch_destroy(leveldb_writebatch_t*);
LEVELDB_EXPORT void leveldb_writebatch_clear(leveldb_writebatch_t*);
LEVELDB_EXPORT void leveldb_writebatch_put(leveldb_writebatch_t*,
@ -155,24 +156,26 @@ LEVELDB_EXPORT void leveldb_writebatch_put(leveldb_writebatch_t*,
LEVELDB_EXPORT void leveldb_writebatch_delete(leveldb_writebatch_t*,
const char* key, size_t klen);
LEVELDB_EXPORT void leveldb_writebatch_iterate(
leveldb_writebatch_t*, void* state,
const leveldb_writebatch_t*, void* state,
void (*put)(void*, const char* k, size_t klen, const char* v, size_t vlen),
void (*deleted)(void*, const char* k, size_t klen));
LEVELDB_EXPORT void leveldb_writebatch_append(
leveldb_writebatch_t* destination, const leveldb_writebatch_t* source);
/* Options */
LEVELDB_EXPORT leveldb_options_t* leveldb_options_create();
LEVELDB_EXPORT leveldb_options_t* leveldb_options_create(void);
LEVELDB_EXPORT void leveldb_options_destroy(leveldb_options_t*);
LEVELDB_EXPORT void leveldb_options_set_comparator(leveldb_options_t*,
leveldb_comparator_t*);
LEVELDB_EXPORT void leveldb_options_set_filter_policy(leveldb_options_t*,
leveldb_filterpolicy_t*);
LEVELDB_EXPORT void leveldb_options_set_create_if_missing(leveldb_options_t*,
unsigned char);
uint8_t);
LEVELDB_EXPORT void leveldb_options_set_error_if_exists(leveldb_options_t*,
unsigned char);
uint8_t);
LEVELDB_EXPORT void leveldb_options_set_paranoid_checks(leveldb_options_t*,
unsigned char);
uint8_t);
LEVELDB_EXPORT void leveldb_options_set_env(leveldb_options_t*, leveldb_env_t*);
LEVELDB_EXPORT void leveldb_options_set_info_log(leveldb_options_t*,
leveldb_logger_t*);
@ -187,10 +190,7 @@ LEVELDB_EXPORT void leveldb_options_set_block_restart_interval(
LEVELDB_EXPORT void leveldb_options_set_max_file_size(leveldb_options_t*,
size_t);
enum {
leveldb_no_compression = 0,
leveldb_snappy_compression = 1
};
enum { leveldb_no_compression = 0, leveldb_snappy_compression = 1 };
LEVELDB_EXPORT void leveldb_options_set_compression(leveldb_options_t*, int);
/* Comparator */
@ -209,8 +209,8 @@ LEVELDB_EXPORT leveldb_filterpolicy_t* leveldb_filterpolicy_create(
char* (*create_filter)(void*, const char* const* key_array,
const size_t* key_length_array, int num_keys,
size_t* filter_length),
unsigned char (*key_may_match)(void*, const char* key, size_t length,
const char* filter, size_t filter_length),
uint8_t (*key_may_match)(void*, const char* key, size_t length,
const char* filter, size_t filter_length),
const char* (*name)(void*));
LEVELDB_EXPORT void leveldb_filterpolicy_destroy(leveldb_filterpolicy_t*);
@ -219,21 +219,21 @@ LEVELDB_EXPORT leveldb_filterpolicy_t* leveldb_filterpolicy_create_bloom(
/* Read options */
LEVELDB_EXPORT leveldb_readoptions_t* leveldb_readoptions_create();
LEVELDB_EXPORT leveldb_readoptions_t* leveldb_readoptions_create(void);
LEVELDB_EXPORT void leveldb_readoptions_destroy(leveldb_readoptions_t*);
LEVELDB_EXPORT void leveldb_readoptions_set_verify_checksums(
leveldb_readoptions_t*, unsigned char);
leveldb_readoptions_t*, uint8_t);
LEVELDB_EXPORT void leveldb_readoptions_set_fill_cache(leveldb_readoptions_t*,
unsigned char);
uint8_t);
LEVELDB_EXPORT void leveldb_readoptions_set_snapshot(leveldb_readoptions_t*,
const leveldb_snapshot_t*);
/* Write options */
LEVELDB_EXPORT leveldb_writeoptions_t* leveldb_writeoptions_create();
LEVELDB_EXPORT leveldb_writeoptions_t* leveldb_writeoptions_create(void);
LEVELDB_EXPORT void leveldb_writeoptions_destroy(leveldb_writeoptions_t*);
LEVELDB_EXPORT void leveldb_writeoptions_set_sync(leveldb_writeoptions_t*,
unsigned char);
uint8_t);
/* Cache */
@ -242,7 +242,7 @@ LEVELDB_EXPORT void leveldb_cache_destroy(leveldb_cache_t* cache);
/* Env */
LEVELDB_EXPORT leveldb_env_t* leveldb_create_default_env();
LEVELDB_EXPORT leveldb_env_t* leveldb_create_default_env(void);
LEVELDB_EXPORT void leveldb_env_destroy(leveldb_env_t*);
/* If not NULL, the returned buffer must be released using leveldb_free(). */
@ -258,13 +258,13 @@ LEVELDB_EXPORT char* leveldb_env_get_test_directory(leveldb_env_t*);
LEVELDB_EXPORT void leveldb_free(void* ptr);
/* Return the major version number for this release. */
LEVELDB_EXPORT int leveldb_major_version();
LEVELDB_EXPORT int leveldb_major_version(void);
/* Return the minor version number for this release. */
LEVELDB_EXPORT int leveldb_minor_version();
LEVELDB_EXPORT int leveldb_minor_version(void);
#ifdef __cplusplus
} /* end extern "C" */
} /* end extern "C" */
#endif
#endif /* STORAGE_LEVELDB_INCLUDE_C_H_ */
#endif /* STORAGE_LEVELDB_INCLUDE_C_H_ */

13
include/leveldb/cache.h
View File

@ -18,7 +18,8 @@
#ifndef STORAGE_LEVELDB_INCLUDE_CACHE_H_
#define STORAGE_LEVELDB_INCLUDE_CACHE_H_
#include <stdint.h>
#include <cstdint>
#include "leveldb/export.h"
#include "leveldb/slice.h"
@ -42,7 +43,7 @@ class LEVELDB_EXPORT Cache {
virtual ~Cache();
// Opaque handle to an entry stored in the cache.
struct Handle { };
struct Handle {};
// Insert a mapping from key->value into the cache and assign it
// the specified charge against the total cache capacity.
@ -95,14 +96,6 @@ class LEVELDB_EXPORT Cache {
// Return an estimate of the combined charges of all elements stored in the
// cache.
virtual size_t TotalCharge() const = 0;
private:
void LRU_Remove(Handle* e);
void LRU_Append(Handle* e);
void Unref(Handle* e);
struct Rep;
Rep* rep_;
};
} // namespace leveldb

6
include/leveldb/comparator.h
View File

@ -6,6 +6,7 @@
#define STORAGE_LEVELDB_INCLUDE_COMPARATOR_H_
#include <string>
#include "leveldb/export.h"
namespace leveldb {
@ -44,9 +45,8 @@ class LEVELDB_EXPORT Comparator {
// If *start < limit, changes *start to a short string in [start,limit).
// Simple comparator implementations may return with *start unchanged,
// i.e., an implementation of this method that does nothing is correct.
virtual void FindShortestSeparator(
std::string* start,
const Slice& limit) const = 0;
virtual void FindShortestSeparator(std::string* start,
const Slice& limit) const = 0;
// Changes *key to a short string >= *key.
// Simple comparator implementations may return with *key unchanged,

27
include/leveldb/db.h
View File

@ -5,17 +5,18 @@
#ifndef STORAGE_LEVELDB_INCLUDE_DB_H_
#define STORAGE_LEVELDB_INCLUDE_DB_H_
#include <stdint.h>
#include <stdio.h>
#include <cstdint>
#include <cstdio>
#include "leveldb/export.h"
#include "leveldb/iterator.h"
#include "leveldb/options.h"
namespace leveldb {
// Update Makefile if you change these
// Update CMakeLists.txt if you change these
static const int kMajorVersion = 1;
static const int kMinorVersion = 20;
static const int kMinorVersion = 23;
struct Options;
struct ReadOptions;
@ -32,11 +33,11 @@ class LEVELDB_EXPORT Snapshot {
// A range of keys
struct LEVELDB_EXPORT Range {
Slice start; // Included in the range
Slice limit; // Not included in the range
Range() = default;
Range(const Slice& s, const Slice& l) : start(s), limit(l) {}
Range() { }
Range(const Slice& s, const Slice& l) : start(s), limit(l) { }
Slice start; // Included in the range
Slice limit; // Not included in the range
};
// A DB is a persistent ordered map from keys to values.
@ -49,8 +50,7 @@ class LEVELDB_EXPORT DB {
// OK on success.
// Stores nullptr in *dbptr and returns a non-OK status on error.
// Caller should delete *dbptr when it is no longer needed.
static Status Open(const Options& options,
const std::string& name,
static Status Open(const Options& options, const std::string& name,
DB** dbptr);
DB() = default;
@ -63,8 +63,7 @@ class LEVELDB_EXPORT DB {
// Set the database entry for "key" to "value". Returns OK on success,
// and a non-OK status on error.
// Note: consider setting options.sync = true.
virtual Status Put(const WriteOptions& options,
const Slice& key,
virtual Status Put(const WriteOptions& options, const Slice& key,
const Slice& value) = 0;
// Remove the database entry (if any) for "key". Returns OK on
@ -85,8 +84,8 @@ class LEVELDB_EXPORT DB {
// a status for which Status::IsNotFound() returns true.
//
// May return some other Status on an error.
virtual Status Get(const ReadOptions& options,
const Slice& key, std::string* value) = 0;
virtual Status Get(const ReadOptions& options, const Slice& key,
std::string* value) = 0;
// Return a heap-allocated iterator over the contents of the database.
// The result of NewIterator() is initially invalid (caller must

1
include/leveldb/dumpfile.h
View File

@ -6,6 +6,7 @@
#define STORAGE_LEVELDB_INCLUDE_DUMPFILE_H_
#include <string>
#include "leveldb/env.h"
#include "leveldb/export.h"
#include "leveldb/status.h"

106
include/leveldb/env.h
View File

@ -13,13 +13,32 @@
#ifndef STORAGE_LEVELDB_INCLUDE_ENV_H_
#define STORAGE_LEVELDB_INCLUDE_ENV_H_
#include <stdarg.h>
#include <stdint.h>
#include <cstdarg>
#include <cstdint>
#include <string>
#include <vector>
#include "leveldb/export.h"
#include "leveldb/status.h"
// This workaround can be removed when leveldb::Env::DeleteFile is removed.
#if defined(_WIN32)
// On Windows, the method name DeleteFile (below) introduces the risk of
// triggering undefined behavior by exposing the compiler to different
// declarations of the Env class in different translation units.
//
// This is because <windows.h>, a fairly popular header file for Windows
// applications, defines a DeleteFile macro. So, files that include the Windows
// header before this header will contain an altered Env declaration.
//
// This workaround ensures that the compiler sees the same Env declaration,
// independently of whether <windows.h> was included.
#if defined(DeleteFile)
#undef DeleteFile
#define LEVELDB_DELETEFILE_UNDEFINED
#endif // defined(DeleteFile)
#endif // defined(_WIN32)
namespace leveldb {
class FileLock;
@ -31,7 +50,7 @@ class WritableFile;
class LEVELDB_EXPORT Env {
public:
Env() = default;
Env();
Env(const Env&) = delete;
Env& operator=(const Env&) = delete;
@ -45,7 +64,7 @@ class LEVELDB_EXPORT Env {
// The result of Default() belongs to leveldb and must never be deleted.
static Env* Default();
// Create a brand new sequentially-readable file with the specified name.
// Create an object that sequentially reads the file with the specified name.
// On success, stores a pointer to the new file in *result and returns OK.
// On failure stores nullptr in *result and returns non-OK. If the file does
// not exist, returns a non-OK status. Implementations should return a
@ -55,7 +74,7 @@ class LEVELDB_EXPORT Env {
virtual Status NewSequentialFile(const std::string& fname,
SequentialFile** result) = 0;
// Create a brand new random access read-only file with the
// Create an object supporting random-access reads from the file with the
// specified name. On success, stores a pointer to the new file in
// *result and returns OK. On failure stores nullptr in *result and
// returns non-OK. If the file does not exist, returns a non-OK
@ -99,15 +118,48 @@ class LEVELDB_EXPORT Env {
// Original contents of *results are dropped.
virtual Status GetChildren(const std::string& dir,
std::vector<std::string>* result) = 0;
// Delete the named file.
virtual Status DeleteFile(const std::string& fname) = 0;
//
// The default implementation calls DeleteFile, to support legacy Env
// implementations. Updated Env implementations must override RemoveFile and
// ignore the existence of DeleteFile. Updated code calling into the Env API
// must call RemoveFile instead of DeleteFile.
//
// A future release will remove DeleteDir and the default implementation of
// RemoveDir.
virtual Status RemoveFile(const std::string& fname);
// DEPRECATED: Modern Env implementations should override RemoveFile instead.
//
// The default implementation calls RemoveFile, to support legacy Env user
// code that calls this method on modern Env implementations. Modern Env user
// code should call RemoveFile.
//
// A future release will remove this method.
virtual Status DeleteFile(const std::string& fname);
// Create the specified directory.
virtual Status CreateDir(const std::string& dirname) = 0;
// Delete the specified directory.
virtual Status DeleteDir(const std::string& dirname) = 0;
//
// The default implementation calls DeleteDir, to support legacy Env
// implementations. Updated Env implementations must override RemoveDir and
// ignore the existence of DeleteDir. Modern code calling into the Env API
// must call RemoveDir instead of DeleteDir.
//
// A future release will remove DeleteDir and the default implementation of
// RemoveDir.
virtual Status RemoveDir(const std::string& dirname);
// DEPRECATED: Modern Env implementations should override RemoveDir instead.
//
// The default implementation calls RemoveDir, to support legacy Env user
// code that calls this method on modern Env implementations. Modern Env user
// code should call RemoveDir.
//
// A future release will remove this method.
virtual Status DeleteDir(const std::string& dirname);
// Store the size of fname in *file_size.
virtual Status GetFileSize(const std::string& fname, uint64_t* file_size) = 0;
@ -143,16 +195,14 @@ class LEVELDB_EXPORT Env {
// added to the same Env may run concurrently in different threads.
// I.e., the caller may not assume that background work items are
// serialized.
virtual void Schedule(
void (*function)(void* arg),
void* arg) = 0;
virtual void Schedule(void (*function)(void* arg), void* arg) = 0;
// Start a new thread, invoking "function(arg)" within the new thread.
// When "function(arg)" returns, the thread will be destroyed.
virtual void StartThread(void (*function)(void* arg), void* arg) = 0;
// *path is set to a temporary directory that can be used for testing. It may
// or many not have just been created. The directory may or may not differ
// or may not have just been created. The directory may or may not differ
// between runs of the same process, but subsequent calls will return the
// same directory.
virtual Status GetTestDirectory(std::string* path) = 0;
@ -250,7 +300,7 @@ class LEVELDB_EXPORT Logger {
virtual ~Logger();
// Write an entry to the log file with the specified format.
virtual void Logv(const char* format, va_list ap) = 0;
virtual void Logv(const char* format, std::va_list ap) = 0;
};
// Identifies a locked file.
@ -266,9 +316,9 @@ class LEVELDB_EXPORT FileLock {
// Log the specified data to *info_log if info_log is non-null.
void Log(Logger* info_log, const char* format, ...)
# if defined(__GNUC__) || defined(__clang__)
__attribute__((__format__ (__printf__, 2, 3)))
# endif
#if defined(__GNUC__) || defined(__clang__)
__attribute__((__format__(__printf__, 2, 3)))
#endif
;
// A utility routine: write "data" to the named file.
@ -285,7 +335,7 @@ LEVELDB_EXPORT Status ReadFileToString(Env* env, const std::string& fname,
class LEVELDB_EXPORT EnvWrapper : public Env {
public:
// Initialize an EnvWrapper that delegates all calls to *t.
explicit EnvWrapper(Env* t) : target_(t) { }
explicit EnvWrapper(Env* t) : target_(t) {}
virtual ~EnvWrapper();
// Return the target to which this Env forwards all calls.
@ -312,14 +362,14 @@ class LEVELDB_EXPORT EnvWrapper : public Env {
std::vector<std::string>* r) override {
return target_->GetChildren(dir, r);
}
Status DeleteFile(const std::string& f) override {
return target_->DeleteFile(f);
Status RemoveFile(const std::string& f) override {
return target_->RemoveFile(f);
}
Status CreateDir(const std::string& d) override {
return target_->CreateDir(d);
}
Status DeleteDir(const std::string& d) override {
return target_->DeleteDir(d);
Status RemoveDir(const std::string& d) override {
return target_->RemoveDir(d);
}
Status GetFileSize(const std::string& f, uint64_t* s) override {
return target_->GetFileSize(f, s);
@ -343,9 +393,7 @@ class LEVELDB_EXPORT EnvWrapper : public Env {
Status NewLogger(const std::string& fname, Logger** result) override {
return target_->NewLogger(fname, result);
}
uint64_t NowMicros() override {
return target_->NowMicros();
}
uint64_t NowMicros() override { return target_->NowMicros(); }
void SleepForMicroseconds(int micros) override {
target_->SleepForMicroseconds(micros);
}
@ -356,4 +404,14 @@ class LEVELDB_EXPORT EnvWrapper : public Env {
} // namespace leveldb
// This workaround can be removed when leveldb::Env::DeleteFile is removed.
// Redefine DeleteFile if it was undefined earlier.
#if defined(_WIN32) && defined(LEVELDB_DELETEFILE_UNDEFINED)
#if defined(UNICODE)
#define DeleteFile DeleteFileW
#else
#define DeleteFile DeleteFileA
#endif // defined(UNICODE)
#endif // defined(_WIN32) && defined(LEVELDB_DELETEFILE_UNDEFINED)
#endif // STORAGE_LEVELDB_INCLUDE_ENV_H_

5
include/leveldb/filter_policy.h
View File

@ -17,6 +17,7 @@
#define STORAGE_LEVELDB_INCLUDE_FILTER_POLICY_H_
#include <string>
#include "leveldb/export.h"
namespace leveldb {
@ -39,8 +40,8 @@ class LEVELDB_EXPORT FilterPolicy {
//
// Warning: do not change the initial contents of *dst. Instead,
// append the newly constructed filter to *dst.
virtual void CreateFilter(const Slice* keys, int n, std::string* dst)
const = 0;
virtual void CreateFilter(const Slice* keys, int n,
std::string* dst) const = 0;
// "filter" contains the data appended by a preceding call to
// CreateFilter() on this class. This method must return true if

19
include/leveldb/iterator.h
View File

@ -77,17 +77,28 @@ class LEVELDB_EXPORT Iterator {
//
// Note that unlike all of the preceding methods, this method is
// not abstract and therefore clients should not override it.
typedef void (*CleanupFunction)(void* arg1, void* arg2);
using CleanupFunction = void (*)(void* arg1, void* arg2);
void RegisterCleanup(CleanupFunction function, void* arg1, void* arg2);
private:
struct Cleanup {
// Cleanup functions are stored in a single-linked list.
// The list's head node is inlined in the iterator.
struct CleanupNode {
// True if the node is not used. Only head nodes might be unused.
bool IsEmpty() const { return function == nullptr; }
// Invokes the cleanup function.
void Run() {
assert(function != nullptr);
(*function)(arg1, arg2);
}
// The head node is used if the function pointer is not null.
CleanupFunction function;
void* arg1;
void* arg2;
Cleanup* next;
CleanupNode* next;
};
Cleanup cleanup_;
CleanupNode cleanup_head_;
};
// Return an empty iterator (yields nothing).

81
include/leveldb/options.h
View File

@ -5,7 +5,8 @@
#ifndef STORAGE_LEVELDB_INCLUDE_OPTIONS_H_
#define STORAGE_LEVELDB_INCLUDE_OPTIONS_H_
#include <stddef.h>
#include <cstddef>
#include "leveldb/export.h"
namespace leveldb {
@ -24,12 +25,15 @@ class Snapshot;
enum CompressionType {
// NOTE: do not change the values of existing entries, as these are
// part of the persistent format on disk.
kNoCompression = 0x0,
kNoCompression = 0x0,
kSnappyCompression = 0x1
};
// Options to control the behavior of a database (passed to DB::Open)
struct LEVELDB_EXPORT Options {
// Create an Options object with default values for all fields.
Options();
// -------------------
// Parameters that affect behavior
@ -42,20 +46,17 @@ struct LEVELDB_EXPORT Options {
const Comparator* comparator;
// If true, the database will be created if it is missing.
// Default: false
bool create_if_missing;
bool create_if_missing = false;
// If true, an error is raised if the database already exists.
// Default: false
bool error_if_exists;
bool error_if_exists = false;
// If true, the implementation will do aggressive checking of the
// data it is processing and will stop early if it detects any
// errors. This may have unforeseen ramifications: for example, a
// corruption of one DB entry may cause a large number of entries to
// become unreadable or for the entire DB to become unopenable.
// Default: false
bool paranoid_checks;
bool paranoid_checks = false;
// Use the specified object to interact with the environment,
// e.g. to read/write files, schedule background work, etc.
@ -65,8 +66,7 @@ struct LEVELDB_EXPORT Options {
// Any internal progress/error information generated by the db will
// be written to info_log if it is non-null, or to a file stored
// in the same directory as the DB contents if info_log is null.
// Default: nullptr
Logger* info_log;
Logger* info_log = nullptr;
// -------------------
// Parameters that affect performance
@ -79,39 +79,30 @@ struct LEVELDB_EXPORT Options {
// so you may wish to adjust this parameter to control memory usage.
// Also, a larger write buffer will result in a longer recovery time
// the next time the database is opened.
//
// Default: 4MB
size_t write_buffer_size;
size_t write_buffer_size = 4 * 1024 * 1024;
// Number of open files that can be used by the DB. You may need to
// increase this if your database has a large working set (budget
// one open file per 2MB of working set).
//
// Default: 1000
int max_open_files;
int max_open_files = 1000;
// Control over blocks (user data is stored in a set of blocks, and
// a block is the unit of reading from disk).
// If non-null, use the specified cache for blocks.
// If null, leveldb will automatically create and use an 8MB internal cache.
// Default: nullptr
Cache* block_cache;
Cache* block_cache = nullptr;
// Approximate size of user data packed per block. Note that the
// block size specified here corresponds to uncompressed data. The
// actual size of the unit read from disk may be smaller if
// compression is enabled. This parameter can be changed dynamically.
//
// Default: 4K
size_t block_size;
size_t block_size = 4 * 1024;
// Number of keys between restart points for delta encoding of keys.
// This parameter can be changed dynamically. Most clients should
// leave this parameter alone.
//
// Default: 16
int block_restart_interval;
int block_restart_interval = 16;
// Leveldb will write up to this amount of bytes to a file before
// switching to a new one.
@ -121,9 +112,7 @@ struct LEVELDB_EXPORT Options {
// compactions and hence longer latency/performance hiccups.
// Another reason to increase this parameter might be when you are
// initially populating a large database.
//
// Default: 2MB
size_t max_file_size;
size_t max_file_size = 2 * 1024 * 1024;
// Compress blocks using the specified compression algorithm. This
// parameter can be changed dynamically.
@ -139,53 +128,43 @@ struct LEVELDB_EXPORT Options {
// worth switching to kNoCompression. Even if the input data is
// incompressible, the kSnappyCompression implementation will
// efficiently detect that and will switch to uncompressed mode.
CompressionType compression;
CompressionType compression = kSnappyCompression;
// EXPERIMENTAL: If true, append to existing MANIFEST and log files
// when a database is opened. This can significantly speed up open.
//
// Default: currently false, but may become true later.
bool reuse_logs;
bool reuse_logs = false;
// If non-null, use the specified filter policy to reduce disk reads.
// Many applications will benefit from passing the result of
// NewBloomFilterPolicy() here.
//
// Default: nullptr
const FilterPolicy* filter_policy;
// Create an Options object with default values for all fields.
Options();
const FilterPolicy* filter_policy = nullptr;
};
// Options that control read operations
struct LEVELDB_EXPORT ReadOptions {
ReadOptions() = default;
// If true, all data read from underlying storage will be
// verified against corresponding checksums.
// Default: false
bool verify_checksums;
bool verify_checksums = false;
// Should the data read for this iteration be cached in memory?
// Callers may wish to set this field to false for bulk scans.
// Default: true
bool fill_cache;
bool fill_cache = true;
// If "snapshot" is non-null, read as of the supplied snapshot
// (which must belong to the DB that is being read and which must
// not have been released). If "snapshot" is null, use an implicit
// snapshot of the state at the beginning of this read operation.
// Default: nullptr
const Snapshot* snapshot;
ReadOptions()
: verify_checksums(false),
fill_cache(true),
snapshot(nullptr) {
}
const Snapshot* snapshot = nullptr;
};
// Options that control write operations
struct LEVELDB_EXPORT WriteOptions {
WriteOptions() = default;
// If true, the write will be flushed from the operating system
// buffer cache (by calling WritableFile::Sync()) before the write
// is considered complete. If this flag is true, writes will be
@ -200,13 +179,7 @@ struct LEVELDB_EXPORT WriteOptions {
// crash semantics as the "write()" system call. A DB write
// with sync==true has similar crash semantics to a "write()"
// system call followed by "fsync()".
//
// Default: false
bool sync;
WriteOptions()
: sync(false) {
}
bool sync = false;
};
} // namespace leveldb

34
include/leveldb/slice.h
View File

@ -15,10 +15,11 @@
#ifndef STORAGE_LEVELDB_INCLUDE_SLICE_H_
#define STORAGE_LEVELDB_INCLUDE_SLICE_H_
#include <assert.h>
#include <stddef.h>
#include <string.h>
#include <cassert>
#include <cstddef>
#include <cstring>
#include <string>
#include "leveldb/export.h"
namespace leveldb {
@ -26,16 +27,16 @@ namespace leveldb {
class LEVELDB_EXPORT Slice {
public:
// Create an empty slice.
Slice() : data_(""), size_(0) { }
Slice() : data_(""), size_(0) {}
// Create a slice that refers to d[0,n-1].
Slice(const char* d, size_t n) : data_(d), size_(n) { }
Slice(const char* d, size_t n) : data_(d), size_(n) {}
// Create a slice that refers to the contents of "s"
Slice(const std::string& s) : data_(s.data()), size_(s.size()) { }
Slice(const std::string& s) : data_(s.data()), size_(s.size()) {}
// Create a slice that refers to s[0,strlen(s)-1]
Slice(const char* s) : data_(s), size_(strlen(s)) { }
Slice(const char* s) : data_(s), size_(strlen(s)) {}
// Intentionally copyable.
Slice(const Slice&) = default;
@ -58,7 +59,10 @@ class LEVELDB_EXPORT Slice {
}
// Change this slice to refer to an empty array
void clear() { data_ = ""; size_ = 0; }
void clear() {
data_ = "";
size_ = 0;
}
// Drop the first "n" bytes from this slice.
void remove_prefix(size_t n) {
@ -78,8 +82,7 @@ class LEVELDB_EXPORT Slice {
// Return true iff "x" is a prefix of "*this"
bool starts_with(const Slice& x) const {
return ((size_ >= x.size_) &&
(memcmp(data_, x.data_, x.size_) == 0));
return ((size_ >= x.size_) && (memcmp(data_, x.data_, x.size_) == 0));
}
private:
@ -92,21 +95,20 @@ inline bool operator==(const Slice& x, const Slice& y) {
(memcmp(x.data(), y.data(), x.size()) == 0));
}
inline bool operator!=(const Slice& x, const Slice& y) {
return !(x == y);
}
inline bool operator!=(const Slice& x, const Slice& y) { return !(x == y); }
inline int Slice::compare(const Slice& b) const {
const size_t min_len = (size_ < b.size_) ? size_ : b.size_;
int r = memcmp(data_, b.data_, min_len);
if (r == 0) {
if (size_ < b.size_) r = -1;
else if (size_ > b.size_) r = +1;
if (size_ < b.size_)
r = -1;
else if (size_ > b.size_)
r = +1;
}
return r;
}
} // namespace leveldb
#endif // STORAGE_LEVELDB_INCLUDE_SLICE_H_

45
include/leveldb/status.h
View File

@ -13,7 +13,9 @@
#ifndef STORAGE_LEVELDB_INCLUDE_STATUS_H_
#define STORAGE_LEVELDB_INCLUDE_STATUS_H_
#include <algorithm>
#include <string>
#include "leveldb/export.h"
#include "leveldb/slice.h"
@ -22,12 +24,14 @@ namespace leveldb {
class LEVELDB_EXPORT Status {
public:
// Create a success status.
Status() : state_(nullptr) { }
Status() noexcept : state_(nullptr) {}
~Status() { delete[] state_; }
// Copy the specified status.
Status(const Status& s);
void operator=(const Status& s);
Status(const Status& rhs);
Status& operator=(const Status& rhs);
Status(Status&& rhs) noexcept : state_(rhs.state_) { rhs.state_ = nullptr; }
Status& operator=(Status&& rhs) noexcept;
// Return a success status.
static Status OK() { return Status(); }
@ -72,13 +76,6 @@ class LEVELDB_EXPORT Status {
std::string ToString() const;
private:
// OK status has a null state_. Otherwise, state_ is a new[] array
// of the following form:
// state_[0..3] == length of message
// state_[4] == code
// state_[5..] == message
const char* state_;
enum Code {
kOk = 0,
kNotFound = 1,
@ -94,18 +91,30 @@ class LEVELDB_EXPORT Status {
Status(Code code, const Slice& msg, const Slice& msg2);
static const char* CopyState(const char* s);
// OK status has a null state_. Otherwise, state_ is a new[] array
// of the following form:
// state_[0..3] == length of message
// state_[4] == code
// state_[5..] == message
const char* state_;
};
inline Status::Status(const Status& s) {
state_ = (s.state_ == nullptr) ? nullptr : CopyState(s.state_);
inline Status::Status(const Status& rhs) {
state_ = (rhs.state_ == nullptr) ? nullptr : CopyState(rhs.state_);
}
inline void Status::operator=(const Status& s) {
// The following condition catches both aliasing (when this == &s),
// and the common case where both s and *this are ok.
if (state_ != s.state_) {
inline Status& Status::operator=(const Status& rhs) {
// The following condition catches both aliasing (when this == &rhs),
// and the common case where both rhs and *this are ok.
if (state_ != rhs.state_) {
delete[] state_;
state_ = (s.state_ == nullptr) ? nullptr : CopyState(s.state_);
state_ = (rhs.state_ == nullptr) ? nullptr : CopyState(rhs.state_);
}
return *this;
}
inline Status& Status::operator=(Status&& rhs) noexcept {
std::swap(state_, rhs.state_);
return *this;
}
} // namespace leveldb

27
include/leveldb/table.h
View File

@ -5,7 +5,8 @@
#ifndef STORAGE_LEVELDB_INCLUDE_TABLE_H_
#define STORAGE_LEVELDB_INCLUDE_TABLE_H_
#include <stdint.h>
#include <cstdint>
#include "leveldb/export.h"
#include "leveldb/iterator.h"
@ -36,13 +37,11 @@ class LEVELDB_EXPORT Table {
// for the duration of the returned table's lifetime.
//
// *file must remain live while this Table is in use.
static Status Open(const Options& options,
RandomAccessFile* file,
uint64_t file_size,
Table** table);
static Status Open(const Options& options, RandomAccessFile* file,
uint64_t file_size, Table** table);
Table(const Table&) = delete;
void operator=(const Table&) = delete;
Table& operator=(const Table&) = delete;
~Table();
@ -60,24 +59,24 @@ class LEVELDB_EXPORT Table {
uint64_t ApproximateOffsetOf(const Slice& key) const;
private:
friend class TableCache;
struct Rep;
Rep* rep_;
explicit Table(Rep* rep) { rep_ = rep; }
static Iterator* BlockReader(void*, const ReadOptions&, const Slice&);
explicit Table(Rep* rep) : rep_(rep) {}
// Calls (*handle_result)(arg, ...) with the entry found after a call
// to Seek(key). May not make such a call if filter policy says
// that key is not present.
friend class TableCache;
Status InternalGet(
const ReadOptions&, const Slice& key,
void* arg,
void (*handle_result)(void* arg, const Slice& k, const Slice& v));
Status InternalGet(const ReadOptions&, const Slice& key, void* arg,
void (*handle_result)(void* arg, const Slice& k,
const Slice& v));
void ReadMeta(const Footer& footer);
void ReadFilter(const Slice& filter_handle_value);
Rep* const rep_;
};
} // namespace leveldb

5
include/leveldb/table_builder.h
View File

@ -13,7 +13,8 @@
#ifndef STORAGE_LEVELDB_INCLUDE_TABLE_BUILDER_H_
#define STORAGE_LEVELDB_INCLUDE_TABLE_BUILDER_H_
#include <stdint.h>
#include <cstdint>
#include "leveldb/export.h"
#include "leveldb/options.h"
#include "leveldb/status.h"
@ -32,7 +33,7 @@ class LEVELDB_EXPORT TableBuilder {
TableBuilder(const Options& options, WritableFile* file);
TableBuilder(const TableBuilder&) = delete;
void operator=(const TableBuilder&) = delete;
TableBuilder& operator=(const TableBuilder&) = delete;
// REQUIRES: Either Finish() or Abandon() has been called.
~TableBuilder();

25
include/leveldb/write_batch.h
View File

@ -22,6 +22,7 @@
#define STORAGE_LEVELDB_INCLUDE_WRITE_BATCH_H_
#include <string>
#include "leveldb/export.h"
#include "leveldb/status.h"
@ -31,11 +32,18 @@ class Slice;
class LEVELDB_EXPORT WriteBatch {
public:
class LEVELDB_EXPORT Handler {
public:
virtual ~Handler();
virtual void Put(const Slice& key, const Slice& value) = 0;
virtual void Delete(const Slice& key) = 0;
};
WriteBatch();
// Intentionally copyable.
WriteBatch(const WriteBatch&) = default;
WriteBatch& operator =(const WriteBatch&) = default;
WriteBatch& operator=(const WriteBatch&) = default;
~WriteBatch();
@ -52,15 +60,16 @@ class LEVELDB_EXPORT WriteBatch {
//
// This number is tied to implementation details, and may change across
// releases. It is intended for LevelDB usage metrics.
size_t ApproximateSize();
size_t ApproximateSize() const;
// Copies the operations in "source" to this batch.
//
// This runs in O(source size) time. However, the constant factor is better
// than calling Iterate() over the source batch with a Handler that replicates
// the operations into this batch.
void Append(const WriteBatch& source);
// Support for iterating over the contents of a batch.
class Handler {
public:
virtual ~Handler();
virtual void Put(const Slice& key, const Slice& value) = 0;
virtual void Delete(const Slice& key) = 0;
};
Status Iterate(Handler* handler) const;
private:

27
issues/issue178_test.cc
View File

@ -3,13 +3,14 @@
// found in the LICENSE file. See the AUTHORS file for names of contributors.
// Test for issue 178: a manual compaction causes deleted data to reappear.
#include <cstdlib>
#include <iostream>
#include <sstream>
#include <cstdlib>
#include "gtest/gtest.h"
#include "leveldb/db.h"
#include "leveldb/write_batch.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace {
@ -17,19 +18,15 @@ const int kNumKeys = 1100000;
std::string Key1(int i) {
char buf[100];
snprintf(buf, sizeof(buf), "my_key_%d", i);
std::snprintf(buf, sizeof(buf), "my_key_%d", i);
return buf;
}
std::string Key2(int i) {
return Key1(i) + "_xxx";
}
class Issue178 { };
std::string Key2(int i) { return Key1(i) + "_xxx"; }
TEST(Issue178, Test) {
// Get rid of any state from an old run.
std::string dbpath = leveldb::test::TmpDir() + "/leveldb_cbug_test";
std::string dbpath = testing::TempDir() + "leveldb_cbug_test";
DestroyDB(dbpath, leveldb::Options());
// Open database. Disable compression since it affects the creation
@ -39,28 +36,28 @@ TEST(Issue178, Test) {
leveldb::Options db_options;
db_options.create_if_missing = true;
db_options.compression = leveldb::kNoCompression;
ASSERT_OK(leveldb::DB::Open(db_options, dbpath, &db));
ASSERT_LEVELDB_OK(leveldb::DB::Open(db_options, dbpath, &db));
// create first key range
leveldb::WriteBatch batch;
for (size_t i = 0; i < kNumKeys; i++) {
batch.Put(Key1(i), "value for range 1 key");
}
ASSERT_OK(db->Write(leveldb::WriteOptions(), &batch));
ASSERT_LEVELDB_OK(db->Write(leveldb::WriteOptions(), &batch));
// create second key range
batch.Clear();
for (size_t i = 0; i < kNumKeys; i++) {
batch.Put(Key2(i), "value for range 2 key");
}
ASSERT_OK(db->Write(leveldb::WriteOptions(), &batch));
ASSERT_LEVELDB_OK(db->Write(leveldb::WriteOptions(), &batch));
// delete second key range
batch.Clear();
for (size_t i = 0; i < kNumKeys; i++) {
batch.Delete(Key2(i));
}
ASSERT_OK(db->Write(leveldb::WriteOptions(), &batch));
ASSERT_LEVELDB_OK(db->Write(leveldb::WriteOptions(), &batch));
// compact database
std::string start_key = Key1(0);
@ -86,7 +83,3 @@ TEST(Issue178, Test) {
}
} // anonymous namespace
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

29
issues/issue200_test.cc
View File

@ -6,35 +6,34 @@
// to forward, the current key can be yielded unexpectedly if a new
// mutation has been added just before the current key.
#include "gtest/gtest.h"
#include "leveldb/db.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
class Issue200 { };
TEST(Issue200, Test) {
// Get rid of any state from an old run.
std::string dbpath = test::TmpDir() + "/leveldb_issue200_test";
std::string dbpath = testing::TempDir() + "leveldb_issue200_test";
DestroyDB(dbpath, Options());
DB *db;
DB* db;
Options options;
options.create_if_missing = true;
ASSERT_OK(DB::Open(options, dbpath, &db));
ASSERT_LEVELDB_OK(DB::Open(options, dbpath, &db));
WriteOptions write_options;
ASSERT_OK(db->Put(write_options, "1", "b"));
ASSERT_OK(db->Put(write_options, "2", "c"));
ASSERT_OK(db->Put(write_options, "3", "d"));
ASSERT_OK(db->Put(write_options, "4", "e"));
ASSERT_OK(db->Put(write_options, "5", "f"));
ASSERT_LEVELDB_OK(db->Put(write_options, "1", "b"));
ASSERT_LEVELDB_OK(db->Put(write_options, "2", "c"));
ASSERT_LEVELDB_OK(db->Put(write_options, "3", "d"));
ASSERT_LEVELDB_OK(db->Put(write_options, "4", "e"));
ASSERT_LEVELDB_OK(db->Put(write_options, "5", "f"));
ReadOptions read_options;
Iterator *iter = db->NewIterator(read_options);
Iterator* iter = db->NewIterator(read_options);
// Add an element that should not be reflected in the iterator.
ASSERT_OK(db->Put(write_options, "25", "cd"));
ASSERT_LEVELDB_OK(db->Put(write_options, "25", "cd"));
iter->Seek("5");
ASSERT_EQ(iter->key().ToString(), "5");
@ -53,7 +52,3 @@ TEST(Issue200, Test) {
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

126
issues/issue320_test.cc Normal file
View File

@ -0,0 +1,126 @@
// Copyright (c) 2019 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 <cstdint>
#include <cstdlib>
#include <iostream>
#include <memory>
#include <string>
#include <vector>
#include "gtest/gtest.h"
#include "leveldb/db.h"
#include "leveldb/write_batch.h"
#include "util/testutil.h"
namespace leveldb {
namespace {
// Creates a random number in the range of [0, max).
int GenerateRandomNumber(int max) { return std::rand() % max; }
std::string CreateRandomString(int32_t index) {
static const size_t len = 1024;
char bytes[len];
size_t i = 0;
while (i < 8) {
bytes[i] = 'a' + ((index >> (4 * i)) & 0xf);
++i;
}
while (i < sizeof(bytes)) {
bytes[i] = 'a' + GenerateRandomNumber(26);
++i;
}
return std::string(bytes, sizeof(bytes));
}
} // namespace
TEST(Issue320, Test) {
std::srand(0);
bool delete_before_put = false;
bool keep_snapshots = true;
std::vector<std::unique_ptr<std::pair<std::string, std::string>>> test_map(
10000);
std::vector<Snapshot const*> snapshots(100, nullptr);
DB* db;
Options options;
options.create_if_missing = true;
std::string dbpath = testing::TempDir() + "leveldb_issue320_test";
ASSERT_LEVELDB_OK(DB::Open(options, dbpath, &db));
uint32_t target_size = 10000;
uint32_t num_items = 0;
uint32_t count = 0;
std::string key;
std::string value, old_value;
WriteOptions writeOptions;
ReadOptions readOptions;
while (count < 200000) {
if ((++count % 1000) == 0) {
std::cout << "count: " << count << std::endl;
}
int index = GenerateRandomNumber(test_map.size());
WriteBatch batch;
if (test_map[index] == nullptr) {
num_items++;
test_map[index].reset(new std::pair<std::string, std::string>(
CreateRandomString(index), CreateRandomString(index)));
batch.Put(test_map[index]->first, test_map[index]->second);
} else {
ASSERT_LEVELDB_OK(
db->Get(readOptions, test_map[index]->first, &old_value));
if (old_value != test_map[index]->second) {
std::cout << "ERROR incorrect value returned by Get" << std::endl;
std::cout << " count=" << count << std::endl;
std::cout << " old value=" << old_value << std::endl;
std::cout << " test_map[index]->second=" << test_map[index]->second
<< std::endl;
std::cout << " test_map[index]->first=" << test_map[index]->first
<< std::endl;
std::cout << " index=" << index << std::endl;
ASSERT_EQ(old_value, test_map[index]->second);
}
if (num_items >= target_size && GenerateRandomNumber(100) > 30) {
batch.Delete(test_map[index]->first);
test_map[index] = nullptr;
--num_items;
} else {
test_map[index]->second = CreateRandomString(index);
if (delete_before_put) batch.Delete(test_map[index]->first);
batch.Put(test_map[index]->first, test_map[index]->second);
}
}
ASSERT_LEVELDB_OK(db->Write(writeOptions, &batch));
if (keep_snapshots && GenerateRandomNumber(10) == 0) {
int i = GenerateRandomNumber(snapshots.size());
if (snapshots[i] != nullptr) {
db->ReleaseSnapshot(snapshots[i]);
}
snapshots[i] = db->GetSnapshot();
}
}
for (Snapshot const* snapshot : snapshots) {
if (snapshot) {
db->ReleaseSnapshot(snapshot);
}
}
delete db;
DestroyDB(dbpath, options);
}
} // namespace leveldb

2
port/README → port/README.md
View File

@ -5,6 +5,6 @@ Code in the rest of the package includes "port.h" from this directory.
"port.h" in turn includes a platform specific "port_<platform>.h" file
that provides the platform specific implementation.
See port_posix.h for an example of what must be provided in a platform
See port_stdcxx.h for an example of what must be provided in a platform
specific header file.

175
port/atomic_pointer.h
View File

@ -1,175 +0,0 @@
// 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.
// AtomicPointer provides storage for a lock-free pointer.
// Platform-dependent implementation of AtomicPointer:
// - If the platform provides a cheap barrier, we use it with raw pointers
// - If <atomic> is present (on newer versions of gcc, it is), we use
// a <atomic>-based AtomicPointer. However we prefer the memory
// barrier based version, because at least on a gcc 4.4 32-bit build
// on linux, we have encountered a buggy <atomic> implementation.
// Also, some <atomic> implementations are much slower than a memory-barrier
// based implementation (~16ns for <atomic> based acquire-load vs. ~1ns for
// a barrier based acquire-load).
// This code is based on atomicops-internals-* in Google's perftools:
// http://code.google.com/p/google-perftools/source/browse/#svn%2Ftrunk%2Fsrc%2Fbase
#ifndef PORT_ATOMIC_POINTER_H_
#define PORT_ATOMIC_POINTER_H_
#include <stdint.h>
#include <atomic>
#ifdef OS_WIN
#include <windows.h>
#endif
#if defined(_M_X64) || defined(__x86_64__)
#define ARCH_CPU_X86_FAMILY 1
#elif defined(_M_IX86) || defined(__i386__) || defined(__i386)
#define ARCH_CPU_X86_FAMILY 1
#elif defined(__ARMEL__)
#define ARCH_CPU_ARM_FAMILY 1
#elif defined(__aarch64__)
#define ARCH_CPU_ARM64_FAMILY 1
#elif defined(__ppc__) || defined(__powerpc__) || defined(__powerpc64__)
#define ARCH_CPU_PPC_FAMILY 1
#elif defined(__mips__)
#define ARCH_CPU_MIPS_FAMILY 1
#endif
namespace leveldb {
namespace port {
// Define MemoryBarrier() if available
// Windows on x86
#if defined(OS_WIN) && defined(COMPILER_MSVC) && defined(ARCH_CPU_X86_FAMILY)
// windows.h already provides a MemoryBarrier(void) macro
// http://msdn.microsoft.com/en-us/library/ms684208(v=vs.85).aspx
#define LEVELDB_HAVE_MEMORY_BARRIER
// Mac OS
#elif defined(__APPLE__)
inline void MemoryBarrier() {
std::atomic_thread_fence(std::memory_order_seq_cst);
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// Gcc on x86
#elif defined(ARCH_CPU_X86_FAMILY) && defined(__GNUC__)
inline void MemoryBarrier() {
// See http://gcc.gnu.org/ml/gcc/2003-04/msg01180.html for a discussion on
// this idiom. Also see http://en.wikipedia.org/wiki/Memory_ordering.
__asm__ __volatile__("" : : : "memory");
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// Sun Studio
#elif defined(ARCH_CPU_X86_FAMILY) && defined(__SUNPRO_CC)
inline void MemoryBarrier() {
// See http://gcc.gnu.org/ml/gcc/2003-04/msg01180.html for a discussion on
// this idiom. Also see http://en.wikipedia.org/wiki/Memory_ordering.
asm volatile("" : : : "memory");
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// ARM Linux
#elif defined(ARCH_CPU_ARM_FAMILY) && defined(__linux__)
typedef void (*LinuxKernelMemoryBarrierFunc)(void);
// The Linux ARM kernel provides a highly optimized device-specific memory
// barrier function at a fixed memory address that is mapped in every
// user-level process.
//
// This beats using CPU-specific instructions which are, on single-core
// devices, un-necessary and very costly (e.g. ARMv7-A "dmb" takes more
// than 180ns on a Cortex-A8 like the one on a Nexus One). Benchmarking
// shows that the extra function call cost is completely negligible on
// multi-core devices.
//
inline void MemoryBarrier() {
(*(LinuxKernelMemoryBarrierFunc)0xffff0fa0)();
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// ARM64
#elif defined(ARCH_CPU_ARM64_FAMILY)
inline void MemoryBarrier() {
asm volatile("dmb sy" : : : "memory");
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// PPC
#elif defined(ARCH_CPU_PPC_FAMILY) && defined(__GNUC__)
inline void MemoryBarrier() {
// TODO for some powerpc expert: is there a cheaper suitable variant?
// Perhaps by having separate barriers for acquire and release ops.
asm volatile("sync" : : : "memory");
}
#define LEVELDB_HAVE_MEMORY_BARRIER
// MIPS
#elif defined(ARCH_CPU_MIPS_FAMILY) && defined(__GNUC__)
inline void MemoryBarrier() {
__asm__ __volatile__("sync" : : : "memory");
}
#define LEVELDB_HAVE_MEMORY_BARRIER
#endif
// AtomicPointer built using platform-specific MemoryBarrier().
#if defined(LEVELDB_HAVE_MEMORY_BARRIER)
class AtomicPointer {
private:
void* rep_;
public:
AtomicPointer() { }
explicit AtomicPointer(void* p) : rep_(p) {}
inline void* NoBarrier_Load() const { return rep_; }
inline void NoBarrier_Store(void* v) { rep_ = v; }
inline void* Acquire_Load() const {
void* result = rep_;
MemoryBarrier();
return result;
}
inline void Release_Store(void* v) {
MemoryBarrier();
rep_ = v;
}
};
// AtomicPointer based on C++11 <atomic>.
#else
class AtomicPointer {
private:
std::atomic<void*> rep_;
public:
AtomicPointer() { }
explicit AtomicPointer(void* v) : rep_(v) { }
inline void* Acquire_Load() const {
return rep_.load(std::memory_order_acquire);
}
inline void Release_Store(void* v) {
rep_.store(v, std::memory_order_release);
}
inline void* NoBarrier_Load() const {
return rep_.load(std::memory_order_relaxed);
}
inline void NoBarrier_Store(void* v) {
rep_.store(v, std::memory_order_relaxed);
}
};
#endif
#undef LEVELDB_HAVE_MEMORY_BARRIER
#undef ARCH_CPU_X86_FAMILY
#undef ARCH_CPU_ARM_FAMILY
#undef ARCH_CPU_ARM64_FAMILY
#undef ARCH_CPU_PPC_FAMILY
} // namespace port
} // namespace leveldb
#endif // PORT_ATOMIC_POINTER_H_

6
port/port.h
View File

@ -10,10 +10,10 @@
// Include the appropriate platform specific file below. If you are
// porting to a new platform, see "port_example.h" for documentation
// of what the new port_<platform>.h file must provide.
#if defined(LEVELDB_PLATFORM_POSIX)
# include "port/port_posix.h"
#if defined(LEVELDB_PLATFORM_POSIX) || defined(LEVELDB_PLATFORM_WINDOWS)
#include "port/port_stdcxx.h"
#elif defined(LEVELDB_PLATFORM_CHROMIUM)
# include "port/port_chromium.h"
#include "port/port_chromium.h"
#endif
#endif // STORAGE_LEVELDB_PORT_PORT_H_

22
port/port_config.h.in
View File

@ -6,9 +6,19 @@
#define STORAGE_LEVELDB_PORT_PORT_CONFIG_H_
// Define to 1 if you have a definition for fdatasync() in <unistd.h>.
#if !defined(HAVE_FUNC_FDATASYNC)
#cmakedefine01 HAVE_FUNC_FDATASYNC
#endif // !defined(HAVE_FUNC_FDATASYNC)
#if !defined(HAVE_FDATASYNC)
#cmakedefine01 HAVE_FDATASYNC
#endif // !defined(HAVE_FDATASYNC)
// Define to 1 if you have a definition for F_FULLFSYNC in <fcntl.h>.
#if !defined(HAVE_FULLFSYNC)
#cmakedefine01 HAVE_FULLFSYNC
#endif // !defined(HAVE_FULLFSYNC)
// Define to 1 if you have a definition for O_CLOEXEC in <fcntl.h>.
#if !defined(HAVE_O_CLOEXEC)
#cmakedefine01 HAVE_O_CLOEXEC
#endif // !defined(HAVE_O_CLOEXEC)
// Define to 1 if you have Google CRC32C.
#if !defined(HAVE_CRC32C)
@ -20,10 +30,4 @@
#cmakedefine01 HAVE_SNAPPY
#endif // !defined(HAVE_SNAPPY)
// Define to 1 if your processor stores words with the most significant byte
// first (like Motorola and SPARC, unlike Intel and VAX).
#if !defined(LEVELDB_IS_BIG_ENDIAN)
#cmakedefine01 LEVELDB_IS_BIG_ENDIAN
#endif // !defined(LEVELDB_IS_BIG_ENDIAN)
#endif // STORAGE_LEVELDB_PORT_PORT_CONFIG_H_

45
port/port_example.h
View File

@ -18,10 +18,6 @@ namespace port {
// TODO(jorlow): Many of these belong more in the environment class rather than
// here. We should try moving them and see if it affects perf.
// The following boolean constant must be true on a little-endian machine
// and false otherwise.
static const bool kLittleEndian = true /* or some other expression */;
// ------------------ Threading -------------------
// A Mutex represents an exclusive lock.
@ -59,46 +55,7 @@ class CondVar {
void Signal();
// Wake up all waiting threads.
void SignallAll();
};
// Thread-safe initialization.
// Used as follows:
// static port::OnceType init_control = LEVELDB_ONCE_INIT;
// static void Initializer() { ... do something ...; }
// ...
// port::InitOnce(&init_control, &Initializer);
typedef intptr_t OnceType;
#define LEVELDB_ONCE_INIT 0
void InitOnce(port::OnceType*, void (*initializer)());
// A type that holds a pointer that can be read or written atomically
// (i.e., without word-tearing.)
class AtomicPointer {
private:
intptr_t rep_;
public:
// Initialize to arbitrary value
AtomicPointer();
// Initialize to hold v
explicit AtomicPointer(void* v) : rep_(v) { }
// Read and return the stored pointer with the guarantee that no
// later memory access (read or write) by this thread can be
// reordered ahead of this read.
void* Acquire_Load() const;
// Set v as the stored pointer with the guarantee that no earlier
// memory access (read or write) by this thread can be reordered
// after this store.
void Release_Store(void* v);
// Read the stored pointer with no ordering guarantees.
void* NoBarrier_Load() const;
// Set va as the stored pointer with no ordering guarantees.
void NoBarrier_Store(void* v);
void SignalAll();
};
// ------------------ Compression -------------------

53
port/port_posix.cc
View File

@ -1,53 +0,0 @@
// 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 "port/port_posix.h"
#include <cstdlib>
#include <stdio.h>
#include <string.h>
namespace leveldb {
namespace port {
static void PthreadCall(const char* label, int result) {
if (result != 0) {
fprintf(stderr, "pthread %s: %s\n", label, strerror(result));
abort();
}
}
Mutex::Mutex() { PthreadCall("init mutex", pthread_mutex_init(&mu_, nullptr)); }
Mutex::~Mutex() { PthreadCall("destroy mutex", pthread_mutex_destroy(&mu_)); }
void Mutex::Lock() { PthreadCall("lock", pthread_mutex_lock(&mu_)); }
void Mutex::Unlock() { PthreadCall("unlock", pthread_mutex_unlock(&mu_)); }
CondVar::CondVar(Mutex* mu)
: mu_(mu) {
PthreadCall("init cv", pthread_cond_init(&cv_, nullptr));
}
CondVar::~CondVar() { PthreadCall("destroy cv", pthread_cond_destroy(&cv_)); }
void CondVar::Wait() {
PthreadCall("wait", pthread_cond_wait(&cv_, &mu_->mu_));
}
void CondVar::Signal() {
PthreadCall("signal", pthread_cond_signal(&cv_));
}
void CondVar::SignalAll() {
PthreadCall("broadcast", pthread_cond_broadcast(&cv_));
}
void InitOnce(OnceType* once, void (*initializer)()) {
PthreadCall("once", pthread_once(once, initializer));
}
} // namespace port
} // namespace leveldb

101
port/port_posix.h → port/port_stdcxx.h
View File

@ -1,11 +1,9 @@
// Copyright (c) 2011 The LevelDB Authors. All rights reserved.
// Copyright (c) 2018 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.
//
// See port_example.h for documentation for the following types/functions.
#ifndef STORAGE_LEVELDB_PORT_PORT_POSIX_H_
#define STORAGE_LEVELDB_PORT_PORT_POSIX_H_
#ifndef STORAGE_LEVELDB_PORT_PORT_STDCXX_H_
#define STORAGE_LEVELDB_PORT_PORT_STDCXX_H_
// port/port_config.h availability is automatically detected via __has_include
// in newer compilers. If LEVELDB_HAS_PORT_CONFIG_H is defined, it overrides the
@ -24,71 +22,80 @@
#endif // defined(LEVELDB_HAS_PORT_CONFIG_H)
#include <pthread.h>
#if HAVE_CRC32C
#include <crc32c/crc32c.h>
#endif // HAVE_CRC32C
#if HAVE_SNAPPY
#include <snappy.h>
#endif // HAVE_SNAPPY
#include <stdint.h>
#include <string>
#include "port/atomic_pointer.h"
#include "port/thread_annotations.h"
#if !HAVE_FDATASYNC
#define fdatasync fsync
#endif // !HAVE_FDATASYNC
#include <cassert>
#include <condition_variable> // NOLINT
#include <cstddef>
#include <cstdint>
#include <mutex> // NOLINT
#include <string>
#include "port/thread_annotations.h"
namespace leveldb {
namespace port {
static const bool kLittleEndian = !LEVELDB_IS_BIG_ENDIAN;
class CondVar;
// Thinly wraps std::mutex.
class LOCKABLE Mutex {
public:
Mutex();
~Mutex();
Mutex() = default;
~Mutex() = default;
void Lock() EXCLUSIVE_LOCK_FUNCTION();
void Unlock() UNLOCK_FUNCTION();
void AssertHeld() ASSERT_EXCLUSIVE_LOCK() { }
Mutex(const Mutex&) = delete;
Mutex& operator=(const Mutex&) = delete;
void Lock() EXCLUSIVE_LOCK_FUNCTION() { mu_.lock(); }
void Unlock() UNLOCK_FUNCTION() { mu_.unlock(); }
void AssertHeld() ASSERT_EXCLUSIVE_LOCK() {}
private:
friend class CondVar;
pthread_mutex_t mu_;
// No copying
Mutex(const Mutex&);
void operator=(const Mutex&);
std::mutex mu_;
};
// Thinly wraps std::condition_variable.
class CondVar {
public:
explicit CondVar(Mutex* mu);
~CondVar();
void Wait();
void Signal();
void SignalAll();
explicit CondVar(Mutex* mu) : mu_(mu) { assert(mu != nullptr); }
~CondVar() = default;
CondVar(const CondVar&) = delete;
CondVar& operator=(const CondVar&) = delete;
void Wait() {
std::unique_lock<std::mutex> lock(mu_->mu_, std::adopt_lock);
cv_.wait(lock);
lock.release();
}
void Signal() { cv_.notify_one(); }
void SignalAll() { cv_.notify_all(); }
private:
pthread_cond_t cv_;
Mutex* mu_;
std::condition_variable cv_;
Mutex* const mu_;
};
typedef pthread_once_t OnceType;
#define LEVELDB_ONCE_INIT PTHREAD_ONCE_INIT
void InitOnce(OnceType* once, void (*initializer)());
inline bool Snappy_Compress(const char* input, size_t length,
::std::string* output) {
std::string* output) {
#if HAVE_SNAPPY
output->resize(snappy::MaxCompressedLength(length));
size_t outlen;
snappy::RawCompress(input, length, &(*output)[0], &outlen);
output->resize(outlen);
return true;
#else
// Silence compiler warnings about unused arguments.
(void)input;
(void)length;
(void)output;
#endif // HAVE_SNAPPY
return false;
@ -99,20 +106,30 @@ inline bool Snappy_GetUncompressedLength(const char* input, size_t length,
#if HAVE_SNAPPY
return snappy::GetUncompressedLength(input, length, result);
#else
// Silence compiler warnings about unused arguments.
(void)input;
(void)length;
(void)result;
return false;
#endif // HAVE_SNAPPY
}
inline bool Snappy_Uncompress(const char* input, size_t length,
char* output) {
inline bool Snappy_Uncompress(const char* input, size_t length, char* output) {
#if HAVE_SNAPPY
return snappy::RawUncompress(input, length, output);
#else
// Silence compiler warnings about unused arguments.
(void)input;
(void)length;
(void)output;
return false;
#endif // HAVE_SNAPPY
}
inline bool GetHeapProfile(void (*func)(void*, const char*, int), void* arg) {
// Silence compiler warnings about unused arguments.
(void)func;
(void)arg;
return false;
}
@ -120,6 +137,10 @@ inline uint32_t AcceleratedCRC32C(uint32_t crc, const char* buf, size_t size) {
#if HAVE_CRC32C
return ::crc32c::Extend(crc, reinterpret_cast<const uint8_t*>(buf), size);
#else
// Silence compiler warnings about unused arguments.
(void)crc;
(void)buf;
(void)size;
return 0;
#endif // HAVE_CRC32C
}
@ -127,4 +148,4 @@ inline uint32_t AcceleratedCRC32C(uint32_t crc, const char* buf, size_t size) {
} // namespace port
} // namespace leveldb
#endif // STORAGE_LEVELDB_PORT_PORT_POSIX_H_
#endif // STORAGE_LEVELDB_PORT_PORT_STDCXX_H_

13
port/thread_annotations.h
View File

@ -13,9 +13,9 @@
#if defined(__clang__)
#define THREAD_ANNOTATION_ATTRIBUTE__(x) __attribute__((x))
#define THREAD_ANNOTATION_ATTRIBUTE__(x) __attribute__((x))
#else
#define THREAD_ANNOTATION_ATTRIBUTE__(x) // no-op
#define THREAD_ANNOTATION_ATTRIBUTE__(x) // no-op
#endif
#endif // !defined(THREAD_ANNOTATION_ATTRIBUTE__)
@ -54,18 +54,15 @@
#endif
#ifndef LOCK_RETURNED
#define LOCK_RETURNED(x) \
THREAD_ANNOTATION_ATTRIBUTE__(lock_returned(x))
#define LOCK_RETURNED(x) THREAD_ANNOTATION_ATTRIBUTE__(lock_returned(x))
#endif
#ifndef LOCKABLE
#define LOCKABLE \
THREAD_ANNOTATION_ATTRIBUTE__(lockable)
#define LOCKABLE THREAD_ANNOTATION_ATTRIBUTE__(lockable)
#endif
#ifndef SCOPED_LOCKABLE
#define SCOPED_LOCKABLE \
THREAD_ANNOTATION_ATTRIBUTE__(scoped_lockable)
#define SCOPED_LOCKABLE THREAD_ANNOTATION_ATTRIBUTE__(scoped_lockable)
#endif
#ifndef EXCLUSIVE_LOCK_FUNCTION

24
port/win/stdint.h
View File

@ -1,24 +0,0 @@
// 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.
// MSVC didn't ship with this file until the 2010 version.
#ifndef STORAGE_LEVELDB_PORT_WIN_STDINT_H_
#define STORAGE_LEVELDB_PORT_WIN_STDINT_H_
#if !defined(_MSC_VER)
#error This file should only be included when compiling with MSVC.
#endif
// Define C99 equivalent types.
typedef signed char int8_t;
typedef signed short int16_t;
typedef signed int int32_t;
typedef signed long long int64_t;
typedef unsigned char uint8_t;
typedef unsigned short uint16_t;
typedef unsigned int uint32_t;
typedef unsigned long long uint64_t;
#endif // STORAGE_LEVELDB_PORT_WIN_STDINT_H_

80
table/block.cc
View File

@ -6,8 +6,10 @@
#include "table/block.h"
#include <vector>
#include <algorithm>
#include <cstdint>
#include <vector>
#include "leveldb/comparator.h"
#include "table/format.h"
#include "util/coding.h"
@ -27,7 +29,7 @@ Block::Block(const BlockContents& contents)
if (size_ < sizeof(uint32_t)) {
size_ = 0; // Error marker
} else {
size_t max_restarts_allowed = (size_-sizeof(uint32_t)) / sizeof(uint32_t);
size_t max_restarts_allowed = (size_ - sizeof(uint32_t)) / sizeof(uint32_t);
if (NumRestarts() > max_restarts_allowed) {
// The size is too small for NumRestarts()
size_ = 0;
@ -51,13 +53,12 @@ Block::~Block() {
// If any errors are detected, returns nullptr. Otherwise, returns a
// pointer to the key delta (just past the three decoded values).
static inline const char* DecodeEntry(const char* p, const char* limit,
uint32_t* shared,
uint32_t* non_shared,
uint32_t* shared, uint32_t* non_shared,
uint32_t* value_length) {
if (limit - p < 3) return nullptr;
*shared = reinterpret_cast<const unsigned char*>(p)[0];
*non_shared = reinterpret_cast<const unsigned char*>(p)[1];
*value_length = reinterpret_cast<const unsigned char*>(p)[2];
*shared = reinterpret_cast<const uint8_t*>(p)[0];
*non_shared = reinterpret_cast<const uint8_t*>(p)[1];
*value_length = reinterpret_cast<const uint8_t*>(p)[2];
if ((*shared | *non_shared | *value_length) < 128) {
// Fast path: all three values are encoded in one byte each
p += 3;
@ -76,9 +77,9 @@ static inline const char* DecodeEntry(const char* p, const char* limit,
class Block::Iter : public Iterator {
private:
const Comparator* const comparator_;
const char* const data_; // underlying block contents
uint32_t const restarts_; // Offset of restart array (list of fixed32)
uint32_t const num_restarts_; // Number of uint32_t entries in restart array
const char* const data_; // underlying block contents
uint32_t const restarts_; // Offset of restart array (list of fixed32)
uint32_t const num_restarts_; // Number of uint32_t entries in restart array
// current_ is offset in data_ of current entry. >= restarts_ if !Valid
uint32_t current_;
@ -112,9 +113,7 @@ class Block::Iter : public Iterator {
}
public:
Iter(const Comparator* comparator,
const char* data,
uint32_t restarts,
Iter(const Comparator* comparator, const char* data, uint32_t restarts,
uint32_t num_restarts)
: comparator_(comparator),
data_(data),
@ -125,23 +124,23 @@ class Block::Iter : public Iterator {
assert(num_restarts_ > 0);
}
virtual bool Valid() const { return current_ < restarts_; }
virtual Status status() const { return status_; }
virtual Slice key() const {
bool Valid() const override { return current_ < restarts_; }
Status status() const override { return status_; }
Slice key() const override {
assert(Valid());
return key_;
}
virtual Slice value() const {
Slice value() const override {
assert(Valid());
return value_;
}
virtual void Next() {
void Next() override {
assert(Valid());
ParseNextKey();
}
virtual void Prev() {
void Prev() override {
assert(Valid());
// Scan backwards to a restart point before current_
@ -162,18 +161,36 @@ class Block::Iter : public Iterator {
} while (ParseNextKey() && NextEntryOffset() < original);
}
virtual void Seek(const Slice& target) {
void Seek(const Slice& target) override {
// Binary search in restart array to find the last restart point
// with a key < target
uint32_t left = 0;
uint32_t right = num_restarts_ - 1;
int current_key_compare = 0;
if (Valid()) {
// If we're already scanning, use the current position as a starting
// point. This is beneficial if the key we're seeking to is ahead of the
// current position.
current_key_compare = Compare(key_, target);
if (current_key_compare < 0) {
// key_ is smaller than target
left = restart_index_;
} else if (current_key_compare > 0) {
right = restart_index_;
} else {
// We're seeking to the key we're already at.
return;
}
}
while (left < right) {
uint32_t mid = (left + right + 1) / 2;
uint32_t region_offset = GetRestartPoint(mid);
uint32_t shared, non_shared, value_length;
const char* key_ptr = DecodeEntry(data_ + region_offset,
data_ + restarts_,
&shared, &non_shared, &value_length);
const char* key_ptr =
DecodeEntry(data_ + region_offset, data_ + restarts_, &shared,
&non_shared, &value_length);
if (key_ptr == nullptr || (shared != 0)) {
CorruptionError();
return;
@ -190,8 +207,15 @@ class Block::Iter : public Iterator {
}
}
// We might be able to use our current position within the restart block.
// This is true if we determined the key we desire is in the current block
// and is after than the current key.
assert(current_key_compare == 0 || Valid());
bool skip_seek = left == restart_index_ && current_key_compare < 0;
if (!skip_seek) {
SeekToRestartPoint(left);
}
// Linear search (within restart block) for first key >= target
SeekToRestartPoint(left);
while (true) {
if (!ParseNextKey()) {
return;
@ -202,12 +226,12 @@ class Block::Iter : public Iterator {
}
}
virtual void SeekToFirst() {
void SeekToFirst() override {
SeekToRestartPoint(0);
ParseNextKey();
}
virtual void SeekToLast() {
void SeekToLast() override {
SeekToRestartPoint(num_restarts_ - 1);
while (ParseNextKey() && NextEntryOffset() < restarts_) {
// Keep skipping
@ -253,7 +277,7 @@ class Block::Iter : public Iterator {
}
};
Iterator* Block::NewIterator(const Comparator* cmp) {
Iterator* Block::NewIterator(const Comparator* comparator) {
if (size_ < sizeof(uint32_t)) {
return NewErrorIterator(Status::Corruption("bad block contents"));
}
@ -261,7 +285,7 @@ Iterator* Block::NewIterator(const Comparator* cmp) {
if (num_restarts == 0) {
return NewEmptyIterator();
} else {
return new Iter(cmp, data_, restart_offset_, num_restarts);
return new Iter(comparator, data_, restart_offset_, num_restarts);
}
}

20
table/block.h
View File

@ -5,8 +5,9 @@
#ifndef STORAGE_LEVELDB_TABLE_BLOCK_H_
#define STORAGE_LEVELDB_TABLE_BLOCK_H_
#include <stddef.h>
#include <stdint.h>
#include <cstddef>
#include <cstdint>
#include "leveldb/iterator.h"
namespace leveldb {
@ -19,24 +20,23 @@ class Block {
// Initialize the block with the specified contents.
explicit Block(const BlockContents& contents);
Block(const Block&) = delete;
Block& operator=(const Block&) = delete;
~Block();
size_t size() const { return size_; }
Iterator* NewIterator(const Comparator* comparator);
private:
class Iter;
uint32_t NumRestarts() const;
const char* data_;
size_t size_;
uint32_t restart_offset_; // Offset in data_ of restart array
bool owned_; // Block owns data_[]
// No copying allowed
Block(const Block&);
void operator=(const Block&);
class Iter;
uint32_t restart_offset_; // Offset in data_ of restart array
bool owned_; // Block owns data_[]
};
} // namespace leveldb

22
table/block_builder.cc
View File

@ -29,35 +29,33 @@
#include "table/block_builder.h"
#include <algorithm>
#include <assert.h>
#include <cassert>
#include "leveldb/comparator.h"
#include "leveldb/table_builder.h"
#include "leveldb/options.h"
#include "util/coding.h"
namespace leveldb {
BlockBuilder::BlockBuilder(const Options* options)
: options_(options),
restarts_(),
counter_(0),
finished_(false) {
: options_(options), restarts_(), counter_(0), finished_(false) {
assert(options->block_restart_interval >= 1);
restarts_.push_back(0); // First restart point is at offset 0
restarts_.push_back(0); // First restart point is at offset 0
}
void BlockBuilder::Reset() {
buffer_.clear();
restarts_.clear();
restarts_.push_back(0); // First restart point is at offset 0
restarts_.push_back(0); // First restart point is at offset 0
counter_ = 0;
finished_ = false;
last_key_.clear();
}
size_t BlockBuilder::CurrentSizeEstimate() const {
return (buffer_.size() + // Raw data buffer
restarts_.size() * sizeof(uint32_t) + // Restart array
sizeof(uint32_t)); // Restart array length
return (buffer_.size() + // Raw data buffer
restarts_.size() * sizeof(uint32_t) + // Restart array
sizeof(uint32_t)); // Restart array length
}
Slice BlockBuilder::Finish() {
@ -74,7 +72,7 @@ void BlockBuilder::Add(const Slice& key, const Slice& value) {
Slice last_key_piece(last_key_);
assert(!finished_);
assert(counter_ <= options_->block_restart_interval);
assert(buffer_.empty() // No values yet?
assert(buffer_.empty() // No values yet?
|| options_->comparator->Compare(key, last_key_piece) > 0);
size_t shared = 0;
if (counter_ < options_->block_restart_interval) {

25
table/block_builder.h
View File

@ -5,9 +5,9 @@
#ifndef STORAGE_LEVELDB_TABLE_BLOCK_BUILDER_H_
#define STORAGE_LEVELDB_TABLE_BLOCK_BUILDER_H_
#include <cstdint>
#include <vector>
#include <stdint.h>
#include "leveldb/slice.h"
namespace leveldb {
@ -18,6 +18,9 @@ class BlockBuilder {
public:
explicit BlockBuilder(const Options* options);
BlockBuilder(const BlockBuilder&) = delete;
BlockBuilder& operator=(const BlockBuilder&) = delete;
// Reset the contents as if the BlockBuilder was just constructed.
void Reset();
@ -35,21 +38,15 @@ class BlockBuilder {
size_t CurrentSizeEstimate() const;
// Return true iff no entries have been added since the last Reset()
bool empty() const {
return buffer_.empty();
}
bool empty() const { return buffer_.empty(); }
private:
const Options* options_;
std::string buffer_; // Destination buffer
std::vector<uint32_t> restarts_; // Restart points
int counter_; // Number of entries emitted since restart
bool finished_; // Has Finish() been called?
std::string last_key_;
// No copying allowed
BlockBuilder(const BlockBuilder&);
void operator=(const BlockBuilder&);
const Options* options_;
std::string buffer_; // Destination buffer
std::vector<uint32_t> restarts_; // Restart points
int counter_; // Number of entries emitted since restart
bool finished_; // Has Finish() been called?
std::string last_key_;
};
} // namespace leveldb

19
table/filter_block.cc
View File

@ -16,8 +16,7 @@ static const size_t kFilterBaseLg = 11;
static const size_t kFilterBase = 1 << kFilterBaseLg;
FilterBlockBuilder::FilterBlockBuilder(const FilterPolicy* policy)
: policy_(policy) {
}
: policy_(policy) {}
void FilterBlockBuilder::StartBlock(uint64_t block_offset) {
uint64_t filter_index = (block_offset / kFilterBase);
@ -62,7 +61,7 @@ void FilterBlockBuilder::GenerateFilter() {
tmp_keys_.resize(num_keys);
for (size_t i = 0; i < num_keys; i++) {
const char* base = keys_.data() + start_[i];
size_t length = start_[i+1] - start_[i];
size_t length = start_[i + 1] - start_[i];
tmp_keys_[i] = Slice(base, length);
}
@ -77,14 +76,10 @@ void FilterBlockBuilder::GenerateFilter() {
FilterBlockReader::FilterBlockReader(const FilterPolicy* policy,
const Slice& contents)
: policy_(policy),
data_(nullptr),
offset_(nullptr),
num_(0),
base_lg_(0) {
: policy_(policy), data_(nullptr), offset_(nullptr), num_(0), base_lg_(0) {
size_t n = contents.size();
if (n < 5) return; // 1 byte for base_lg_ and 4 for start of offset array
base_lg_ = contents[n-1];
base_lg_ = contents[n - 1];
uint32_t last_word = DecodeFixed32(contents.data() + n - 5);
if (last_word > n - 5) return;
data_ = contents.data();
@ -95,8 +90,8 @@ FilterBlockReader::FilterBlockReader(const FilterPolicy* policy,
bool FilterBlockReader::KeyMayMatch(uint64_t block_offset, const Slice& key) {
uint64_t index = block_offset >> base_lg_;
if (index < num_) {
uint32_t start = DecodeFixed32(offset_ + index*4);
uint32_t limit = DecodeFixed32(offset_ + index*4 + 4);
uint32_t start = DecodeFixed32(offset_ + index * 4);
uint32_t limit = DecodeFixed32(offset_ + index * 4 + 4);
if (start <= limit && limit <= static_cast<size_t>(offset_ - data_)) {
Slice filter = Slice(data_ + start, limit - start);
return policy_->KeyMayMatch(key, filter);
@ -108,4 +103,4 @@ bool FilterBlockReader::KeyMayMatch(uint64_t block_offset, const Slice& key) {
return true; // Errors are treated as potential matches
}
}
} // namespace leveldb

24
table/filter_block.h
View File

@ -9,10 +9,11 @@
#ifndef STORAGE_LEVELDB_TABLE_FILTER_BLOCK_H_
#define STORAGE_LEVELDB_TABLE_FILTER_BLOCK_H_
#include <stddef.h>
#include <stdint.h>
#include <cstddef>
#include <cstdint>
#include <string>
#include <vector>
#include "leveldb/slice.h"
#include "util/hash.h"
@ -30,6 +31,9 @@ class FilterBlockBuilder {
public:
explicit FilterBlockBuilder(const FilterPolicy*);
FilterBlockBuilder(const FilterBlockBuilder&) = delete;
FilterBlockBuilder& operator=(const FilterBlockBuilder&) = delete;
void StartBlock(uint64_t block_offset);
void AddKey(const Slice& key);
Slice Finish();
@ -38,20 +42,16 @@ class FilterBlockBuilder {
void GenerateFilter();
const FilterPolicy* policy_;
std::string keys_; // Flattened key contents
std::vector<size_t> start_; // Starting index in keys_ of each key
std::string result_; // Filter data computed so far
std::vector<Slice> tmp_keys_; // policy_->CreateFilter() argument
std::string keys_; // Flattened key contents
std::vector<size_t> start_; // Starting index in keys_ of each key
std::string result_; // Filter data computed so far
std::vector<Slice> tmp_keys_; // policy_->CreateFilter() argument
std::vector<uint32_t> filter_offsets_;
// No copying allowed
FilterBlockBuilder(const FilterBlockBuilder&);
void operator=(const FilterBlockBuilder&);
};
class FilterBlockReader {
public:
// REQUIRES: "contents" and *policy must stay live while *this is live.
// REQUIRES: "contents" and *policy must stay live while *this is live.
FilterBlockReader(const FilterPolicy* policy, const Slice& contents);
bool KeyMayMatch(uint64_t block_offset, const Slice& key);
@ -63,6 +63,6 @@ class FilterBlockReader {
size_t base_lg_; // Encoding parameter (see kFilterBaseLg in .cc file)
};
}
} // namespace leveldb
#endif // STORAGE_LEVELDB_TABLE_FILTER_BLOCK_H_

48
table/filter_block_test.cc
View File

@ -4,11 +4,11 @@
#include "table/filter_block.h"
#include "gtest/gtest.h"
#include "leveldb/filter_policy.h"
#include "util/coding.h"
#include "util/hash.h"
#include "util/logging.h"
#include "util/testharness.h"
#include "util/testutil.h"
namespace leveldb {
@ -16,18 +16,16 @@ namespace leveldb {
// For testing: emit an array with one hash value per key
class TestHashFilter : public FilterPolicy {
public:
virtual const char* Name() const {
return "TestHashFilter";
}
const char* Name() const override { return "TestHashFilter"; }
virtual void CreateFilter(const Slice* keys, int n, std::string* dst) const {
void CreateFilter(const Slice* keys, int n, std::string* dst) const override {
for (int i = 0; i < n; i++) {
uint32_t h = Hash(keys[i].data(), keys[i].size(), 1);
PutFixed32(dst, h);
}
}
virtual bool KeyMayMatch(const Slice& key, const Slice& filter) const {
bool KeyMayMatch(const Slice& key, const Slice& filter) const override {
uint32_t h = Hash(key.data(), key.size(), 1);
for (size_t i = 0; i + 4 <= filter.size(); i += 4) {
if (h == DecodeFixed32(filter.data() + i)) {
@ -38,12 +36,12 @@ class TestHashFilter : public FilterPolicy {
}
};
class FilterBlockTest {
class FilterBlockTest : public testing::Test {
public:
TestHashFilter policy_;
};
TEST(FilterBlockTest, EmptyBuilder) {
TEST_F(FilterBlockTest, EmptyBuilder) {
FilterBlockBuilder builder(&policy_);
Slice block = builder.Finish();
ASSERT_EQ("\\x00\\x00\\x00\\x00\\x0b", EscapeString(block));
@ -52,7 +50,7 @@ TEST(FilterBlockTest, EmptyBuilder) {
ASSERT_TRUE(reader.KeyMayMatch(100000, "foo"));
}
TEST(FilterBlockTest, SingleChunk) {
TEST_F(FilterBlockTest, SingleChunk) {
FilterBlockBuilder builder(&policy_);
builder.StartBlock(100);
builder.AddKey("foo");
@ -69,11 +67,11 @@ TEST(FilterBlockTest, SingleChunk) {
ASSERT_TRUE(reader.KeyMayMatch(100, "box"));
ASSERT_TRUE(reader.KeyMayMatch(100, "hello"));
ASSERT_TRUE(reader.KeyMayMatch(100, "foo"));
ASSERT_TRUE(! reader.KeyMayMatch(100, "missing"));
ASSERT_TRUE(! reader.KeyMayMatch(100, "other"));
ASSERT_TRUE(!reader.KeyMayMatch(100, "missing"));
ASSERT_TRUE(!reader.KeyMayMatch(100, "other"));
}
TEST(FilterBlockTest, MultiChunk) {
TEST_F(FilterBlockTest, MultiChunk) {
FilterBlockBuilder builder(&policy_);
// First filter
@ -99,30 +97,26 @@ TEST(FilterBlockTest, MultiChunk) {
// Check first filter
ASSERT_TRUE(reader.KeyMayMatch(0, "foo"));
ASSERT_TRUE(reader.KeyMayMatch(2000, "bar"));
ASSERT_TRUE(! reader.KeyMayMatch(0, "box"));
ASSERT_TRUE(! reader.KeyMayMatch(0, "hello"));
ASSERT_TRUE(!reader.KeyMayMatch(0, "box"));
ASSERT_TRUE(!reader.KeyMayMatch(0, "hello"));
// Check second filter
ASSERT_TRUE(reader.KeyMayMatch(3100, "box"));
ASSERT_TRUE(! reader.KeyMayMatch(3100, "foo"));
ASSERT_TRUE(! reader.KeyMayMatch(3100, "bar"));
ASSERT_TRUE(! reader.KeyMayMatch(3100, "hello"));
ASSERT_TRUE(!reader.KeyMayMatch(3100, "foo"));
ASSERT_TRUE(!reader.KeyMayMatch(3100, "bar"));
ASSERT_TRUE(!reader.KeyMayMatch(3100, "hello"));
// Check third filter (empty)
ASSERT_TRUE(! reader.KeyMayMatch(4100, "foo"));
ASSERT_TRUE(! reader.KeyMayMatch(4100, "bar"));
ASSERT_TRUE(! reader.KeyMayMatch(4100, "box"));
ASSERT_TRUE(! reader.KeyMayMatch(4100, "hello"));
ASSERT_TRUE(!reader.KeyMayMatch(4100, "foo"));
ASSERT_TRUE(!reader.KeyMayMatch(4100, "bar"));
ASSERT_TRUE(!reader.KeyMayMatch(4100, "box"));
ASSERT_TRUE(!reader.KeyMayMatch(4100, "hello"));
// Check last filter
ASSERT_TRUE(reader.KeyMayMatch(9000, "box"));
ASSERT_TRUE(reader.KeyMayMatch(9000, "hello"));
ASSERT_TRUE(! reader.KeyMayMatch(9000, "foo"));
ASSERT_TRUE(! reader.KeyMayMatch(9000, "bar"));
ASSERT_TRUE(!reader.KeyMayMatch(9000, "foo"));
ASSERT_TRUE(!reader.KeyMayMatch(9000, "bar"));
}
} // namespace leveldb
int main(int argc, char** argv) {
return leveldb::test::RunAllTests();
}

11
table/format.cc
View File

@ -21,8 +21,7 @@ void BlockHandle::EncodeTo(std::string* dst) const {
}
Status BlockHandle::DecodeFrom(Slice* input) {
if (GetVarint64(input, &offset_) &&
GetVarint64(input, &size_)) {
if (GetVarint64(input, &offset_) && GetVarint64(input, &size_)) {
return Status::OK();
} else {
return Status::Corruption("bad block handle");
@ -62,10 +61,8 @@ Status Footer::DecodeFrom(Slice* input) {
return result;
}
Status ReadBlock(RandomAccessFile* file,
const ReadOptions& options,
const BlockHandle& handle,
BlockContents* result) {
Status ReadBlock(RandomAccessFile* file, const ReadOptions& options,
const BlockHandle& handle, BlockContents* result) {
result->data = Slice();
result->cachable = false;
result->heap_allocated = false;
@ -86,7 +83,7 @@ Status ReadBlock(RandomAccessFile* file,
}
// Check the crc of the type and the block contents
const char* data = contents.data(); // Pointer to where Read put the data
const char* data = contents.data(); // Pointer to where Read put the data
if (options.verify_checksums) {
const uint32_t crc = crc32c::Unmask(DecodeFixed32(data + n + 1));
const uint32_t actual = crc32c::Value(data, n + 1);

41
table/format.h
View File

@ -5,8 +5,9 @@
#ifndef STORAGE_LEVELDB_TABLE_FORMAT_H_
#define STORAGE_LEVELDB_TABLE_FORMAT_H_
#include <cstdint>
#include <string>
#include <stdint.h>
#include "leveldb/slice.h"
#include "leveldb/status.h"
#include "leveldb/table_builder.h"
@ -21,6 +22,9 @@ struct ReadOptions;
// block or a meta block.
class BlockHandle {
public:
// Maximum encoding length of a BlockHandle
enum { kMaxEncodedLength = 10 + 10 };
BlockHandle();
// The offset of the block in the file.
@ -34,9 +38,6 @@ class BlockHandle {
void EncodeTo(std::string* dst) const;
Status DecodeFrom(Slice* input);
// Maximum encoding length of a BlockHandle
enum { kMaxEncodedLength = 10 + 10 };
private:
uint64_t offset_;
uint64_t size_;
@ -46,30 +47,24 @@ class BlockHandle {
// end of every table file.
class Footer {
public:
Footer() { }
// Encoded length of a Footer. Note that the serialization of a
// Footer will always occupy exactly this many bytes. It consists
// of two block handles and a magic number.
enum { kEncodedLength = 2 * BlockHandle::kMaxEncodedLength + 8 };
Footer() = default;
// The block handle for the metaindex block of the table
const BlockHandle& metaindex_handle() const { return metaindex_handle_; }
void set_metaindex_handle(const BlockHandle& h) { metaindex_handle_ = h; }
// The block handle for the index block of the table
const BlockHandle& index_handle() const {
return index_handle_;
}
void set_index_handle(const BlockHandle& h) {
index_handle_ = h;
}
const BlockHandle& index_handle() const { return index_handle_; }
void set_index_handle(const BlockHandle& h) { index_handle_ = h; }
void EncodeTo(std::string* dst) const;
Status DecodeFrom(Slice* input);
// Encoded length of a Footer. Note that the serialization of a
// Footer will always occupy exactly this many bytes. It consists
// of two block handles and a magic number.
enum {
kEncodedLength = 2*BlockHandle::kMaxEncodedLength + 8
};
private:
BlockHandle metaindex_handle_;
BlockHandle index_handle_;
@ -91,17 +86,13 @@ struct BlockContents {
// Read the block identified by "handle" from "file". On failure
// return non-OK. On success fill *result and return OK.
Status ReadBlock(RandomAccessFile* file,
const ReadOptions& options,
const BlockHandle& handle,
BlockContents* result);
Status ReadBlock(RandomAccessFile* file, const ReadOptions& options,
const BlockHandle& handle, BlockContents* result);
// Implementation details follow. Clients should ignore,
inline BlockHandle::BlockHandle()
: offset_(~static_cast<uint64_t>(0)),
size_(~static_cast<uint64_t>(0)) {
}
: offset_(~static_cast<uint64_t>(0)), size_(~static_cast<uint64_t>(0)) {}
} // namespace leveldb

73
table/iterator.cc
View File

@ -7,58 +7,67 @@
namespace leveldb {
Iterator::Iterator() {
cleanup_.function = nullptr;
cleanup_.next = nullptr;
cleanup_head_.function = nullptr;
cleanup_head_.next = nullptr;
}
Iterator::~Iterator() {
if (cleanup_.function != nullptr) {
(*cleanup_.function)(cleanup_.arg1, cleanup_.arg2);
for (Cleanup* c = cleanup_.next; c != nullptr; ) {
(*c->function)(c->arg1, c->arg2);
Cleanup* next = c->next;
delete c;
c = next;
if (!cleanup_head_.IsEmpty()) {
cleanup_head_.Run();
for (CleanupNode* node = cleanup_head_.next; node != nullptr;) {
node->Run();
CleanupNode* next_node = node->next;
delete node;
node = next_node;
}
}
}
void Iterator::RegisterCleanup(CleanupFunction func, void* arg1, void* arg2) {
assert(func != nullptr);
Cleanup* c;
if (cleanup_.function == nullptr) {
c = &cleanup_;
CleanupNode* node;
if (cleanup_head_.IsEmpty()) {
node = &cleanup_head_;
} else {
c = new Cleanup;
c->next = cleanup_.next;
cleanup_.next = c;
node = new CleanupNode();
node->next = cleanup_head_.next;
cleanup_head_.next = node;
}
c->function = func;
c->arg1 = arg1;
c->arg2 = arg2;
node->function = func;
node->arg1 = arg1;
node->arg2 = arg2;
}
namespace {
class EmptyIterator : public Iterator {
public:
EmptyIterator(const Status& s) : status_(s) { }
virtual bool Valid() const { return false; }
virtual void Seek(const Slice& target) { }
virtual void SeekToFirst() { }
virtual void SeekToLast() { }
virtual void Next() { assert(false); }
virtual void Prev() { assert(false); }
Slice key() const { assert(false); return Slice(); }
Slice value() const { assert(false); return Slice(); }
virtual Status status() const { return status_; }
EmptyIterator(const Status& s) : status_(s) {}
~EmptyIterator() override = default;
bool Valid() const override { return false; }
void Seek(const Slice& target) override {}
void SeekToFirst() override {}
void SeekToLast() override {}
void Next() override { assert(false); }
void Prev() override { assert(false); }
Slice key() const override {
assert(false);
return Slice();
}
Slice value() const override {
assert(false);
return Slice();
}
Status status() const override { return status_; }
private:
Status status_;
};
} // namespace
Iterator* NewEmptyIterator() {
return new EmptyIterator(Status::OK());
}
} // anonymous namespace
Iterator* NewEmptyIterator() { return new EmptyIterator(Status::OK()); }
Iterator* NewErrorIterator(const Status& status) {
return new EmptyIterator(status);

54
table/iterator_wrapper.h
View File

@ -16,10 +16,8 @@ namespace leveldb {
// cache locality.
class IteratorWrapper {
public:
IteratorWrapper(): iter_(nullptr), valid_(false) { }
explicit IteratorWrapper(Iterator* iter): iter_(nullptr) {
Set(iter);
}
IteratorWrapper() : iter_(nullptr), valid_(false) {}
explicit IteratorWrapper(Iterator* iter) : iter_(nullptr) { Set(iter); }
~IteratorWrapper() { delete iter_; }
Iterator* iter() const { return iter_; }
@ -35,18 +33,46 @@ class IteratorWrapper {
}
}
// Iterator interface methods
bool Valid() const { return valid_; }
Slice key() const { assert(Valid()); return key_; }
Slice value() const { assert(Valid()); return iter_->value(); }
bool Valid() const { return valid_; }
Slice key() const {
assert(Valid());
return key_;
}
Slice value() const {
assert(Valid());
return iter_->value();
}
// Methods below require iter() != nullptr
Status status() const { assert(iter_); return iter_->status(); }
void Next() { assert(iter_); iter_->Next(); Update(); }
void Prev() { assert(iter_); iter_->Prev(); Update(); }
void Seek(const Slice& k) { assert(iter_); iter_->Seek(k); Update(); }
void SeekToFirst() { assert(iter_); iter_->SeekToFirst(); Update(); }
void SeekToLast() { assert(iter_); iter_->SeekToLast(); Update(); }
Status status() const {
assert(iter_);
return iter_->status();
}
void Next() {
assert(iter_);
iter_->Next();
Update();
}
void Prev() {
assert(iter_);
iter_->Prev();
Update();
}
void Seek(const Slice& k) {
assert(iter_);
iter_->Seek(k);
Update();
}
void SeekToFirst() {
assert(iter_);
iter_->SeekToFirst();
Update();
}
void SeekToLast() {
assert(iter_);
iter_->SeekToLast();
Update();
}
private:
void Update() {

42
table/merger.cc
View File

@ -24,15 +24,11 @@ class MergingIterator : public Iterator {
}
}
virtual ~MergingIterator() {
delete[] children_;
}
~MergingIterator() override { delete[] children_; }
virtual bool Valid() const {
return (current_ != nullptr);
}
bool Valid() const override { return (current_ != nullptr); }
virtual void SeekToFirst() {
void SeekToFirst() override {
for (int i = 0; i < n_; i++) {
children_[i].SeekToFirst();
}
@ -40,7 +36,7 @@ class MergingIterator : public Iterator {
direction_ = kForward;
}
virtual void SeekToLast() {
void SeekToLast() override {
for (int i = 0; i < n_; i++) {
children_[i].SeekToLast();
}
@ -48,7 +44,7 @@ class MergingIterator : public Iterator {
direction_ = kReverse;
}
virtual void Seek(const Slice& target) {
void Seek(const Slice& target) override {
for (int i = 0; i < n_; i++) {
children_[i].Seek(target);
}
@ -56,7 +52,7 @@ class MergingIterator : public Iterator {
direction_ = kForward;
}
virtual void Next() {
void Next() override {
assert(Valid());
// Ensure that all children are positioned after key().
@ -82,7 +78,7 @@ class MergingIterator : public Iterator {
FindSmallest();
}
virtual void Prev() {
void Prev() override {
assert(Valid());
// Ensure that all children are positioned before key().
@ -111,17 +107,17 @@ class MergingIterator : public Iterator {
FindLargest();
}
virtual Slice key() const {
Slice key() const override {
assert(Valid());
return current_->key();
}
virtual Slice value() const {
Slice value() const override {
assert(Valid());
return current_->value();
}
virtual Status status() const {
Status status() const override {
Status status;
for (int i = 0; i < n_; i++) {
status = children_[i].status();
@ -133,6 +129,9 @@ class MergingIterator : public Iterator {
}
private:
// Which direction is the iterator moving?
enum Direction { kForward, kReverse };
void FindSmallest();
void FindLargest();
@ -143,12 +142,6 @@ class MergingIterator : public Iterator {
IteratorWrapper* children_;
int n_;
IteratorWrapper* current_;
// Which direction is the iterator moving?
enum Direction {
kForward,
kReverse
};
Direction direction_;
};
@ -169,7 +162,7 @@ void MergingIterator::FindSmallest() {
void MergingIterator::FindLargest() {
IteratorWrapper* largest = nullptr;
for (int i = n_-1; i >= 0; i--) {
for (int i = n_ - 1; i >= 0; i--) {
IteratorWrapper* child = &children_[i];
if (child->Valid()) {
if (largest == nullptr) {
@ -183,14 +176,15 @@ void MergingIterator::FindLargest() {
}
} // namespace
Iterator* NewMergingIterator(const Comparator* cmp, Iterator** list, int n) {
Iterator* NewMergingIterator(const Comparator* comparator, Iterator** children,
int n) {
assert(n >= 0);
if (n == 0) {
return NewEmptyIterator();
} else if (n == 1) {
return list[0];
return children[0];
} else {
return new MergingIterator(cmp, list, n);
return new MergingIterator(comparator, children, n);
}
}

Some files were not shown because too many files have changed in this diff Show More